Preparation of zeolite thin films for gas...

PREPARATION OF ZEOLITE THIN FILMS FOR GAS PURIFICATION

Francisco José Varela Gandía

http://www.ua.es/

http://www.eltallerdigital.com/

UNIVERSIDAD DE ALICANTE

Instituto Universitario de Materiales de Alicante

Departamento de Química Inorgánica

PREPARATION OF ZEOLITE THIN FILMS FOR GAS PURIFICATION

Memoria presentada pasa aspirar al grado de

DOCTOR EN QUÍMICA CON

MENCIÓN DE DOCTOR INTERNACIONAL

FRANCISCO JOSÉ VARELA GANDIA

Directores del trabajo

Diego Cazorla Amorós Dolores Lozano Castelló

Catedrático de Química Inorgánica Profesora Titular de Química Inorgánica

Alicante, Julio 2012

Diego Cazorla Amorós, Catedrático de Química Inorgánica, Dolores Lozano

Castelló, Profesora Titular de Química Inorgánica, ambos de la Universidad de

Alicante.

CERTIFICAN:

D. Francisco José Varela Gandía, Licenciado en Ciencias Químicas ha realizado en

el Departamento de Química Inorgánica de la Facultad de Ciencias, bajo nuestra

dirección, el trabajo que lleva por título: PREPARATION OF ZEOLITE THIN

FILMS FOR GAS PURIFICATION, que constituye su Memoria para aspirar al

grado de Doctor en Ciencias Químicas, reuniendo, a nuestro juicio, las condiciones

necesarias para ser presentada y defendida ante el tribunal correspondiente.

Y para que conste a efectos oportunos, en cumplimiento de la legislación vigente,

firmamos el presente certificado en a Alicante a 23 de Julio de 2012.

Dr. Diego Cazorla Amorós Dra. Dolores Lozano Castelló

Catedrático de Química Inorgánica Profesora Titular de Química Inorgánica

A mi familia

A Esther

Agradecimientos.

Septiembre de 2007, aquí comencé mi Tesis Doctoral. Desde entonces han pasado 5

años, pero parece que fue ayer cuando entraba muy ilusionado en el departamento

para empezar en un proyecto de zeolitas.

En primer lugar, quisiera agradecer a mis directores de Tesis, Dr. Diego Cazorla y

Dra. Dolores Lozano la oportunidad de haber podido realizar la Tesis, una de mis

ilusiones y sentirme así más realizado. Durante estos años me habéis enseñado a

trabajar con precisión, siempre con una buena idea para continuar, a ver más a allá de

unos simples resultados, a comprender lo que tengo entre manos y a superarme en

todas aquellas dificultades encontradas. Debo decir, que contagiáis vuestras ganas de

trabajar e ilusión a quien está a vuestro lado.

Además, quisiera agradecer al Dr. Ángel Berenguer la ayuda que me has prestado,

pues has contribuido como un director más a esta Tesis. Muchos han sido los

consejos y los ratos de café que me han ayudado a que todo saliese bien.

Dentro del departamento, me gustaría valorar el trabajo realizado por los técnicos y

secretarias del departamento, el cual es esencial para que todo funcione. Gracias por

vuestra ayuda.

A todos los compañeros del departamento, tanto a los que ya terminaron (Sonia, Juan

Antonio, María…) como los que continúan (David, Moha, Aroldo, Cristina, Noelia,

Vero…). Hemos compartido grandes momentos dentro y fuera del racó durante todos

estos años. A Guido, por los momentos buenos que hemos pasado y los cafés

compartidos. Mención especial quisiera hacer a Miriam e Izaskun, pues son muchos

los momentos de ayuda, de apoyo, de trabajo compartiendo equipos… que han

forjado una gran amistad. Fuimos compañeros de clase, compañeros de trabajo y

ahora grandes amigos.

Quisiera agradecer al Dr. Stuart Taylor por acogerme el verano pasado en su grupo

de investigación en Cardiff, para realizar una estancia de investigación. A Dave, por

acogerme, integrarme en el departamento y ayudarme con todo lo que hacia falta de

la estancia. Y sobre todo, a todos los amigos que hice (Raimon, Marco, Pili, Lino,

Rocío, Mauricio,…), me incluisteis en vuestro grupo de amigos, me sentí como uno

más de la familia que sois.

A mi familia, por vuestro apoyo a lo largo de estos años. Y sobretodo, a mi abuelo,

por sus grandes palabras y alegría de ver a su nieto evolucionando en la universidad.

Siempre me has dado ánimos para seguir por este camino.

A Esther, mi amor, gracias por estar todos estos años a mi lado, durante la carrera y

durante la tesis. Me has dado la fuerza necesaria para seguir adelante en los buenos y

malos momentos. Cuando llegaba mal o enfadado por cualquier motivo, me hacías

ver el problema desde otro punto de vista, y me has animado a seguir para adelante.

Gracias por venirte el año pasado a Cardiff, fueron dos meses maravillosos.

General Index

Capítulo 0. Introducción, objetivos y estructura de la Tesis Doctoral. 1

Chapter 1. Introduction and objectives. 13

Capítulo 2. Técnicas de caracterización, materiales y métodos de

preparación.

35

Chapter 3. Hydrogen purification for PEM fuel cells using membranes

prepared by ion-exchange of Na-LTA/carbon membranes.

71

Chapter 4. Zeolite A/carbon membranes for H2 purification from a

simulated gas reformer mixture.

97

Chapter 5. Hydrocarbon traps based on zeolites for gasoline vehicle

emission control tested under cold start conditions.

121

Chapter 6. Total oxidation of naphthalene using palladium

nanoparticles supported on BETA, ZSM-5, SAPO-5 and

alumina powders.

137

Chapter 7. Total oxidation of naphthalene at low temperatures using

palladium nanoparticles supported on inorganic oxide-

coated cordierite honeycomb monoliths.

161

Chapter 8. Preferential oxidation of CO catalyzed by palladium

nanoparticles supported on inorganic oxides and on

inorganic oxide-coated cordierite honeycomb monoliths.

183

Chapter 9. General conclusions. 193

Resumen en castellano. Preparación de películas delgadas de zeolita

para la purificación de gases.

197

Contents

Capítulo 0: Introducción, objetivos y estructura de la Tesis Doctoral.

1. Introducción general.

1.1. Zeolitas, membranas de zeolita y películas delgadas de zeolitas.

1.2. Purificación de H2 para pilas de combustible.

1.3. Reducción de las emisiones de hidrocarburos durante el arranque en

frío en vehículos.

1.4. Reducción de las emisiones de compuestos aromáticos policíclicos

(PAH).

2. Objetivos de la tesis doctoral.

3. Estructura de la tesis doctoral.

Chapter 1: Introduction and objectives.

1.1. Zeolites, zeolite membranes and zeolites films.

1.2. Hydrogen purification for PEM fuel cells.

1.3. Removal of hydrocarbons under cold start conditions in vehicles

emissions.

1.4. Removal of polycyclic aromatic hydrocarbons (PAH) by catalytic

oxidation.

PhD Thesis objectives.

References.

Capítulo 2: Técnicas de caracterización, materiales y métodos de preparación.

2.1. Técnicas de caracterización.

2.1.1. Adsorción física de gases.

2.1.2. Análisis térmico. Análisis termogravimétrico.

2.1.3. Espectroscopia de emisión de plasma por acoplamiento

inductivo.

2.1.4. Espectroscopia Infrarroja de reflectancia difusa con

transformada de Fourier.

2.1.5. Difracción de Rayos X.

2.1.6. Espectroscopia fotoelectrónica de rayos X.

2.1.7. Microscopía electrónica de transmisión, barrido y

espectroscopia de fluorescencia de rayos X basada en

dispersión de energía.

2.1.8. Sistema experimental de análisis de membranas: célula de

permeación Wicke-Kallenbach.

2.2. Materiales empleados.

2.2.1 Monolitos cerámicos de estructura celular (“honeycomb”).

2.2.2 Discos de grafito macroporoso.

2.3. Métodos de preparación de membranas de zeolita por crecimiento

secundario: Sembrado de soportes mediante depósito electroforético.

2.4. Preparación de películas delgadas de zeolita y silicoaluminofosfato

sobre monolitos de cordierita.

Bibliografía.

Chapter 3: Hydrogen purification for PEM fuel cells using membranes

prepared by ion-exchange of Na-LTA/carbon membrane.

3.1. Introduction.

3.2. Experimental.

3.2.1. LTA/carbon membrane preparation.

3.2.2. Membrane characterization.

3.3. Results and discussion.

3.3.1. Crystal structure, morphology and composition.

3.3.2. Single gas permeation tests.

3.3.3. H2/CO mixtures permeation tests in LTA/carbon membranes.

3.4. Conclusions.

References.

Chapter 4. Zeolite A/carbon membranes for H2 purification from a simulated

gas reformer mixture.

4.1. Introduction.

4.2. Experimental.




4.3.1. Effect of CO2 concentration in the membrane permeation

properties.

4.3.2. Study of the CO2-zeolite interaction.

4.3.2.1. DRIFTS studies.

4.3.2.2. TGA and CO2 isotherms analysis.

4.3.3. Membrane permeation properties in a simulated reformer

mixture.

4.4. Conclusions.

References

Chapter 5. Hydrocarbon traps based on zeolites for gasoline vehicle emission

control tested under cold start conditions.

5.1. Introduction.

5.2. Experimental.

5.2.1. Synthesis of zeolite supported on cordierite honeycomb

monoliths.

5.2.2. Characterization.

5.2.3. Cold Start Tests.


5.3.1. Characterization of coated monoliths.

5.3.2. Cold start test.

5.4. Conclusions.

References.

Chapter 6. Total oxidation of naphthalene using palladium nanoparticles

supported on BETA, ZSM-5, SAPO-5 and alumina powders.

6.1. Introduction.

6.2. Experimental.

6.2.1. Catalyst preparation.

6.2.2. Catalyst characterization.

6.2.3. Catalyst testing.


6.3.1. Catalysts characterization.

6.3.2. Catalytic performance for naphthalene oxidation.

6.4. Conclusions.

References.

Chapter 7. Total oxidation of naphthalene at low temperatures using palladium

nanoparticles supported on inorganic oxide-coated cordierite honeycomb

monoliths.

7.1. Introduction.

7.2. Experimental.

7.2.1. Coating of BETA, ZSM-5 and SAPO-5.

7.2.2. Coating of γ-Al2O3.


7.2.4. Characterization of the catalysts.

7.2.5. Catalytic reaction


7.3.1. Coated monoliths characterization.

7.3.2. Monolithic catalysts characterization.

7.3.3. Catalytic performance in the oxidation of naphthalene.

7.4. Conclusions.

References.

Chapter 8. Preferential oxidation of CO catalyzed by palladium nanoparticles

supported on inorganic oxides and on inorganic oxide-coated cordierite

honeycomb monoliths.

8.1. Introduction.

8.2. Experimental.

8.2.1. Catalyst preparation and characterization..

8.2.2. Catalytic reaction.


8.4. Conclusions.

References.

Chapter 9. General conclusions.

Resumen en castellano

Table Index

Table 1.1. Advantages and disadvantages of zeolite membranes compared with organic membranes.

Table 1.2. Seeding methods reported in the literature.

Tabla 2.1. Métodos atómicos de emisión.

Table 3.1. Porous texture characterization results

Table 4.1. CO2 adsorbed on the Na-LTA and Cs-LTA zeolite measured by TGA experiments.

Table 4.2. Na-LTA/carbon and Cs-LTA/carbon membranes permeance properties on dry conditions and humid conditions using a simulated reformer mixture.

Table 5.1.Weight increase (%) and porous texture results of BETA coated monoliths.

Table 5.2. Temperature for HC desorption (propene and toluene) for the three coated monoliths tested

Table 6.1. Porous texture characterization of the supports and prepared catalyst.

Table 6.2. Nanoparticles sizes obtained from TEM data and metal loading of the different catalysts.

Table 6.3. Temperature required for naphthalene total oxidation for the different catalysts for the four oxidation cycles tested and coke deposition after 4 cycles.

Table 6.4. Turnover frequency (TOF) at 125ºC and apparent activation energy for the four catalysts at first reaction cycle.

Table 6.5. CO2 conversion at 150ºC for the four catalysts and during the four cycles tested.

Table 6.6. Results of Pd nanoparticle analysis by TEM for catalysts after testing in naphthalene oxidation.

Table 6.7. XPS analysis of the four catalysts used, before and after use for naphthalene oxidation.

Table 7.1. Monolith weight increase (wt%) for single and two steps coatings.

Table 7.2. Porous texture characterization results of the coated monoliths.

Table 7.3. Porous texture characterization results of the monolithic catalysts.

Table 7.4. Temperature required for naphthalene total oxidation for the different catalysts for the three oxidation cycles tested

Table 8.1. Catalytic results in PrOx reaction. Maximum selectivity, maximum conversion and selectivity at 25% of CO conversion are specified for each sample. TOF(s-1) calculated at 150ºC is also included.

Table 8.2. Catalytic results in PrOx reaction. Maximum selectivity and maximum conversion and selectivity at 25% of CO conversion for each monolithic catalyst after the first and second cycle.

Figure Index

Figure 1.1. Scheme of a membrane acting as a separation barrier.

Figure 1.2. Scheme of a zeolite membrane (left side) and a magnification of the zeolite membrane (right side) formed by a support (black part) and the zeolite membrane (grey).

Figure 1.3. Scheme of an idealized monolith channel of a zeolite based catalyst with monolithic configuration.

Figure 1.4. Scheme of the process for H2 production by steam hydrocarbon reforming.

Figure 1.5. Processes for H2 purification for PEM fuel cell by selective membranes or PrOx-CO.

Figure 1.6. Scheme for the HC trap module added to the TWC converter.

Figura 2.1. Representación de las isotermas de fisisorción según la IUPAC [3].

Figura 2.2. Equipo volumétrico de adsorción de gases Autosorb 6.

Figura 2.3. Termobalanza TG-MS acoplada a un espectrómetro de masa.

Figura 2.4. Espectrofotómetro Perkin Elmer 4300DV.

Figura 2.5. Espectrofotómetro FTIR acoplado a un espectrómetro de masas.

Figura 2.6. Difractómetro de rayos X JSO-DEBYEFLEX 2002.

Figura 2.7. Esquema de la interacción de un haz de electrones con la muestra.

Figura 2.8. Equipo de XPS: Espectrómetro VG-microtech-Multilab 3000.

Figura 2.9. Microscopio electrónico de transmisión JEOL modelo JEM-2010.

Figura 2.10. Microscopio electrónico de barrido HITACHI S-3000N.

Figura 2.11. Esquema de una célula de Wicke-Kallenbach

Figura 2.12. Monolito original con estructura celular "honeycomb".

Figura 2.13. Imágenes de SEM correspondientes a: A) vista superior monolito, B) corte transversal del monolito.

Figura 2.14. Espectro de Difracción de rayos X de la cordierita.

Figura 2.15. Isoterma de adsorción/desorción N2 (-196 ºC).

Figura 2.16. Imagen de SEM de la superficie del disco de carbón.

Figura 2.17. Espectro de difracción de rayos X del disco de carbón.

Figura 2.18. Montaje experimental de una célula de EDP.

Figura 2.19. Esquema del sistema de rotación empleado para preparar películas delgadas de zeolita sobre monolitos de cordierita con estructura celular.

Figure 3.1. (A) X-ray diffractogram of an as-synthesized LTA/carbon membrane and (B) X-ray diffractograms of ion-exchanged zeolite A powders.

Figure 3.2. Top view and cross-sectional view (inset) of the zeolite A supported on the carbon disc.

Figure 3.3. Energy-dispersive X-ray spectrometry (EDX) analysis corresponding to ion-exchanged LTA/carbon membranes: (A) Na-LTA/carbon; (B) K-LTA/carbon; (C) Rb-LTA/carbon; (D) Cs-LTA/carbon.

Figure 3.4. (A) H2 adsorption isotherms at 298 K corresponding to Na-LTA and ion-exchanged LTA powders and (B) CO adsorption isotherm at 298 K for Na-LTA powder.

Figure 3.5. Single gas permeance: (A) H2 permeance and (B) CO permeance.

Figure 3.6. H2/CO mixtures permeance: a) H2 permeance; and b) CO permeance.

Figure 4.1. H2/CO/CO2 mixture permeance in Na-LTA/carbon membrane: (A) H2 permeance, (B) CO permeance and (C) CO2 permeance.

Figure 4.2. H2/CO/CO2 mixture permeance in Cs-LTA/carbon membrane: (A) H2 permeance and (B) CO2 permeance. Note: No CO permeation was detected at any temperature using this membrane.

Figure 4.3. Representative spectra of CO2 adsorbed on the zeolite Na-LTA; variation of the ν3 vibration mode with time: (A) spectra collected at 303 K; (B) spectra collected at 398 K; (C) spectra collected at 423 K.

Figure 4.4. Representative spectra of CO2 adsorbed on the ion exchanged Cs-LTA zeolite. Spectra collected at 303 K after treatment with CO2 at different temperatures (303 K, 398 K and 423 K) for 2 h, heat treatment up to 523 K under Helium for 1 h. Variation of the ν1 + ν3 vibration mode with temperature of CO2 treatment (A) region of 3710 cm-1 (B) region around 2350 cm-1.

Figure 4.5. CO2 adsorption isotherms at 303 K, 398 K and 423 K corresponding to: (A) Na-LTA powder zeolite and (B) ion-exchanged Cs-LTA powder zeolite.

Figure 5.1. XRD diffractograms of cordierite, coated monoliths and powder zeolite.

Figure 5.2. Nitrogen adsorption/desorption isotherms at -196ºC.

Figure 5.3. SEM images of Top view (left side) and cross sectional (right side) view for the prepared monoliths: (a) and (b) MBETA1; (c) and (d) MBETA2; (e) and (f) MBETA3.

Figure 5.4. SEM images of cross sectional of BETA zeolite monolith with an internal opening full filled with BETA zeolite.

Figure 5.5. Experimental results of the cold start tests for MBETA1, MBETA2and MBETA3 monoliths after the first and third cycle.

Figure 6.1. XRD diffractograms of the supports.

Figure 6.2. N2 adsorption/desorption isotherms (-196ºC) of the selected supports.

Figure 6.3. TEM images and particle size distribution of: (A) colloid; (B) Pd/BETA; (C) Pd/ZSM-5; (D) Pd/SAPO-5; (E) Pd/γ-Al2O3.

Figure 6.4. Variation of the catalytic activity for naphthalene oxidation (expressed as yield to CO2) as a function of reaction temperature over the four catalysts.(A) Pd/BETA; (B) Pd/ZSM-5; (C) Pd/SAPO-5; (D) Pd/γ-Al2O3.

Figure 6.5. Variation of the amount of naphthalene adsorbed as a function of temperature on the three catalysts.

Figure 6.6. TEM images and particle size distribution of catalystsusedafter four cycles of naphthalene oxidation: (A) Pd/BETA; (B) Pd/ZSM-5; (C) Pd/SAPO-5; (D) Pd/γ-Al2O3.

Figure 6.7.Conversion of naphthalene to CO2 as a function of time-on-line at 250oC for the Pd/BETA catalyst.

Figure 7.1. XRD diffractograms of cordierite, coated monoliths and powder zeolites.

Figure 7.2. Nitrogen adsorption/desorption isotherms at -196 °C.

Figure 7.3. SEM images of top view for the four prepared monoliths after the first coating step (left) and the second coating step (right)

Figure 7.4. SEM images of cross sectional for the four prepared monoliths after the second coating step.

Figure 7.5. Variation of the catalytic activity for naphthalene total oxidation (expressed as yield to CO2) as a function of reaction temperature over the four catalysts. (A) Pd/MBETA; (B) Pd/MZSM-5; (C) Pd/MSAPO-5; (D) Pd/Mγ-Al2O3

Figure 7.6. Conversion of naphthalene to CO2 as a function of time-on-line at 250ºC for the four monolithic catalysts.

Figure 8.1. Variation of the catalytic activity for CO oxidation and the selectivity towards CO oxidation as function of reaction temperature over the three powder catalysts.

Figure 8.2. Variation of the catalytic activity for CO oxidation and the selectivity towards CO oxidation as function of reaction temperature over the three monolithic catalysts for the first reaction cycle.

Scheme Index

Scheme 3.1. LTA structure of A zeolite.

Scheme 4.1. Experimental system employed to do the permeation measurements.

Scheme 4.2. FTIR experimental conditions. Full dots indicate when the spectra were collected.

Scheme 5.1. BEA structure of BETA zeolite.

Scheme 6.1. MFI and AFI structures of ZSM-5 zeolite and SAPO-5 molecular sieve.

Capítulo 0. Introducción, objetivos y

estructura de la Tesis Doctoral.

Capítulo 0

1

1. Introducción general1.

El presente trabajo de Tesis Doctoral ha sido motivado por el creciente interés de

nuestro grupo de investigación (Materiales Carbonosos y Medio Ambiente, del

Instituto Universitario de Materiales y del Departamento de Química Inorgánica), en

el desarrollo de materiales basados en películas delgadas de zeolitas para

aplicaciones emergentes, dentro de las cuales, la purificación de corrientes de gases

está teniendo cada vez más importancia debido a su repercusión medioambiental y

energética.

Esta memoria de Tesis Doctoral se centra en la preparación y caracterización de

películas delgadas de zeolitas con el fin de sintetizar membranas de zeolita, o para

conseguir recubrimientos de zeolita soportados sobre monolitos de cordierita de

estructura celular y catalizadores estructurados, que se utilicen en la purificación de

gases en diferentes aplicaciones: (i) purificación de H2 para pilas de combustible, (ii)

retención de hidrocarburos durante el arranque en frío de motores de combustión

interna y (iii) para la eliminación total de compuestos aromáticos policíclicos.

1.1. Zeolitas, membranas de zeolita y películas delgadas de zeolitas.

Las zeolitas son una familia de aluminosilicatos cristalinos con una estructura

tridimensional que presenta un sistema regular de poros bien definidos formados por

canales y cavidades de dimensiones moleculares. Además de las zeolitas clásicas,

formadas por Al y Si, también se pueden emplear otros elementos químicos como P,

B, Ga, Fe, Mn, entre otros, y a estos materiales se les conoce como zeotipos. Un

ejemplo de estos zeotipos son los silicoaluminofosfatos (SAPO). En general, las

1NOTA: En este apartado no se incluyen referencias bibliográficas, ofreciéndose tan solo un resumen acerca de la Tesis Doctoral. Se puede encontrar una revisión bibliográfica exhaustiva en el Capítulo 1.

Introducción, objetivos de la Tesis Doctoral y estructura

2

zeolitas presentan un elevado número de aplicaciones en distintas áreas como son el

intercambio iónico, la catálisis o la separación de compuestos químicos.

Aunque las zeolitas se obtienen principalmente en forma de polvo, éstas se pueden

preparar también como películas sobre una gran variedad de soportes tanto orgánicos

como inorgánicos. De esta forma, se pueden preparar películas delgadas continuas y

homogéneas tipo membrana para diferentes aplicaciones. Así, en una configuración

tipo membrana se podrían transferir las propiedades de las zeolitas (adsorción,

catálisis, reconocimiento molecular y difusión) a una estructura bidimensional con

aplicaciones en diversas áreas como reactores químicos, separaciones o análisis

químico.

Por otro lado, es necesario señalar que las zeolitas en polvo presentan tamaños de

partícula muy pequeños, lo que provoca que tanto su manejo como su separación

sean problemáticos, y que, además, cuando se emplean estos materiales en corrientes

de gases, puedan causar importantes caídas de presión. Este problema podría

solventarse con la fabricación de materiales compuestos de morfología apropiada.

Así, una opción es la preparación de películas delgadas y continuas de zeolita sobre

sistemas estructurados, como monolitos cerámicos. Estas zeolitas soportadas se

pueden aplicar de forma apropiada en procesos de separación o catalíticos en los que

se emplean corrientes de gases.


En la actualidad se está considerando la posibilidad de utilizar el hidrógeno como un

vector energético que pueda reemplazar los combustibles fósiles. El hidrógeno tiene

diversas ventajas respecto a los combustibles fósiles convencionales, siendo una de

las más importantes el hecho de que su combustión no genera contaminantes, tales

como partículas, óxidos de nitrógeno, óxidos de azufre, monóxido de carbono y

Capítulo 0

3

dióxido de carbono. Así, el hidrógeno es un combustible que se caracteriza por ser

limpio y energéticamente eficiente.

En el caso de la industria del automóvil, el hidrógeno se utilizaría como combustible

de las pilas de combustible, tales como las basadas en membranas de intercambio de

protones (pilas tipo “PEM”), que trabajan transformando la energía química en

energía eléctrica, Así, las pilas de combustible son una opción interesante para

suplementar o, incluso, sustituir a los motores convencionales. Sin embargo, uno de

los principales inconvenientes es el almacenamiento del hidrógeno y el conseguir una

alimentación de hidrógeno suficientemente puro a la pila de combustible.

En la actualidad el método más empleado para la producción de H2 es el reformado

de hidrocarburos. Así, un vehículo puede llevar en su interior un reformador de

hidrocarburos que produciría una corriente de hidrógeno. Sin embargo, en este

proceso se producen ciertos compuestos en muy baja concentración (sulfuros y

monóxido de carbono) que envenenan el electrocatalizador del ánodo. Después del

proceso de reformado se puede llevar a cabo una reacción de desplazamiento del gas

de agua (water gas shift, WGS) que reduce la concentración de CO a valores entre

1000 y 10000 ppm, por lo que se requiere de otro paso para disminuir la

concentración de CO hasta valores cercanos a 10 ppm.

Se han desarrollado varias tecnologías avanzadas para la purificación de H2, siendo

las más estudiadas la purificación mediante membranas, o la oxidación selectiva de

CO. En cuanto a la purificación de H2 mediante membranas selectivas de zeolita,

éstas son una alternativa prometedora debido a sus propiedades excepcionales. La

otra opción interesante para reducir la concentración de CO a los límites deseados es

la oxidación selectiva de CO (PrOx), que tiene la finalidad de oxidar selectivamente

el CO de la corriente de gases (generando CO2), intentando minimizar en la medida

de lo posible la oxidación paralela de H2, mediante el uso del catalizador adecuado.


4

1.3. Reducción de las emisiones de hidrocarburos durante el

arranque en frío en vehículos.

Durante el arranque en frío de un motor de gasolina se producen alrededor del 80 %

de las emisiones de hidrocarburos (HC). Estas emisiones ocurren como consecuencia

de que el catalizador de tres vías (CTV) no es activo hasta que alcanza temperaturas

de trabajo de 200ºC o superiores. Así, es necesario evitar dichas emisiones de HC, y

una posible solución es el uso de una trampa de HC. Una trampa de HC deber tener

buenas propiedades como adsorbente, ser capaz de retener los HC adsorbidos hasta

temperaturas superiores a 200ºC, ser resistente térmicamente (temperaturas

superiores a 750 ºC) y debe presentar reversibilidad en el proceso de adsorción-

desorción de HC. De esta forma, debido a las características tan exigentes del

proceso, las zeolitas se presentan como un candidato interesante para esta aplicación.

1.4. Reducción de las emisiones de compuestos aromáticos

policíclicos (PAH).

Uno de los tipos de compuestos orgánicos volátiles (VOC) son los hidrocarburos

aromáticos policíclicos (PAH) y se emiten principalmente durante la combustión de

la materia orgánica. Hoy en día, existe una gran variedad de fuentes de emisión de

PAH como la combustión incompleta de hidrocarburos, diesel o gasolina, motores de

combustión interna, plantas de transformación de asfalto o de producción de energía

a partir de carbón mineral. La necesidad de erradicar las emisiones de PAH se debe a

que ha sido identificado como un grupo de compuestos peligrosos para la salud y el

medio ambiente. De todos ello, se suele escoger el naftaleno como molécula modelo

de este grupo de contaminantes.

En los últimos años, se ha incrementado el desarrollo de nuevas tecnologías para la

reducción de PAH. A pesar de la existencia de numerosas técnicas, la oxidación

Capítulo 0

5

catalítica para la producción de CO2 y H2O es la más prometedora debido a que

permite trabajar a temperaturas moderadas y posee una elevada selectividad a la

producción de CO2. El empleo de catalizadores estructurados basados en zeolitas y

manopartículas metálicas son una opción interesante para esta aplicación.

2. Objetivos de la tesis doctoral.

El objetivo principal de la presente memoria de investigación es preparar zeolitas y

películas delgadas de zeolita para la purificación de corrientes de gases.

En cuanto la purificación de H2, los objetivos son:

• Preparar membranas de Na-LTA soportadas sobre discos porosos de grafito,

modificar su porosidad mediante intercambio iónico con cationes alcalinos

(K, Rb o Cs) y estudiar su comportamiento de separación de las mezclas H2 y

CO y de mezclas que simulen los gases emitidos por un reformador.

• Preparar catalizadores estructurados basados en nanopartículas de Pd

soportadas en películas delgadas de zeolita crecidas en monolitos de

cordierita. Estudiar la oxidación la oxidación selectiva de CO en corrientes

ricas en H2.

En referencia a la eliminación de las emisiones de hidrocarburos, los objetivos

específicos han sido:

• Preparar y caracterizar películas delgadas de zeolita BETA crecidas en

monolitos de cordierita de estructura celular.

• Estudiar sus propiedades de eliminación de hidrocarburos en condiciones que

simulan el arranque en frio de un motor de combustión interna.


6

En relación a la eliminación de PAHs, los objetivos concretos son:

• Preparar catalizadores en polvo basados en nanopartículas de Pd soportadas

sobre las zeolitas BETA y ZSM-5, el tamiz molecular SAPO-5 y γ-alúmina

para estudiar la oxidación selectiva de naftaleno.

• Preparar y caracterizar catalizadores estructurados basados en nanopartículas

de Pd depositadas en películas delgadas de zeolita BETA, ZSM-5, un tamiz

molecular SAPO-5 mediante síntesis in-situ y γ-alúmina, mediante el método

de inmersión, y estudiar la oxidación total de naftaleno.

3. Estructura de la tesis doctoral.

La presente memoria de Tesis Doctoral está dividida en 9 Capítulos cuyo contenido

es el siguiente: el Capítulo 1 incluye una extensa revisión bibliográfica de los temas a

tratar durante la presente Tesis Doctoral. El Capítulo 2 muestra las técnicas

instrumentales utilizadas, los materiales empleados como soporte y además se presta

atención a técnicas de preparación relacionadas con la presente Tesis Doctoral.

Los Capítulos 3 y 4 presentan los resultados relacionados con la síntesis y

caracterización de membranas de zeolita soportadas sobre materiales carbonosos para

su uso en la purificación de hidrógeno. El Capítulo 5 incluye la preparación y

caracterización de la zeolita BETA soportada sobre monolitos de cordierita para su

aplicación como trampa adsorbente para la reducción de las emisiones de

hidrocarburos en condiciones de arranque en frío en vehículos. En los Capítulos 6 y 7

se presenta la preparación de catalizadores en polvo y estructurados basados en

zeolitas, tamices moleculares y alúmina como soporte de nanopartículas de Pd para

su aplicación en la eliminación de los compuestos orgánicos volátiles. En el Capítulo

8 se muestran los resultados del empleo de los catalizadores estructurados descritos

en los capítulos anteriores para la purificación de hidrógeno mediante oxidación

Capítulo 0

7

selectiva de CO (PrOx-CO). Finalmente, en el Capítulo 9 se recopilan las

conclusiones finales de esta Tesis Doctoral.

La presente memora de Tesis Doctoral opta a la mención de “Doctor Internacional”,

por lo que ha sido redactada en una Lengua Oficial de la Unión Europea (Inglés).

Con el fin de cumplir con la normativa, al final de la memoria de Tesis Doctoral se

incluye un resumen que contiene una introducción general, los objetivos y un

resumen de los resultados obtenidos destacando las conclusiones más relevantes, en

una de las dos lenguas oficiales de la Comunidad Valenciana, en este caso en

castellano.

Esta investigación ha sido realizada en el grupo Materiales Carbonosos y Medio

Ambiente de la Universidad de Alicante y se ha complementado con una estancia en

un centro de investigación extranjero, en concreto en la Universidad de Cardiff

(Reino Unido).

A continuación se resume brevemente el contenido de cada uno de los capítulos:

Capítulo 1. Introducción general.

En este capítulo se presenta una revisión bibliográfica sobre los temas abarcados en

esta Tesis Doctoral: preparación y caracterización de películas delgadas de zeolita

para purificación de gases. En primer lugar, se hace una revisión acerca de las

zeolitas, membranas de zeolitas y zeolitas estructuradas. A continuación, se ha

descrito el estado actual de los temas de interés medioambiental en los que se ha

centrado esta Tesis Doctoral: (i) purificación de H2 para pilas de combustible; (ii)

retención de hidrocarburos durante el arranque en frio de motores de combustión

interna; y (iii) eliminación total de compuestos aromáticos policíclicos. Esta revisión

facilita la lectura y comprensión de los capítulos posteriores.


8

Capítulo 2. Técnicas de caracterización, materiales y métodos de preparación

Durante el desarrollo del presente trabajo se ha utilizado una amplia gama de

técnicas experimentales destinadas a la identificación y caracterización de los

materiales sintetizados. En este capítulo se describen las características de las

técnicas empleadas. Además se incluye una caracterización de los materiales

empleados como soportes para la preparación de películas delgadas, así como los

métodos de preparación empleados.

Capítulo 3. Purificación de hidrógeno para pilas de combustible tipo PEM

mediante membranas preparadas por intercambio iónico de membranas de Na-

LTA/Carbón.

En este capítulo se presenta la purificación de H2 mediante membranas de zeolita. Se

ha realizado un estudio de la preparación de membranas de zeolita A (referida como

Na-TLA) soportada sobre discos porosos de grafito. Además se ha modificado el

tamaño de la porosidad para la obtención de una porosidad más estrecha. Para ello,

estas membranas de zeolita, se han modificado mediante un método simple y

reproducible como es el intercambio iónico con sales metales alcalinos (K, Rb y Cs).

Se han estudiado las propiedades de permeación de las membranas descritas para

mezclas binarias de hidrógeno y monóxido de carbono.

Este capítulo ha sido publicado en la revista Journal of Membrane Science, siendo la

referencia:

• Francisco J. Varela-Gandía, Angel Berenguer-Murcia, Dolores Lozano-

Castelló, Diego Cazorla-Amorós, Journal of Membrane Science 351 (2010)

123-130.

Además, se ha publicado el método de preparación en el libro:

Capítulo 0

9

• F.J. Varela-Gandía, A. Berenguer-Murcia, A. Linares-Solano, E. Morallón,

D. Cazorla Amorós, Electrophoretic Deposition for The Synthesis of

Inorganic Membranes, 381-393, in Membranes for Membrane Reactors.

Preparation, Optimization and Selection, A. Basile and F. Gallucci. Ed. John

Wiley & Sons (2011).

Capítulo 4: Purificación de H2 mediante membranas de zeolita A/carbón en

mezclas que simulan las emisiones de un reformador.

En este capítulo se analizan las propiedades de la membrana original de Na-LTA/C y

de la mejor membrana obtenida en el estudio anterior, Cs-LTA/C, en mezclas que

simulan las gases emitidos por un reformador de hidrocarburos. Se ha llevado a cabo

un estudio detallado de cómo afecta la concentración de dióxido de carbono y la

presencia de agua en las propiedades finales de permeación de las membranas

descritas.

Este capítulo ha sido publicado en la revista Journal of Membrane Science, siendo la

referencia:

• Francisco J. Varela-Gandía, Ángel Berenguer-Murcia, Dolores Lozano-

Castelló, Diego Cazorla-Amorós, Journal of Membrane Science 378 (2011)

407-414

Capítulo 5: Preparación y caracterización de zeolita BETA soportada en

monolitos de cordierita de estructura celular para la retención de hidrocarburos

en condiciones de arranque en frio.

En el presente capítulo se ha optimizado la síntesis para la preparación de zeolita

BETA soportada sobre monolitos de cordierita de estructura celular mediante el

método de síntesis in-situ, para su uso como trampa de hidrocarburos en condiciones

de arranque en frío de vehículos. Los hidrocarburos que componen estas emisiones


10

se dividen en hidrocarburos ligeros (compuesto modelo propeno) e hidrocarburos

pesados (compuesto modelo tolueno). La tendencia observada es que mientras que

los hidrocarburos pesados se encuentran atrapados adecuadamente, los hidrocarburos

ligeros se emiten a temperaturas bajas. Por tanto, se necesita optimizar una trampa de

hidrocarburos que retenga los hidrocarburos ligeros hasta temperaturas de al menos

200ºC.

El estudio preliminar realizado con el que se ha concluido que la zeolita BETA es un

material óptimo para la preparación de trampas de hidrocarburos, ha sido publicado

en la revista Microporous and Mesoporous Materials, siendo la referencia:

• J.M. López, M.V. Navarro, T. García, R. Murillo, A.M. Mastral, F.J. Varela-

Gandía, D. Lozano-Castelló, A. Bueno-López, D. Cazorla-Amorós,

Microporous and Mesoporous Materials 130 (2010) 239-247.

Capítulo 6: Oxidación total de naftaleno mediante nanopartículas de paladio

soportadas en BETA, ZSM-5, SAPO-5 y alúmina en polvo.

En el presente capítulo, se han preparado diferentes catalizadores basados en

nanopartículas de Pd soportadas (1% peso carga nominal) en las zeolitas BETA,

ZSM-5, el tamiz molecular SAPO-5 (todos ellos preparados mediante síntesis

hidrotérmica) y γ-alúmina para su empleo en la oxidación catalítica de naftaleno.

Los resultados correspondientes a este capítulo, se han enviado a publicar a la revista

Applied Catalysis B, siendo la referencia:

• Francisco J Varela-Gandía, Ángel Berenguer-Murcia, Dolores Lozano-

Castelló, Diego Cazorla-Amorós, David R Sellick; Stuart H. Taylor, Applied

Catalysis B: Environmental, enviado para publicación.

Capítulo 0

11

Capítulo 7: Oxidación total de naftaleno a bajas temperaturas mediante

nanopartículas de paladio soportadas en monolitos de cordierita de estructura

celular recubiertos de óxidos inorgánicos.

En este capítulo se ha estudiado la preparación de los catalizadores monolíticos

recubiertos con películas de distintos sólidos cristalinos (BETA, ZSM-5, SAPO-5 y

γ-Al2O3). En primer lugar, se ha realizado un estudio detallado para la preparación de

películas delgadas de los sólidos anteriores con el fin de cubrir la superficie del

monolito. Posteriormente, se han preparado los catalizadores monolíticos, empleando

como fase activa nanopartículas de Pd. Todos estos catalizadores monolíticos se han

empleado en la oxidación total de naftaleno.

Capítulo 8: Oxidación selectiva de CO catalizada por nanopartículas de paladio

soportadas en óxidos inorgánicos en polvo y óxidos inorgánicos soportados

sobre monolitos de cordierita de estructura celular.

En este último capítulo se ha llevado a cabo el estudio de la purificación de H2

mediante la oxidación selectiva de CO. El estudio se ha llevado a cabo empleando

los catalizadores en polvo y monolíticos descritos en los capítulos 7 y 8.

Capítulo 9: Conclusiones generales.

En este capítulo, se recogen las conclusiones generales más importantes que se

derivan del presente trabajo de Tesis Doctoral.

Chapter 1. Introduction and objectives.

Introduction

13

1.1. Zeolites, zeolite membranes and zeolites films.

Zeolites are an important family of crystalline aluminosilicates with a three

dimensional network containing a well defined regular pore system formed by

channels and cavities of molecular dimensions [1,2]. Crystalline structures of the

zeolite are formed by TO4 tetrahedra (T = Si, Al) linked to each other by sharing all

the oxygens [2]. Apart from the classical zeolites, transition metals and many other

elements such as P, B, Ga, Fe, Cr, Ti, V, Mn, Co, Zn, Be, Cu, etc can be also used

for its synthesis, and they are denoted as zeotypes including in this family, among

others, AlPO4, SAPO (silicoaluminophosphate) and MeAPO (metal

aluminophosphate) molecular sieves [3-6].

The most interesting features of zeolites lie in the variable chemical compositions, as

well as the tuneable pore diameter and pore geometries. Zeolites can be classified as

having small, medium, large and extra-large pore structures with pore windows

delimited by 8, 10, 12, and more than 12 T-atoms, respectively [7]. Zeolites are

suitable for a wide range of applications in several fields such as, ion-exchange,

catalysis and separation [1,2,8-11].

Although most zeolite users still think of zeolites as powdered materials, in fact

zeolites can be grown as films on a variety of supports, both inorganic (e.g. glass,

ceramic, metal, porous graphite) and organic (e.g. plastics, cellulose, wood). This

allows to transfer some of the characteristic properties of zeolites, such as adsorption,

catalysis, molecular recognition and diffusion, to a 2-dimensional structure with

potential applications in areas ranging from reactor engineering to molecular

separations and chemical analysis [12]. Therefore, the possibility of having zeolite

properties in a membrane configuration is an attractive option, with many potential

applications.

Chapter 1

14

In general, a membrane is an intervening phase separating two phases and/or acting

as an active or passive barrier to the transport of matter between phases adjacent to it

under a driving force [13]. Figure 1.1 shows a scheme of membrane acting as a

separation barrier. There are several kinds of membranes which can be organized

into three categories [14]: (i) polymeric, (ii) metallic and (iii) inorganic membranes

like zeolite membranes. Polymer membranes have several advantages like having a

low cost and not causing significant pressure drops. However, mechanical strength

problems and high sensitivity to swelling and compacting reduce their usefulness for

this purpose [15]. The second type, metallic membranes, has an excellent hydrogen

permeance but suffer from hydrogen embrittlement at low temperatures [16]. This is

eliminated by using alloys but the product is more expensive. The latter, zeolite

membranes, combine the general advantages of inorganic membranes like

temperature stability and solvent resistance with those of polymeric membranes as

they are composed of a thin homogeneous layer.

Figure 1.1. Scheme of a membrane acting as a separation barrier.

A great progress in the science of zeolite membrane synthesis has been made during

the last 25 years but, although more than 201 zeolitic framework type codes have

been indexed by the International Zeolite Association (IZA), only about 20 structure

types have been prepared as membranes. Typical examples include SOD, CHA,

LTA, DDR, FER, MEL, MFI, ERI-OFF, MOR, FAU and BEA types [17]. The

remarkable progress made in the development of zeolite membranes can be

Introduction

15

summarized in some milestones. The first membrane based on zeolites was coined by

Kulprathinja [18] when a mixed matrix membrane was prepared. However, the

concept and the preparation of a continuous layer of intergrown zeolite crystals was

introduced by Suzuki in 1987 [19] opening the way for the preparation of new types

of zeolite membrane. Toda et al. [20] prepared a zeolite A membrane with separation

properties in 1989 and Matsukata et al. [21] reported the first zeolite membrane

reactor. Although, one of the greatest advances in the synthesis of zeolite thin films

was performed by Kita et al. [22], who established the secondary growth method to

prepare zeolite A membranes. Focusing the attention on the industrial

implementation of zeolite membranes, the first large-scale pervaporation plant for

alcohol dehydration and solvent dewatering was put into industrial operation at the

end of the 1990s by Mitsui Engineering and Shipbuilding Co. Ltd using Na-A zeolite

membranes tubes [17].

Table 1.1. Advantages and disadvantages of zeolite membranes compared with organic membranes. Advantages Disadvantages Long-term stability at high temperatures High capitals cost Resistance to harsh environments Brittleness Resistance to high pressure drops Low membrane surface per

module volume Inertness to microbiological degradation Difficulty in industrial scale-up Easy cleaning after fouling Low permeability of the high

selective (dense) membranes Easy catalytic activation Module-to-membrane sealing

problems at high temperature

In general, a zeolite membrane combines the common advantages of inorganic

membranes with a perfect shape selectivity. Table 1.1 summarizes the advantages

and disadvantages of zeolite membranes compared with organic membranes. Zeolite

membranes show specific advantages, such as stability at high temperatures and

against solvent. These promising advantages give the driving force to solve some

Chapter 1

16

open problems related to the industrial implementation of zeolite membranes [23]

such as high production costs and unsolved problems of module development.

The synthesis of a zeolite membrane requires the development of a continuous,

defect-free, nearly bidimensional layer of zeolite crystals, so that only transport

through the zeolite pores takes place. The usual procedure for developing zeolite

membranes consists in depositing zeolite crystals onto a previously existing support,

which confers the necessary mechanical strength and allows the development of a

more extensive structure [24]. Figure 1.2 shows a scheme of a zeolite membrane (left

side) and a magnification of the zeolite membrane (right side) formed by a support

(black part) and the zeolite membrane (grey).

Figure 1.2. Scheme of a zeolite membrane (left side) and a magnification of the zeolite membrane (right side) formed by a support (black part) and the zeolite

membrane (grey).

Concerning the zeolite membrane synthesis, “direct synthesis” on a given support

was the first methodology. Two main steps are involved in this protocol: (i) a zeolite

synthesis solution is prepared and the support is immersed in it, (ii) the mixture is

submitted to hydrothermal treatment under the zeolite synthesis conditions [25,26].

Several reports have dealt with this topic, despite its disadvantages such as the

necessity of long synthesis times and several synthesis steps which in turn lead to

thick zeolite layers [27-28]. Recently, a new procedure called “seeded (or secondary)

growth” [2,30-34], has been developed for the preparation of zeolite membranes. The

main advantages are: (i) the control of thin oriented layers [27]; (ii) improved zeolite

Introduction

17

growth on different supports [30,32]; and (iii) the need for only one hydrothermal

synthesis step.

On the whole, zeolite synthesis can be separated in two main steps: nucleation and

crystal growth. The former is the rate-limiting step in the synthesis [29] and it

determines the subsequent crystal growth. Bearing in mind the two aforementioned

methods (direct and seeded growth) the former needs long synthesis times and there

is a limited intergrowth between the crystals which form the membrane. The latter

‘skips’ the nucleation step and then a thin zeolite film can be obtained with

comparatively higher quality crystals and better intergrowth with reduced synthesis

times. Therefore, decoupling film deposition from crystal growth provided a large

flexibility for tailoring the film microstructure and ease scale up [35]. Figure 1.2

(right side) correspond to the schematic representation of a zeolite membrane

prepared by secondary growth method in which seeded crystals (small grey dots)

have been deposited onto the surface of the support and the zeolite membrane (grey

squares) has been grown later. Concerning the seeding procedure, different

techniques have been developed in the last years. Some of the methods are

summarized in Table 1.2. The typical supports for the preparation of zeolite

membranes are porous inorganic tubes or plates made of alumina [36,37], stainless

steel [38] or based on carbon materials [32].

Table 1.2. Seeding methods reported in the literature. Seeding method Reference Metal Surface Modification with organic polymer [33, 39] Dip-coating [2, 31, 35,39-42] Rubbing [43, 44] Suction [11, 45] Ultrasound [46] Spin-coating [47] Electrophoretic Deposition (EPD) [32,38,48]

Chapter 1

18

Another type of zeolite-based materials consists of zeolite films supported on

structured materials. As reported by Pina et al. [12], the experience in supported

zeolites gained in recent years allows the production of zeolite films on almost any

type of support, with unprecedented control on their morphology, orientation and

degree of crystal intergrowth. Thus, characteristics of zeolite films can be very

different depending on the intended application, and the preparation procedures have

to be tailored accordingly.

Different supports have been used for the preparation of structured materials:

ceramic monoliths (cordierite monoliths [49-51] or pure silica [52]), metal wire

gauze packings [52], stainless steel grids [53], sintered metal fibers [54], stainless

steel microchannel reactor [55], Mo-based plates [56], Si wafers [57], quartz

microfibers [58] or carbon materials (cloths, fibers, etc) [59].

From an engineering point of view, the main reason for using a layer of a supported

zeolite instead of powder materials is because the direct use of materials with small

particle size, like zeolites, causes some problems like high pressure drops and makes

difficult the separation and handling operations. To solve these problems zeolite

crystals are usually mixed with a binder to prepare pellets. However, the use of a

binder may increase diffusional problems and reduce the accessible porosity and the

catalytic activity of the zeolitic phase [59]. Therefore, low pressure drops, uniform

flow distribution, avoidance of hot spots and good tolerance to plugging by dust are

essential requisites that lead the use of monolith configuration [60].

In the application of a zeolite supported on monoliths, one should determine which

are the requirements for the support. In this sense, ceramic honeycomb monoliths are

standard material supports for gas phase applications. In reference to ceramic

monoliths, the support structure is made of a non-catalytic, thermally resistant

material, typically cordierite (2MgO·2Al2O3·5SiO2), onto which a catalytic layer is

deposited [50]. A cordierite monolith has excellent properties such as low pressure

Introduction

19

drops in the exhaust system, good thermal resistance, refractoriness, good washcoat

adherence and compatibility with wahscoat and catalyst [49,61].

Concerning the deposition method, two scenarios are proposed for coating a

cordierite monolith with a zeolite material. The first one in the so-called dip-coating

or wash-coating, which consist in the deposition from a slurry of zeolite particles

followed by a stabilizing thermal treatment [60]. The second one is the use of

hydrothermal synthesis methods like (i) direct hydrothermal synthesis (also called in-

situ growth) or (ii) secondary growth which is the hydrothermal synthesis after

seeding the surface of the substrate [51].

Related to the applications, zeolitic monoliths can be applied to adsorption or

catalytic processes. In terms of adsorption applications, these monoliths have been

found to be very interesting as adsorbents of volatile organic compounds [62,63]. As

catalytic systems, metal ion-exchanged supported zeolites have performed efficiently

on gas phase reaction systems, such as NOx abatement in diesel exhaust [49]. Figure

1.3 shows a scheme of a reaction in an idealized monolith channel.

Figure 1.3. Scheme of an idealized monolith channel of a zeolite based catalyst with

monolithic configuration.

MONOLITH WALL

MONOLITH WALL

ZEOLITE CATALYST

ZEOLITE CATALYST

HC+CO+O2 CO2+H2O

Chapter 1

20

1.2. Hydrogen purification for PEM fuel cells.

The gases generated by the combustion of petroleum and coal have a strong

contribution to the greenhouse effect and constitute a serious environmental issue

[64]. This environmental aspect and the progressive depletion of these reserves have

focused the attention on hydrogen as a new and clean energy vector. Therefore, its

use as a suitable energy carrier for replacing gasoline and other fossil fuels has been

widely discussed as a way to implement hydrogen economy [65]. More specifically,

its application in portable power systems has generated substantial interest in the last

years, with applications ranging from the automotive industry to portable electronics

such as laptops or mobile telephones [66].

Focusing our attention on mobile applications, conventional power sources could be

replaced by fuel cells operating with hydrogen. In this sense, polymer electrolyte

membrane (PEM) fuel cells can be an alternative since these devices can transform

chemical energy into electrochemical energy, avoiding the mechanical requirements

and thermodynamic limitations of such devices [67]. Thus, hydrogen fuel cells are an

attractive candidate to replace conventional devices. One of the main issues in using

hydrogen in PEM fuel cell is that it is difficult to find a way to produce pure enough

hydrogen in a portable device [68]. Up to now, hydrocarbon reforming is the most

prominent industrial process to produce hydrogen [67]. The advantages of steam

reforming of hydrocarbons like methanol or ethanol are a very low process

temperature, a good H2/CO ratio (≈3/1) for hydrogen production, and that oxygen is

not required [69]. However, the CO concentration exiting the reformer (even after

desulphuration and water-gas shift (WGS) reactions) is between 1000 and 10000

ppm. This gas poisons the Pt electrocatalyst in the anode, thereby becoming a

substantial obstacle toward fuel cell implementation. Therefore, concentrations lower

than 10 ppm become mandatory in order to avoid fuel cell degradation [67]. Thus, a

previous purification step of the gaseous stream coming out from the reformer is

Introduction

21

required. Figure 1.4 shows a simplified description for generating hydrogen from

hydrocarbons or alcohols.

Figure 1.4. Scheme of the process for H2 production by steam hydrocarbon

reforming.

It is well known that hydrogen purification has been identified as one of the

bottlenecks in the development of advanced hydrogen technologies. Consequently, it

must be highlighted that there are various alternatives to achieve the purification of

hydrogen, being the two mostly studied the separation by membranes and the

preferential oxidation of CO by a suitable catalyst [67].

One alternative to purify hydrogen is the use of hydrogen selective membranes due

to their easy preparation, low energy consumption and cost effectiveness at low gas

volumes [70]. Among them, zeolite membranes are an interesting alternative for the

possibility of being used as separators and catalytic membrane reactors [71]. More

details about this issued were described in section 1.1.

On the other hand, the catalyzed preferential oxidation of CO (PrOx-CO) is one of

the most efficient technologies brought forth to reduce the CO levels to the desired

values [67]. PrOx-CO is an exothermic catalytic process which uses molecular O2 as

oxidant. However, during the PrOx-CO process, the H2 present in the reformate

should not be consumed at all to ensure reasonable fuel processor efficiency. The

involved reactions are:

(1) 2 CO + O2 2 CO2

(2) 2 H2 + O2 2 H2O

Desulfuration Steam Reforming

Water-gas-shift(400 + 200ºC)

Hydrocarbons Alcohols

H2, CO2, H2O

CO (0.1-1%) CH4 + H2O CO +3H2

T > 400ºCCO + H2O CO2 +H2

CO (10%) H2O

Chapter 1

22

Therefore, both active and selective catalysis, taking into account the very low CO

concentration compared to the H2 one (from 1/30 to 1/60), are mandatory to convert

CO to CO2 with minimal involvement of reaction (2) [73].

Different catalysts have been analysed (e.g., Pt, Au, Pd, Ru or Cu), being noble

metals those most intensively studied [67,74-77]. It must be noted that the

preparation of catalysts in the majority of the cases found in the literature, is based in

the conventional methods of heterogeneous catalysis, such as impregnation or ion

exchange. However, a new class of catalysts have been developed over the past few

decades, which employ metal nanoparticles synthesized by colloidal methods. The

advantages of nanoparticles include the fact that can be synthesised with tuneable

particle size, shapes and compositions, which might affect on catalytic activity and

selectivity [78]. In this sense, Pt [78], Rh [79] or Pd [80] nanoparticles have been

investigated for CO oxidation. Furthermore, according to Kotokubi et al. [81] zeolite

supported catalysts are expected to promote CO oxidation selectively due to the

physicochemical “molecular sieve effect”.

In summary, Figure 1.5 includes the two possibilities mentioned above for H2

purification before reaching the PEM fuel cell: (i) H2 selective membranes; or (ii)

preferential oxidation of CO. The present PhD Thesis has focused in the design and

synthesis of zeolite thin films for H2 purification, using these two approaches.

Figure 1.5. Processes for H2 purification for PEM fuel cell by selective membranes

or PrOx-CO.

H2H2

H2

H2H2

H2

CO 10000ppmH2

H2

H2H2

H2

H2H2

H2

H2

H2

H2H2

H2

H2H2

H2

CO 10000ppmH2

H2

H2 flow after reforming and WGS

Purified H2 flow[CO]< 10ppm

PEM Fuel Cell

H2

H2

H2

H+

H+

e-H+

H+

e-

e-

O2

H2O

Anode CathodePEM

Electron flow

Introduction

23

1.3. Removal of hydrocarbons under cold start conditions in vehicles

emissions.

A significant effort has been paid to develop a method to treat automobile exhausts

for the regulated emissions on CO, hydrocarbons (HC) and NOx [82,83]. Thus, in

order to protect global environment, the efficiency of the automobile fuel

consumption must be improved and the exhaust must be purified as much as

possible. One of the concerns in the automotive emission control is the HC emitted

from gasoline automobiles during the cold start [84]. Vehicles are equipped with a

conventional three way catalytic (TWC) converter which consists of precious metals

supported on washcoat layer deposited on a monolithic carrier [83,85]. Normally,

TWC for gasoline emission control usually start working at 170ºC in the case of

fresh catalyst and around 200-225ºC in the aged catalyst. Cold start is the period of

time (1-2 min) required by the TWC to reach the working temperature; during this

time, up to 80% of HC are released in a drive cycle. The whole spectrum of HCs

emitted upstream of a TWC for a gasoline engine, which is included elsewhere [83],

can be divided mainly in two groups: the heavier exhaust HCs (e.g. aromatics),

where toluene is the most abundant (mass emitted during start-up 130.1 mg), and the

light HC components of the exhaust stream, where ethene, propane, and propene are

the most abundant and the ones usually considered as representative of this group

(during start-up mass emitted 117.4, 100, 72.8 mg, respectively). The emissions

control during this ‘‘cold start” period is essential to reduce the environmental

impact of gasoline engines.

Consequently, several potential solutions to the cold start problem have been

proposed. These can tentatively be divided into two groups. The first one is based

upon methods of quickly bringing the catalyst to working temperature such as (i)

placing TWC closer to the engine, (ii) electric heating at the start up, (iii) exhaust-gas

ignition, which is the running of the engine under very rich conditions air-to-fuel,

Chapter 1

24

and (iv) combustion of heated catalyst, which consists on heating the TWC by the

exothermic combustion of hydrogen and oxygen prior to starting the engine. The

second group of cold-start solutions involves trapping HC during the cold start for

their release after the TWC has reached operating temperature [86]. In the last one,

the HC trap adsorbs HC during the start-up, when the gas is cold. As the exhaust

heats up, the trap temperature increases, causing the hydrocarbon, to be desorbed and

then, converted by the TWC. Figure 1.6 shows a scheme of the TWC position in an

automobile and the development of a HC trap module added to the TWC converter.

Figure 1.6. Scheme for the HC trap module added to the TWC converter.

Although several techniques have been disclosed for HC traps, each has been

reported to have its own strong and weak points [86,87]. Among them, to achieve

this goal, the use of HC adsorbents, ‘‘hydrocarbon traps”, before the TWC seems to

be the most relevant from a scientific-technological point of view. The critical factors

for any emission trap are: (i) high adsorption capacity of hydrocarbons at low

temperatures, (ii) desorption starting at temperatures higher than 200 ºC, (iii) a

reversible adsorption process and (iv) solid material resistant at temperatures higher

than 750 ºC. The maximum values of these factors are limited by properties of the

solids like pore volume, pore structure, chemical nature of the solid, extra-framework

atoms, etc. Zeolites have been found to be the preferred adsorbents for these

Three ways catalyst

Adsorbent

Introduction

25

applications, mainly due to their stability under severe process conditions. The

general finding in different hydrocarbon traps is that while the heavier exhaust HCs

(e.g. aromatics, such as toluene) are adequately trapped by the zeolites in many of the

proposed systems (even in the presence of 10% steam), the light HC components of

the exhaust stream, such as propene, often desorb from the HC trap before the

catalyst has reached its light-off temperature [88-90].

Considering that for automotive applications, ceramic monoliths are often the

material of choice due to their excellent properties mentioned in Section 1.1 and the

fact that the zeolites are the preferred material as cold-start trap, this PhD Thesis has

focused on the preparation of cordierite honeycomb monoliths coated with zeolites.

1.4. Removal of polycyclic aromatic hydrocarbons (PAH) by

catalytic oxidation.

PAHs are a family of volatile organic compounds (VOCs) mainly produced during

combustion processes of organic matter [91,92]. PAH emissions originate from a

great variety of sources such as incomplete fuel combustion, diesel or gasoline,

internal combustion engines, asphalt transformation plants or coal and wood powered

plants [93,94]. According to the Environmental Protection Agency (EPA), a list

containing 16 PAHs has been established as major pollutants whose emission must

be controlled [95]. PAH compounds are mainly formed by several aromatic rings

such as naphthalene (two-rings PAH), phenanthrene (three-rings PAH), pyrene or

chrysene (four-rings PAH) or benzo(g,h,i)perylene (five-rings PAH). PAH

concentration composition differs depending on the emission source. Thus, PAHs

found in diesel emissions mainly consist of two to five aromatic rings such as

naphthalene, phenanthrene and pyrene/fluoranthene [96,97]. The main reason to

reduce PAH emissions is that they have now been identified as a serious

environmental and health risk, due to their carcinogenic and/or mutagenic properties

Chapter 1

26

[98,99]. Naphthalene is the simplest and least toxic of the PAHs and is thus used as a

model compound for this group of pollutants.

As previously mentioned, the development of technologies to reduce atmospheric

pollution has increased in the last years. Methods such as biodegradation, high-

energy electron beam irradiation, ozonization, adsorption, thermal incineration or

catalytic oxidation has been employed to reduce the level of PAHs [92,100-105].

Despite the numerous options, catalytic oxidation to produce carbon dioxide and

water is the most promising technology due to its lower operating temperature and/or

higher selectivity to CO2 [94,106].

Several studies have been performed to investigate the catalytic combustion of PAHs

using different catalyst, such as noble metal supported on γ-alumina [96], metal

oxide based on Co, Mn, Cu, Ce or Ti [107], mesoporous cerium oxides [94] or metal

exchanged zeolites [108]. Despite all these previous studies, supported

nanoestructured metal nanoparticles have not been studied yet for this reaction. In

general, nanostructured metal nanoparticles are isolable particles of sizes between 1

and 50 nm which, in order to prevent their agglomeration, have to be stabilized by

ligand molecules or a whole plethora of “protecting shells”. The resulting metal

colloids can be redispersed in water (‘‘hydrosols’’) or organic solvents

(“organosols”) [109]. During the last ten years, significant efforts have been

dedicated to the preparation of nanoparticles. Their preparation, structure,

characterization and applications are issues of great current interest [110], with

particular attention paid to the importance of their size and structure. Focusing on Pd

nanoparticles, several authors [111-113] have dealt with the fundamental aspects of

monometallic palladium nanoparticles, their synthesis and structural control, and the

system has played an important role as a catalyst in numerous organic reactions.

Therefore in this PhD Thesis the properties of zeolites and Pd nanoparticles have

been combined in order to achieve interesting results in the abatement of PAHs.

Introduction

27

PhD Thesis objectives. The overall objective of the present work is the preparation of zeolite thin films in

order to obtain zeolite membranes and supported zeolites on honeycomb monoliths,

for three different applications: (i) hydrogen purification for PEM fuel cells; (ii)

removal of hydrocarbons emissions under cold start conditions in vehicles; and (iii)

catalytic oxidation of polycyclic aromatic hydrocarbons.

Regarding hydrogen purification, the specific objectives are:

• To study the synthesis of Na-LTA membranes supported on a carbon material

by hydrothermal treatment following the secondary growth method, as well

as to prepare LTA membranes by means of ion-exchange. To test the gas

permeation properties of the prepared composites using small molecules (H2

and CO) at different temperatures and try to understand their usefulness for

hydrogen purification for PEM fuel cells.

• To use the Na-LTA and ion-exchange membranes under simulated reformer

gas mixtures to understand the separation properties of these zeolite

membranes.

• To study hydrogen purification by means of preferential oxidation of CO in a

H2-rich feed by metal nanoparticles supported on zeolite coated monoliths.

Regarding removal of HC emissions under cold start conditions, the specific

objectives are:

• To choose among different zeolites and a silicoaluminophosphate the most

suitable for the retention of HCs under cold start conditions.

• To prepare zeolite coated monoliths and to study the HC adsorption under

cold start conditions.

Regarding the removal of PAHs, the specific objectives are:

• To study the naphthalene total oxidation by catalysts based on Pd

nanoparticles supported on powder inorganic oxides and inorganic oxides

coated monoliths.

Chapter 1

28

References.

[1] A. Corma, J. Catal. 216 (2003) 298-312. [2] H. Van Bekkum, E.M. Flanigen, P.A. Jacobs, J.C. Jansen, Introduction to zeolite science and practice, Ed. Elsevier (2001). [3] S.T. Wilson, B.M. Lok, C.A. Messina, T.R. Cannan, E.M. Flanigen, J. Am. Chem. Soc. 104 (1982) 1146-1147. [4] B.M. Lok, C.A. Messina, R.L. Patton, R.T. Gajek, T.R. Cannan, E.M. Flanigen, J. Am. Chem. Soc. 106 (1984) 6092-6093. [5] S.T. Wilson, E.M. Flanigen, US Patent No 4567029 (1986). [6] M. Estermann, L.B. McCusker, C. Baerlocher, A. Merrouche, H. Kessler, Nature 352 (1991) 320-323. [7] J. Jiang, J. Yu, A. Corma, Angew. Chem. Int. Ed. 49 (2010) 3120-3145. [8] A. Corma, F. Rey, J. Rius, M.J. Sabater, S. Valencia, Nature 431 (2004) 287-290. [9] S.M. Kuznicki, V.A. Bell, S. Nair, H.W. Hillhouse, R.M. Jacubinas, C.M. Braunbarth, B.H. Toby, M. Tsapatsis, Nature 412 (2001) 720-724. [10] P.B. Venuto, Microporous Materials 2 (1994) 297-411. [11] A. Corma, Chem. Rev. 95 (1995) 559-614. [12] M.P. Pina, R. Mallada, M. Arruebo, M. Urbiztondo, N. Navascués, O. de la Iglesia, J. Santamaría, Microporous Mesoporous Mater. 144 (2011) 19-27. [13] A. Tavolaro, E. Drioli, Adv. Mater. 11 (1999) 975-996. [14] A. Basile, F. Galluci, Membranes for Membrane Reactors. Preparation, optimization and selection. Ed. John Wiley & Sons (2011). [15] S. Adhikari, S. Fernando, Ind. Eng. Chem. Res. 45 (2006) 875-881. [16] F. Roa, M.J. Block, J.D. Way, Desalination 147 (2002) 411-416. [17] A. Julbe, Stud. Surf. Sci. Cat. 168 (2007) 181-219. [18] S. Kulprathipanja, US Patent 6726744 (1986). [19] H. Suzuki, US Patent 4699892 (1987). [20] A. Ishikawa, T.H. Chiang, F. Toda, J. Chem. Soc., Chem. Commun. 12 (1989) 764-765. [21] M. Matsukata, N. Nishiyama, K. Ueyama, Stud. Surf. Sci. Cat. 84 (1994) 1183-1190. [22] H. Kita, K. Horii, Y. Ohtoshi, K. Tanaka, K.I. Okamoto, J. Mater. Sci. Lett. 14 (1995) 206-208. [23] J. Caro, M. Noack, P. Kölsch, R. Schäfer, Microporous Mesoporous Mater. 38 (2000) 3-24.

Introduction

29

[24] R. Mallada, M. Menéndez, Membrane Science and Technology, N 13. Inorganic membranes, synthesis, characterization and applications, Ed. Elsevier (2008). [25] Y. Yan, M.E. Davis, G.R. Gavalas, Ind. Eng. Chem. Res. 34 (1995) 1652-1661. [26] A. Berenguer-Murcia, J. Garcia-Martinez, D. Cazorla-Amorós, A. Linares-Solano, A.B. Fuertes, Microporous Mesoporous Mater. 59 (2003) 147-159. [27] B.J. Schoeman, A. Erdem-Senatalar, J. Hedlund, J. Sterte, Zeolites 19 (1997) 21-28. [28] Z.A.E.P. Vroon, K. Keizer, A.J. Burggraaf, H. Verweij, J. Membr. Sci. 144 (1998) 65-76. [29] J.H. Koegler, H. Van Bekkum, J.C. Jansen, Zeolites 19 (1997) 262-269. [30] Y. Takata, T. Tsuru, T. Yoshioka, M. Asaeda, Microporous Mesoporous Mater. 54 (2002) 257-268. [31] S. Mintova, J. Hedlund, V. Valtchev, B. Schoeman, J. Sterte, J. Mater. Chem. 8 (1998) 2217-2221. [32] Á. Berenguer-Murcia, E. Morallón, D. Cazorla-Amorós, A. Linares-Solano, Microporous Mesoporous Mater. 66 (2003) 331-340. [33] J. Hedlund, J. Sterte, Nanostruct. Mater. 12 (1999) 413-416. [34] M.P. Bernal, G. Xomeritakis, M. Tsapatsis, Catal. Today 67 (2001) 101-107. [35] M.C. Lovallo, M. Tsapatsis, T. Okubo, Chem. Mater. 8 (1996) 1579-1583. [36] S.M. Holmes, C. Markert, R.J. Plaisted, J.O. Forrest, J.R. Agger, M.W. Anderson, C.S. Cundy, J. Dwyer, Chem. Mater. 11 (1999) 3329-3332. [37] S. Mintova, T. Bein, Adv. Mater. 13 (2001) 1880-1883. [38] S. Domínguez-Domínguez, Á. Berenguer-Murcia, E. Morallón, Á. Linares-Solano, D. Cazorla-Amorós, Microporous Mesoporous Mater. 115 (2008) 51-60. [39] S. Mintova, J. Hedlund, B. Schoeman, V. Valtchev, J. Sterte, Chem. Commun 1 (1997) 15-16. [40] J. Hedlund, M. Noack, P. Kolsch, D. Creaser, J. Caro, J. Sterte, J. Membr. Sci. 159 (1999) 263-273. [41] M.C. Lovallo, M. Tsapatsis, AIChE J. 42 (1996) 3020-3029. [42] K. Weh, M. Noack, I. Sieber, J. Caro, Microporous Mesoporous Mater. 54 (2002) 27-36. [43] K. Kusakabe, T. Kuroda, S. Morooka, J. Membr. Sci. 148 (1998) 13-23. [44] I. Lee, J. Buday, H. Jeong, Microporous Mesoporous Mater. 122 (2009) 288-293. [45] C. Algieri, P. Bernardo, G. Barbieri, E. Drioli, Microporous Mesoporous Mater. 119 (2009) 129-136.

Chapter 1

30

[46] W. Shan, Y. Zhang,W. Yang, C. Ke, Z. Gao, Y. Ye, Y. Tang, Microporous Mesoporous Mater. 69 (2004) 35-42. [47] C.B. Ahlers, J.B. Talbot, Electrochim. Acta 45 (2000) 3379-3387. [48] T.W. Kim, M. Sahimi, T.T. Tsotsis, Ind. Eng. Chem. Res. 47 (2008) 9127-9132. [49] A. Bueno-López, D. Lozano-Castelló, I. Such-Basáñez, J.M. García Cortés, M.J. Illán-Gómez, C. Salinas-Martínez de Lecea, Appl. Catal. B 58 (2005) 1-7. [50] T.A. Nijhuis, A.E. W. Beers, T. Vergunst, I. Hoek, F. Kapteijn, J.A. Moulijn, Catalysis Reviews 43 (2001) 345-380. [51] M.A. Ulla, R. Mallada, J. Coronas, L. Gutierrez, E. Miró, J. Santamaría, Appl. Catal. A 253 (2003) 257-269. [52] A.E.W. Beers, T.A. Nijhuis, N. Aalders, F. Kapteijn, J.A. Moulijn, Appl. Catal. A 243 (2003) 237-250. [53] B. Louis, L. Kiwi-Minsker, P. Reuse, A. Renken, Ind. Eng. Chem. Res. 40 (2001) 1454-1459. [54] I. Yuranov, A. Renken, L. Kiwi-Minsker, Appl. Catal. A 281 (2005) 55-60. [55] E.V. Rebrov, G.B.F. Seijger, H.P.A. Calis, M.H.J.M. de Croon, C.M. van den Bleek, J.C. Schouten, Appl. Catal. A 206 (2001) 125-143. [56] M.J.M. Mies, J.L.P. Van den Bosch, E.V. Rebrov, J.C. Jansen, M.H.J.M. de Croon, J.C. Schouten, Catal. Today 110 (2005) 38-46. [57] Y.S.S. Wan, J.L.H. Chau, A. Gavriilidis, K.L. Yeung, Microporous Mesoporous Mater. 42 (2001) 157-175. [58] A.R. Pradhan, M.A. Macnaughtan, D. Raftery, J. Am. Chem. Soc. 122 (2000) 404-405. [59] J. García-Martínez, D. Cazorla-Amorós, A. Linares-Solano, Y.S. Lin, Microporous Mesoporous Mater. 42 (2001) 255-268. [60] J.M. Zamaro, M.A. Ulla, E.E. Miró, Chem. Eng. J. 106 (2005) 25-33. [61] J.L. Williams, Catal. Today 69 (2001) 3-9. [62] F. Shiraishi, S. Yamaguchi, Y. Ohbuchi, Chem. Eng. Sci. 58 (2003) 929-934. [63] Y. Zhang, Q. Su, Z. Wang, Y. Yang, Y. Xin, D. Han, X. Yang, H. Wang, X. Gao, Z. Zhang, Chem. Eng. Technol. 31 (2008) 1856-1862. [64] T. Letcher, Future Energy, Ed. Elsevier (2008) pp. 1-24. [65] M. Fan, C.P. Huang, A. Bland, Z. Wang, R. Slimane, I. Wright, Environanotechnology, Ed. Elsevier (2010) pp. 221-222. [66] B.A. Wilhite, S.E. Weiss, J.Y. Ying, M.A. Schmidt, K.F. Jensen, Adv. Mater. 18 (2006) 1701-1704. [67] R. Farrauto, S. Hwang, L. Shore, W. Ruettinger, J. Lampert, T. Giroux, Y. Liu, O. Ilinich, Annu. Rev. Mater. Res. 33 (2003) 1-27.

Introduction

31

[68] R. Masel, Nature 442 (2006) 521-522. [69] J.D. Holladay, J. Hu, D.L. King, Y. Wang, Catal. Today 139 (2009) 244-260. [70] M. Freemantle, Chem. Eng. News 83 (2005) 49-57. [71] T.Q. Gardner, J.G. Martinek, J.L. Falconer, R.D. Noble, J. Membr. Sci. 304 (2007) 112-117. [72] C. Algieri, P. Bernardo, G. Barbieri, E. Drioli, Microporous Mesoporous Mater. 119 (2009) 129-136. [73] C. Galletti, S. Specchia, G. Saracco, V. Specchia, Chem. Eng. J. 154 (2009) 246-250. [74] O. Korotkikh, R. Farrauto, Catal. Today 62 (2000) 249-254. [75] I. Rosso, C. Galletti, G. Saracco, E. Garrone, V. Specchia, Appl. Catal. B 48 (2004) 195-203. [76] A. Wootsch, C. Descorme, D. Duprez, J. Catal. 225 (2004) 259-266. [77] M. Kotobuki, A. Watanabe, H. Uchida, H. Yamashita, M. Watanabe, J. Catal. 236 (2005) 262-269. [78] J.Y. Park, C. Aliaga, J.R. Renzas, H. Lee, G.A. Somorjai, Catal. Lett. 129 (2009) 1-6. [79] M.E. Grass, S.H. Joo, Y. Zhang, G.A. Somorjai, J. Phys. Chem. C 113 (2009) 8616-8623. [80] I. Miguel-García, A. Berenguer-Murcia, D. Cazorla-Amorós, Appl. Catal. B 98 (2010) 161-170. [81] M. Kotobuki, A. Watanabe, H. Uchida, H. Yamashita., M. Watanabe, Appl. Catal. A 307 (2006) 275-283. [82] S.P. Elangovan, M. Ogura, S. Ernst, M. Hartmann, S. Tontisirin, M.E. Davis, T. Okubo, Microporous Mesoporous Mater. 96 (2006) 210-215. [83] J.C. Guibet, E. Faure-Birchem, Fuels and Engines: Technology, Energy, Environment, vol. 2, Ed. Technip (1999). [84] T.H. Yeon, H.S. Han, E.D. Park, J.E. Yie, Microporous Mesoporous Mater. 119 (2009) 349-355. [85] S.P. Elangovan, M. Ogura, M.E. Davis, T. Okubo, J. Phys. Chem. B 108 (2004) 13059-13061. [86] D.S. Lafyatis, G.P. Ansell, S.C. Bennett, J.C. Frost, P.J. Millington, R.R. Rajaram, A.P. Walker, T.H. Ballinger, Appl. Catal., B 18 (1998) 123-135. [87] Z. Sarshar, M.H. Zahedi-Niaki, Q. Huang, M. Eić, S. Kaliaguine, Appl. Catal. B 87 (2009) 37-45. [88] S.W. Baek, J.R. Kim, S.K. Ihm, Catal. Today 93-95 (2004) 575-581. [89] N.R. Burke, D.L. Trimm, R.F. Howe, Appl. Catal. B 46 (2003) 97-104.

Chapter 1

32

[90] A. Iliyas, M.H. Zahedi-Niaki, M. Eic, S. Kaliaguine, Microporous Mesoporous Mater. 102 (2007) 171-177. [91] B. Solsona, T. Garcia, R. Murillo, A.M. Mastral, E.N. Ndifor, C.E. Hetrick, M.D. Amiridis, S.H. Taylor, Top. Catal. 52 (2009) 492-500. [92] M.H. Yuan, C.Y. Chang, J.L. Shie, C.C. Chang, J.H. Chen, W.T. Tsai, J. Hazard. Mater. 175 (2010) 809-815. [93] F. Diehl, J. Barbier Jr., D. Duprez, I. Guibard, G. Mabilon, Appl. Catal. B 95 (2010) 217-227. [94] B. Puertolas, B. Solsona, S. Agouram, R. Murillo, A.M. Mastral, A. Aranda, S.H. Taylor, T. Garcia, Appl. Catal. B 93 (2010) 395-405. [95] A.M. Mastral, M.S. Callen, Environ. Sci. Technol. 34 (2000) 3051-3057. [96] X. Zhang, S. Shen, L.E. Yu, S. Kawi, K. Hidajat, K.Y. Simon Ng, Appl. Catal. A 250 (2003) 341-352. [97] J.I. Park, J.K. Lee, J. Miyawaki, S.H. Yoon, I. Mochida, J. Ind. Eng. Chem. 17 (2011) 271-276. [98] E. N. Ndifor, T. Garcia, S. H. Taylor, Catal. Lett. 110 (2006) 125-128. [99] S.C. Kim, S.W. Nahm, W.G. Shim, J.W. Lee, H. Moon, J. Hazard. Mater 141 (2007) 305-314. [100] E. Ntaunjua N., S.H. Taylor, Top. Catal. 52 (2009) 528-541. [101] C.C. Chang, C.Y. Chiu, C.Y. Chang, C.F. Chang, Y.H. Chen, D.R. Ji, Y.H. Yu, P.C. Chiang, J. Hazard. Mater 161 (2009) 287-293. [102] L. Yu, X. Li, X. Tu, Y. Wang, S. Lu, J. Yan, J. Phys. Chem. A 114 (2010) 360-368. [103] M.H. Yuan, Y.Y. Lin, C.Y. Chang, C.C. Chang, J.L. Shie, C.H. Wu, IEEE Transactions On Plasma Science 39 (2009) 1092-1098. [104] S.C. Marie-Rose, T. Belin, J. Mijoin, E. Fiani, M. Taralunga, F. Nicol, X. Chaucherie, P. Magnoux, Appl. Catal. B 90 (2009) 489-496. [105] A. Bampenrat, V. Meeyoo, B. Kitiyanan, P. Rangsunvigit, T. Rirksomboon, Catal. Commun. 9 (2008) 2349-2352. [106] T. Garcia, B. Solsona, D. Cazorla-Amorós, Á. Linares-Solano, S.H. Taylor, Appl. Catal. B 62 (2006) 66-76. [107] T. Garcia, B. Solsona, S.H. Taylor, Appl. Catal. B 66 (2006) 92-99. [108] A. K. Neyestanaki, L.E Lindfors, T. Ollonqvist, J. Väyryen, Appl. Catal. A 196 (2000) 233-246. [109] H. Bonnemann, K.S. Nagabhushana, Metal Nanoclusters: Synthesis and Strategies for their Size Control, pages 21-48, in B. Corain, G. Schmid and N.

Introduction

33

Toshima, Metal Nanoclusters in Catalysis and Materials Science, Ed. Elsevier (2007). [110] M. Moreno-Mañas, R. Pleixats, Acc. Chem. Res. 36 (2003) 638-643. [111] T. Teranishi, M. Miyake, Chem. Mater. 4756 (1998) 594-600. [112] S. Domínguez-Domínguez, Á. Berenguer-Murcia, D. Cazorla-Amorós, Á. Linares-Solano, J. Catal. 243 (2006) 74-81. [113] L. Kiwi-Minsker, N. Semagina, A. Renken, Size Controlled Pd Nanoparticles Anchored to Carbon Fiber Fabrics: Novel Structured Catalyst Effective for Selective Hydrogenation, pages 293-299 in B. Corain, G. Schmid and N. Toshima, Metal Nanoclusters in Catalysis and Materials Science, Ed. Elsevier (2007).

Capítulo 2. Técnicas de

caracterización, materiales y métodos de preparación.

Capítulo 2

35

2.1. Técnicas de caracterización.

2.1.1. Adsorción física de gases.

Cuando un gas o vapor entra en contacto con la superficie de un material sólido,

parte de sus moléculas son atraídas y se asocian a dicha superficie, recubriéndola en

forma de capas moleculares [1]. Este fenómeno se debe a que las fuerzas de cohesión

en la superficie del sólido no están compensadas, es decir, los átomos de su

superficie no están unidos al mismo número de átomos que un átomo en el interior

del sólido. A este fenómeno se le denomina adsorción [2], al sólido, adsorbente y al

gas, adsorbato (si no está adsorbido) o adsortivo si se encuentra adsorbido.

La adsorción se puede dividir en dos grupos, adsorción por fisisorción o adsorción

por quimisorción. En el caso de la fisisorción o adsorción física, esta presenta

interacciones adsorbente-adsorbato débiles (Qads < 40 kJ/mol), son de tipo Van der

Waals, pudiéndose adsorber en forma de multicapa y se encuentra favorecida a

temperaturas bajas. En cambio, para que se produzca la quimisorción, debe

producirse enlaces químicos (Qads > 40 kJ/mol) y sólo se forma una monocapa.

Para la caracterización del sólido poroso (adsorbente) es de gran importancia conocer

cuál es su porosidad y qué tipo de poros tiene. La IUPAC ha propuesto una

clasificación de la porosidad en función del tamaño de poro basándose en el

mecanismo de adsorción de los mismos [3]:

• Macroporos: un tamaño de poro superior a 50 nm (0.05 µm).

• Mesoporos: un tamaño de poro entre 2 nm y 50 nm.

• Microporos: un tamaño de poro inferior a 2 nm.

Supermicroporos: un tamaño de poro entre 0.7 nm y 2 nm.

Ultramicroporos: un tamaño de poro inferior a 0.7 nm.

Técnicas de caracterización, materiales y métodos de preparación

36

De forma general, en un sólido poroso, los macroporos actúan como poros de

transporte, permitiendo que las moléculas del adsortivo alcancen los poros más

pequeños situados en el interior del sólido poroso. Los mesoporos, actúan como

unión entre los microporos y los macroporos, produciéndose la condensación capilar.

Los microporos, constituyen la mayor parte de la porosidad interna del sólido poroso

y debido a su tamaño, se llenan a presiones relativas del adsorbato bajas [4].

Actualmente, se emplean una gran cantidad de gases como adsorbatos (N2, CO2, Ar,

He, CH4, H2O…) para obtener información acerca de la porosidad de un sólido [5].

Para nuestro estudio, se ha empleado la adsorción de N2 a -196 ºC y CO2 a 0 ºC [3-

6]. La principal desventaja que posee la adsorción de N2 (-196 ºC) para la

determinación de la porosidad es la existencia de problemas difusionales en la

entrada de la porosidad estrecha (alrededor de 0.4 nm) [7]. Sin embargo, la adsorción

de CO2 (0 ºC) en la porosidad estrecha ocurre más rápidamente. Por lo tanto,

podemos decir que la adsorción de CO2 (0 ºC) es un complemento a la adsorción de

N2 para la determinación de la porosidad estrecha (tamaño menor de 0.7 nm) [6-9].

Figura 2.1. Representación de las isotermas de fisisorción según la IUPAC [3].

Capítulo 2

37

A partir de la adsorción física de gases se obtiene de forma gráfica lo que se conoce

como isoterma de adsorción, que consiste en la representación de la cantidad

adsorbida de un gas por un sólido en función de la presión de equilibrio a

temperatura constante [5]. La mayoría de las isotermas de adsorción pueden ser

agrupadas dentro de uno de los 6 tipos de isotermas expuestos en la Figura 2.1.

La isoterma Tipo I, se ajusta a la isoterma de tipo Langmuir. Este tipo de adsorción

es típico de sólidos microporosos con una superficie externa pequeña (carbones

activados, zeolitas y algunos óxidos porosos), cuyo límite máximo de adsorción está

gobernado por el volumen de microporos accesibles en vez del área superficial

interna.

La isoterma Tipo II es la isoterma obtenida cuando un sólido no poroso o

macroporoso se utiliza como adsorbente. Esta isoterma representa una adsorción no

restrictiva en forma de monocapa-multicapa, es decir, el punto B (ver Figura 2.1,

isoterma tipo II) se toma como el punto donde se ha adsorbido la monocapa

completamente y a partir de él, se completa la adsorción en multicapa.

La isoterma Tipo III es una curva convexa que no exhibe el punto B. Este tipo de

isotermas no es común y se debe a que la interacción adsorbato-adsorbente es débil.

A modo de ejemplo, podemos citar la adsorción de N2 en polietileno.

La isoterma Tipo IV presenta un ciclo de histéresis característico, que está asociado

a la condensación capilar en los mesoporos. La parte inicial de la isoterma se asocia a

la adsorción en monocapa al igual que en la isoterma Tipo II. Este tipo de isoterma es

característica de sólidos mesoporosos.

La isoterma Tipo V es poco frecuente. En este caso, al igual que ocurre en la

isoterma Tipo II, la interacción adsorbato-adsorbente es débil pero presenta un ciclo

de histéresis debido a la presencia de mesoporosidad en el sólido.


38

La isoterma Tipo VI presenta una forma escalonada. Este tipo de adsorción es típica

de sólidos con superficies muy homogéneas y con varios tamaños de poro muy bien

definidos. La adsorción en un tipo de poro comienza cuando prácticamente se han

llenado los poros de diámetro anterior.

En la bibliografía existen varios modelos [10,11] que se pueden utilizar para

interpretar los datos de las isotermas de adsorción y extraer información acerca de la

textura porosa del sólido, como por ejemplo, el área superficial específica, el

volumen de porosidad o la distribución del tamaño de poros. Para la determinación

de estos parámetros se han empleado la teoría de Brunauer-Emmett-Teller (B.E.T.) y

la teoría de Dubinin-Radushkevich (DR) que se resumen a continuación.

Teoría B.E.T.

El modelo de adsorción de gases de BET [10] se ha convertido en el procedimiento

estándar para la determinación del área superficial de materiales porosos finamente

divididos a pesar de la gran simplificación del modelo en el que está basada la

teoría. Esta teoría es una aproximación semiempírica que supone la adsorción en

multicapas, sin limitación en el número de capas de gas que se pueden adsorber. La

isoterma de adsorción según el método BET es de la forma:

1

1 1 (2.1)

Donde “n” es el número de moles adsorbidos a una presión relativa P/Po; “nm” es la

capacidad de la monocapa y C es una constante relacionada con el calor de

adsorción.

A partir de esta ecuación, representando el término de la izquierda de la ecuación

frente a la presión relativa (P/Po) se obtiene una línea recta de cuya pendiente y

ordenada se calculan los parámetros característicos C y nm. Finalmente, a partir del

Capítulo 2

39

valor de la monocapa, se puede calcular la superficie específica del sólido aplicando

la ecuación:

10 ⁄ (2.2)

Donde “am” es el área que ocupa una sola molécula de adsorbato (nm2/molécula): NA

es el número de Avogadro (6.022·1023 moléculas/mol) y nm (mol/gramo).

Ecuación de Dubinin-Radushkevich.

La ecuación de Dubinin-Radushkevich [11] se utiliza ampliamente en la

caracterización de sólidos microporosos. Esta ecuación está basada en la teoría del

potencial de Polanyi, en la que se supone que la condensación del líquido del gas en

los microporos es en forma de capas equipotenciales [2]. La ecuación en la que se

basa este modelo es:

(2.3)

donde “V” es el volumen adsorbido a una presión P, “Vo” es el volumen de

microporos del sólido, “Eo” es la energía característica dependiente de la estructura

del poro, “β” es el coeficiente de afinidad el cual es característico del adsortivo y

“Po” es la presión de saturación del adsortivo a la temperatura de trabajo. Finalmente,

representando el LnV frente a Ln (Po/P)2 se obtiene el volumen de microporos Vo.

Para la determinación de la porosidad y la superficie de un sólido mediante adsorción

física, se pueden emplear técnicas gravimétricas y volumétricas [3]. En las técnicas

gravimétricas se determina midiendo, a cada presión relativa, el aumento de peso

experimentado por el sólido debido a la adsorción. En el caso de las técnicas

volumétricas, se introduce una presión del gas, y mediante la medida de la

disminución de la misma se determina la cantidad adsorbida. Para este fin, el equipo


40

de adsorción volumétrico utilizado para la realización de las isotermas de N2 (-196

ºC) y CO2 (0 ºC) ha sido un Autosorb 6. Previamente, el sólido poroso ha sido

desgasificado a 250 ºC durante 4 horas mediante una bomba de vacío con el fin de

eliminar las impurezas adsorbidas por la muestra. En la Figura 2.2 se muestran los

equipos de adsorción volumétricos.

Figura 2.2. Equipo volumétrico de adsorción de gases Autosorb 6.

2.1.2. Análisis térmico. Análisis termogravimétrico (TGA) [12].

El término “Análisis Térmico” engloba una serie de técnicas en las cuales se mide

una propiedad física de una muestra y/o sus productos de reacción en función de la

temperatura cuando esta se somete a una variación controlada de la temperatura.

Los efectos debidos al aumento de temperatura sobre la muestra pueden ser varios y

producir cambios en muchas de sus propiedades. Así, los cambios en peso

configurarán la base de la termogravimetría (TG), mientras que la medida de los

cambios de energía constituye la base del Análisis Térmico Diferencial (ATD) y de

la Calorimetría Diferencial de Barrido (DSC).

La termogravimetría es una técnica basada en la variación de la masa de una muestra

(ganancia o pérdida) cuando es sometida a un tratamiento térmico y a una atmosfera

Capítulo 2

41

controlada. De forma general, existen tres formas distintas de realizar un análisis de

TG [12]:

(i) Isotermo. Se analiza la variación de la masa cuando la muestra es sometida a

una temperatura constante.

(ii) Cuasi-isotermo. La muestra se somete a un tratamiento térmico hasta masa

constante en cada una de las etapas de una serie de incrementos de

temperatura.

(iii) Dinámico. La muestra se somete a un tratamiento térmico (calentamiento,

enfriamiento o combinación) a una velocidad constante de forma lineal.

La atmósfera puede ser estática o dinámica con un caudal determinado y los gases

más habituales empleados son N2, aire, Ar o CO2, y en ocasiones H2, Cl2 ó SO2. Una

característica fundamental de la técnica TG es que sólo permite detectar procesos en

los que se produce una variación de peso. Por tanto, las aplicaciones más importantes

de la técnica TG son el análisis de la descomposición de distintas sustancias

(inorgánicas, orgánicas o poliméricas), corrosión de metales en distintas atmósferas a

temperaturas elevadas, reacciones en estado sólido, calcinación de materiales,

destilación o evaporación de líquidos, pirolisis de carbón, petróleo o madera,

determinación de humedad, contenido en volátiles o cenizas de un carbón y

velocidades de evaporación o sublimación de materiales.

Además, la detección y análisis de los gases emitidos va a proporcionar una

información complementaria a la técnica de TG. De esta forma, la introducción de un

espectrómetro de masas (MS) a la salida de los gases del TG, permite detectar y

cuantificar la cantidad de gases que están siendo emitidos en cada momento.

Para el análisis mediante TG-MS se ha llevado a cabo en un dispositivo experimental

que consta de un equipo de termogravimetría (TA Instruments, SDT 2960) acoplado


42

a un espectrómetro de masas (Thermostar, Balzers, GSD 300T3). En la Figura 2.3 se

muestra el equipo experimental empleado.

Figura 2.3. Termobalanza TG-MS acoplada a un espectrómetro de masa.

2.1.3. Espectroscopia de emisión de plasma por acoplamiento

inductivo (ICP-OES) [13,14].

Las técnicas espectroscópicas atómicas consisten en transformar la muestra en

átomos en estado vapor (atomización) y medir la radiación electromagnética

adsorbida o emitida por dichos átomos. Un aspecto interesante de los espectros

atómicos es que están constituidos por picos estrechos (teóricamente lineales) y bien

definidos, originados por transiciones entre distintos niveles de energía electrónica.

Debido a ello, se explica la gran selectividad que suelen presentar este tipo de

técnicas. Asimismo, la sensibilidad también suele ser elevada y depende del número

de átomos en estado fundamental (técnicas de absorción) y en estado excitado

(técnicas de emisión). En cuanto a la obtención del vapor atómico, pueden utilizarse

distintas fuentes, tales como una llama, energía eléctrica o un plasma, dando lugar a

distintas técnicas.

De forma breve, la absorción atómica es el proceso que ocurre cuando átomos de un

elemento en estado fundamental absorben energía radiante a una longitud de onda

específica.

Capítulo 2

43

En cuanto la emisión atómica se basa en la media de la radiación emitida por los

átomos de una muestra, previamente excitados. La energía utilizada en el proceso de

excitación puede proceder de diferentes fuentes, dando lugar a las distintas técnicas

como se muestra en la Tabla 2.1.

Tabla 2.1. Métodos atómicos de emisión. Fuente de energía Técnica Llama Fotometría de llama Radiación electromagnética Fluorescencia atómica Eléctrica Espectrometría de emisión Plasma ICP Rayos X Fluorescencia de rayos X

Un plasma se define como un gas ionizado, es decir, una mezcla gaseosa que

contiene una concentración significativa de cationes y de electrones, siendo el Argón

el gas mas empleado para la obtención del plasma. Para originar un plasma es preciso

un aporte externo de energía que provoque la ionización del gas y la mantenga

estacionaria. En función de cómo se aporte esta energía externa, se han desarrollado

tres tipos de fuentes de alimentación: (i) una fuente continua (DCP), consistente en

dos electrodos sumergidos en la corriente de gas argón, (ii) uso de potentes campos

de microondas (MIP) e (iii) de radiofrecuencia (ICP), que es la más interesante desde

un punto de vista analítico.

La fuente de ICP o antorcha consta de tres tubos concéntricos de cuarzo, que se

encuentran rodeados en su extremo superior por tres o cuatro anillos de una bobina

de inducción alimentada por un generador de radiofrecuencias. El argón para la

generación del plasma se introduce por el tubo central y se ionizan por una descarga

producida por una bobina Tesla. El plasma una vez formado, se automantiene, y el

resultado es un gas altamente ionizado con temperaturas entre 6000 y 10000 K.

Como la temperatura obtenida es muy elevada, es necesario aislar térmicamente el

plasma para evitar el sobrecalentamiento del tubo de cuarzo. Para ello, se introduce


44

una corriente de He (10-15mL/min) a través del tubo exterior permitiendo, no solo el

enfriamiento del tubo, sino que también estabiliza y centra el plasma.

Finalmente, la muestra es aspirada por un sistema nebulizador y trasportada por el

tubo interior por el gas portador a una velocidad relativamente pequeña (1 mL/min).

En estas condiciones, los átomos presentes en la muestra son ionizados/excitados. Al

volver a su estado fundamental, estos iones o átomos excitados emiten radiaciones de

una longitud de onda que es característica de cada elemento. Esta radiación pasa a

través de un sistema óptico que separa la radiación según su longitud onda. A

continuación un detector mide la intensidad de cada una de las radiaciones

relacionando ésta con la concentración de cada elemento en la muestra. En la Figura

2.4 se muestra el equipo experimental empleado.

Figura 2.4. Espectrofotómetro Perkin Elmer 4300DV.

2.1.4. Espectroscopia Infrarroja de reflectancia difusa con

transformada de Fourier (DRIFTS) [1].

El rango espectral de la zona infrarroja (IR) está comprendido entre 10 y 12800

cm-1. Tradicionalmente, la región IR se divide en tres partes:

Capítulo 2

45

(i) IR lejano (10-200cm-1). En esta región tienen lugar movimientos rotacionales,

en general, y traslaciones en sólidos.

(ii) IR medio (200-4000cm-1). Esta es la región más común y en ella tiene lugar

tanto las vibraciones moleculares o de grupos atómicos (OH-, CO3=), como

las vibraciones de red de los sólidos.

(iii) IR próximo (4000-12800cm-1). Es esta región tienen lugar absorciones

debidas a sobretonos y modos de combinación.

El principio básico en el que se basa la espectroscopia IR es la excitación de los

modos vibracionales y rotacionales de los enlaces cuando la muestra se irradia con

luz infrarroja. Cada molécula, de acuerdo a las características de sus enlaces,

absorberá una radiación característica permitiendo su identificación.

A día de hoy, existen varias técnicas analíticas basadas en espectroscopia infrarroja,

siendo la espectroscopia infrarroja por reflectancia difusa idónea para trabajar con

muestras en polvo finamente divididas. Cuando se irradia una muestra en polvo, la

radiación penetra a través de la superficie de las partículas, excitando los modos de

vibración de las diferentes especies presentes en la muestra. La luz infrarroja se

absorbe y refleja en todas las direcciones. A través de un conjunto de espejos, la luz

reflejada se conduce hacia el detector donde se mide, obteniendo el espectro

infrarrojo.

Los experimentos de in-situ DRIFTS se realizaron empleando un espectrofotómetro,

Mattson modelo MI60 equipado con un accesorio de reflectancia difusa

(SpectraTEch Collector) y una cámara de ambiente controlado con cristales de CaF2

que permite el control, tanto de los gases introducidos, como de la temperatura (ver

Figura 2.5). Durante todos los experimentos, la composición del gas se siguió con un

espectrómetro de masas (OmniStar Pfeiffer Vacuum).


46

Figura 2.5. Espectrofotómetro FTIR acoplado a un espectrómetro de masas.

2.1.5. Difracción de Rayos X.

Dentro de los métodos físicos utilizados para la caracterización de sólidos, las

técnicas basadas en la utilización de Rayos X forman un grupo especialmente

importante, tanto por la variedad de técnicas como por la información que

suministran [1]. Las técnicas basadas en radiación X se pueden dividir en función del

fenómeno que se produce cuando la radiación incide sobre el sólido. Por lo tanto,

podemos hablar de 3 grupos:

• Absorción de Rayos X. Se observa cuando la radiación incidente es atenuada

parcialmente por el sólido y el resto de la radiación atraviesa el sólido sin que

se produzcan cambios. La medida de la intensidad transmitida es la que

proporciona información del sólido.

• Técnicas basadas en el efecto fotoeléctrico. En este grupo de técnicas la

absorción fotoeléctrica se traduce en la emisión, por parte de la muestra

irradiada, de radiación X y electrones. Por lo tanto las técnicas basadas en el

efecto fotoeléctrico se encargan del estudio tanto del espectro de rayos X

como de los electrones emitidos.

• Difracción de Rayos X.

Capítulo 2

47

La radiación de Rayos X es una radiación electromagnética de onda corta

comprendida entre 10-5 y 100 Å aproximadamente. El origen de la difracción de

rayos X es consecuencia de la interacción de una onda electromagnética de rayos X

con la nube electrónica de los átomos de un sólido cristalino, cuyos parámetros de

celda son del orden de magnitud de la longitud de onda de la radiación incidente.

Parte de esta radiación es absorbida y posteriormente devuelta en forma de radiación

dispersada en todas las direcciones del espacio. Las distintas radiaciones dispersadas

sufren fenómenos de interferencia que, únicamente, son constructivas en direcciones

muy bien definidas, dando lugar al difractograma del cristal.

Las condiciones necesarias para que se produzca la difracción vienen determinadas

por la Ley de Bragg [15]. Se asume que una sustancia cristalina se puede considerar

como distintas familias de planos paralelos y equidistantes entre sí. Cada una de estas

familias tiene designado un índice de Miller (hkl) y un espaciado dhkl. Si sobre estos

planos incide un haz de Rayos X monocromático, con una longitud de onda , en una

dirección que forma un ángulo θ con la superficie de los planos, sólo se producirá

difracción cuando el ángulo de incidencia, la longitud de onda de la radiación y el

espaciado de la familia de planos cumplan la relación de la ley de Bragg:

2 sen (2.4)

donde n es un número entero llamado orden de difracción.

A partir de la ecuación de Scherrer se puede determinar el tamaño medio de los

cristales. Esta ecuación relaciona el tamaño medio de los cristales, t, con la anchura a

mitad de altura de los picos de difracción debido al tamaño finito de los cristales, β:

cos (2.5)

donde, la constante K depende de la forma de los cristales y se denomina constante

de Scherrer, siendo su valor cercano a la unidad. Como se ha mencionado


48

anteriormente, la ecuación postulada por Scherrer predice que el tamaño medio de

partícula está relacionado con la anchura del pico de difracción a mitad de altura. En

el caso de las zeolitas unidimensionales, se ha realizado una evaluación crítica para

su aplicación [16] que ha revelado errores significativos en la medida obtenida.

El método más empleado para la determinación de la cristalinidad de un sólido es el

método del polvo cristalino. Este método se basa en dos condiciones experimentales

básicas [1]: el empleo de una radiación monocromática y un sólido constituido por

un polvo o agregado cristalino. Este sólido debe de estar integrado por un número

elevado de fragmentos muy pequeños de cristales idealmente orientados al azar, de

forma que no exista ningún tipo de correlación en la orientación. La difracción del

haz de rayos X monocromático en este tipo de muestras produce una serie de conos

de rayos difractados coaxiales con la dirección del haz incidente. Cada cono de

difracción representa una solución de la ecuación de Bragg para cada valor

específico de los espaciados dhkl de las distintas familias de planos cristalinos. De

acuerdo con la ecuación de Bragg, y como es conocida, si podemos medir el

ángulo θ, podemos conocer los distintos valores de los espaciados del sólido que

difracta.

Este método se emplea para la identificación de sustancias cristalinas y el análisis

cuantitativo de estas, asignación de índices hkl o determinación de los parámetros de

celdilla, entre otros.

En el presente trabajo de investigación la técnica de Rayos X se ha utilizado para la

determinación de las estructuras cristalinas de las zeolitas preparadas, tanto en polvo

como soportadas sobre los distintos materiales empleados. Para tal fin, se ha

utilizado el equipo de difracción de rayos X disponible en los servicios técnicos de

investigación de la Universidad de Alicante de la marca Seifert modelo JSO-

DEBYEFLEX 2002 (ver Figura 2.6) con una radiación Cu-Kα ( = 0.1542 nm), un

Capítulo 2

49

filtro de níquel y una velocidad de escaneo de 2º/min en un ángulo comprendido

entre 2-50º.

Figura 2.6. Difractómetro de rayos X JSO-DEBYEFLEX 2002.

2.1.6. Espectroscopia fotoelectrónica de rayos X (XPS) [1].

En general, las técnicas espectroscópicas de análisis se basan en la medida de

partículas o radiación que emite el sólido cuando es irradiado con fotones, electrones

o partículas pesadas (iones, átomos, etc.). En la Figura 2.7 se presenta el esquema del

proceso de la interacción de un haz de electrones con la muestra.

Figura 2.7. Esquema de la interacción de un haz de electrones con la muestra.

En función del tipo de partículas analizadas, podemos dividir las técnicas en dos

grupos: (i) técnicas espectroscópicas de electrones (XPS, AES); y (ii) técnicas

espectrométricas de iones (ISS, RBS, SIMS).

iones

electronesPartículas retrodispersadas

Partículas neutras

Fotones

muestra


50

En concreto, la espectroscopia de fotoelectrones (XPS), es una técnica de

caracterización de superficies, basada en la ionización de los átomos que componen

la muestra mediante radiación monoenergética de rayos X (aproximadamente

1.4KeV) y el análisis de la energía cinética de los electrones excitados. A partir de la

energía cinética y de la energía de radiación utilizada, se puede calcular la energía de

ligadura de dicho electrón al átomo.

Una vez identificados los diferentes elementos presentes en la muestra la medida de

las intensidades de los diferentes picos permite determinar la concentración. Además,

puesto que la energía de ligadura depende del entorno químico en que se encuentra el

átomo, el XPS proporciona también información sobre los estados de oxidación de

los átomos ionizados. Esta técnica puede llegar a ser semicuantitativa, siendo capaz

de dar resultados fiables de porcentajes atómicos en superficie.

Figura 2.8. Equipo de XPS: Espectrómetro VG-microtech-Multilab 3000.

El equipo de XPS empleado es un espectrómetro VG-Microtech Multilab 3000 (Ver

Figura 2.8) equipado con un analizador de electrones semiesférico con 9

channeltrons (con una energía de paso de 2-200eV) y una fuente de radiación de

rayos X con ánodos de Mg y Al. Se ha empleado la radiación de MgKα (1253,6 eV)

trabajando con el detector en el modo de energía constante con una energía de paso

de 50 eV). La precisión en la determinación de los valores de la energía de ligadura

(BE) y energía cinética (KE) es de ±0.2 y ±0.3 eV, respectivamente. Los valores de

Capítulo 2

51

BE y KE fueron obtenidos utilizando el Peak-fit Program incorporado en el

programa informático de control del espectrómetro.

2.1.7. Microscopía electrónica de transmisión (TEM), barrido (SEM)

y espectroscopia de fluorescencia de rayos X basada en

dispersión de energía (EDX) [1,17,18].

La microscopía electrónica y las técnicas analíticas derivadas de ellas son

herramientas muy eficaces para la identificación y caracterización de sólidos, ya que

permiten conocer detalles microestructurales y de composición de las muestras.

La microscopía nos permite obtener una imagen amplificada de un objeto,

dividiéndose en dos tipos: microscopía óptica y microscopía electrónica. Ambas

técnicas obtienen la imagen del objeto a partir de la interacción de una onda

electromagnética con la muestra. Sin embargo, la principal diferencia radica en la

naturaleza del haz que se hace incidir sobre la muestra (junto con su instrumentación

y características propias). En el caso de la microscopía óptica se utiliza luz visible

para irradiar la muestra (longitud de onda comprendida entre 400 y 700 nm),

mientras que la microscopía electrónica emplea un haz de electrones como radiación

que tiene una longitud de onda entre 0.001 y 0.01 nm.

En la microscopía electrónica, el haz de electrones se obtiene a partir de la emisión

termoiónica producida por el calentamiento a 2700 K de un filamento de wolframio

que actúa como cátodo. Estos electrones emitidos termoiónicamente se aceleran

rápidamente hacia el ánodo mediante una diferencia de potencial controlada,

comprendida entre decenas y centenas de kilovoltios. Por lo tanto, podemos decir

que se puede seleccionar la energía de los electrones, y en consecuencia, su longitud

de onda.


52

Para el caso de la microscopía óptica, se ignoran las interacciones de la luz con la

materia porque es suficiente con la luz transmitida o reflejada para que la imagen se

pueda observar con nitidez. En cambio, en la microscopia electrónica no se puede

ignorar la interacción de los electrones con la materia porque produce diversas

consecuencias. En primer lugar, la columna del microscópico electrónico debe de

estar a vacío para evitar la interacción de los electrones con las moléculas de gas en

el aire y así evitar la dispersión de los electrones. En segundo lugar, hay que tener en

cuenta el calentamiento de la muestra por el haz de electrones que puede, en

ocasiones, producir cambios químicos o estructurales de esta. Cuando el haz de

electrones incide sobre la muestra pueden ocurrir dos tipos de fenómenos (i)

Dispersión elástica; y (ii) Dispersión inelástica.

En el caso de la dispersión elástica, no hay un cambio en su energía. Es la resultante

de la interacción entre los electrones y los átomos de la muestra. Esta dispersión es

característica de la microcopia electrónica de transmisión (TEM).

En la dispersión inelástica, se produce una pérdida de energía apreciable en el haz de

electrones. Existe una gran cantidad de procesos de interacción que producen una

pérdida de energía en el haz primario de electrones debido a la detención del electrón

por la muestra, con el consecuente aumento de calor en la muestra. De toda la

energía que incide sobre la muestra, sólo una pequeña porción de electrones es

emitida por la muestra en forma de rayos X o como electrones secundarios.

La Microscopía Electrónica de Transmisión (TEM) [1,17], consiste en irradiar una

fina película de muestra (no más de 100 nm de grosor) con un haz de electrones de

densidad de corriente uniforme, con una energía elevada de 100KeV o superior.

Parte de estos electrones son trasmitidos, otra parte son dispersados y otra parte da

lugar a interacciones que producen distintos fenómenos como emisión de luz,

electrones secundarios y Auger, rayos X, etc. Todas estas señales se pueden emplear

para obtener información sobre la naturaleza de la muestra (morfología,

Capítulo 2

53

composición, estructura cristalina y estructura electrónica, etc.). En el caso del TEM,

se emplea la transmisión/dispersión de los electrones para formar imágenes, la

difracción de los electrones para obtener información acerca de la estructura

cristalina y la emisión de rayos X característicos para conocer la composición de

elemental de la muestra. Por tanto, los electrones que se emplean en TEM para

análisis de la muestra son aquellos que la atraviesan.

Un microscopio electrónico de transmisión proporciona dos tipos de información

complementaria:

(i) Obtención de la imagen directa del material por análisis de los eletrones

transmitidos no dispersados.

(ii) Obtención del llamado diagrama de difracción a partir del análisis de la

distribución espacial de los electrones dispersados elásticamente.

En la Figura 2.9 se muestra el microscopio electrónico empleado (JEOL modelo

JEM-2010), usando un voltaje de aceleración de 200 KeV, y con una resolución

espacial estructural de 0.5 nm.

Figura 2.9. Microscopio electrónico de transmisión JEOL modelo JEM-2010.


54

Los electrones secundarios son los que se utilizan para formar las imágenes en el

microscopio electrónico de barrido (SEM) [1,18]. Son aquellos electrones que

escapan de la muestra con energías por debajo de unos 50 eV. Pueden ser electrones

primarios que al final de su trayectoria alcancen la superficie con una energía inferior

a ésta, pero, normalmente, son electrones que han recibido una transferencia de

energía mediante algún proceso de dispersión inelástica a una distancia pequeña de la

superficie.

La emisión de rayos X característicos, como resultado de la interacción de los

electrones con la materia, permite una de las aplicaciones más importantes en los

microscopios electrónicos que es analizar la composición “in situ” de la muestra a la

vez que observamos su imagen real. Por definición, la emisión de rayos X es

consecuencia de la relajación del átomo excitado al ocupar un electrón la posición

que otro ha dejado. Esta puede ser de dos tipos: (i) catodoluminiscencia; que consiste

en la emisión de un fotón y que tiene lugar cuando el electrón vacante pertenecía a

una capa externa del átomo; (ii) emisión característica; que tiene lugar cuando el

electrón vacante pertenece a una capa interna del átomo. En este caso, cuando un

electrón pasa a ocupar una órbita más interna, emite una energía característica de

cada átomo. Este fenómeno es el que se emplea en la espectroscopia de fluorescencia

de rayos X basada en dispersión de energía (EDX). El EDX permite realizar un

microanálisis de distintas partes de la muestra, pudiendo realizar un barrido o análisis

puntuales de las distintas zonas observadas. De esta forma podemos conocer la

composición elemental de las zonas estudiadas.

Para el análisis de todas las muestras estudiadas en el presente trabajo de

investigación se han empleado los equipos de microscopía electrónica de barrido

disponibles en servicios técnicos de investigación de la Universidad de Alicante,

JEOL JSM-840 e HITACHI S-3000N. A modo de ejemplo, en la Figura 2.10 se

presenta una imagen del equipo de SEM HITACHI S-3000N.

Capítulo 2

55

Figura 2.10. Microscopio electrónico de barrido HITACHI S-3000N.

2.1.8. Sistema experimental de análisis de membranas: célula de

permeación Wicke-Kallenbach.

La célula de permeación Wicke-Kallenbach se ha empleado para la determinación de

las propiedades de las membranas preparadas en la presente Tesis Doctoral. Dentro

de las propiedades de transporte de un sólido poroso, se deben definir cinco

magnitudes [19].

• Permeación. Se define como los moles de una especie que atraviesan la

membrana por unidad de tiempo y por unidad de área de la membrana

(mol/m2·s).

• Permeancia. Se define como los moles de una especie que atraviesan la

membrana por unidad de tiempo, por unidad de área de la membrana y por

unidad de presión (mol/m2·s·P), es decir, se trata de la permeación por unidad

de presión.

• Permeabilidad. Se define como la permeancia multiplicada por el grosor de la

membrana (mol·m/m2·s·P).

• Selectividad ideal. Este tipo de selectividad se obtiene cuando se miden por

separado los flujos de permeación de dos componentes individuales y se

dividen para obtener la relación entre ellos. Para ello, se hace permear un gas


56

A a través de la membrana, se determina su flujo X. A continuación, una vez

purgada la membrana, se hace permear el gas B determinando el flujo Y. La

selectividad obtenida en este caso es X/Y.

• Selectividad. En este tipo de selectividad (también llamada selectividad real)

se miden los flujos de permeación a la vez haciendo permear una mezcla de

los dos componentes A y B, obteniendo sus respectivos flujos X e Y. En este

caso, la selectividad se define como [(X/αx)/Y/αy)], donde αN es la fracción

del componente N en la mezcla.

Cuando se mide el flujo de un gas, debemos de distinguir entre medidas de

permeación realizadas bajo un gradiente de presión constante (y conocido) y aquellas

que trabajan a presión total constante a ambos lados de la membrana. En este último

caso están basadas las células de Wicke-Kallenbach, donde el flujo de un

determinado componente está regido por la diferencia de concentración entre las

caras de la membrana bajo una presión total constante a ambos lados de la

membrana. A modo de ejemplo, se muestra en la Figura 2.11 el sistema experimental

empleado:

Figura 2.11. Esquema de una célula de Wicke-Kallenbach.

En la célula de Wicke-Kallenbach [20], el gas se suministra por uno de los lados de

la membrana (cara de alimentación) y para evitar la acumulación de la fracción no

permeada, se dispone de una salida de gas. Las moléculas del gas que atraviesan la

membrana (fracción permeada) son recogidas por un gas de barrido, siendo en este

Capítulo 2

57

caso el gas empleado He y llevadas al detector (espectrómetro de masas) para

analizar posteriormente su concentración. Operando de esta forma, se mantiene el

gradiente de concentraciones entre la entrada y la salida de los gases a estudiar

durante todo el experimento, manteniendo la presión a ambos lados de la membrana

constante.

El número de Knudsen (Kn) es un parámetro característico que define diferentes

regiones de flujo de gases en sistemas porosos. Este depende de la proporción entre

el número de colisiones entre las moléculas ( : recorrido libre medio de las

moléculas de un gas) y el número de colisiones entre las moléculas y las paredes de

los poros (dp: tamaño de poro):

(2.6)

El valor de Kn separa cuatro tipo de regímenes de flujo en la difusión gaseosa:

• Flujo viscoso. Kn << 1; << dp.

• Flujo (difusión) de Knudsen. Kn > 1; > dp.

• Flujo de transición. Kn = 1; = dp.

• Difusión activada. Kn >> 1; >> dp.

De esta forma se pueden diferenciar tres mecanismos de transporte de gases o

vapores a través de las membranas.

• Flujo viscoso. Cuando el número de colisiones intermoleculares es mucho

mayor que el número de colisiones molécula-pared del poro (Kn << 1). Este

tipo de mecanismo se produce en poros grandes, por tanto, no se aplica a

membranas porosas.

• Difusión de Knudsen. Cuando el número de colisiones entre las moléculas y

las paredes del poros son predominantes (Kn > 1).


58

• Difusión activada. Se trata de un caso limitante dentro del régimen de

Knudsen. En este caso el diámetro de poro es muy parecido al tamaño de la

molécula (Kn >> 1; >> dp). Este proceso va a ser dependiente de la

temperatura. La dependencia con la energía de activación (y con la

temperatura) también puede ser debida a lo que se conoce como difusión

superficial que ocurre cuando la temperatura o la presión del gas es tal que la

adsorción sobre las paredes de los poros es importante. En este último caso,

los resultados experimentales muestran que los mecanismos anteriores (flujo

viscoso y Knudsen) no son válidos.

En el caso de transporte a través de membranas de zeolita, debemos considerar los

siguientes puntos:

• El transporte dominante es mediante difusión superficial y/o activada,

siguiendo una ecuación tipo Arrhenius.

• Puede darse en algunos casos el mecanismo de tamizado molecular.

2.2. Materiales empleados.

2.2.1. Monolitos cerámicos de estructura celular (“honeycomb”).

De forma general, los soportes monolíticos de estructura celular se definen como

estructuras de cerámica, metálicas o de plástico formadas por celdas o canales

interconectados que se repiten en todas las direcciones [21].

Hoy en día, el mayor uso de los monolitos de estructura celular es en la industria del

automóvil para la fabricación de los catalizadores de tres vias. La aplicación de los

monolitos en otras reacciones en la industria se encuentra muy limitada excepto en

reacciones de multifase, donde los reactores monolíticos poseen claras ventajas sobre

los reactores convencionales (lecho fijo y de mezcla). A modo de ejemplo, la

Capítulo 2

59

producción a escala industrial de peróxido de hidrógeno a partir del proceso de la

antraquinona se realiza mediante catalizadores monolíticos [22].

En el caso de las aplicaciones en la industria del automóvil, los monolitos empleados

son cerámicos y fabricados de cordierita cuya fórmula molecular es

2MgO·2Al2O3·5SiO2. La cordierita es una fase cristalina altamente anisotrópica que a

su vez, tiene una expansión térmica anisotrópica, lo que produce una orientación

durante el proceso de extrusión y un bajo coeficiente de expansión térmico [23].

Los monolitos de cordierita se preparan a nivel industrial mediante un proceso de

extrusión. Para ello, inicialmente los precursores sólidos se mezclan. A continuación

se mezclan con agua para que la masa fluya y se extrusionan en forma de barras. A

partir de estas barras se preparan, nuevamente por extrusión, los monolitos con la

geometría deseada. Finalmente, se secan y calcinan a 1400ºC para completar la

reacción en estado sólido, transformando la mezcla en cordierita. La geometría de la

cordierita se caracteriza por la densidad de celda (200-1200 cpsi -celdas/pulg2-) y el

grosor de la pared (0.051-0.27 mm).

Los monolitos de cordierita poseen una serie de características únicas con respecto a

otros materiales: (i) bajas caídas de presión en sistemas de gases; (ii) buena

resistencia térmica; (iii) porosidad y distribución de poros que le da buenas

propiedades de mojado; (iv) compatibilidad con suspensiones de catalizadores y (v)

alto índice de refracción [24]. Sin embargo, una desventaja es que la cantidad de

catalizador que se deposita en las paredes del monolito, es mucho menor que el

volumen de catalizador de un lecho fijo. Por lo tanto, los monolitos catalíticos no

contienen la suficiente cantidad de catalizador para ser utilizados en una serie de

reacciones en la industria [25].

Los monolitos de cordierita que se han empleado en el presente trabajo de

investigación los ha suministrado la empresa Corning®, y poseen una estructura


60

celular tipo panal de abeja (“honeycomb”) de 400 cpsi, un diámetro (d) de 14 cm y

altura (h) de 14 cm. Estos monolitos iniciales se han tallado hasta obtener monolitos

de dimensiones menores que permitieran preparar los sólidos soportados (d= 1.4 cm

y h= 1.6 cm). Una vez cortados, se han soplado en aire y calcinado a 800 ºC para

eliminar restos de polvo y aceite del proceso de cortado. En la Figura 2.12 se

presenta una foto del monolito original junto con la ampliación de la estructura tipo

“honeycomb”.

Figura 2.12. Monolito original con estructura celular "honeycomb".

Tras el tratamiento térmico, los monolitos de cordierita han sido caracterizados

mediante SEM, DRX y adsorción física de N2 (-196 ºC) y CO2 (0 ºC) con la finalidad

de conocer sus propiedades, que serán consideradas en la caracterización de las

muestras de zeolita soportada. A modo de ejemplo, la Figura 2.13 presenta la imagen

de SEM de los monolitos. En la Figura 2.13A se observa la vista superior de una

pared interna del monolito en la que se aprecian los huecos superficiales que serán de

gran utilidad para soportar zeolitas. En la Figura 2.13B se muestra un corte

transversal del monolito que posee paredes de 150 µm de grosor.

Capítulo 2

61

Figura 2.13. Imágenes de SEM correspondientes a: A) vista superior monolito, B)

corte transversal del monolito.

En la Figura 2.14 se presenta el espectro de DRX de polvo del monolito con los picos

característicos de la cordierita que aparecen a ángulos 2θ de 10.4, 18.3, 21.7, 26.3,

28.5 y 29.4 º, que están de acuerdo con los picos característicos descritos en la

bibliografía para este material [24,26]. Es de gran importancia la identificación de los

mismos a la hora de discernir, posteriormente, entre los picos de la cordierita y la

zeolita en los materiales soportados. Finalmente, en la Figura 2.15 se presenta la

isoterma de adsorción de N2 (-196 ºC). Se puede observar que la isoterma es

característica de sólidos no porosos dando lugar a un área superficial muy baja (SBET

~ 1 m2/g).

Figura 2.14. Espectro de Difracción de rayos X de la cordierita.


62

Figura 2.15. Isoterma de adsorción/desorción N2 (-196 ºC).

2.2.2. Discos de grafito macroporoso.

Las membranas de zeolita se pueden soportar en materiales diferentes, como metales

[27], polímeros [28], cerámicas [29] o materiales carbonosos [30]. De los anteriores,

los más ampliamente utilizados son los soportes cerámicos (alúmina) o poliméricos.

Los soportes poliméricos tienen un bajo coste, pero son muy sensibles al aumento de

temperatura, que produce cambios de volumen (hinchamiento y contracción) y

poseen una baja estabilidad química. Los soportes cerámicos (alúmina) son frágiles y

con un alto coste. Además, puede producirse una disolución superficial que puede

alterar la composición de la zeolita resultante [31]. Los materiales carbonosos han

sido estudiados relativamente poco, debido a su carácter hidrofóbico (debido a la

baja cantidad de grupos oxigenados superficiales), que produce una interacción

carbón-gel de síntesis muy débil y dificulta el crecimiento de zeolitas sobre la

superficie del carbón [31]. Las ventajas que presentan los soportes carbonosos son su

elevada estabilidad térmica y química, facilidad para preparar distintas formas

(cilindros, discos) y una porosidad y química superficial fácil de modificar [32].

En la presente Tesis Doctoral se han utilizado soportes de carbón poroso para crecer

membranas de zeolita LTA. Estos soportes de carbón poroso los ha suministrado la

empresa Pocco Graphite en forma de láminas de grosor de 0.3 mm y tamaño de poro

Capítulo 2

63

medio de 0.7 µm. En la Figura 2.16, se muestra una imagen de SEM del disco de

grafito macroporoso y en la imagen de SEM insertada, se puede observar con detalle

la superficie del disco donde se deposita la película de zeolita.

Figura 2.16. Imagen de SEM de la superficie del disco de carbón.

En la Figura 2.17 se muestra el espectro de difracción de rayos del disco de carbón,

donde se observan los picos principales a 26.2 y 42.1 º junto con dos picos de menor

intensidad a 23.2 y 28.9 º.

Figura 2.17. Espectro de difracción de rayos X del disco de carbón.


64

2.3. Métodos de preparación de membranas de zeolita por

crecimiento secundario: Sembrado de soportes mediante

depósito electroforético.

Como se ha descrito en el Capítulo 1 el procedimiento empleado para la preparación

de membranas de zeolita, ha sido el denominado crecimiento secundario en el que se

introduce una etapa de sembrado del soporte previa al tratamiento hidrotérmico [27].

Hasta la fecha se han desarrollado diferentes técnicas para producir un sembrado

sobre el soporte (ver Capítulo 1). En este trabajo se ha realizado mediante depósito

electroforético (EPD) [30,33].

Una zeolita es un aluminosilicato cristalino con una estructura abierta regular en 3D

[14]. Esta estructura característica está basada en tetraedros (TO4) de aluminio y

silicio, con los cationes ocupando su centro y los oxígenos en los vértices que están

siempre compartidos por dos tetraedros adyacentes. Una zeolita se prepara

normalmente en presencia de un plantilla que actúa como agente director de la

estructura hacia la formación de la estructura zeolítica final. Esta plantilla, se elimina

normalmente por descomposición térmica.

En este sentido, moléculas orgánicas como cationes de tetrametilamonio (TMA+), se

encuentran atrapados en las cajas β de la estructura LTA [33] de la zeolita A o iones

tetrapropilamonio (TPA+), en la silicalita [30]. Estas moléculas le confieren una

carga positiva a las zeolitas preparadas, lo que tiene una relevancia importante en la

preparación de soportes sembrados mediante EPD.

El EPD como método de sembrado de zeolitas, es un método novedoso que permite

el recubrimiento de un soporte seleccionado con las semillas de la zeolita

seleccionada por medio de la aplicación de un potencial eléctrico a la suspensión de

la zeolita. El método de EPD puede ser aplicado fácilmente mediante el uso de una

célula electroquímica simple, obteniendo recubrimientos homogéneos. La cantidad

Capítulo 2

65

de semillas depositadas sobre el soporte se puede controlar fácilmente ajustando el

tiempo de sembrado o la intensidad de corriente aplicada sobre la célula

electroquímica. Normalmente, en un proceso de sembrado, las interacciones

electrostáticas entre la superficie del soporte y los cristales de zeolita son los

responsables del sembrado, aunque en el caso de EPD, gracias al control sobre la

densidad de corriente y el tiempo de depósito es más fácil controlar el depósito de

semillas.

En relación a la metodología empleada, el equipamiento necesario para la

preparación de soportes sembrados por medio de EPD, es una célula electroquímica

[30,32,34]. En principio, el procedimiento se puede utilizar cuando el soporte es un

conductor eléctrico, aunque esta premisa no es obligatoria, pues, el soporte puede

estar localizado muy cerca del electrodo. En primer lugar, hay que preparar una

suspensión coloidal de la zeolita seleccionada. A continuación, se introduce el

soporte en la disolución, como electrodo de trabajo, y un contraelectrodo (Pt).

Entonces, se aplica una corriente (DC) durante un tiempo determinado para la

obtención del soporte sembrado.

Figura 2.18. Montaje experimental de una célula de EDP.

En la Figura 2.18, se muestra un esquema del montaje experimental para el sembrado

de soportes carbonosos. En primer lugar, se prepara la suspensión de la zeolita. Para

ello se preparan semillas de zeolita A siguiendo el procedimiento experimental

v

A

Contra electrodo Pt

Disco carbón

Cátodo

Ánodo

Suspensión de LTA


66

descrito por Hendlund y col. [35] para la obtención de nanocristales de 10nm. La

composición molar empleada es 0.22Na2O·5.0SiO2·1.0Al2O3·8.0(TMA)2O·400H2O,

siendo (TMA)2O óxido de tretrametilamonio, y las condiciones de cristalización 63

ºC durante 63 horas en un autoclave de 10 ml y ocupando el 75 % del volumen. Una

vez obtenidos los cristales de la zeolita nanocristalina, estos se recogen por

centrifugación (20000 r.p.m. durante 3 min) y a continuación se re-dispersan en agua

para su uso (20g/l). El pH de la disolución resultante es entre 9-9.5. A continuación,

se introduce el disco de carbón, que actúa como electrodo de trabajo unido a un

electrodo de Cu y como contraelectrodo se usa un hilo de Pt. Seguidamente, se aplica

una densidad de corriente constante referida al área geométrica del disco (0.57 cm2),

que es el área expuesta a la suspensión de zeolita A. En nuestro caso, se ha aplicado

una intensidad de corriente de 9 mA/cm2 durante 30 min de acuerdo a resultados

previos [32]. La intensidad de corriente y el voltaje a través del circuito se ha seguido

en todo momento. Tras el sembrado, se aplica la etapa de crecimiento hidrotérmico

para obtener la película de zeolita.

2.4. Preparación de películas delgadas de zeolita y

silicoaluminofosfato sobre monolitos de cordierita.

Como se ha descrito anteriormente, el método empleado para la síntesis de zeolitas

en polvo es la síntesis hidrotérmica. Para ello, el gel de síntesis preparado con los

precursores de la zeolita se ha introducido en un autoclave (50 mL), ocupando el

80% del volumen, Una vez cerrado herméticamente, se ha introducido en una estufa

y se ha calentado a la temperatura necesaria (que genera una presión interna) durante

un tiempo estimado para la obtención de la zeolita.

Por otro lado, para la preparación de zeolitas soportadas se han utilizado como

soporte monolitos de cordierita suministrados por la empresa Corning®. Este tipo de

monolito tiene una estructura de 400 cpsi (celdas por pulgada cuadrada) y

co

(2

pr

di

Pa

m

hi

(z

ha

cr

au

re

la

de

es

omposición

2MgO·2Al2O

reviamente

iámetro de 1

ara llevar a

monolito en

idrotérmica

zeolita BET

an sido: (i)

recimiento

utoclave de

ealización d

as propiedad

el sistema

stufa.

Figura 2.1delgad

química

O3·5SiO2). L

(sección 2.2

1.4 cm y un

a cabo la s

el autoclav

. Las cond

A y tamiz m

rotación de

homogéneo

e 10 mL (

de 1, 2 o 3 e

des de la cap

experiment

9. Esquemadas de zeolit

basada e

Las caracter

2.1). Las di

na altura de

síntesis in-s

ve junto co

diciones gen

molecular d

e los autocl

o de la ca

(la relación

etapas de c

pa deposita

tal de rotac

a del sistemata sobre mo

67

en óxidos

rísticas de e

imensiones

1.6 cm.

situ de sól

on el gel de

nerales de

de SAPO-5)

laves (v = 4

apa de sóli

n “gel de

cristalización

ada [24]. En

ción emple

a de rotacióonolitos de c

de silicio

este tipo de

de los mon

lidos soport

e síntesis y

síntesis in

) soportados

4 rpm), con

do poroso

síntesis/ma

n consecuti

n la Figura 2

eado que se

ón empleadcordierita co

o, magnes

monolito h

nolitos empl

tados se ha

se ha real

-situ de só

s en monoli

n el objetivo

[24], (ii)

sa monolit

ivas, con el

2.19 se mue

e encuentra

do para prepon estructur

Capít

sio y alum

han sido des

leados pose

a introduci

lizado la sín

ólidos crista

itos de cord

o de produc

un volume

to” = 4) y

fin de con

estra un esq

a dentro de

parar pelícura celular.

tulo 2

minio

critas

en un

do el

ntesis

alinos

dierita

cir un

en de

y (iii)

ntrolar

quema

e una

ulas


68

Bibliografía.

[1] J.M. Abella, A.M. Cintas, T. Miranda, J.M. Serratosa, Introducción a la ciencia de materiales. Técnicas de preparación y caracterización, Ed. CSIC (1993). [2] F. Rouquerol, J. Rouquerol, K. Sing, Adsorption by powders and porous solids. Principles, methodology and applications, Ed. Academic Press (1999). [3] K.S.W. Sing, D. H. Everet, R.A.W. Haul, L. Moscou, R.A. Pierotti, J. Rouquerol, T. Siemieniewska, Pure & Applied Chemistry 57 (1985) 603-619. [4] F. Rodriguez-Reinoso, A. Linares-Solano, Chemistry and Physics of Carbon, 21, 1, P.A. Thrower, Ed. Marcel Dekker (1988). [5] S.J. Gregg, K.S.W. Sing, Adsorption, Surface Science and Porosity, Ed. Academic Press (1982). [6] D. Cazorla-Amorós, J. Alcañiz-Monge, M. A. de la Casa-Lillo, A. Linares-Solano, Langmuir 14 (1998) 4589-4596 [7] D. Lozano-Castelló, D. Cazorla-Amorós, A. Linares-Solano, D.F. Quinn, J. Phys. Chem. B 106 (2002) 9372-9379. [8] D. Cazorla-Amorós, J. Alcañiz-Monge, A. Linares-Solano, Langmuir 12 (1996) 2820-2824. [9] J. García-Martínez, D. Cazorla-Amorós, A. Linares-Solano, Stud. Surf. Sci. Cat. 128 (2000) 485-494. [10] S. Brunauer, P.H. Emmett, E. Teller, J. Am. Chem. Soc. 60 (1938) 309-319. [11] M.M. Dubinin, Chem. Rev. 60 (1960) 235-241. [12] P.J. Elving, J.D. Winefordner, Chemical Anaysis, Volume 19: W.W. Wendlant, Thermal Analysis 3rd, Ed. John Wiley&Sons (1986). [13] L. Hernández Hernández, C. González Pérez, Introducción al análisis instrumental Ed. Ariel Ciencia (2002). [14] D.C. Harris, Análisis Químico cuantitativo 3ª edición, Ed. Reverté (2007). [15] J.F. Shackelford, Introduction to Materials Science for Engineers. Ed. Prentice-Hall (2000). [16] A.W. Burton, K. Ong, T. Rea, I.Y. Chan, Microporous Mesoporous Mater. 117 (2009) 75-90. [17] L. Reimer, “Transmission Electron Microscopy”, Berlin Heidelberg, Ed. Springer-Verlag (1984). [18] J.I. Goldstein, D.E. Newbury, P. Echlin, D.C. Joy, A.D. Romig Jr., C.E. Lyman, C. Fiori, E. Lifshin, “Scanning Electrón Microscopy and X-Ray Microanálisis”, Ed. Plenum Press (1992).

Capítulo 2

69

[19] Ángel Berenguer Murcia. Tesis doctoral. Universidad de Alicante (2005). [20] R.E. Cunningham, R.J.J. Williams, “Diffusion in Gases and Porous Media”, Ed. Plenum Press (1983). [21] R.M. Heck, S. Gulati, R.J. Farrauto, Chem. Eng. J. 82 (2001) 149-156. [22] T.A. Nijhuis, A.E.W. Beers, T. Vergunst, I. Hoek, F. Kapteijn, J.A. Moulijn, Catal. Rev. Sci. Eng. 43 (2001) 345-380. [23] J.L. Williams, Catal. Today 69 (2001) 3-9. [24] A. Bueno-López, D. Lozano-Castello, I. Such-Basáñez, J.M. García-Cortés, M.J. Illán-Gómez, C. Salinas-Martínez de Lecea, Appl. Catal. B 58 (2005) 1–7. [25] M.A. Ulla, R. Mallada, J. Coronas, L. Gutierrez, E. Miró, J. Santamaría, Appl. Catal. A 253 (2003) 257–269. [26] L. Li, B. Xue, J. Chen, N. Guan, F. Zhang, D. Liu, H. Feng, Appl. Catal. A 292 (2005) 312-321. [27] H. Negishi, A. Endo, T. Ohraori, K. Sakaki, Ind. Eng. Chem. Res. 47 (2008) 7236-7241. [28] C.B. Ahlers, J.B. Talbot, Electrochim. Acta 45 (2000) 3379-3387. [29] M. Abdollahi, S.N. Ashrafizadeh, A. Malekpour. Microporous Mesoporous Mater. 106 (2007) 192-200. [30] Á. Berenguer-Murcia, E. Morallón, D. Cazorla-Amorós, A. Linares-Solano, Microporous Mesoporous Mater. 66 (2003) 331-340. [31] J. García-Martínez, D. Cazorla-Amorós, A. Linares-Solano, Y.S. Lin, Microporous Mesoporous Mater. 42 (2001) 255-268. [32] S. Domínguez-Domínguez, A. Berenguer-Murcia, E. Morallón, A. Linares-Solano, D. Cazorla-Amorós, Microporous Mesoporous Mater. 115 (2008) 51-60. [33] F.J. Varela-Gandía, A. Berenguer-Murcia, A. Linares-Solano, E. Morallón, D. Cazorla Amorós, Electrophoretic Deposition for The Synthesis of Inorganic Membranes, 381-393, in Membranes for Membrane Reactors. Preparation, Optimization and Selection, A. Basile and F. Gallucci. Ed. John Wiley & Sons (2011). [34] W. Shan, Y. Zhang, W. Yang, C. Ke, Z. Gao, Y. Ye, Y. Tang, Microporous Mesoporous Mater. 69 (2004) 35-42. [35] S. Mintova, J. Hedlund, B. Schoeman, V. Valtchev, J. Sterte, Chem. Comm. 13 (1997) 1193-1194.

Chapter 3. Hydrogen purification for

PEM fuel cells using membranes prepared by

ion-exchange of Na-LTA/carbon

membranes.

Chapter 3

71

3.1. Introduction.

As it has been described in Chapter 1, hydrogen economy is expected to be

implemented in the near future, due to the necessity of a clean, efficient energy

source and the growing energy crisis [1]. As mentioned, hydrogen has several

advantages as a fuel compared to conventional fossil fuels. Its combustion does not

produce pollutants such as carbon dioxide, nitrogen oxides, particles or carbon

monoxide and thus presents itself as an interesting alternative. It can be used for

mobile and stationary devices mainly in transport vehicles. Nevertheless, one of the

most significant shortcomings is the storage of the produced hydrogen, so a more

convenient source is needed as a transient solution in order to use it in polymer

electrolyte membrane (PEM) fuel cells. The main advantage of PEMs is their being

twice as fuel efficient as an internal combustion engine. They operate transforming

chemical energy into electrochemical energy, avoiding the mechanical requirements

and thermodynamic limitations of conventional engines [2]. Hydrogen and oxygen

react electrochemically and water is produced as remainder so it is a clean process.

Ideally, hydrogen may be obtained from renewable sources like water by electrolysis

but this technology is not sufficiently developed. Nowadays, hydrocarbon reforming

is the most prominent industrial process to produce hydrogen. Vehicles may carry an

onboard reformer which would produce a hydrogen stream from the reforming of

hydrocarbons like ethanol or methanol. However, in the hydrogen produced there are

some compounds that poison the platinum electrocatalyst in the anode, more

specifically sulphur and carbon monoxide. Thus, concentrations lower than 10 ppb

and 10 ppm, respectively are needed to avoid poisoning. After the reforming step and

water gas-shift reaction, the sulphur concentration is reduced to desirable levels but

another step is needed to reduce the CO concentration from 1000 to 10000 ppm (the

Hydrogen purification by zeolite membranes

72

concentration of the gaseous stream leaving the reformer) to values around 10ppm

[2].

One alternative to separate H2 and CO is the use of membranes, among which zeolite

membranes can be a good candidate. In the literature, there are many reports on

zeolite membranes, but the studies have been focused from a different point of view.

Caro and Noack [3] have presented a review about zeolite membranes focused on

separating gases like CO2 or CH4 from H2. Noble et al. [4] have studied hydrogen

purification using a SAPO-34 membrane with the aim to remove CO2 and CH4.

However, the separation of H2/CO mixtures has been solved by means of other

methods. In this case, the use of palladium and reactive membranes has been studied

[5,6]. Thus, Drioli et al. [5] described the effect produced by CO concentration in the

H2 permeation properties at different temperatures towards the design of a hydrogen

purification unit. This group [6] has also studied Pt-Y zeolite tubular membranes

prepared by the secondary growth method. This reactive membrane is used for the

preferential oxidation of CO, reducing its concentration from 10000 ppm to 10-50

ppm. However, the use of zeolite membranes in the separation of H2/CO mixtures is

a new, although extremely important matter, which has not been studied in detail yet.

In relation to zeolite membrane performance, the single gas permeation properties of

different gases have been widely studied. Nitrogen, which has a similar kinetic

diameter to carbon monoxide, has been extensively studied in several papers [7, 8].

Aoki et al. [9] studied the single gas permeation properties of a LTA membrane on a

porous α-alumina substrate obtaining a better permeance for hydrogen than for

nitrogen. Finally, Kusakabe et al. [10] have studied the permeance of different gases

in ion-exchanged LTA zeolite membranes. LTA zeolite (also known as zeolite A) is

formed by sodalite cages (β) connected through double four member rings (MR)

which form a larger cavity denominated α supercage. The former cage presents 6MR

openings of 0.2nm in size and the inner cavity that is formed has a diameter of 0.66

Chapter 3

73

nm. The latter cage is connected by 8MR openings of 0.41nm in size and the inner

cavity formed has a diameter of 1.12nm [11]. The Scheme 3.1 shows the LTA

structure of the A zeolite.

Scheme 3.1. LTA structure of A zeolite.

The original pore size (0.41 nm) may be fine-tuned by means of substituting the

cations incorporated in the structure (and which bring forth charge balance) by a

simple ion-exchange process. According to Breck et al. [11], the LTA zeolite can be

ion exchanged with different cations, such as alkaline ions. Depending on the cation,

the pore size may be changed. Initially the synthesized zeolite has sodium as

counterion (because Na+ also acts as templating agent for LTA zeolite) which

produces an effective pore size of 0.38 nm. When this zeolite is ion-exchanged with

potassium, the effective pore size is reduced to 0.28 nm. Nevertheless, a partial ion-

exchange with a calcium salt produces a larger pore size of 0.51 nm. Bearing in mind

the modification of the pore size, if the Na-form is ion-exchanged with either Rb or

Cs (which are substantially larger cations), the expected pore size would be lower

than that of the K-form. Therefore, by changing the cations, the zeolite A can be used

for different purposes.

This Chapter reports the synthesis of Na-LTA membranes supported on a carbon

material by hydrothermal treatment following the secondary growth method, as well


74

as K-, Rb- and Cs-LTA/carbon membranes prepared by means of ion-exchange.

Carbon discs were used as support due to their outstanding properties, such as

thermal and chemical stability, tunable shape, porosity and surface chemistry that

provide advantages comparing it with metal and ceramic supports [12]. The gas

permeation properties of the prepared composites were tested using small molecules

(H2 and CO) at different temperatures with the objective to understand their

usefulness for hydrogen purification for PEM fuel cells.

3.2. Experimental.


LTA/carbon membranes were prepared by the secondary growth following the

methodology described in our previous works [12]. Initially, the carbon support,

consisting of macroporous carbon sheets (thickness = 0.3 mm, mean pore size

0.7µm, geometric area 1.54 cm2) provided by Poco Graphite (DFP-1), was cut into

pieces and subjected to an acid treatment which consists in immersing the carbon

pieces in boiling concentrated nitric acid for 12 h in order to create surface oxygen

groups. Secondly, the treated carbon support was seeded by means of electrophoretic

deposition (EPD) using a LTA colloidal suspension for 30 min. The employed seeds

were prepared by hydrothermal treatment using a synthesis solution with the

following molar composition: 0.22Na2O·SiO2·1.0Al2O3·8.0(TMA)2O·400H2O. The

hydrothermal conditions were a synthesis temperature of 336 K for 63 h. Finally, a

hydrothermal synthesis was carried out over the seeded support and a thin zeolite

film was grown. The molar composition of this LTA synthesis solution was

50Na2O·5.0SiO2·1.0Al2O3·1000H2O and the hydrothermal conditions were a

synthesis temperature of 373 K for 4.5 h.

The prepared Na-LTA/carbon membranes were ion-exchanged according to the

method described by Rakoczy and Traa [13]. The membrane was immersed into a

Chapter 3

75

0.1M alkaline ion nitrate solution and heated at 333 K for 1 h under mild agitation

(60 rpm) to prevent damage to the zeolite layer.

It is important to note that Na-LTA powder zeolite has been synthesized and ion-

exchanged in the same manner, due to the fact that it is impossible to use the zeolite

on the membrane support to characterize its adsorption properties.


Firstly, LTA/carbon membranes and LTA zeolite powder (obtained from the same

autoclave) were characterized by X-ray diffraction (XRD), using a SEIFERT 2002

power diffractometer with a Cu Kα radiation. The scanning rate was 2º/min in the 5-

50º angle range.

The LTA/carbon membranes were also characterized by scanning electron

microscopy (SEM) in a HITACHI S-3000N microscope and the morphology and

thickness of the thin film prepared was studied. Energy-dispersive X-ray

spectrometry (EDX) was used to ascertain if the ion-exchange method had been

carried out successfully. Incident electron beam energies from 3 keV to 30 keV were

used.

Finally, the recovered powders of the synthesized zeolites were used to study H2 and

CO adsorption at 298K and up to ambient pressure (Micromeritics ASAP2020) with

the aim to assess their adsorptive properties. Moreover, porous texture

characterization of these zeolites was carried out by means of the adsorption of N2 at

77 K (Micromeritics ASAP2020) and CO2 at 273 K (Autosorb 6, Quantachrome).

Prior to the adsorption measurements, the samples were outgassed in vacuum at

523K for 4 h to remove any adsorbed impurities.

The permeation measurements were performed in a Wicke–Kallenbach (WK) cell.

The membrane was mounted in a stainless steel module and sealed by means of


76

polymeric O-rings to prevent leaks. The final exposed area in the WK was 0.28cm2.

Firstly, the membrane was cleaned by heating it at 423K for 6 h under a He flow of

100 ml/min on both the feed side and the permeate side with the aim of removing

impurities. The permeation tests were carried out between 298K and 423 K, with a

total pressure of 1 bar. In the single gas permeation tests, the feed gases used were

diluted in He (50% H2 or 1.25% CO). In the H2/CO mixture permeation tests, the

binary mixture was diluted in He with a composition of 50% H2 and 1.25% CO. For

both types of experiments, the total flow rate was 100 ml/min. A sweep gas (He,

flow rate of 100 ml/min) was used in order to remove the gas/gases that flow through

the membrane, hence the system is always in pseudo-equilibrium. The permeate

stream was analyzed with a mass spectrometer (Balzer, Thermostar GSD 301T). The

detection limit of this mass spectrometer is 1 ppm. For those samples where the CO

permeance is below the detection limit of the mass spectrometer (i.e. permeance

values of 0), additional measurements were done reducing the He flow rates to the

minimum rate allowed (30 ml/min), but CO concentration still remained under the

detection limit of our mass spectrometer. From the permeation experiments,

permeances were calculated considering the partial pressure difference between the

retentate and permeate sides [14].


3.3.1. Crystal structure, morphology and composition.

To confirm the structure of the synthesized membranes as that of LTA zeolite, XRD

analyses were performed on the samples. Moreover, XRD patterns of the powders

recovered by filtration from the reaction autoclaves were also collected. Figure 3.1

presents the XRD spectra of an as-synthesized Na-LTA/carbon membrane (Figure

3.1A) and powder LTA zeolites (Figure 3.1B). The characteristic diffraction peaks

for the crystalline Na-LTA zeolite are observed in Figure 3.1B (see peaks marked

Chapter 3

77

with +). The XRD pattern corresponding to the as synthesized Na-LTA/carbon

membrane (Figure 3.1A) shows that pure LTA crystals have grown on the surface of

the carbon support (see peaks marked with +). Moreover, peaks from the carbon

support were also detected (see peaks marked with * in Figure 3.1A); the sharp, very

intense peak appearing at 26.2º and the broad peak at 42.1º correspond to the

graphite support, as it has been described elsewhere [12].

Figure 3.1. (A) X-ray diffractogram of an as-synthesized LTA/carbon membrane and (B) X-ray diffractograms of ion-exchanged zeolite A powders.

Similar to the Na-LTA materials (carbon supported membrane and zeolite powder),

the spectra corresponding to the K-LTA materials showed most of the characteristic

peaks of the LTA structure. Nevertheless, the Rb- and Cs-forms displayed a marked

decrease in both, the number and intensity of the XRD peaks. To observe the effect

of ion-exchange in crystallinity when Na+ is exchanged with other alkaline cations,

Figure 3.1B includes the diffractograms for the ion-exchanged powder zeolite A. It

can be seen that there is a substantial decrease in crystallinity when Na+ is exchanged

with other alkaline cations. In this respect, some peaks disappeared from the

spectrum. In the case of the Rb-LTA sample, the diffractogram did not show any

peaks at all. It should be noted that when the ion exchange took place in the case of

Rb and Cs, the employed salts contain trace amounts of Ba2+, which has been

reported to significantly alter the LTA structure [11,15]. According to those results,

full exchange with Rb+ results in the zeolite becoming amorphous, with Ba2+ being


78

found on the outer edge of the crystals, leaving a “Ba-free core” [15]. This

phenomenon, however, did not seem to significantly affect the membrane properties

(vide infra). According to previous reports, the inclusion of a larger cation on the

zeolite network can partially distort the structure because it must be accommodated

into the structure [16], which may explain the shifting of peaks or change in their

relative intensities. Nevertheless, these structural changes need further research that

will be done in the future. In spite of the fact that LTA crystallinity is reduced by

ion-exchange (especially for Rb), for simplicity, we will refer to these materials as

M-LTA/carbon membranes, where M is the cation loaded by ion-exchange.

The equation postulated by Scherrer predicts that the average crystal size is related to

the diffraction peak half-width. Although a critical evaluation of its application to

zeolites with one-dimensional pore systems has revealed significant errors [17], in

the case of LTA zeolites the equation applies satisfactorily. This equation can be

expressed as follows:

cos (3.1)

where t is the averaged dimension of crystals, K is a constant near unity which

depends on the particle shape, taken as 1.00 [17] in our case, λ is the X-ray

wavelength, β is defined as the width at half maximum of the peak and θ is the

position (angle) of the peak in radians. Peaks of moderately high intensity are needed

to obtain good results. Therefore, using this equation, the crystal size of the powder

zeolite obtained during the membrane synthesis has been estimated to be between

0.5µm and 0.6µm.

SEM images of both the top view and the cross-section were taken in order to obtain

the general view of the membranes and to estimate the thickness of the deposited

zeolite layer. Figure 3.2 shows, as an example, the SEM images corresponding to the

as-synthesized Na-LTA/carbon membrane. It can be seen from the top view that all

Chapter 3

79

the carbon surface is homogeneously covered with zeolite A crystals and from the

cross-section view (inset), the estimated zeolite layer thickness is 2-3µm.

Figure 3.2. Top view and cross-sectional view (inset) of the zeolite A supported on

the carbon disc.

Finally, EDX analyses were performed to complete the chemical characterization of

the different membranes and to confirm that the ion-exchange method had been

carried out successfully. The prepared Na-LTA/carbon membranes were ion-

exchanged for 1 h to obtain their corresponding alkaline ion form. The extent to

which the ion-exchange had been carried out could be measured semi-quantitatively

by means of EDX analyses of the zeolite crystals supported on different membrane

areas. Figure 3.3 includes representative EDX spectra for each sample. It is shown

how the peak appearing at 1.040 keV corresponding to the Kα line of sodium

significantly decreases its intensity when the sample is ion-exchanged with alkaline

ions. Despite the fact that in some cases overlapping between the analysis lines is

observed (for instance, between the L1 line of rubidium and the K line of Si at 1.83

keV), quantification of the signals was used in order to determine the extent to which

ion-exchange had occurred. An average taken from different measurements indicates


80

that the exchange had been performed to a large extent, reaching almost 100% in

some cases.

Figure 3.3. Energy-dispersive X-ray spectrometry (EDX) analysis corresponding to

ion-exchanged LTA/carbon membranes: (A) Na-LTA/carbon; (B) K-LTA/carbon; (C) Rb-LTA/carbon; (D) Cs-LTA/carbon.

In all the cases, the maximum expected degree of ion-exchange was achieved

successfully. It was about 96% in the case of the K-form, 70% in the Rb-form and

70% for the Cs-form. According to Breck et al. [11], the maximum ion-exchange is

related to the size of the different cations and their respective location in the zeolite

cages, but this point will be extensively discussed later, in the permeation results.

Other authors have studied the ion-exchange with Rb salts [18,19]. Yan and Bein

[18] have studied the ion-exchange with different cations in a thin film made by LTA

crystals and silica over a gold surface modified with siloxane. The ion-exchange

quantification was monitored by the mass change in a quartz crystal microbalance,

Chapter 3

81

obtaining the same ion-exchange for Rb than that obtained in the present study.

Barrer et al. [19] have used powder LTA zeolite carrying out ion-exchange with a Rb

salt, obtaining an ion-exchange of 70%.

From a practical point of view, it is of utmost importance to determine whether our

starting hypothesis (i.e. that ion-exchanged LTA membranes are able to separate CO

from H2 whereas Na-LTA is not) is accomplished and, then, the characterization of

the adsorption properties becomes relevant. In order to study the effect of the ion-

exchange in the H2 and CO adsorption properties, H2 and CO adsorption isotherms at

298K and up to ambient pressure were performed for all the materials. Figure 3.4

contains the obtained isotherms. Regarding H2 adsorption (Figure 3.4A), it can be

seen that, in all the zeolites studied in the present work, the amount of H2 adsorbed at

298K and up to 1 atm is the amount expected according to their micropore volume

[20,21]. It can be also observed that, despite reducing the effective pore size in the

K- and Rb-exchanged LTA zeolites, hydrogen adsorption has been not affected

significantly, showing approximately the same hydrogen adsorption properties than

the Na-LTA zeolite. However, the H2 adsorption properties of the Cs-LTA zeolite

have been affected noticeably, decreasing the adsorbed amount by almost 50%. This

observation will be discussed in detail in a subsequent section. Figure 3.4B contains

the CO adsorption isotherm corresponding only to the Na-LTA zeolite. The CO

adsorption isotherms corresponding to K-, Rb- and Cs-exchanged LTA zeolites are

not included because these materials did not present any CO adsorption under the

studied conditions. This indicates that the pores of these samples are too small to host

carbon monoxide, which is in direct agreement with the starting principle of this

work.


82

Figure 3.4. (A) H2 adsorption isotherms at 298 K corresponding to Na-LTA and ion-

exchanged LTA powders and (B) CO adsorption isotherm at 298 K for Na-LTA powder.

The microporous texture was characterized by N2 adsorption at 77 K and CO2

adsorption at 273 K. Table 3.1 contains the narrow micropore volume (VDR (CO2))

(pore size smaller than around 0.7 nm) obtained by applying the Dubinin-

Radushkevich (DR) equation to the CO2 (273 K) adsorption data. The density used

for adsorbed CO2 at 273 K was 1.023 g/ml, and the affinity coefficient used was 0.35

[22,23]. Results obtained from N2 adsorption at 77 K have not been included in

Table 3.1 because the kinetics of N2 adsorption was very slow at 77 K for these

samples, and extremely long times were necessary to reach the equilibrium at each

point of the isotherm, resulting in VDR values of zero or within its observed

experimental error. That was expected considering that the Na-LTA zeolite pore size

is already smaller than the critical value which N2 molecules can access at 77 K

(around 0.4 nm) [24]. However, in this narrow porosity CO2 adsorption at 273 K

Chapter 3

83

occurred more readily. As it has been shown in many previous studies using porous

carbon materials and also zeolites, CO2 adsorption at 273 K is an interesting

complement to N2 adsorption for the assessment of the narrow microporosity, where

diffusional problems of the N2 molecules (77 K) inside the narrow porosity range

(size < 0.7 nm) occur [22-25]. For this type of samples, CO2 becomes significantly

more useful than N2 (77 K) since it may access the ultramicropores present in Na-

LTA and ion-exchanged LTA zeolites. According to the results presented in 3.3.1,

the ion-exchange performed in order to modify the pore size has a notorious

influence in the CO2 adsorption properties at 273 K. In this respect, the results

indicate that the Na-LTA and K-LTA samples have the same micropore volume

obtained using the DR equation, whereas, the Rb-, Cs-LTA zeolites do not present

any appreciable micropore volume. This indicates that ion-exchange with bulky

cations renders the zeolite porosity inaccessible to CO2 adsorption at 273 K.

Table 3.1. Porous texture characterization results Sample VDR CO2 (cm3/g) Na-LTA 0.16 K-LTA 0.18 Rb-LTA 0.03 Cs-LTA -

3.3.2. Single gas permeation tests.

The single gas permeation tests for H2 and CO were performed separately for each

membrane at three different temperatures (298 K, 398 K and 423 K). These tests

were carried out for 2 h (time after which a quasi-steady state was reached). Prior to

the permeation experiments, the membranes were heated up to 423 K for 6 h under a

He flow of 100 ml/min in order to dehydrate the membrane and remove any

contaminants and/or adsorbed species. Moreover, a dwell time of 2 h was always

applied every time the temperature was increased up to 398 K or 423 K in order to

have a homogeneous temperature profile in the membrane before the permeation


84

tests. Figure 3.5 includes the values of permeances of H2 and CO for the different

membranes of the present study (as-prepared Na-LTA/carbon membrane and the ion-

exchanged LTA/carbon membranes).

From Figure 3.5A, it becomes apparent that H2 permeance is not considerably

affected by the ion-exchange applied to the as-prepared Na-LTA membrane. These

results are quite concordant with those presented in Figure 3.4A, where it was seen

that all powdered samples showed appreciable H2 adsorption capacities. However, a

closer look to the results points out that the differences shown in Figure 3.4A

between the four samples are much more important compared to the relatively

similar H2 permeance values presented in Figure 3.5A. This fact can be a

consequence of the Wicke-Kallenbach cell operating in a quasi-equilibrium regime.

Thus, a sample that has a low adsorption capacity but fast kinetics of adsorption may

display a permeation behavior similar to samples having a more pronounced

adsorption capacity but with much slower kinetics of adsorption.

Regarding the effect of temperature, H2 permeance suffers some decrease in the case

of the Na-LTA membrane when the temperature increases from 298K to 398 K,

which may be interpreted as permeance being dominated by adsorption at room

temperature in Na-LTA zeolite. When the temperature is increased, this process is

hindered noticeably. At temperatures between 398K and 423K the adsorption and

diffusion mechanisms seem to become equally dominant, and thus permeance in that

temperature range is not affected significantly. Nevertheless, this is a minor

contribution being the main mechanism responsible for the separation the molecular

sieve. These results are in concordance with those obtained by Aoki et al. [9], who

studied the hydrogen permeance behavior at different temperatures using Na-LTA

membranes. This tendency, however, seems to become compensated in the case of its

ion-exchanged counterparts, in which the changes in single gas permeation behaviors

with temperature become less dramatic.

Chapter 3

85

Figure 3.5. Single gas permeance: (A) H2 permeance and (B) CO permeance.

The results of CO adsorption isotherms at 298K pointed out that K-, Rb- and Cs-

exchanged LTA zeolites have too small pores to adsorb CO, while Na-LTA zeolite

adsorbs a considerable amount of CO (see Figure 3.4B). These aspects are further

corroborated by the CO permeance results included in Figure 3.5B. Thus, in the case

of the Na-LTA/carbon membrane, CO permeates through the membrane with little

impediment, while as the size of the exchanged cation increases, CO permeance

seems to decrease. In the case of K-LTA and Rb-LTA, CO permeance is lower than

in the Na-LTA membrane. In agreement with Figure 3.4A, the negligible CO

adsorption in the zeolite pores might indicate a material impermeable to CO. Thus,

this low permeance might be due to the presence of defects like pinholes in the

membrane. The case of the K-LTA membrane will be discussed below. For the Cs-

LTA membrane, no CO was observed in the permeate side of the membrane at high


86

temperatures, hinting at its possessing better properties when the mixture H2/CO will

be used.

Morooka et al. have studied in detail the single gas permeance properties of LTA

membranes [9,10,26,27]. Our permeation results are in agreement with their single

gas permeance values [10]. The LTA membranes reported in their work were

prepared on porous α-alumina tubes. The aim of that research was improving the

separation factor of H2 over n-C4H10. This was achieved by repeating the synthetic

procedure, producing a thicker zeolite layer which restricted the H2 flow through the

membrane, obtaining a H2 permeance of 0.8 x 10-8 molm-2s-1Pa-1 to 2.6 x 10-8

molm-2s-1Pa-1 at 308 K. Bearing in mind these values, our permeance results are in

agreement with the LTA membrane prepared by Aoki et al. after 4 syntheses. In our

case, the optimization of the permeance values has been performed with only one

synthesis and the modification of the pore size opening with different alkaline

cations without dramatic changes in the H2 permeance values. It should be mentioned

that ion-exchange was also applied in Ref. [10] with potassium and calcium salts

with views towards the improvement of the ideal separation factor. The introduction

of a larger cation which can completely exchange the Na+ ions (such as K+) tended to

produce a considerable reduction in the permeance values.

3.3.3. H2/CO mixtures permeation tests in LTA/carbon membranes.

In order to test the prepared membranes under realistic conditions, a gas stream

containing both H2 and CO (H2:CO = 50ml/min:1.25 ml/min, rest He, total flow 100

ml/min) was fed to the membrane. Each permeation experiment was performed for

several hours. Firstly, the prepared membrane was tested at 298K until quasi-

stationary steady state was reached. After that, the temperature was increased up to

398K and the permeation properties were monitored for 2 h. The temperature was

increased once again up to 423 K and the membrane was tested under these

Chapter 3

87

conditions. Figure 3.6 summarizes the permeance results of the prepared membranes

(Na-, K-, Rb- and Cs-LTA).

Figure 3.6. H2/CO mixtures permeance: a) H2 permeance; and b) CO permeance.

The as-synthesized Na-LTA/carbon membrane, presents a pore size of 0.38 nm. The

total number of cations presents in the LTA structure is 12, located in the sodalite

cages and in the supercage α. In the case of the dehydrated zeolite, there are 8 Na

cations located in the β-cage, 3 cations located in the opening of the 8MR which

results in a partial blocking of the aperture and the remaining cation is located inside

the supercage near the 4MR [11]. These results are in agreement with simulations

performed in recent years [28,29]. In these studies, molecular dynamics simulations

have been used to optimize the zeolite structure parameters and the positioning of the

sodium ions for the LTA zeolite unit cell (Na96Al96Si96O384). It should be noted that

when the ratio of pore diameter to molecule diameter is near unity, permeation is

affected by the molecular structure and the affinity between permeating molecules


88

and the pore walls [10]. Moreover, in agreement with the results presented in Figure

3.6, using Na-LTA/carbon membranes it is not possible to separate H2 from CO. This

material has a separation factor (Sf) of 1.0 at 298 K. As shown in Figure 3.4, CO can

be adsorbed in the Na-LTA zeolite pores more strongly than H2. Considering that its

(pore size/molecule diameter) ratio is higher than 1 in the case of both H2 and CO, in

principle it is not possible to separate them at room temperature. After that, when the

temperature is increased to 398 K, the CO adsorption is lower and the diffusion

becomes more relevant, so there is an increase in the permeance value. In relation to

the H2 permeation, the results are in agreement with those of Guan et al. [10]. In this

case, the H2 is adsorbed weakly, so the diffusion through the zeolite pores is

predominant at all temperatures.

The as-synthesized membranes prepared were ion-exchanged with alkaline salts. In

the case of the K-LTA membrane, all the cations present in the structure are ion-

exchanged (the degree of ion-exchange was about 96% in the case of the K-form, as

indicated in Section 3.3.1). At this point, the location of the exchanged cations

becomes of critical importance in order to understand how the permeation properties

may be affected. The potassium cations are situated in the same position as the

corresponding Na+ ions and, therefore, the zeolite pore size is reduced to 0.28 nm.

This statement is in agreement with the results obtained in the permeation tests. In

this case, the reduction in the pore size causes the H2 and CO permeance to be

reduced in comparison with the Na-form. When the experiment is performed at room

temperature, it is possible to separate H2 from CO, not observing any CO permeation

under the conditions in which the permeation tests were carried out. This fact can be

explained considering that the H2 and CO minimum radii are 0.24nm and 0.34 nm,

respectively. Bearing in mind that the zeolite effective pore size and the CO radius

are similar, the effect of temperature in CO diffusion through the membrane pores

becomes relevant, as it can be seen from the results obtained at 398K and 423 K: the

Sf value is reduced drastically down to 3.0 at 398 K and 0.6 at 423 K and clearly

Chapter 3

89

these values are insufficient to reach the target of purifying H2 flow with a

concentration of CO lower than 10 ppm. These values are in agreement with the

affirmation of Gaffney [30] that the potassium exchanged adsorbents (e.g. zeolite X)

are good for low temperature adsorptive applications as oxygen purification.

Consequently, this kind of membranes would only become useable at room

temperature, when the membrane can effectively separate H2 from CO.

With the aim to obtain a membrane which is able to separate H2 from CO, the Na-

LTA/carbon membrane was ion-exchanged with Rb and Cs salts. One of the first

studies reported in the literature [11] concluded that the Na+ cations located in the

supercages are exchangeable with larger cations such as Rb or Cs. The concentration

of alkaline cations was determined by flame spectrophotometry after the zeolite was

dissolved by acid treatment. Other authors [31,32] have used IR spectroscopy to

establish the ring associations and the approximate positions of some of the larger

univalent cations in LTA zeolite. Their studies concluded that only those Na+ ions

that were loosely bound may be exchanged by Rb+ or Cs+. These cations, in turn

interact relatively weakly with the zeolite framework, and they tend not to be

associated with the smaller openings (6-rings) of the zeolite framework. Further

studies by means of XRD analysis [15,31] have revealed that Rb and Cs can be

exchanged in both the sodalite cage and in the supercage. In both cases, the alkaline

cations seem to substitute Na+ to a significant extent. Vance and Seff [31] concluded

that Cs+ ions are located within the supercages and the porous channels, and not only

on the external surface, as it was originally thought. Recent studies [33] have

analyzed the ion-exchange ability of LTA zeolite by incorporating Cs+ in its

framework and then exchanging it with a series of different cations. Their results

showed that Cs+ occupy the internal cages and the channels. Early studies on the

location of Rb+ in LTA zeolites [34] determined that up to 11 out of 12 Na+ cations

are exchanged in the framework when subjected to a flow of 0.1M aqueous RbOH.

Thus, the Rb+ cations may be found practically in all available positions, even inside


90

the sodalite cages. Other reports [15] suggested that only 6 Na+ ions are exchanged

by Rb+ due to 2 Ba2+ ions being incorporated in the zeolite structure, with the

aforementioned distortion to the zeolite structure. Recent studies performed by

means of synchrotron XRD analysis on Rb-exchanged zeolite Y agree that Rb+ ions

may be found within the sodalite cage structure [35].

The results of the Rb-LTA membrane (see Figure 3.6) are very similar to those of the

K-LTA membrane. From this observation it would appear that the Rb+ cation

location in the zeolite framework structure is similar to that in the K-LTA membrane.

Therefore, as only 70% of the Na+ cations are exchanged, a part of the Rb+ cations

may be occupying the sodalite cages [15,33]. This statement is in agreement with the

H2 permeance results being similar to the K-LTA membrane. In the Rb-LTA

membrane, the Rb+ cations produce a reduction in the permeation flow but the main

advantage is the suppression of the CO flow under all temperatures tested. As a

result, the Rb-LTA/carbon membrane is suitable to separate H2 from CO at all

temperatures tested with a Sf close to ∞ (i.e. no CO permeance can be observed, thus

a pure H2 stream may be obtained). It should be mentioned that considering the

structure changes in the LTA framework after Rb-ion-exchange, the separation might

not occur exclusively through the LTA structure.

In order to obtain H2 flows with permeance values close to those obtained with the

Na-LTA/carbon membrane, the Cs-LTA membrane was tested. In this case, it must

be noted that from our permeation results it appears the Cs+ cations are located

mostly in the supercages, with only a few positions in the sodalite cages being

occupied by Cs+ ions. Thus, the porosity in the sodalite cages is not significantly

affected. On the other hand, the supercages are partially blocked due to the

introduction of a cation larger than Na or K, as our results show. Due to the larger

size of Cs+ ions, they partially block the porosity, and as a result there is no CO flow

through the membrane, whereas H2 permeation is unaffected. In this case, ion-

Chapter 3

91

exchange with Cs+ causes H2 to be adsorbed more weakly in the zeolite than its

sodium counterpart (see Figure 3.4A). Therefore, the process that mainly determines

permeation behavior in the case of Cs-LTA/carbon membranes is diffusion. The Cs-

LTA membrane behavior may also be related to distortion of the zeolite structure due

to the ion-exchange process with Cs+ (see Figure 3.1B and its discussion), which

severely hinders CO permeance. So, it is possible to separate H2 from CO with the

Cs-LTA membrane and obtain a purified H2 flow with similar permeance values as

those observed for the Na-LTA/carbon membrane.

Analyzing H2 permeation in detail, it can be seen that it is significantly influenced by

the cation present in the structure. Firstly, the Na-form can be taken as a reference

with a H2 permeance average of 1.7×10-8 molm-2s-1Pa-1 at 298 K. When all the Na

cations are replaced by K cations, the pore size is reduced drastically and the

tortuosity of the channels is larger due to the channel opening having been reduced to

0.28nm resulting in a more difficult pathway for the molecules to pass through the

membrane pores. As a result, H2 permeance becomes reduced considerably due to its

passing through the membrane pores being hindered and the resulting average value

obtained is 1.1×10-8 molm-2s-1Pa-1 at 298 K. Finally, the Rb and Cs replace partially

(although to a considerable extent) the Na cations. In the case of Rb+, H2 permeance

is reduced in a way similar to that in the K-form. H2 can pass through both

membranes without an appreciable difference between the two different composites.

However, the presence of Rb+ instead of K+ permits that the CO cannot pass through

the membrane pores, and the Rb-form can operate at higher temperatures. The

average permeance observed is 1.2×10-8m-2s-1Pa-1 at 298 K. In the case of Cs-LTA

membrane, the average permeance observed is 1.4×10-8 molm-2s-1Pa-1 at 298 K

whereas no CO permeance is observed. The difference between these values is in the

partial ion exchange with Rb+ or Cs+ and the different cation accommodation in the

structure that results in a larger H2 permeance in the Cs than the Rb-form at all

temperatures tested. From these results it can be said that the materials obtained in


92

the present study are promising, especially in terms of selectivity, although further

efforts need to be undertaken in order to improve the flux through the composite

materials.

Chapter 3

93

3.4. Conclusions.

Zeolite layers based on zeolite LTA deposited on porous carbon discs have been

successfully prepared following an established approach. Subsequent ion-exchange

carried out by a simple, reproducible methodology allowed to modify the as-grown

Na-LTA zeolite into K-, Rb-, Cs-forms, tailoring the porosity of the final material.

Our permeance results show good agreement with previous literature reports referred

to Na- and K-LTA zeolite. Single gas permeation tests reveal that the Cs-LTA

membrane can separate H2 from CO at 423 K, not observing any CO permeance. In

the case of Rb-LTA membranes, at all temperatures tested, and Cs-LTA membranes,

at room temperature and 398 K, the presence of a low CO permeance value seems to

be due to the presence of some defects (intercrystalline spaces, pinholes) in the

zeolite membrane.

Concerning the permeation tests performed with mixtures, our as-prepared Na-

LTA/carbon membrane is unsuitable for a H2 purification process. The K-LTA

membrane can be used at room temperature, but the H2 flow obtained is lower that

the Na-LTA membrane. The best results have been obtained with the Rb-LTA and

Cs-LTA/carbon membrane working in a Wicke-Kallenbach cell at all temperatures

tested. Nevertheless, the main results obtained are those referred to the Cs-LTA

membrane. In this case, the H2 permeation flow obtained is similar to the Na-

LTA/carbon membrane but without CO permeation. Thus, a H2 rich flow with a CO

concentration lower than 10 ppm, which is mandatory in order to use it in PEM fuel

cell, can be reached with the Cs-LTA/carbon membrane.

The new materials prepared in this study, Rb-LTA and Cs-LTA/carbon membranes,

are promising materials for its use as alternative in the H2 purification towards their

use in PEM fuel cells, especially in terms of selectivity, although further efforts need

to be undertaken in order to improve the flux through the composite materials.


94

References

[1] N.W. Ockwing, T.M. Nenoff, Chem. Rev. 107 (2007) 4078-4110. [2] R. Farrauto, S. Hwang, L. Shore, W. Ruettinger, J. Lampert, T. Giroux, Y. Liu, O. Ilinch, Annu. Rev. Mater. Res. 33 (2003) 1-27. [3] J. Caro, M. Noack, Micropor. Mesopor. Mater. 115 (2008) 215-233. [4] M. Hong, S. Li, J.L. Falconer, R. Noble, J. Membr. Sci. 307 (2008) 277-283. [5 P. Bernardo, C. Algieri, G. Barbieri, E. Drioli, Sep. Purif. Technol. 62 (2008) 629-635. [6] F. Scura, G. Barbieri, G. De Luca, E. Drioli, Int. J. Hydrogen Energy 33 (2008) 4183-4192. [7] J. Zah, H.M. Krieg, J.C. Breytenbach, J. Membr. Sci. 287 (2007) 300-310. [8] Y. Cui, H. Kita, K.-I. Okamoto, J. Mater. Chem. 14 (2004) 924-932. [9] K. Aoki, K. Kusakabe, S. Morooka, J. Membr. Sci. 141 (2008) 197-205. [10] G. Guan, K. Kusakabe, S. Morooka, Sep. Sci. Technol. 36 (2001) 2233-2245. [11] D.W. Breck, W.G. Eversole, R.M.Milton, T.B. Reed, T.L. Thomas, J. Am. Chem. Soc. 78 (1956) 5963-5972. [12] S. Dominguez-Dominguez, A. Berenguer-Murcia, E. Morallon, A. Linares-Solano, D. Cazorla-Amoros, Micropor. Mesopor. Mater. 115 (2008) 51-60. [13] R.A. Rakoczy, Y. Traa, Micropor. Mesopor. Mater. 60 (2003) 69-78. [14] T. Seike, M. Matsuda, M. Miyake, J. Mater. Chem. 12 (2002) 366-368. [15] J.J. Pluth, J.V. Smith, J. Am. Chem. Soc. 105 (1983) 2621-2624. [16] R. Khaleghian-Moghadam, F. Seyedeyn-Azad, Micropor. Mesopor. Mater. 120 (2009) 285-293. [17] A.W. Burton, K. Ong, T. Rea, I.Y. Chan, Micropor. Mesopor. Mater. 117 (2009) 75-90. [18] Y. Yan, T. Bein, J. Am. Chem. Soc. 117 (1995) 9990-9994. [19] R.M. Barrer, L.V.C. Rees, D.J. Ward, Proc. R. Soc. Lond. A 273 (1963) 180-197. [20] M. Jorda-Beneyto, F. Suarez-Garcia, D. Lozano-Castello, D. Cazorla-Amoros, A. Linares-Solano, Carbon 45 (2007) 293-303. [21] M. Jorda-Beneyto, M. Kunowsky, D. Lozano-Castello, F. Suarez-Garcia, D. Cazorla-Amoros, A. Linares-Solano, Hydrogen storage in carbon materials, in: A.P. Terzyk, P.A. Gaude, P. Kowalczyk (Eds.), Carbon Materials: Theory and Practice, Research Signpost, India, 2008, pp. 245–281. [22] D. Cazorla-Amoros, J. Alcañiz-Monge, A. Linares-Solano, Langmuir 12 (1996) 2820-2824.

Chapter 3

95

[23] D. Cazorla-Amoros, J. Alcañiz-Monge, M.A. De la Casa-Lillo, A. Linares-Solano, Langmuir 14 (1998) 4589-4596. [24] D. Lozano-Castello, D. Cazorla-Amoros, A. Linares-Solano, D.F. Quinn, J. Phys. Chem. B 106 (2002) 9372-9379. [25] J. Garcia-Martinez, D. Cazorla-Amoros, A. Linares-Solano, Stud. Surf. Sci. Catal. 128 (2000) 485-494. [26] K. Aoki, K. Kusakabe, S. Morooka, AIChE J. 46 (2000) 221-224. [27] K. Aoki, K. Kusakabe, S. Morooka, Ind. Eng. Chem. Res. 39 (2000) 2245-2251. [28] D.A. Faux, W. Smith, T.R. Forester, J. Phys. Chem. B 101 (1997) 1762-1768. [29] E. Jamarillo, M. Chandross, J. Phys. Chem. B 108 (2004) 20155-20159. [30] T.R. Gaffney, Curr. Opin. Solid State Mater. 1 (1996) 69-75. [31] T.B. Vance Jr., K. Seff, J. Phys. Chem. 79 (1975) 2163-2167. [32] I.E. Maxwell, A. Baks, The influence of exchangeable cations on zeolite framework vibrations, in: W.M. Meier, J.B. Uytterhoeven (Eds.), Molecular Sieves, Ed. ACS (1973) pp. 87-95. [33] J. Mon, Y. Deng, M. Flury, J.B. Harsh, Micropor. Mesopor. Mater. 86 (2005) 277-286. [34] R.L. Firor, K. Seff, J. Am. Chem. Soc. 99 (1977) 1112-1117. [35] G.L. Marra, A.N. Fitch, A. Zecchina, G. Ricchiardi, M. Salvalaggio, S. Bordiga, C. Lamberti, J. Phys. Chem. B 101 (1997) 10653-10660.

Chapter 4. Zeolite A/carbon

membranes for H2 purification from a

simulated gas reformer mixture.

Chapter 4

97

4.1. Introduction.

In Chapter 3, LTA zeolite supported on carbon membranes were used to study their

viability and usefulness in H2 purification by filtering off the CO in a gaseous stream.

Na-LTA/carbon membranes were ion-exchanged by a simple, reproducible

methodology, tailoring the porosity of the final membrane. Our results showed that

the Na-LTA/carbon membrane was unsatisfactory for H2 purification processes. On

the contrary, the Cs-form has the advantage that its permeation values are similar to

those reported for the Na-form but without any apparent CO permeation. Therefore,

the Cs-form was deemed suitable for H2 purification processes.

To gain a real understanding about the separation properties of membranes it

becomes desirable to use a feed as similar as possible to the composition of the gas

mixture in the intended application. Therefore, this work deals with the use of the

Na-LTA and Cs-LTA/carbon membranes under simulated reformer gas mixtures.

Thus, a feed stream containing 50% H2, 1.25% of CO, and a CO2 concentration

between 2 and 20% was used for the permeation experiments. In order to further test

the prepared materials, a simulated reformer mixture, containing 50% H2, 1.25% CO,

20% CO2 and 5% H2O in He was used to study the permeation properties of the

selected membranes. With the aim to gain a better understanding of the obtained

results, the interaction between CO2 and the zeolite layer was studied by infrared

spectroscopy.

4.2. Experimental.


Na-LTA/carbon membranes were prepared following the preparation method

described in Chapter 3. Briefly, the treated carbon support was seeded by means of

electrophorethic deposition, EPD, (see Chapter 2) using a LTA colloidal suspension.


98

Secondly, a LTA zeolite thin film was grown on the seeded carbon support by using

the secondary growth method. The as-made membranes were ion exchanged with a

Cs salt following the methodology described in [1]. The cesium form of the zeolite

membranes was obtained by their immersion in a 0.1 M CsNO3 solution and heated

at 333 K for 1 h under mild agitation (60 rpm) to prevent damage to the zeolite layer.

Furthermore, it is important to mention that the zeolite deposited on the carbon

surface cannot be used to characterize the zeolite properties by different techniques

without damaging the integrity of the layer and the amount of powder zeolite

recovered for each synthesis was insufficient for characterization purposes. For that

reason, commercial LTA zeolite powder (molecular sieve 4A powder < 5 µm,

Aldrich, which possessed an identical XRD pattern to the zeolite powder recovered

from the autoclaves) has been used and ion-exchanged in the same manner to use it

in the study of the interaction between the zeolite and CO2.


The permeation measurements were performed by duplicate to assess the

reproducibility of the results in a Wicke–Kallenbach (WK) cell which has been

explained in detail in Refs. [2,3] and Chapter 2. Scheme 1 shows the experimental

system used for these experiments. In this case, the feed side has been modified in

order to have a dry or humid stream. Prior to the experiments, and following our

results obtained analyzing powder Na-LTA samples in TG-MS experiments [2], the

membrane was out-gassed and dried by heating it at 423 K for 6 h under a constant

He flow of 100 ml/min on both the feed and permeate side. The permeation

experiments were carried out at three different temperatures, 303, 398 and 423 K,

with a system pressure of 1 bar. Permeation of individual gases (H2 and CO) was

performed as in Chapter 3. Additionally, two different kinds of experiments were

carried out to test the membrane permeation properties:

Chapter 4

99

(i) To study the effect of CO2 concentration in the feed (feed composition of

50% H2, 1.25% CO and n% CO2 diluted in He, where n% = 2, 5, 10, 15 and

20% CO2).

(ii) To study the effect of the use of a simulated reformed mixture as feed

(composition of 50% H2, 1.25% CO, 20% CO2 and 5% H2O, balance He).

In order to obtain a feed flow with the desired amount of water vapour the feed gases

were bubbled through a Dreschler bottle (see Scheme 4.1), which contained water

and was kept at 306 K by means of an oil bath. For both types of experiments the

total flow rate was 100 ml/min. The gases were swept with He (100 ml/min) at the

permeate side and analyzed with a mass spectrometer (Balzer, Thermostar GSD

301T) and a gas chromatograph (Agilent Technologies 6890N equipped with a

carbon-PLOT column operating at 303 K and a TCD detector). The detection limit of

this mass spectrometer is 1 ppm. For those samples where the CO permeance is

below the detection limit of the mass spectrometer (i.e. permeances values of 0),

additional measurements were done reducing the He flow rates to the minimum rate

allowed (30 ml/min), but CO concentration still remained under the detection limit of

our mass spectrometer. From the permeation experiments, permeances were

calculated considering the partial pressure difference between the retentate and

permeate sides.

Scheme 4.1. Experimental system employed to do the permeation measurements.


100

To explain the CO2-zeolite interaction several experiments were performed. First, the

CO2-zeolite interaction was studied combining in situ diffuse reflectance infrared

Fourier transform spectroscopy (DRIFTS) (with controlled environment chamber)

and thermogravimetric analysis (TGA). A FTIR spectrophotometer (model Infinity

MI60 from Mattson) with a diffuse reflectance accessory (model Collector from

SpectraTech) was used. The resolution was 4 cm-1 and 64 scans were collected for

each spectrum. Prior to all DRIFTS experiments, all the powder samples were

outgassed by a treatment under a He flow of 30 ml/min at 423 K for 6 h. A sequence

of heat treatments up to the temperature of study (T = 303 K, 398 K or 423 K) under

different atmosphere (He or CO2) has been performed for all the powder samples.

Scheme 4.2 includes this sequence with the details of the experimental conditions

(full dots indicate when the DRIFT spectra were collected). Firstly, the sample was

heat treated up to the temperature of study (T) for 4 h and DRIFT spectra were

collected at this temperature and also at 303 K (before and after heating). Afterwards,

the sample was heat treated up to the same temperature (T) under CO2 atmosphere

and the temperature was held for 2 h. Then, the gas was switched to He and after 30

min four spectra were collected every 30 min. The last stage was a heat treatment up

to 573 K for 1 h. The last DRIFT spectrum was collected at room temperature in He

after the last thermal treatment.

Scheme 4.2. FTIR experimental conditions. Full dots indicate when the spectra were

collected.

Tem

pera

ture

TimeT = 303K, 398K or 423K

4 h

T

573K

5K/min

5K/min

1 h

1 h

CO2He He

4 h

T

5K/min

Chapter 4

101

The TGA experiments were performed in a Thermogravimetric Analyzer (TA

Instruments, model SDT 2960). In these experiments, approximately 10 mg of the

zeolite were treated. Firstly, the sample was heat-treated at 423 K for 2 h with a

heating rate of 5 K/min, under a constant flow rate of 80 ml/min in He. At this point,

the temperature was equilibrated at 303, 398 and 423 K, depending on the

experiment. After 1 h the sample was treated with a CO2 flow (60 ml/min) for 60 min

and then a He stream was flowed again for 30 min.

Finally, with the aim to assess the adsorption properties, CO2 adsorption at 303, 398

and 423 K up to ambient pressure (Micromeritics ASAP2020) was used. Prior to the

experiments, the samples were out-gassed in vacuum at 523 K for 4 h to remove any

adsorbed impurities.


4.3.1. Effect of CO2 concentration in the membrane permeation

properties.

The effect of CO2 concentration in the permeation of H2 and CO was performed at

five different CO2 concentrations (2, 5, 10, 15, and 20% of CO2 in the mixture) and

at three different temperatures (303 K, 398 K, and 423 K). The H2 and CO

concentrations in the feed were constant for the whole test (50% H2 and 1.25% CO).

Each of the experiments with a different CO2 concentration was studied for 2 h

(enough time to reach a stable reading in the MS signals that were analyzed), time

after which the CO2 concentration was increased according to the intervals described

previously until reaching a final CO2 concentration of 20%. Figure 4.1 summarizes

the permeance results of the prepared membrane (Na-LTA). It is important to note

that the H2 (Figure 4.1A) and CO (Figure 4.1B) permeance values presented for 0%

CO2 obtained in this study are very similar to those obtained in Chapter 3.

Furthermore, this observation is in agreement with the results previously exposed in


102

that Chapter, in which the permeation of single components in the feed was studied.

Our results clearly indicated that while H2 permeation was not affected significantly

in the zeolite/C membrane, CO permeation was diminished dramatically when Rb or

Cs-exchanged membranes were used.

From Figure 4.1A and B, it becomes apparent that H2 and CO permeances are

affected differently by the introduction of CO2 in the feed stream. Focusing on the H2

permeation values, taking into consideration its small size, very low polarizability

and high diffusivity, and the fact that H2 may also diffuse through the six-member

rings in the sodalite (β) cages [4], through which neither CO nor CO2 may diffuse,

permeation of the latter affects them to a minor extent compared to the CO

permeation values, which are significantly influenced by the CO2 presence. In this

sense, the CO permeance value at 303 K and 398 K decreases considerably when

CO2 is introduced in the feed stream. At these temperatures, its permeance is

considerably inhibited throughout the whole CO2 concentration range in the feed

stream, indicating that there exists competitive adsorption between CO and CO2 in

the membrane pores. At higher temperature this effect is diminished and CO

permeance at 423 K increases. Note that in Figure 4.1B the error bars are

substantially larger that those obtained at 303 K and 398 K. This is due to the fact

that the results are presented as permeances (permeation divided by the partial

pressure gradient between the feed and the permeate side) and that CO is present in a

small concentration as compared to H2 or CO2. Furthermore, with 20% CO2, the CO

permeance value is higher than that obtained in absence of CO2 at the same

temperature. In any case, the results were sufficient to deem the Na-LTA/C

membrane as unsuitable for H2/CO separation under all the conditions studied. On

the other hand, as the concentration of CO2 in the feed stream was increased, the

permeance of CO2 decreased as shown in Figure 4.1C. It should be mentioned that

despite the fact that CO2 permeation increased together with its concentration in the

Chapter 4

103

feed, the permeance (i.e. permeation divided by the difference in partial pressures

between the feed and permeate of the gas) decreased steadily.

Figure 4.1. H2/CO/CO2 mixture permeance in Na-LTA/carbon membrane: (A) H2

permeance, (B) CO permeance and (C) CO2 permeance.

0

0.5

1

1.5

2

2.5

3

3.5

0 5 10 15 20 25

Perm

eanc

e (m

ol/m

2 sPa

)·108

[CO2] (%)

H2 permeance

T = 303K T = 398K T = 423K

(A)

0

1

2

3

4

5

6

7

0 5 10 15 20 25

Perm

eanc

e (m

ol/m

2 sPa

)·108

[CO2] (%)

CO permeance

T = 303K T = 398K T = 423K

(B)

0.00.51.01.52.02.53.03.54.04.5

0 5 10 15 20 25

Perm

eanc

e (m

ol/m

2 sPa

)·108

[CO2] (%)

CO2 permeance

T = 303K T = 398K T = 423K

(C)


104

Bearing in mind the Na-LTA zeolite pore size, 0.38 nm [5] and the minimum

molecular dimensions of the three gases (0.24 nm, 0.34 nm, and 0.33 nm for H2, CO,

and CO2, respectively), it is expected that all the gases used in this study may pass

through the membrane pores. Therefore, with the aim to explain the observed

behaviour, it is important to focus the attention on their chemical properties.

According to the literature [6-8], CO2 permeance is affected at low temperatures. Sea

et al. [6] suggest that CO2 permeance is enhanced by surface diffusion. In this sense,

the CO2 molecules are first adsorbed on the external surface of the pores of the LTA

zeolite membrane layer and secondly, migrate into the pores by surface diffusion.

Therefore, CO2 molecules produce two effects:

(i) Inhibition of CO adsorption molecules on the surface and pores due to the

competitive adsorption between CO and CO2.

(ii) The adsorbed molecules reduce the openings of micropores.

Focusing the attention on the first point, CO2 adsorption at 303 K is more significant

(143 mg CO2/g zeolite (73 cm3/g STP) see Section 4.3.2.2) than CO adsorption (9

mg CO/g zeolite (7 cm3/g STP) at 303 K) as reported in Chapter 3. As a consequence,

the CO2 molecules inhibit CO adsorption in the zeolite at low temperature.

On the other hand, CO2 adsorption, being an exothermic process, is diminished with

increasing temperature (see below, Section 4.3.2.2) minimizing the competitive

adsorption effect. Therefore, CO permeance at 423 K (see Figure 4.1) increased with

respect to room temperature and increased with CO2 concentration. Nonetheless, the

least desirable permeation properties (i.e. higher CO permeances) are obtained at

high temperatures, precisely when this parameter becomes the main factor affecting

CO permeation. At this temperature, the interaction between CO2 and the zeolite

framework is hindered and permits CO diffusion through the membrane with a value

of 1.6x10-8 mol/m2 s Pa (for a CO2 concentration of 2%), which is identical to the

“original” CO permeance value for a H2/CO mixture (See Chapter 3). Finally, the

Chapter 4

105

worse CO permeation properties are obtained with a 20% of CO2 where CO

permeance is higher than the initial value for a H2/CO mixture. These results seem to

suggest that at high temperatures the interaction between both CO and CO2 and the

zeolite layer is very small. As a result both gases can diffuse quickly.

From the H2 permeance analysis, it seems possible that its low polarizability and low

critical temperature is responsible for its poor adsorption [9]. Thus, H2 permeation is

hardly affected by CO or CO2. The maximum permeance loss was 25% at 303 K

with 20% of CO2 in the feed stream. The less promising permeation values were

obtained at 423 K with 20% CO2, where H2 permeance dropped by 20%. In this case,

the H2/CO separation factor was 0.4, far below the high values required in order to

successfully separate H2 from CO. Therefore, the Na-LTA/carbon membranes are

not suitable for their use as H2 purification systems.

In order to obtain a CO-free H2 flow, the Cs-LTA/carbon membrane was used in the

same study. Figure 4.2 shows the H2 and CO2 permeances at the three temperatures

studied and the different CO2 concentrations. Contrary to Na-LTA/carbon

membranes, despite using high CO2 concentrations, CO does not diffuse through the

membrane pores in all the studies performed. This result corroborates the results

reported in Chapter 3, where CO could not diffuse through the membrane pores due

to the pore size in the Cs-LTA/carbon membrane. In this case, H2 permeance is

hardly affected by the presence of CO2 (see Figure 4.2A).

Nevertheless, due to the reduction of the zeolite pore diameter to values smaller than

the molecular size of CO2, it was expected that CO2 permeation would be hindered.

However, not only are its values similar to those obtained for the Na-LTA/carbon

membrane, but they also follow a similar trend over the studied temperature range as

the CO2 concentration in the feed stream is increased (see discussion above). To

analyze this aspect, commercial powder Na-LTA zeolite was used and ion exchanged

to obtain Cs-LTA zeolite. These two zeolites, namely Na-LTA and Cs-LTA were


106

used to perform DRIFTS studies (see Figures 4.3 and 4.4, respectively), CO2

isotherms at different temperatures (see Figure 4.5) and TGA experiments (see Table

4.1). These results are discussed below.

Figure 4.2. H2/CO/CO2 mixture permeance in Cs-LTA/carbon membrane: (A) H2 permeance and (B) CO2 permeance. Note: No CO permeation was detected at any

temperature using this membrane.

4.3.2. Study of the CO2-zeolite interaction.

4.3.2.1. DRIFTS studies.

Before studying the interaction between CO2 and the zeolite, it is important to note

that CO2 is a linear molecule with four vibration modes [10]. The first one is the

0.00

0.50

1.00

1.50

2.00

2.50

3.00

3.50

0 5 10 15 20 25

Perm

eanc

e (m

ol/m

2 sPa

)·108

[CO2] (%)

H2 permeance

T = 303K T = 398K T = 423K

(A)

0.00

0.50

1.00

1.50

2.00

2.50

3.00

3.50

0 5 10 15 20 25

Perm

eanc

e (m

ol/m

2 sPa

)·108

[CO2] (%)

CO2 permeance

T = 303K T = 398K T = 423K

(B)

Chapter 4

107

symmetric stretching vibration mode (ν1) at 1480 cm-1 which is not infrared active.

The second one, O–C–O bending (ν2) at 667 cm-1, is infrared active and it is equal to

the fourth one (ν4) and the main difference between them is that one bending is in the

molecule plane and the other one is out of-plane. The third mode is the asymmetric

stretching mode (ν3) at 2349 cm-1 which is infrared active due to there being a change

in the molecular dipole moment during this vibration. These are the major vibration

modes of the molecules but it is possible to observe the combination of them with a

minor intensity in the infrared spectrum. The last one, ν3, is the main vibration mode.

Thus, with the aim to study the CO2 interaction with the zeolites, several authors

[10,11] have focused their attention on its variation when the CO2 molecules interact

with the zeolite surface.

Therefore, according to the literature, CO2 can interact with the LTA structure in

different ways. Montanari and Busca [10] have studied this interaction and their main

conclusion was that CO2 is adsorbed on the zeolite mostly on the surface of the

cavities, forming a linear OCO-Na+ complex that also involves the perturbation of

the Si-O-Al bonds. The band which was assigned to adsorbed CO2 by FTIR is

observed at 2351cm-1. Other authors [11] have pointed out that for zeolite Y the CO2

adsorbed on the zeolite surface of the cavities presents a weak band at about 3710

cm-1 (combination mode, ν1 + ν3). Therefore, the issue is to study the variation in the

ν3 vibration mode and to ensure that the CO2 is adsorbed on the zeolite with the

appearance of the combination mode (ν1 + ν3).

Figure 4.3 shows the DRIFTS study of the interaction between CO2 and Na-LTA

powder zeolite. Only the band appearing at lower wavenumbers is shown because the

band located at 3710 cm-1 is not observed. It is seen that, under all temperatures

tested, this band decreases gradually with exposure time to a He flow, after CO2 had

flowed through the sample chamber for 2 h. Moreover, if at this point the sample is


108

heated up to 573 K, this band completely disappears. These results point out that CO2

is adsorbed weakly on the zeolite.

Figure 4.3. Representative spectra of CO2 adsorbed on the zeolite Na-LTA; variation

of the ν3 vibration mode with time: (A) spectra collected at 303 K; (B) spectra collected at 398 K; (C) spectra collected at 423 K.

Chapter 4

109

Figure 4.4. Representative spectra of CO2 adsorbed on the ion exchanged Cs-LTA

zeolite. Spectra collected at 303 K after treatment with CO2 at different temperatures (303 K, 398 K and 423 K) for 2 h, heat treatment up to 523 K under Helium for 1 h.

Variation of the ν1 + ν3 vibration mode with temperature of CO2 treatment (A) region of 3710 cm-1 (B) region around 2350 cm-1.

The interaction between CO2 and the ion-exchanged Cs-LTA powder zeolite was

also analyzed thoroughly by DRIFTS and the results obtained can be summarized as

follows: (i) for the three temperatures under study, two bands were observed

corresponding to the CO2 adsorbed on the zeolite surface, in agreement with previous

results [10,11]; (ii) no difference was observed between peaks collected at different

times after switching the gas flow from CO2 to He (results not shown); (iii) both

peaks are still observed after heating the sample up to 573 K. As an example, Figure


110

4.4 shows the spectra corresponding to Cs-LTA powder zeolite after treatment with

CO2 at different temperatures (303 K, 398 K and 423 K) for 2 h, heat treatment up to

523 K under Helium for 1 h and cooling down to 303 K. The existence of both peaks

even after heat treatment indicates a strong interaction between the zeolite and CO2.

It can be observed that both peaks increase in intensity as the temperature of CO2

treatment increases, indicating that the CO2-zeolite interaction is an activated process

and limited by the kinetics of the adsorption process.

4.3.2.2. TGA and CO2 isotherms analysis.

In order to establish the amount of CO2 adsorbed on both zeolites, CO2 adsorption

isotherms were obtained at different temperatures and TGA measurements have been

performed. Figure 4.5A shows the CO2 adsorption isotherms obtained at three

different temperatures for the sample Na-LTA. As expected, the amount of CO2

adsorbed by the sample decreased importantly with increasing temperature. These

results are in agreement with those previously presented for the Na-LTA/carbon

membrane, which point out that this CO2 adsorption, taking place to a very

significant degree, inhibits CO adsorption and its permeation through the membrane

at low temperatures. As the temperature is increased CO can permeate through the

membrane pores due to the reduction of the amount of CO2 competitively adsorbed

on the zeolite.

On the other hand, Figure 4.5B shows the CO2 isotherm at the same temperatures for

the Cs-LTA zeolite. The low CO2 uptake values with respect to Na-LTA zeolite

indicate that its pore diameter has decreased considerably and it makes very difficult

for CO2 to enter its framework being the adsorption kinetics very slow. This means

that the isotherms are not at equilibrium conditions, which causes the Cs-LTA

sample to adsorb a similar small amount of CO2 at all the temperatures tested under

the experimental conditions used. Since the kinetics of CO2 adsorption on the Cs-

LTA sample is very slow, as the temperature is increased the decrease in the amount

Chapter 4

111

adsorbed seems to be compensated by the increase in adsorption rate, so the two

processes counterbalance each other, leading to very similar CO2 adsorption

isotherms. This phenomenon indicates that a small quantity of CO2 is strongly

adsorbed on the zeolite surface (as it was shown in the DRIFTS results). From our

permeation results, it appears that this CO2 chemisorbed can diffuse from the zeolite

surface to the zeolite pores.

Figure 4.5. CO2 adsorption isotherms at 303 K, 398 K and 423 K corresponding to:

(A) Na-LTA powder zeolite and (B) ion-exchanged Cs-LTA powder zeolite.

Table 4.1 summarizes the amount of CO2 adsorbed on the two zeolites Na-LTA and

Cs-LTA determined by TGA experiments. In the case of the Na-LTA zeolite, the

trend observed for the amount of CO2 adsorbed in the sample is the same as that

020406080

100120140160

0 100 200 300 400 500 600 700Ads

orbe

d C

O2

(mg

CO

2/gze

olite

)

Pressure (mmHg)

Na-LTA

T = 303K T = 398K T = 423K

(A)

0

0.4

0.8

1.2

1.6

2

0 100 200 300 400 500 600 700Ads

orbe

d C

O2

(mg

CO

2/gze

olite

)

Pressure (mmHg)

Cs-LTA

T = 303K T = 398K T = 423K

(B)


112

observed from the isotherms (see Figure 4.5). This further corroborates the affinity

between CO2 and the zeolite. As it has been presented previously, the Cs-LTA

sample can strongly adsorb CO2 but to a much minor extent compared to zeolite Na-

LTA. In this case, the zeolite adsorbs approximately 1 mg of CO2 per gram of sample

at 398 and 423 K (the data at room temperature could not be obtained due to the

substantial fluctuation of the equipment signal during the measurements). Therefore,

the maximum information that it is obtained with this kind of experiments is that

both zeolites can adsorb growing amounts of CO2 as the cation size is decreased.

Table 4.1. CO2 adsorbed on the Na-LTA and Cs-LTA zeolite measured by TGA experiments.

Temperature (K) Sample

Na-LTA, m (mgCO2/gzeolite) Cs-LTA, m (mgCO2/gzeolite) 303 18.5 - 398 6.1 0.9 423 5.9 0.2

It becomes possible to obtain conclusions on how CO2 interacts with the two zeolites

and, in turn the resulting permeation properties. In the case of the Na-LTA zeolite,

the CO2 adsorbed on the zeolite hinders CO permeation at low temperatures. As the

temperature increases adsorption is hindered and CO permeates through the

membrane.

In the case of the Cs-LTA powder zeolite, according to the DRIFTS experiments,

CO2 is adsorbed strongly on the zeolite but this adsorption is mainly restricted to the

outer surface under the studied adsorption conditions. The isotherms and TG studies

indicate that only a small amount is adsorbed. For that reason, we propose that the

mechanism that governs CO2 permeation is CO2 adsorption on the zeolite surface

and migration by surface diffusion through the membrane pores and small pinholes

due to the concentration gradient between both sides of the membrane. This

conclusion combined with no CO permeation indicates that it is a high quality

membrane with few/small defects.

Chapter 4

113

4.3.3. Membrane permeation properties in a simulated reformer

mixture.

The best way to know if the prepared membranes, Na-LTA/C and Cs-LTA/C

membranes are useful as H2 purification systems is to test them under realistic

conditions that simulate the gases emitted from the reformer unit. Namely, a gas

stream containing H2, CO, CO2 and H2O (50% H2, 1.25% CO, 20% CO2, 5% H2O

and rest He, total flow rate 100 ml/min) was used for these experiments.

Table 4.2 presents the permeance values corresponding to the Na-LTA/C membrane

for a simulated reformer mixture on dry and humid conditions. As it was explained

previously, the as-synthesized membrane, Na-LTA/carbon, does not show CO

permeation at 303 K. However, CO permeance increases with temperature due to a

higher diffusivity of the gases through the membrane pores. It is important to note

that CO permeance changes markedly in the presence of water (See Table 4.2), while

H2 and CO2 permeance remain similar to the permeance values on dry basis. Based

on X-ray diffraction experiments, Sorenson et al. [12] concluded that water contracts

NaA zeolite crystals by approximately 0.22 vol.% at 300 K and 348 K for H2O

concentrations below 6%. Therefore, focusing our attention on the permeation

properties in the presence of water, intercrystalline spaces may be generated during

zeolite crystal contraction in the presence of low concentrations of water at 303 K.

Thus, CO permeation may occur via these new defects. Nevertheless, it must be

noted that the rest of gases maintain its permeance values almost unaltered. For this

reason, other factors must be taken into account in the experiments that are affecting

CO permeation. According to Montanari et al. [13], CO interacts with the zeolite by

the formation of C-bonded species over a Na+ ion mainly, affecting the Al-O-Si

bonds. This interaction was found to be strong at low temperatures (at 130 K), but at

temperatures close to ambient this interaction is relatively weak. This conclusion is

in line with our hypothesis that CO and CO2 show a similar interaction with the


114

zeolite framework at 423 K. Beta et al. [14] expressed that water molecules form an

extensive network of water-water and water-framework oxygen hydrogen bonds.

Furthermore, simulation studies of water adsorbed on LTA zeolite [15,16] have

explained that the water molecules are adsorbed and form different structures over

the Na+ ions including those located on the α and β cages of the LTA zeolite.

Therefore, we have considered that the CO molecules can interact with the zeolite

(C-bonded) or with water (H-bonded). Therefore, at 303 and 398 K CO permeance is

enhanced in the presence of water due to its ability to interact with both the zeolite

and water. These results are in agreement with the binary mixture experiment

(CO/H2O) reported by Zhu et al. at 398 K [17].

Table 4.2. Na-LTA/carbon and Cs-LTA/carbon membranes permeance properties on dry conditions and humid conditions using a simulated reformer mixture.

Sample Na-LTA

Permeance (10-8mol/m2·s·Pa) Cs-LTA

Permeance (10-8mol/m2·s·Pa)

T (K) Gas Dry

conditions Humid

conditionsDry

conditionsHumid

conditions

303

H2 1.3±0.1 0.8±0.1 1.3±0.1 1.4±0.1

CO 0 4.0±2.8 0 0

CO2 1.4±0.5 1.1±0.1 0.6±0.3 0.7±0.1

398

H2 1.1±0.1 0.8±0.1 1.3±0.1 1.4±0.1

CO 0.8±0.6 4.8±2.2 0 0

CO2 1.0±0.7 1.3±0.1 0.8±0.5 0.8±0.1

423

H2 1.3±0.1 0.8±0.1 1.4±0.1 1.4±0.1

CO 4.7±1.8 2.9±0.1 0 0

CO2 0.6±0.2 1.3±0.4 0.8±0.5 0.8±0.1

In a different trend as to that observed with CO, CO2 permeance remains unchanged

through all experiments performed with water in the feed stream. The reason for this

observation might be due to the fact that both diffusion and adsorption contribute to a

similar extent to CO2 permeation in our case. Thus, as adsorption is hindered by the

Chapter 4

115

presence of moisture in the feed stream, diffusion in the gas phase through the

membrane pores becomes the dominant mechanism for CO2 permeation, resulting in

similar permeance values. On the other hand, despite the presence of water in the

feed stream, neither H2 nor CO2 permeation were affected for all the conditions

studied which constitutes a remarkable observation.

The effect of water on the permeation properties has been discussed by different

authors from a variety of perspectives [19-22]. Sebastián et al. [19,20] attributed the

observed increase in the H2/N2 separation factor using a Ti-silicate umbite to a

change in the pore size of the microporous umbite as its hydration level changed

when the membrane was exposed to a feed stream containing water vapour. Poshusta

et al. [21] established that, in the case of SAPO-34 membranes, the presence of water

in the feed stream could be a good indicator of the presence of defects in the zeolite

layer and a means to estimate the fraction of transport through non-zeolite pores.

Aoki et al. [22] concluded that water was adsorbed on the zeolite A membrane

surface that they used in their studies, and could diffuse through the zeolitic and non

zeolitic pores. As a result, permeation of small hydrophobic species (H2 or He)

showed smaller permeances, whereas small hydrophilic gases, such as O2 or CO2

showed higher permeances. In all these cases, however, the supports used were

alumina tubes with a pore size between 5 and 200 nm. On the one hand this pore size

is, in some cases, at least one order of magnitude smaller than the pores present in

our carbon support, and the chemical nature of the aforementioned supports and that

presented in this study is markedly different. In our case, this may be a consequence

of using a hydrophobic support for the synthesized membranes and would highlight

the characteristics of the support used in the membrane performance. We are

currently conducting studies in order to further analyze the effect of the support and

its porosity in hydrogen purification experiments.


116

Finally, the most important aspect of the CO permeation properties is that at 423 K

the CO permeation value is close to those obtained in dry conditions. This fact is in

agreement with the zeolite membrane becoming de-hydrated again and thus, the Na-

LTA/C zeolite membrane recovers its permeation properties under dry conditions

(see Table 4.2). Finally, H2 is a molecule with a weak adsorption affinity but with the

highest diffusivity which can furthermore diffuse through the sodalite cages in the

zeolite structure [17]. Therefore, its permeance is only slightly affected by the

presence of water (see Table 4.2). However, despite these interesting properties, this

material is not useful to use as a H2 purification system due to the high CO

concentration at the permeate under all the temperatures tested.

The Cs-LTA/C membranes prepared in this study were tested under the same

conditions. Table 4.2 includes all the permeance values obtained for these

membranes for a simulated reformer mixture on dry and humid conditions. The most

important aspect is that despite the presence of water, there is no CO permeation at

all temperatures tested, in agreement with our previous results reported in Chapter 3,

in which CO molecules cannot pass through the Cs-exchanged membrane. These

also agree with the results presented previously with different CO2 concentrations.

Furthermore, this observation indicates that there are few defects in the prepared

membrane, that the defects present are of small size, or that the crystals do not

contract in the presence of water, which in turn would allow the CO molecules to

diffuse through the membrane. In this respect, Cs+ may stabilize the zeolite network

due to its bulky size, preventing contraction of the crystals. Concerning CO2

permeation, as it has been explained previously, it can diffuse through the membrane

due to its strong adsorption on the zeolite surface and subsequent diffusion in the

membrane pores. Finally, due to its low affinity for the zeolite surface and high

diffusivity, H2 permeance remains practically constant in the presence of water.

The results of this study, in which a Cs-exchanged LTA membrane is capable of

delivering a high H2 permeance while no CO is detected on the permeate side should

Chapter 4

117

be regarded as highly promising, especially considering that the membrane

performance is unaffected by the presence of other gases (CO2) or vapours (water).

Finally, we consider that the Cs-LTA/carbon membrane is a high quality membrane

with a relatively small amount of defects and stable over time, since each experiment

has been performed for 10 h. In total, the working time was over 60 h with no

reduction of its excellent properties, even in the presence of water.


118

4.4. Conclusions.

Na-LTA/carbon membranes and Cs-LTA/carbon membranes have been used to study

the effect of CO2 and H2O in their membrane permeation properties and in conditions

that simulated the stream of a reformer mixture.

Concerning the effect of CO2 in the permeation properties, the as prepared Na-

LTA/carbon membrane shows no CO permeation at room temperature due to the

competitive adsorption produced by the presence of CO2 in the feed streams.

However, this phenomenon becomes less pronounced as the temperature increases.

In this sense, the least promising permeation properties are obtained at 423 K and

20% CO2 in the mixture in which CO permeation reaches a maximum. Furthermore,

the H2 permeation values are not affected by the presence of CO2. In reference to the

Cs-LTA/carbon membrane, CO2 affects neither CO permeation nor H2 permeation.

In the case of CO2 permeation it has been proposed that its permeation mechanism is

governed by its strong adsorption on the zeolite followed by migration on the zeolite

surface.

In the case of a simulated reformer mixture, Na-LTA/carbon membranes show CO

permeation up to 423 K. These results are worse than those obtained at room

temperature on dry basis, in which no CO was detected on the permeate side of the

membrane. The Cs-LTA/carbon membrane, on the other hand, is a high quality and

stable membrane (no performance loss at working times over 60 h). The permeations

of neither CO nor H2 are affected by the presence of water, making these membranes

very promising candidates for their prospective use in H2 purification devices.

Chapter 4

119

References.

[1] R.A. Rakoczy, Y. Traa, Microporous Mesoporous Mater. 60 (2003) 69-78. [2] S. Dominguez-Dominguez, A. Berenguer-Murcia, E. Morallon, A. Linares-Solano, D. Cazorla-Amoros, Microporous Mesoporous Mater. 115 (2008) 51-60. [3] A. Berenguer-Murcia, L. Gora, W. Zhu, J.C. Jansen, F. Kapteijn, D. Cazorla-Amorós, A. Linares-Solano, Ind. Eng. Chem. Res. 46 (2007) 3997-4006. [4] A.W.C. van den Berg, S.T. Bromley, J.C. Wojdel, J.C. Jansen, Microporous Mesoporous Mater. 87 (2006) 235-242. [5] D.W. Breck, W.G. Eversole, R.M. Milton, T.B. Reed, T.L. Thomas, J. Am. Chem. Soc. 78 (1956) 5963-5971. [6] B.-K. Sea, K. Kusakabe, S. Morooka, J. Membr. Sci. 130 (1997) 41-52. [7] S. Li, J.L. Falconer, R.D. Noble, Adv. Mater. 18 (2006) 2601-2603. [8] J. Motuzas, R. Mikutaviciute, E. Gerardin, A. Julbe, Microporous Mesoporous Mater. 128 (2010) 136-143. [9] J. Lindmark, J. Hedlund, Stud. Surf. Sci. Catal. 170 (2007) 975-980. [10] T. Montanari, G. Busca, Vib. Spectrosc. 46 (2008) 45-51. [11] A. Otero Arean, M. Rodríguez Delgado, Appl. Surf. Sci. 256 (2010) 5259-5262. [12] S.G. Sorenson, E.A. Payzant, W.T. Gibbons, B. Soydas, H. Kita, R.D. Noble, J.L. Falconer, J. Membr. Sci. 366 (2011) 413-420. [13] T. Montanari, I. Salla, G. Busca, Microporous Mesoporous Mater. 109 (2008) 216-222. [14] I.A. Beta, B. Hunger, W. Böhlmann, H. Jobic, Microporous Mesoporous Mater. 79 (2005) 69-78. [15] V. Crupi, F. Longo, D. Majolino, V. Venuti, J. Chem. Phys. 123 (2005), 154702-1-154702-11. [16] T. Kyotani, T. Ikeda, J. Saito, T. Nakane, T. Hanaoka, F. Mizukami, Ind. Eng. Chem. Res. 48 (2009) 10870-10876. [17] W. Zhu, L. Gora, A.W.C. Van Der Berg, F. Kapteijn, J.C. Jansen, J.A. Moulijn, J. Membr. Sci. 253 (2005) 57-66. [19] V. Sebastián, Z. Lin, J. Rocha, C. Téllez, J. Santamaría, J. Coronas, Chem. Mater. 18 (10) (2006) 2472-2479. [20] V. Sebastián, Z. Lin, J. Rocha, C. Tellez, J. Santamaria, J. Coronas, J. Membr. Sci. 323 (2008) 207-212. [21] J.C. Poshusta, R.D. Noble, J.L. Falconer, J. Membr. Sci. 186 (2001) 25-40. [22] K. Aoki, K. Kusakabe, S. Morooka, Ind. Eng. Chem. Res. 39 (2000) 2245-2251.

Chapter 5. Hydrocarbon traps based

on zeolites for gasoline vehicle emission control

tested under cold start conditions.

Chapter 5

121

5.1. Introduction.

Cold Start is the period of time (1-2 minutes) required by the Three Ways Catalyst

(TWC) to reach the working temperature; during this time, up to 80% of

hydrocarbons are released in a drive cycle [1]. Therefore, the emission control during

this “cold start” period is essential to reduce the environmental impact of gasoline

engines. Among all the solutions studied up to now (see Chapter 1), the use of HC

adsorbents, ‘‘hydrocarbon traps”, before the TWC seems to be the most relevant

from a scientific-technological point of view. The critical factors for any

hydrocarbon trap are: (i) High adsorption capacity of hydrocarbons at low

temperatures, (ii) hydrocarbon desorption starting at temperatures higher than 200

ºC, (iii) a reversible adsorption process and (iv) a solid material resistant at

temperatures higher than 750 ºC.

The maximum values of these factors are limited by properties of the solids like pore

volume, pore structure (3D or 1D), chemical nature of the solid or extra-framework

atoms. Zeolites have been found to be preferred adsorbents for these applications,

mainly due to their stability under severe process conditions. Thus, zeolites with

three dimensional pore structures, such as BETA [2] or SZZ-33 [3], present a high

adsorption and then, are very promising for this application. Other researchers

indicated that one-dimensional (1D) zeolites as MOR or EOU [4] or

silicoaluminophosphates molecular sieves as SAPO-5 or SAPO-11, SAPO-36 or

SAPO-41 [5] are interesting materials for the retention of light hydrocarbon

components (i.e. propene). They concluded that, in the 1-D zeolites EUO and MOR,

the toluene molecules (molecule representative of heavy hydrocarbons) block the

diffusive movement of the propane molecules (light hydrocarbon), thus trapping the

propane molecules inside. Propane molecules can only be desorbed to the gas phase,

once toluene begins to desorb. This leaded to an increase in the propane desorption

Hydrocarbon traps based on zeolites

122

temperature of 35ºC over the single-component value. In addition, the Si/Al ratio is

going to play an important role in the HC trapping [2]. The chemical and thermal

stability of zeolites is strongly dependent on the Si/Al ratio, being more stable for

high ratios. On the contrary, zeolites with low Si/Al ratio have a large HC adsorption,

and the HC emission is produced at higher temperatures, but present a low stability

[2,5]. Then, to obtain the best material suitable as hydrocarbon trap, all these

aforementioned properties, among others, must be taken into account.

In a previous work done by our research group, a screening of different zeolites and

silicoaluminophosphates for the retention of propene under cold start conditions was

carried out [6]. To cover a wide range of properties, six materials (zeolites and

silicoaluminophosphate molecular sieves) with distinctive framework structures (3D

and 1D) and a variety of Si/Al ratios were prepared. These materials included four

zeolites (BETA, ZSM-5, Silicalite-1 and Mordenite) and two silicoaluminophosphate

molecular sieves (SAPO-5 and SAPO-41). Propene adsorption isotherms and

temperature programmed desorption experiments were performed on those materials.

Moreover, for a better evaluation of the solids in cold trap application, the retention

of propene under cold start conditions (cold start tests (CST)) were carried out by

doing measurements at conditions that more closely mimic the automotive ‘‘cold

start” (increasing the feed concentration simultaneously with the start of the

temperature ramp). The main observations reported in our previous work [6] were

that, for a possible application of zeolites and/or silicoaluminophosphates as propene

trap, it is necessary to develop materials presenting simultaneously: (i) high

equilibrium propene adsorption capacity at 30 ºC, where propene should be mainly

chemisorbed and; (ii) kinetic restrictions. From these results, among the adsorbents

analyzed, the best material for propene retention in a CST was BETA zeolite,

followed by ZSM-5 and SAPO-5 adsorbents. The Scheme 5.1 shows the BEA

structure of the BETA zeolite.

Chapter 5

123

Scheme 5.1. BEA structure of BETA zeolite.

From a practical point of view, the use of powder samples has several drawbacks,

specifically the high pressure drops generated when it is used in gaseous flows. Then,

for a practical system, it is required to have the adsorbent supported on a structured

material. Honeycomb monoliths made from synthetic cordierite

(2MgO·2Al2O3·5SiO2), as it has been mentioned in Chapter 2, is the selected material

to support the TWC catalyst due to their excellent characteristics (see Chapter 1).

Thus, the objectives of the present study are: (i) to support BETA zeolite on

cordierite honeycomb monolith using in-situ synthesis; and (ii) to test these materials

as hydrocarbon trap under simulated cold start conditions (cold start tests (CST),

where several hydrocarbons (propene and toluene), O2 and water are present.

5.2. Experimental.

5.2.1. Synthesis of zeolite supported on cordierite honeycomb

monoliths.

Cylindrical cordierite monoliths (see Chapter 2) were used (400 cpsi, length (l): 1.6

cm, diameter (d): 1.4 cm). Prior to coating, a calcination in static air in a furnace at

800ºC for 2 hours (heating rate: 6 ºC/min) was performed with the aim to remove

organic impurities.


124

The growth of BETA zeolite was carried out by in-situ synthesis onto the cordierite

monolith avoiding the use of binders. The general experimental procedure for the

synthesis onto the monolith was:

(i) The cordierite honeycomb monoliths were wrapped with Teflon tape to

prevent zeolite growth on the outer surface.

(ii) The “teflonated” monoliths were placed in a stainless steel autoclave with a

Teflon liner (V: 10 mL). A precursor gel/monolith (volume/weight) ratio of 4

mL/g was used for coating the monoliths.

(iii) The autoclaves are placed horizontally in a convection oven and rotated

during the whole growth process (v = 4rpm) to ensure a homogeneous growth

of the solid film and prevent blocking of the monolith channels [7]. Four

monoliths were performed per batch.

(iv) The crystallization step is carried out at the desired temperature and time.

(v) After the hydrothermal synthesis, the autoclaves were submitted to “fast

cooling” with running tap water.

(vi) After the synthesis, the coated monoliths were washed with abundant distilled

water and dried at 100 ºC overnight. All the monoliths were submitted to

sonication for 6 hours to remove loosely adhered crystal and then, dried again

overnight in an oven at 100 ºC.

One, two and three-consecutive synthesis steps were performed. After finishing the

procedure the first-synthesis step (see below), the samples were place again in the

autoclave for a second-synthesis step using a fresh synthesis solution (points (ii) to

(vi)). In the case of the third synthesis step it was performed in the same manner.

After preparing the final coated monoliths after one, two or three consecutive

synthesis steps, the removal of the template was carried out by combustion in a static

air in a furnace at 500 ºC (heating rate 1 ºC/min) during 6 h. Finally, the weight

increase after the coating steps was measured (wt. %).

Chapter 5

125

Coating of the monoliths with zeolite BETA was carried out following the

methodology reported by Bueno-Lopez et al. [7] but it was adapted to our larger

monolith size. The synthesis gel was prepared at room temperature by mixing two

solutions. Solution A was prepared by adding slowly 7.390g of Fume silica (Sigma-

Aldrich) to a transparent aqueous solution of 25.701g of TEAOH (35% aqueous

solution of tetraethylammonium hydroxide, Sigma-Aldrich), 0.367 g NaCl (98%,

Sigma-Aldrich ) and 0.132g KCl (99, L.T. Baker) in distilled water (11.585g),

stirring until complete dissolution of silica. Solution B was prepared by adding

slowly 0.448g of sodium aluminate (Riedel-de Haën) to a basic solution (0.083g

NaOH-99.99%, Sigma-Aldrich- in 5g of distilled water) and it was stirred until a

clear solution was obtained. Solution B was added drop by drop to the solution A and

was stirred during 15min until it was completely mixed. The optimum conditions

found for the synthesis of zeolite BETA over cordierite monoliths were synthesis

steps with a crystallization temperature of 132ºC for 48 hours. The BETA coated

monoliths, were named MBETA1, MBETA2 and MBETA3 for one, two or three

consecutive synthesis steps, respectively.

5.2.2. Characterization.

The coated monoliths were characterized by XRD, using a 2002 Seifert powder

diffractometer. The scanning rate was 2º/min and Cu-Kα radiation was used.

Scanning measurements from 2º up to 50º were done.

The prepared monoliths were carefully cut with parallel cuts to the monolith axis.

This working procedure allowed the analysis of the surface of the monolith channels

and the evaluation of the thickness and homogeneity of the coated layer by scanning

electron microscope (JEOL microscope model JSM-80).

The textural characterization of the coated monoliths with BETA zeolite was carried

out by N2 adsorption at -196 ºC and CO2 at 0 ºC (Autosorb 6, Quantachrome). Prior


126

to the adsorption measurements, the adsorbents were out-gassed in vacuum at 250 ºC

for 4 h to remove any adsorbed impurities. Surface area was calculated from nitrogen

adsorption isotherms using the BET equation (SBET). Total micropore volume (VDR

(N2)) and narrow micropore volume (VDR (CO2)) were calculated applying the

Dubinin-Radushkevich (DR) equation to the N2 adsorption data at -196 ºC and the

CO2 adsorption data at 0 ºC, respectively. CO2 adsorption at 0 ºC is an interesting

complement to N2 adsorption for the assessment of the narrow microporosity, where

diffusional problems of the N2 molecules (-196 ºC) inside the narrow porosity (pore

size <0.7 nm) occur [8,9].

5.2.3. Cold Start Tests.

The zeolite coated monoliths performance as HC trap was tested under simulated

cold start conditions by adsorption-desorption tests. It is worth commenting that a

CST procedure allows the investigation of competitive adsorption and desorption

from an initially empty zeolite, which is closer to the conditions that would

experience an in-line HC trap in an automobile [6]. During a simulated CST,

molecules first diffuse into an “empty” adsorbent at low temperatures. Then, as the

temperature increases the driving force for diffusion changes direction, causing

molecules to diffuse back out of the solid and into the gas stream.

The inlet gas composition used in the CST experiments was 100 ppmv propene, 87

ppmv toluene, 1 % v/v oxygen, 10 % v/v steam and Ar balance under a total flow of

30 ml/min. Steam was generated in an auxiliary reactor with temperature control, and

was introduced into the main stream. The experiments were run in a fixed bed reactor

externally heated and the gases emitted were analyzed by a Mass Spectrometer.

Under CST conditions, the reactor temperature was increased from 30 ºC to 600 ºC

(heating rate 50 ºC/min), keeping the sample at the maximum temperature for 30

minutes. In order to study the zeolite ageing, three consecutive CST were performed.

Chapter 5

127


5.3.1. Characterization of coated monoliths.

Table 5.1 collects the percentage of weight increase of the coated monoliths after the

first, second and third coating steps. According to Bueno-Lopez et al. [7] with the

aim to improve the zeolite loading and to get a thin and homogeneous zeolite layer, a

second and a third synthesis steps are necessary. It is worth mentioning that

supported zeolite crystals (after the first step) or the zeolite layers (after the second

and third steps) have a high stability and reproducibility even after sonication,

indicating a good adherence of the zeolite on the monoliths surface. Therefore, the

error in the estimation of the weight increase after the synthesis steps is reasonably

low considering the complex process under study, including the crystallization step,

the cleaning step and the template removal. Furthermore, another key factor of the

reproducibility obtained is autoclave rotation during the synthesis step since the

rotation favors crystallization of the zeolite homogeneously on the whole cordierite

support [7].

X-ray diffraction was used to confirm the phase purity and crystallinity of the zeolite

layer coated on the cordierite monoliths. Figure 5.1 compiles the diffractograms of

the uncoated cordierite and the coated monoliths. Powder BETA zeolite

diffractogram (extracted from [6]) has been added to confirm the crystallinity of the

coated zeolite layer. The coated monolith shows the characteristic peaks of cordierite

and those of the BETA zeolite except for sample MBETA1. Regarding this sample,

because of the low quantity of zeolite deposited, the diffraction peaks are not clearly

observed.


128

Figure 5.1. XRD diffractograms of cordierite, coated monoliths and powder zeolite.

Nitrogen adsorption was performed to analyze the porous texture of the as-prepared

materials. Figure 5.2 shows the nitrogen adsorption isotherms at -196ºC for the

uncoated monolith and the BETA coated monoliths after 1, 2 and 3 synthesis steps.

From the nitrogen isotherms (see Figure 5.2), it is possible to confirm that cordierite

exhibits a Type II isotherm (see Chapter 2), typical of macroporous solids and a very

low adsorption caducity. Therefore, the contribution of the cordierite support to the

adsorption properties of the coated monoliths is negligible. Focusing the attention on

the zeolite coated materials, all of them have a Type I isotherm, which according to

IUPAC classification [10], is typical of microporous solids.

Figure 5.2. Nitrogen adsorption/desorption isotherms at -196ºC.

5 10 15 20 25 30 35 40 45 502θ (º)

BETA

Cordierite

MBETA3

MBETA2

MBETA1

0

10

20

30

40

50

60

0 0.2 0.4 0.6 0.8 1

Am

ount

ads

orbe

d (S

TP)

(c

m3 /g

)

P/Po

Cordierite MBETA1 MBETA2 MBETA3

Chapter 5

129

Table 5.1 summarizes the BET surface areas (SBET), the total micropore volume

(VDR(N2)) and narrow micropore volume (VDR(CO2)). The porous texture results for

the BETA zeolite powder [6] have been added for comparison purposes. As

expected, the N2 volume adsorbed increases with the number of zeolite synthesis.

However, the increase in adsorption capacity is not proportional with the amount of

zeolite loaded in the cordierite (see weight increase in Table 5.1 compared to SBET

(m2/gzeolite)). In this sense, the increase in the amount of nitrogen adsorbed is higher

in the second synthesis step than in the third when the data are calculated with

respect to the amount of zeolite loaded. These results suggest that the sample

MBETA3 presents N2 diffusion limitations (at -196ºC) during the adsorption.

Table 5.1.Weight increase (%) and porous texture results of BETA coated monoliths.

Sample Weight increase

(%)

SBET (m2/gmonolith)

SBET (m2/gzeolite)

VDR(N2) (cc/g)

VDR(CO2) (cc/g)

BETApowder - - 570 0.25 0.21

MBETA1 3.6±0.4 28 -* 0.01 0.01 MBETA2 17.7±2.0 100 600 0.05 0.03 MBETA3 24.9±2.0 117 460 0.06 0.05

(*) not possible to be calculated due to the low amount of zeolite.

A detailed study of the cordierite monoliths and the zeolite-loaded materials was

carried out by SEM. The original uncoated monolith (see Figure 2.13 in Chapter 2) is

a continuous support with macroscopic openings along the entire surface (See Top

View, Figure 2.13A and Chapter 2). From the cross sectional view (Figure 2.13B,

Chapter 2), it can be established that the monoliths wall are 150 µm wide and have a

macroporous character.

Figure 5.3 shows the top view and the cross sectional view of the BETA coated

monoliths, MBETA1, MBETA2 and MBETA3. In all the cases, the typical spherical

shape of the BETA zeolite crystals is observed. After the first step, the MBETA1


130

monolith is formed by almost individual crystals of the zeolite BETA that cover

partially the monolith surface, existing uncovered areas. With a second and third

synthesis steps (MBETA2 and MBETA3 samples, respectively) the entire surface of

the monolith is covered by a continuous zeolite layer of intergrowth crystals. From

the cross sectional view, the zeolite layer thickness is estimated and it can be

observed an increase with the number of synthesis from 2 to 4 µm approximately.

Figure 5.3. SEM images of Top view (left side) and cross sectional (right side) view for the prepared monoliths: (a) and (b) MBETA1; (c) and (d) MBETA2; (e) and (f)

MBETA3.

A

1μm

C

1μm

E

1μm

B

1μm

2 μm

D

10μm

3 μm

F

1μm

4μm

Chapter 5

131

To conclude with the SEM analysis, the small crystal size of the BETA zeolite

permitted to reach the internal opening of the monoliths wall (see Figure 5.4), thus

giving rise to a zeolite layer formed by crystals grown in the internal and the external

surface which are continuously connected, resulting in a high mechanical stability

[7].

Figure 5.4. SEM images of cross sectional of BETA zeolite monolith with an internal

opening full filled with BETA zeolite.

5.3.2. Cold start test.

Figure 5.5 includes the evolution of propene and toluene concentration in the cold

start test for BETA coated monoliths. In general, the profiles obtained in CST can be

divided in three different parts. In the first one, during the first minutes of the CST,

the HC are adsorbed. The second one corresponds to HC desorption, and starts when

the trap is saturated with HC and/or the temperature is so high, that adsorption is not

longer favored and desorption dominates the process. At this moment, HC signals

start to increase. If HC catalytic combustion takes place, the HC concentration starts

to decrease and a maximum can be observed in the CST. These three characteristic

areas are observed for the signal of propene in the CST of BETA coated monoliths

and also for toluene in the case of samples prepared with one and two synthesis

steps. However, it is not observed in the toluene profile corresponding to the samples

Monolith

BETA layer

BETA crystals

Opening filled

10 µm


132

prepared with three synthesis steps (MBETA3), because in this sample toluene

desorbs at higher temperatures than propene, when the zeolite is catalytically active

and toluene combustion takes place. For all the samples, it is observed that, while the

heavier HC (e.g. toluene) is adequately trapped by BETA supported zeolite (even in

the presence of 10% steam), the light HC (propene) desorbs from the HC trap at

lower temperature. To follow the activity of BETA supported zeolite as HC

oxidation catalyst, CO2 profiles during the CST were obtained (results not shown for

brevity). The evolution of CO2 for the three monoliths starts at around 440 ºC, and a

high concentration of this gas was observed, which agrees with the low concentration

of HC at the end of the cold start tests for all coated monoliths (see Figure 5.5). It

should be mentioned that coke deposition in zeolite BETA was evidenced, since grey

samples were recovered after use.

Analyzing the CST profiles for the MBEA1 and MBEA2 samples (see Figure 5.5A

and 5.5B) it is seen that both monoliths are able to adsorb propene and toluene

during the first minutes of the cold start test. Moreover, from the data included in

Table 5.2 (propene and toluene desorption temperature during CST), it is clearly seen

that desorption of both HC occurs at temperatures close to the temperature required

for the TWC to be active (around 200 ºC).

Table 5.2. Temperature for HC desorption (propene and toluene) for the three coated monoliths tested

T desorption (ºC) MBETA1 MBETA2 MBETA3 Propene Toluene Propene Toluene Propene Toluene Cycle 1 200 400 220 470 290 - Cycle 3 190 400 200 450 270 -

(-) not observed in all the experiment tested.

However, the results are much better for the coated monolith with three syntheses

(MBETA3), which has higher amount of supported BETA zeolite (see Table 5.1).

Propene desorption takes place at much higher temperatures and evolution of toluene

Chapter 5

133

is not observed, because, as mentioned before, toluene adsorption is so effective that,

when it desorbs, its combustion takes place. The high performance of BETA coated

monolith (MBETA3) as HC trap can be due to important diffusional problems for the

adsorbed molecules, considering the higher thickness of the layer and the effective

intergrown crystals that may increase the tortuosity of the material. This may delay

significantly the desorption temperature of the HC.

Figure 5.5. Experimental results of the cold start tests for MBETA1, MBETA2and

MBETA3 monoliths after the first and third cycle.

0100200300400500600700

00.0020.0040.0060.008

0.010.0120.0140.016

0 5 10 15 20 25 30

Tem

pera

ture

(ºC

)

HC

conc

entr

atio

n (%

)

Time (min)

MBETA3

First cycle

Third cycle

First cycle

Third cycle

Tol

uene

Prop

ene

C

0

100

200

300

400

500

600

700

00,0020,0040,0060,008

0,010,0120,0140,016

0 5 10 15 20 25 30

Tem

pera

ture

(ºC

)

HC

conc

entr

atio

n (%

)

Time (min)

MBETA1

First cycle

Third cycle

First cycle

Third cycleT

olue

nePr

open

e

A

0

100

200

300

400

500

600

700

00,0020,0040,0060,008

0,010,0120,0140,016

0 5 10 15 20 25 30

Tem

pera

ture

(ºC

)

HC

conc

entr

atio

n (%

)

Time (min)

MBETA2

First cycle

Third cycle

First cycle

Third cycle

Tol

uene

Prop

ene

B


134

In a first approximation, these diffusional problems can be an advantage for this

application. Moreover, the diffusion coefficient of propene is higher than that of

toluene. In this sense, the propene molecules diffuse faster than the toluene

molecules through the adsorbent layer, and thus it is expected that the propane

molecules may reach more internal adsorption sites within the zeolite layer during

the adsorption. Once the desorption temperature of propene is achieved and the

propene molecules leave the adsorption sites, these molecules have to diffuse

through the adsorbent layer to reach the gas phase. Since toluene molecules desorb at

high temperatures, these molecules may provide affective diffusion restrictions to

propene molecules.

The ageing process of the materials has been also analyzed in this work by doing

three cycles of the CST (adsorption/desorption). Figure 5.5 and Table 5.2 also

include the results of CST corresponding to the third cycles. It can be seen that

BETA supported zeolites do not suffer an important aging (only a decrease in the

desorption temperature of propene and toluene of around 20 ºC is observed, keeping,

in the case of the best sample (MBETA3) the complete retention of toluene), and all

preserve their performance after three cycles despite the coke deposition on the

zeolitic surface.

CST results point out that BETA coated monolith (MBETA3) is a highly effective

HC trap for the abatement of cold start emissions. This material shows 100% of

toluene retention, and accomplishes the requested performance as a HC trap

(desorbing propene at temperatures close to 300 ºC and being stable after cycling.

Chapter 5

135

5.4. Conclusions

Cordierite honeycomb monoliths were coated with BETA zeolite by in-situ

crystallization with different zeolite loading after one, two or three consecutive

synthesis steps. Depending on the number of synthesis the properties of the final

zeolite coated material are different. The optimum conditions found are after three

consecutive synthesis steps reaching a high zeolite loading and a stable and compact

zeolite layer. This high loading of zeolite and the effective intergrowth between

crystals obtained by in-situ crystallization, has a positive effect on the supported

zeolite performance as HC trap in cold start conditions. Propene diffusion is affected

by the thickness of the supported zeolite layer, and the coated monolith prepared

after three synthesis steps, MBETA3, has the best performance.

Thus, a highly effective HC trap for the abatement of cold start emissions has been

synthesized and tested in the laboratory. This material shows 100% of toluene

retention, and accomplishes the requested performance as a HC trap, desorbing

propene at temperatures close to 300 ºC, and being stable after cycling.


136

References.

[1] J.H. Park, S.J. Park, I.S. Nam, G.K. Yeo, J.K. Kil, Y.K. Youn, Microporous Mesoporous Mater. 101 (2007) 264-270. [2] N.R. Burke, D.L. Trimm, R.F. Howe, Appl. Catal. B 46 (2003) 97-104. [3] S.P. Elangovan, M. Ogura, M.E. Davis, T. Okubo, J. Phys. Chem. B 108 (2004) 13059-13061. [4] K.F. Czaplewski, T.L. Reitz, Y.J. Kim, R.Q. Snurr, Microporous Mesoporous Mater. 56 (2002) 55-64. [5]A. Iliyas, M.H. Zahedi-Niaki, M. Eic, S. Kaliaguine, Microporous Mesoporous Mater. 102 (2007) 171-177. [6] J.M. López, M.V. Navarro, T. García, R. Murillo, A.M. Mastral, F.J. Varela-Gandía, D. Lozano-Castelló, A. Bueno-López, D. Cazorla-Amorós, Microporous Mesoporous Mater. 351 (2010) 239-247. [7] A. Bueno-Lopez, D. Lozano-Castelló, I. Such-Basáñez, J.M. García-Cortés, M.J. Illán-Gómez, C. Salinas-Marinez de Lecea, App. Catal. B 58 (2005) 1-7. [8] D. Cazorla-Amorós, J. Alcañiz-Monge, A. Linares-Solano, Langmuir 12 (1996) 2820-2824. [9] D. Cazorla-Amorós, J. Alcañiz-Monge, M.A. de La Casa-Lillo, A. Linares-Solano, Langmuir 14 (1998) 4589-4596. [10] K. S. W. Sing, D. H. Everett, R. A. W. Haul, L. Moscou, R. A. Pierotti, J. Rouquerol, T. Siemieniewska, Pure & App. Chem. 57 (1985) 603-619.

Chapter 6. Total oxidation of naphthalene using

palladium nanoparticles supported on BETA, ZSM-5,

SAPO-5 and alumina powders.

Chapter 6

137

6.1. Introduction.

Polycyclic aromatic hydrocarbons (PHAs), such as naphthalene, are environmentally

unacceptable compounds produced as a result of the incomplete combustion or

pyrolysis of organic material [1-9]. As it has been mentioned in Chapter 1,

naphthalene is considered the least toxic and simplest molecule and thus, it is used as

a model compound for this group of pollutants. In the literature, a range of different

technologies have been studied to reduce PAHs emissions [2, 10-15]. Among them,

catalytic combustion is the most promising for the removal of PAHs from polluted

air streams due to its lower operating temperature and higher selectivity towards CO2

[4,16].

Dealing with catalytic oxidation, the type of catalysts used to a large extent is based

on metals such as Pd, Pt, Ru or Co supported on γ-alumina by impregnation [6],

metal oxide catalysts [16,17], mesoporous cerium oxides [4], Pt supported on

mesoporous aluminosilicates [7] or zeolites ion-exchanged with Pt [18].

As mentioned in Chapter 1, the preparation, structure and characterization of

nanoparticles are issues of current interest [19] and Teranishi et al. [20] dealt with

the fundamental aspects of monometallic Pd nanoparticles. Within this field,

unsupported Pd based nanoparticles have been used successfully in our group in the

semi-hydrogenation of phenylacetylene [21], supported on carbon [22] or inorganic

materials [23]. Recently, Pd nanoparticles supported on γ-alumina have been

successfully used in the PrOx reaction [24], indicating the benefits of using PVP

polymer protected Pd nanoparticles instead of Pd catalysts prepared by classical

methods such as impregnation with a Pd compound and further reduction in

hydrogen atmosphere.

Removal of PAHs by catalytic oxidation

138

The objective of this chapter is the use of powder catalysts based on Pd nanoparticles

supported on two different zeolites (BETA and ZSM-5), a molecular sieve SAPO-5,

and a more conventional γ-alumina support for the total oxidation of naphthalene.

Special attention has been paid to the recyclability and stability of the catalyst over

extended time-on-stream experiments. Catalysts have been characterized before and

after use and this is important for assessing catalyst lifetime and stability.

6.2. Experimental.


Four supported catalysts were prepared by deposition of palladium nanoparticles

onto selected inorganic supports. Pd nanoparticles protected by polyvinylpirrolidone

(PVP) were synthesized using the reduction-by-solvent method previously described

[24]. Four inorganic supports were used:

(1) The powder γ-alumina support was provided from Alfa Aesar 99.97% (Ref.

039812).

(2) Two zeolites, BETA and ZSM-5. The procedure followed for the synthesis of

these materials has been described elsewhere [25].

(3) A silicoaluminophosphate molecular sieve SAPO-5 was prepared following

the method described by Campelo et al. [26].

The Scheme 6.1 shows the MFI and AFI structure of the ZSM-5 zeolite and the molecular sieve SAPO-5, respectively.

Chapter 6

139

Scheme 6.1. MFI and AFI structures of ZSM-5 zeolite and SAPO-5 molecular sieve.

The catalysts used in this work were prepared by the impregnation method, as

reported previously by our group [22-24]. In a typical procedure, 3 g of support was

mixed with the appropriate amount of the purified colloidal suspension to yield 1 wt.

% of metallic loading. The solution was stirred for two days at room temperature, in

order to ensure a similar loading and dispersion for the different catalysts, and

transferred to an oven, where it was left at 60ºC in order to evaporate the solvent.

The catalysts were then washed several times with a cold mixture of H2O/EtOH

(50:50, v/v), and left to dry overnight at 60 ºC.

6.2.2. Catalyst characterization.

ZSM-5 and BETA zeolites and SAPO-5 silicoaluminophosphate molecular sieve

were characterized by XRD, using a 2002 Seifert powder diffractometer. The

scanning rate was 2º/min and Cu-Kα radiation was used.

TEM images of the catalysts were recorded in a JEOL (JEM-2010) electron

microscope equipped with an EDS analyzer (OXFORD, model INCA Energy TEM

100) operating at 200 kV with a space resolution of 0.24 nm. For the analysis, a

MFI AFI


140

small amount of the catalyst was suspended in a few drops of methanol, and

sonicated for a few minutes. A drop of this suspension was then deposited onto a

300mesh Lacey copper grid and left to dry at room temperature. The diameter of the

nanoparticles and the dispersion of the metal over the catalyst were determined as

described previously [24].

The textural characterization of the catalysts was carried out by means of the

adsorption of N2 at -196 ºC and CO2 at 0 ºC (Autosorb 6, Quantachrome). Prior to

the adsorption measurements, the catalysts were outgassed under vacuum (10-2 mbar)

at 250 ºC for 4 h to remove any adsorbed impurities. Surface area was calculated

from nitrogen adsorption isotherms using the BET equation (SBET). Total micropore

volume (VDR (N2)) and narrow micropore volume (VDR (CO2)) were calculated by

applying the Dubinin-Radushkevich (DR) equation to the N2 adsorption data at -196

ºC and the CO2 adsorption data at 0 ºC, respectively [27].

The metal loading of the catalysts was analyzed by inductively coupled plasma-

optical emission spectroscopy (ICP-OES), in a PerkinElmer Optima 4300 system.

The extraction of the metal was made by an oxidative treatment of the samples with

aqua regia, followed by filtration of the remaining solid using a nylon membrane

filter (average pore diameter of 400 nm) and dilution of the resulting metal solution

using a volumetric flask.

XPS measurements were made on an Omicron ESCA+ photoelectron spectrometer

using a non-monochromatized MgKαX-ray source (hν = 1253.6 eV). An analyser

pass energy of 50 eV was used for survey scans and 20 eV for detailed scans.

Binding energies are referenced to the C1s peak assumed to have a binding energy of

284.5 eV.

Thermogravimetric analysis (TGA) experiments were performed in a

Thermogravimetric Analyzer (TA Instruments, model SDT 2960). In these experiments,

Chapter 6

141

approximately 10 mg of the catalyst (fresh and used) were treated. The catalyst was

heated up to 900ºC (heating rate of 10ºC/min) and equilibrated for 1 h, under a constant

air flow rate of 100 ml/min.

6.2.3. Catalyst testing.

Catalytic activity tests for naphthalene oxidation were carried out in a fixed bed

reactor (diameter: 1.6cm). The feed stream consisted, in all cases, of 100 ppmv

naphthalene in a mixture of 20% O2 and 80% He. The total flow (F) was set to

50ml/min and a catalyst powder weight (m) of 200 mg (F/m = 15 L/g·h). Analysis of

reactants and reaction products was performed by on-line gas chromatography using

thermal conductivity and flame ionization detectors. Catalytic activity was measured

over the temperature range 100-200 ºC in incremental steps and temperatures were

measured by a thermocouple placed in the catalyst bed connected to a PID controller.

Data were collected at each temperature after a stabilization time of 20 minutes.

Three analyses were made at each temperature. Oxidation activity is expressed as a

yield of carbon dioxide. Furthermore, time-on-line experiments for long-term use of

these catalysts were performed with the aim to probe possible deactivation. For this

purpose, the catalysts were tested using the same reaction conditions as described

above but at a higher temperature of 250ºC for 48 h.


6.3.1. Catalysts characterization.

Firstly, the as synthesized materials, BETA, ZSM-5 and SAPO-5 were characterized

by XRD, confirming the typical structure for each material [25] (see Figure 6.1).

Moreover, after the impregnation method, the catalysts were analysed by XRD and

no change was observed in the prepared catalysts (results not shown), indicating that


142

the process of introducing the Pd nanoparticles did not compromise the structure of

the supports.

Figure 6.1. XRD diffractograms of the supports.

Nitrogen adsorption was performed to investigate the porous texture of the as

prepared materials, and any changes that may have occurred due to nanoparticle

deposition during catalyst preparation. From the N2 isotherms (see Figure 6.2), it is

possible to confirm that BETA, ZSM-5 and SAPO-5 exhibit a Type I isotherm,

which according to IUPAC classification [28] is typical of microporous solids. γ-

Al2O3 shows a Type II isotherm [28] typical of macroporous materials. After

preparing the catalyst, the adsorption/desorption isotherms of the catalysts did not

exhibit any noticeable change in the isotherm shape, but the adsorbed amounts

decreased for all the prepared catalysts, with respect to the original support what is a

consequence of some blockage of porosity by the Pd nanoparticles.

Table 6.1 summarizes the surface areas (SBET), the total micropore volume (VDR(N2))

and narrow micropore volume (VDR(CO2)). Analysing the results in depth, it can be

established that the catalyst preparation method has affected the porosity of the

supports. According to the N2 adsorption isotherms the amount of N2 adsorbed on the

0 10 20 30 40 502θ/º

SAPO-5

ZSM-5

BETA

Chapter 6

143

catalysts diminished in all cases with respect to the supports alone (See SBET and

VDR(N2) in Table 6.1). Pd/BETA and Pd/ZSM-5 showed reduced adsorption

properties to a minor extent. Nevertheless, the Pd/SAPO-5 catalyst showed a

considerably reduced apparent surface area while still being a microporous material

as mentioned previously. The Pd/γ-Al2O3catalyst had reduced specific BET surface

area, although it is a macroporous material. Therefore, it can be ascertained that the

catalyst preparation method affects the porous texture, especially the microporosity.

This occurred to a minor degree for BETA and ZSM-5 supports, but to a more

significant extent for the SAPO-5 support.

Figure 6.2. N2 adsorption/desorption isotherms (-196 ºC) of the selected supports.

Table 6.1. Porous texture characterization of the supports and prepared catalyst. Sample SBET (m2/g) VDR(N2) (ml/g) VDR(CO2) (ml/g)

BETA 570 0.25 0.21 Pd/BETA 510 0.22 0.18 ZSM-5 290 0.14 0.11 Pd/ZSM-5 270 0.12 0.11 SAPO-5 210 0.10 0.18 Pd/SAPO-5 105 0.07 0.13 γ-alumina 70 0.03 0.03 Pd/γ-alumina 50 0.02 0.02

050

100150200250300350400450

0 0,2 0,4 0,6 0,8 1

Am

ount

ads

orbe

d (S

TP)

(cm

3 /g)

P/Po

BETAZSM-5SAPO-5γ-Al2O3γ-Al2O3


144

Figure 6.3 shows the TEM images of the Pd colloid together with the four catalysts

prepared for this work. As can be extracted from the TEM micrographs of the

colloid, the synthesised colloid shows very low polidispersity for the size of the

particles. From the different images of the different catalysts prepared for this work,

the deposition of polymer-protected nanoparticles results in a noticeable change in

Pd particle size, which agrees with our previous studies [22-24]. Therefore, TEM

micrographs show an effect on the size of the particles during catalyst preparation, as

the mean particle size slightly increases from the starting colloid to the final catalyst.

Table 6.2 shows the diameter of the prepared particles (d) before and after deposition

on the supports and the dispersion (D) of the nanoparticles, calculated according to a

procedure described in detail elsewhere, assuming a spherical particle shape [29].

Metal loading (wt %) calculated by ICP analyses are included in Table 6.2. The Pd

loading of the catalysts varies by nearly a factor of 2. The reason for such a

difference is not straightforward and must be related to both the surface roughness of

the supports and to the interaction among the polymer-protected Pd nanoparticles and

the support.

Table 6.2. Nanoparticles sizes obtained from TEM data and metal loading of the different catalysts.

Sample d TEM (nm) D TEM (%)

Metal loading (wt%)

Pd colloid 1.7 ± 0.3 - - Pd/BETA 3.1 ± 0.5 29 0.54 Pd/ZSM-5 2.4 ± 0.4 38 0.74 Pd/SAPO-5 3.1 ± 0.6 29 0.74 Pd/γ-Al2O3 2.6 ± 0.2 35 0.38

Chapter 6

145

Figure 6.3. TEM images and particle size distribution of: (A) colloid; (B) Pd/BETA;

(C) Pd/ZSM-5; (D) Pd/SAPO-5; (E) Pd/γ-Al2O3.


146

6.3.2. Catalytic performance for naphthalene oxidation.

The catalytic activity for naphthalene oxidation was measured for the four catalysts.

Figure 6.4 shows the variation of the catalytic activity (expressed as conversion to

CO2) for naphthalene oxidation, as a function of the reaction temperature. Four

cycles of increasing and then reducing reaction temperature were performed with

each catalyst in order to ensure the stability of the sample.

Figure 6.4. Variation of the catalytic activity for naphthalene oxidation (expressed as yield to CO2) as a function of reaction temperature over the four catalysts.(A) Pd/BETA; (B) Pd/ZSM-5; (C) Pd/SAPO-5; (D) Pd/γ-Al2O3.

The main reaction product observed for naphthalene oxidation was CO2.

Nevertheless, by-products such as phthalic anhydride, naphthanol, ketone, benzoic

acid and 1,2-dibenzoic acid were also detected in very low concentrations under

certain reaction conditions when conversion was relatively low. These compounds

were trapped in acetone using an acetone/ice cold trap at the exit of the reaction

0.0

20.0

40.0

60.0

80.0

100.0

50 100 150 200

Con

vers

ion

to C

O2

(%)

Temperature (ºC)

Pd/BETA

CYCLE 1CYCLE 2CYCLE 3CYCLE 4

A

0.0

20.0

40.0

60.0

80.0

100.0

50 100 150 200

Con

vers

ion

to C

O2

(%)

Temperature (ºC)

Pd/ZSM-5


B

0.0

20.0

40.0

60.0

80.0

100.0

50 100 150 200

Con

vers

ion

to C

O2

(%)

Temperature (ºC)

Pd/SAPO-5


C

0.0

20.0

40.0

60.0

80.0

100.0

50 100 150 200

Con

vers

ion

to C

O2

(%)

Temperature (ºC)


Pd/γ-Al2O3

D

Chapter 6

147

system and then analyzed by GC-MS. The reaction temperature was fixed between

150-165 ºC, depending on the catalyst, and products collected for 2 h. In addition, it

is worth noting that CO was not detected as a reaction product in any of the

experiments performed.

Table 6.3 shows the temperature required for total naphthalene oxidation to CO2 for

each catalyst over the four temperature cycles. Excellent catalytic performance has

been observed for each of the four catalysts. Analyzing these results thoroughly,

Pd/BETA and Pd/ZSM-5, show a lower temperature for total naphthalene

conversion, whilst Pd/SAPO-5 and Pd/γ-Al2O3 are able to reach full naphthalene

conversion at slightly higher temperatures. For comparative purposes, the activity of

the support for the most active catalyst was measured. BETA zeolite was only active

at temperatures above 300 ºC, reaching a conversion to CO2 of 64% at 350 ºC. Hence

the zeolite support without the Pd nanoparticles was totally inactive over the

temperature range where the Pd-based catalyst showed total naphthalene conversion.

Table 6.3. Temperature required for naphthalene total oxidation for the different catalysts for the four oxidation cycles tested and coke deposition after 4 cycles.

T(ºC) for total conversion to CO2 Coke deposited (%) after 4

cycles Sample Cycle

1 Cycle

2 Cycle

3 Cycle

4 Pd/BETA 180 180 165 165 3.3 Pd/ZSM-5 165 165 165 165 0.8 Pd/SAPO-5 195 180 180 180 1.2 Pd/γ-Al2O3 180 180 180 195 0

The Arrhenius plots for the first reaction cycle for all catalysts were made (results

not shown). The linearity of these plots at low conversions (<20%) indicated that the

reaction is in the chemical kinetics controlled regime. Table 6.4 shows the calculated

apparent activation energies (Ea) and the values obtained indicated that there were no

diffusional limitations for the reaction tested. Table 6.4 includes the turnover

frequency (TOF) for the four catalysts during the first catalytic reaction cycle. For


148

comparison purposes, the TOF for a 0.5% Pt/SiO2 catalyst tested previously for the

same reaction, under very similar conditions, [30] has been included. Interestingly,

despite using the same Pd-PVP nanoparticles in all the catalysts tested, the TOF and

Ea were different depending on the support, demonstrating Ea values for the Pd/γ-

Al2O3 which are four times greater than that found for the Pd/BETA catalyst.

Therefore, close inspection of the influence of the support is necessary to explain the

different activity data for the four catalysts tested. Comparing our results for the

different catalysts tested with one prepared by classical impregnation, Pt/SiO2 [30], it

is possible to observe that the Pd-PVP nanoparticles based catalysts have much

greater TOF than a standard silica supported Pt-based catalyst, indicating the

potential of these kind of nanoparticles which can be easily tuned during the

preparation method to adapt their properties to the desired reaction. Hence, with the

aim of explaining the different behaviour of the catalysts special attention needs to be

paid to the important influence of the support on the catalytic activity. In this respect,

supports which strong surface acidity may favour the naphthalene oxidation reaction

when used in conjunction with Pd-PVP nanoparticles.

Table 6.4. Turnover frequency (TOF) at 125ºC and apparent activation energy for the four catalysts at first reaction cycle.

Sample Data from first cycle TOF (s-1) (T=125ºC) Ea (kJ/mol)

Pd/BETA 0.075 46 Pd/ZSM-5 0.056 55 Pd/SAPO-5 0.010 126 Pd/γ-Al2O3 0.019 166 0.5%Pt/SiO2 [30] 0.0007 --

The higher activity of the ZSM-5 and BETA catalysts may be associated with the

porosity of the catalyst. Thus, Pd/BETA and Pd/ZSM-5 both present a significant

surface area that may have an influence on the activity during the total oxidation, and

therefore, there may be a relationship between the apparent surface area of the

Chapter 6

149

catalyst and its capacity to oxidize naphthalene. Moreover, the pore size distribution

and channel interconnectivity may also have an important contribution, in agreement

with Puertolas et al. [4], who studied the use of different mesoporous cerium oxides

prepared through a nanocasting method for the total oxidation of naphthalene,

mentioning that wider pore size distributions and pore interconnectivity could

facilitate the total oxidation of naphthalene. In this sense, the zeolite supports BETA

and ZSM-5, which have a three-dimensional pore size distribution, present a high

interconnectivity, in contrast to SAPO-5 acting as a support which presents a one-

dimensional porosity with a larger pore size than zeolites BETA or ZSM-5.

Considering the pore size of the selected supports, it is clear that the Pd nanoparticles

will be deposited solely on the outer surface of the microporous zeolites and

silicoaluminophosphate, whereas in the case of the macroporous support (γ-Al2O3)

the nanoparticles will be deposited both in the inner and outer surface. The location

of the particles indicates that while the inner surface area of the selected supports

does not play a significant role in the oxidation of naphthalene as such, it clearly

influences the outcome of the catalytic reaction. The supports with a larger surface

area, in this respect, may adsorb larger amounts of naphthalene in their porous

structure, thus acting as “naphthalene reservoirs” which feed the Pd particles where

the catalytic oxidation takes place, which could help to explain our results.

It is well known that BETA zeolite has a relatively high external surface area

resulting from its structure being the combination of two intergrown polymorphs.

This high external surface area may result in increased accessibility for naphthalene

molecules to reach the Pd nanoparticles of the catalyst, as well as a possible reservoir

for naphthalene molecules. As a consequence, total oxidation of naphthalene to CO2

can be reached at lower temperatures. It is also possible that the adsorption of

naphthalene onto the catalysts could be significant and could provide a reservoir for

naphthalene. In order to verify this, the Pd/BETA, Pd/ZSM-5 and Pd/SAPO-5


150

catalysts were tested under the same conditions as the aforementioned oxidation

experiments, but without O2 in the reactant flow, in order to study the contribution of

the catalyst towards naphthalene adsorption. The catalyst Pd/γ-Al2O3 was excluded

from this experiment, because from the catalytic oxidation experiments it became

evident that the Pd/γ-Al2O3 catalyst did not exhibit any significant naphthalene

adsorption during the reaction time of about 1h and at the lowest temperature used.

That is, naphthalene in the gas stream leaving the reactor at this temperature

contained 96 ppm of naphthalene, which corresponds to 96% of the initial

concentration in the inlet stream.

Figure 6.5 shows the amount of naphthalene adsorbed by the different catalysts

during a time similar to that used in the experiments in the presence of oxygen and

over the temperature range studied. According to these results, the amount of

adsorbed naphthalene in Pd/SAPO-5 is lower when compared with the zeolite based-

catalysts. Furthermore, there is no appreciable naphthalene adsorption above 150 ºC.

On the contrary, Pd/BETA and Pd/ZSM-5 present a much larger adsorption capacity

for naphthalene at higher temperatures. This adsorbed naphthalene may act as a

“reservoir” to feed the catalysts in the initial stages of the reaction, thus increasing

the efficiency of the catalyst. Considering this feature, Pd/BETA and Pd/ZSM-5 may

permit naphthalene oxidation at lower temperatures than Pd/SAPO-5. Therefore, the

three dimensional pore structure and the high adsorption capacity seem to play an

important role in the performance of the catalyst.

Chapter 6

151

Figure 6.5. Variation of the amount of naphthalene adsorbed as a function of

temperature on the three catalysts.

Although Pd/BETA was the most active catalyst after four cycles, another

outstanding feature was the reduction of the temperature at which total conversion to

CO2 was achieved as the number of cycles was increased, whilst the temperature for

total oxidation for Pd/ZSM-5 remains unchanged after four cycles. This could be

explained by considering that in the case of the Pd/BETA catalyst there was coke

deposition during the first cycle (evident by a change of colour from grey to black

upon completion of the first catalytic cycle). As the catalyst was used coke

deposition was reduced, producing naphthalene oxidation to CO2 at lower

temperatures. In this sense, our results reveal that the amount of coke deposited after

4 cycles (see Table 6.3) does not affect the catalytic activity of the studied systems in

terms of deactivation.

Another parameter that can have an influence on the catalytic activity is the stability

of the samples. According to these results, the catalytic activity for Pd/BETA,

Pd/ZSM-5 and Pd/SAPO-5 remained unchanged after four cycles and maintained

high naphthalene oxidation activity towards the formation of CO2 (see Figure 6.4). In

contrast, Pd/γ-Al2O3 showed an increase in the temperature for total oxidation after

0.0

20.0

40.0

60.0

80.0

100.0

50 100 150 200

Nap

htha

lene

ads

orpt

ion

(%)

Temperature (ºC)

Pd/BETAPd/ZSM-5Pd/SAPO-5


152

the completion of the four cycles and this can tentatively be attributed to the

deactivation of the catalyst. With the aim to distinguish differences between

deactivation of these three catalysts, it is necessary to analyse their behaviour at

intermediate temperatures. Therefore, in order to probe potential deactivation,

catalytic activity at 150ºC is summarized in Table 6.5. Despite appearing to be

relatively stable after four temperature cycles, the yield of CO2 during incomplete

conversion at 150ºC can provide further information on which is the most suitable

for naphthalene total oxidation. Consequently, focusing our attention on these values

the Pd/BETA catalyst is not affected after four cycles, while Pd/ZSM-5 and

Pd/SAPO-5 showed reduced overall conversion after four catalytic cycles, this

decrease being more moderate for Pd/SAPO-5 and more pronounced for Pd/ZSM-5.

Table 6.5. CO2 conversion at 150ºC for the four catalysts and during the four cycles tested.

CO2 conversion (%) (T=150 ºC) Sample Cycle 1 Cycle 2 Cycle 3 Cycle 4 Pd/BETA 16 49 51 56 Pd/ZSM-5 24 59 31 28 Pd/SAPO-5 12 34 42 33 Pd/γ-Al2O3 29 48 37 34

TEM analyses were performed on the used catalysts after reaction, with the aim to

analyse any significant effect of the reaction conditions on the catalyst. Figure 6.6

illustrates the TEM micrographs of the four used catalysts and the diameter of the Pd

particles (d) and the dispersion of the catalysts (D) after the reaction cycles are

included in Table 6.6. From the TEM analysis it can be concluded that all the Pd

nanoparticles in the catalysts have increased in size. Despite this increase in particle

size, only the Pd/alumina catalyst has shown a significant increase in the temperature

for total oxidation after four cycles (see Table 6.3). Therefore, the rest of the

catalysts have demonstrated high stability for oxidation of naphthalene and the

increase of Pd particle size has not had a major detrimental effect on activity.

cy

X

re

th

cu

co

pr

Figure 6.6cycles of nap

XPS analyse

esults corres

he catalyst s

urve fitting

orrespondin

resence of P

6. TEM imagphthalene o

s were perf

sponding to

urface and

g. The XP

ng to the 3d

Pd(II) on th

ges and paroxidation: (A

formed on t

o the atomic

the binding

S binding

d3/2 and 3d5/

e surface of

153

rticle size dA) Pd/BETA

Pd/γ-Al2O

the fresh an

c percentage

g energy of e

energies r

/2 transition

f the Pd nan

distribution oA; (B) Pd/ZSO3.

nd used cat

e of the diff

each of the

revealed th

s. In the ca

noparticles

of catalystsSM-5; (C) P

talysts. Tabl

ferent Pd sp

species det

he presence

ase of prepa

is attributed

Chap

usedafter fo

Pd/SAPO-5;

le 6.7 show

pecies prese

termined thr

e of two p

ared sample

d to the carb

pter 6

our ; (D)

ws the

ent on

rough

peaks

s, the

bonyl


154

function of the protecting polymer interacting with the metallic palladium, removing

electron density from the surface, as has been established in our previous work [24]

and by Somorjai et al. for Pt nanoparticles capped with this kind of agent [31-33]. In

short, what appears in this case as Pd(II) in the XPS spectra, is attributed to the

palladium atoms on the surface of the nanoparticle, which are electron deficient due

to their interaction with the protecting polymer. Following on, the greater amount of

Pd (II) present in the Pd/BETA with respect to the other catalysts could be due to a

stronger interaction of Pd with PVP than for PVP with the support, which in turn

may affect the catalytic activity. In contrast, Pd/ZSM-5, Pd/SAPO-5 and Pd/γ-Al2O3

catalysts have a stronger PVP-support interaction that produces a reduction in the

electronic density withdrawn and therefore, the quantity of Pd (0) is higher.

Table 6.6. Results of Pd nanoparticle analysis by TEM for catalysts after testing in naphthalene oxidation.

Sample d TEM (nm) D TEM (%) Pd/BETA 4.1 ± 1.0 22 Pd/ZSM-5 5.0 ± 1.5 18 Pd/SAPO-5 4.1 ± 1.0 22 Pd/γ-Al2O3 3.6 ± 0.7 25

Once all the catalysts were tested, the amount of Pd (II) increased in all the catalysts

except on the surface of the Pd/BETA catalyst, indicating that the Pd/nanoparticles

were partially oxidised, producing Pd (II) species on the Pd nanoparticle surface. On

the one hand it could be stated that the Pd nanoparticle is protected by the PVP due

to a stronger interaction of Pd with PVP and this makes the Pd nanoparticle more

resistant towards oxidation. On the other hand, the Pd/BETA catalyst increased the

intensity of the Pd (0) signal. This phenomenon could be attributed to the Pd/BETA

catalyst being more stable due to the aforementioned stronger interaction between the

Pd nanoparticles and the protecting polymer. The importance of the protection can be

better assessed upon analyzing the conversion of naphthalene at 150ºC (see Table

6.5). Thus, despite having the lowest Pd(0)/Pd(II) ratio in the fresh catalyst (see

Chapter 6

155

Table 6.7), the PVP-Pd interaction explains the absence of deactivation of this

catalyst as the sample is submitted to successive reaction cycles. On the contrary, the

rest of the prepared catalysts, which have a higher initial Pd(0)/Pd(II) ratio, have a

lower conversion at 150 ºC (see Table 6.5) because the interaction between the Pd

nanoparticles and the protecting polymer is less intense than the Pd/BETA catalysts,

rendering them more susceptible to surface oxidation. Another possible reason that

could be preventing Pd nanoparticle oxidation is associated with the coke deposition

produced on the Pd/BETA catalyst. Apart from zeolite BETA acting as a cracking

catalyst, the Pd (II) produced on the surface is being reduced due to this enhanced

effect between the Pd nanoparticles and BETA zeolite that can promote the reduction

of Pd (II) at the surface to Pd (0) and the oxidation of part of this deposited coke.

Therefore, the Pd/BETA catalyst has different behaviour, making it more stable than

the rest of the catalysts.

Table 6.7. XPS analysis of the four catalysts used, before and after use for naphthalene oxidation.

Sample BE Pd 3d5/2 Pd percentage (%) Prepared Tested Prepared Tested

Pd/BETA Pd(0) 335.1 335.3 47 52 Pd(II) 336.4 336.7 53 48

Pd/ZSM-5 Pd(0) 334.9 335.1 64 40 Pd(II) 336.9 337 36 60

Pd/SAPO-5 Pd(0) 335.2 335.3 80 65 Pd(II) 337.1 337.2 20 35

Pd/γ-Al2O3 Pd(0) 335.6 335.8 69 54 Pd(II) 336.8 337.0 31 46

The Pd/BETA catalyst was identified as the most effective metal-supported catalyst

for naphthalene total oxidation. Therefore, long-term use of this catalyst, in order to

start to evaluate the catalyst life time, was performed. Earlier studies indicated that

activity was stable and the catalyst was able to produce total naphthalene conversion

at low temperatures between 165-180 ºC. An additional accelerated ageing stability

study was conducted at 250 ºC for 48 h. Figure 6.7 shows the time-on-line data for


156

the Pd/BETA catalyst. It was evident that the catalyst was stable for at least 48 h, and

there was no evidence for any catalyst deactivation over the test period. The yield of

CO2 at the beginning of the experiment was 100% and it was not diminished after

48h. Although the Pd/BETA was the best catalyst, the time-on-line activity of

Pd/ZSM-5 was also analyzed in the same manner and it showed equally stable 100%

conversion to CO2 over the reaction time (results not shown).

With respect to naphthalene total oxidation, in the literature it is possible to find

several research studies focused on using metal supported catalysts, prepared by

conventional methods, or metal oxide catalysts for naphthalene oxidation.

Neyestanaki et al. [18] have studied the use of ZSM-5 impregnated with different

concentrations of Pd to remove the emission of pollutants from the combustion of

biofuels. In this study, the catalysts were used for four consecutive cycles, with three

different gas mixtures varying the composition of different polluting gases including

always 50 ppm of naphthalene. In terms of naphthalene conversion, it can be seen

that our catalyst based on Pd nanoparticles exhibits a naphthalene total conversion

between 165-180 ºC. Meanwhile, the catalyst prepared by the impregnation method

from the aforementioned work needed a temperature of 400 ºC to achieve total

conversion for the Na-form of the ZSM-5 zeolite. On the other hand, the H-form of

the ZSM-5 reduced the temperature for total conversion to 200 ºC, however, catalyst

deactivation was observed. Also using a Pd-based catalyst, Zhang et al. [6] studied

the oxidative decomposition of naphthalene (100 ppm) by a supported Pd/γ-Al2O3

catalyst, and it demonstrated 90% conversion to CO2 at 285 ºC, but total conversion

was not reached until approximately 400 ºC. Furthermore, Garcia et al. [16] studied

the total oxidation of naphthalene (100 ppm) with a Pd/TiO2 catalyst, which

produced a total conversion to CO2 at approximately 270oC, which was a lower

temperature than Zhang et al. [6]. In addition, it was reported that naphthalene

physisorption on the catalyst had an important role in the catalyzed reaction with

Pd/TiO2. Therefore, our catalysts supported on zeolites based on PVP stabilized Pd

Chapter 6

157

nanoparticles appear to be more stable than those prepared by conventional

impregnation methods, and the total conversion to CO2 is achieved at lower

temperatures.

Figure 6.7. Conversion of naphthalene to CO2 as a function of time-on-line at 250oC

for the Pd/BETA catalyst.

Comparing our results with those obtained with other kinds of catalysts, it is possible

to observe that the total oxidation of naphthalene is reached at lower temperatures

with our zeolite supported catalysts. In the case of metal oxide catalysts [7] such as

CeO2 or mixed oxides (CuZnO) full conversion is reached at temperatures between

250 ºC and 270 ºC, respectively. Whilst for the case of metals supported on different

oxides such as Pt/SiO2 [30] the optimum temperature for naphthalene oxidation was

250 ºC. Shie et al. [34] indicated that Pt/γ-alumina decreased the reaction

temperature for 95% naphthalene total oxidation to 207 ºC. On the other hand, Pt

supported on mesoporous materials have also been used. Park et al. [35] have

prepared Pt/MCM41 and Pt/SM41 (SM: molecular sieve) and used them as catalysts

for the total oxidation of naphthalene in dry and wet conditions, establishing the

reaction temperature for total conversion to CO2 at 300ºC. To conclude, the

combination of Pd-based nanoparticles with zeolites, BETA and ZSM-5, permits the

total oxidation of naphthalene at substantially lower temperatures (165ºC) than many

catalysts reported in the literature.

0

20

40

60

80

100

0 5 10 15 20 25 30 35 40 45 50

Con

vers

ion

to C

O2

(%)

Time (h)


158

6.4. Conclusions.

Zeolites BETA and ZSM-5, silicoaluminophosphate molecular sieve, SAPO-5 and γ-

alumina impregnated with Pd nanoparticles have been used as catalysts for the total

oxidation of naphthalene. All the catalysts produced high activity for naphthalene

conversion to CO2, with total conversion taking place between 165 and 180 ºC.

Pd/BETA was the most active catalyst. Furthermore, all the catalysts have high

stability because their properties remain largely unchanged after testing for several

oxidation cycles. Time-on-line experiments have been used to test the stability of the

catalysts. The Pd/ZSM-5 and Pd/BETA samples were stable after accelerated ageing

time-on-line for 48 at 250 ºC. Therefore, the use of Pd-based nanoparticles supported

on zeolites (BETA or ZSM-5) could be potential options for the abatement of PAHs

by catalytic oxidation.

Chapter 6

159

References

[1] B. Solsona, T. Garcia, R. Murillo, A.M. Mastral, E. NtainjuaNdifor, C. E. Hetrick, M. D. Amiridis, S. H. Taylor, Top. Catal. 52 (2009) 492-500. [2] M.H. Yuan, C.Y. Chang, J.L. Shie, C.C. Chang, J.H. Chen, W.T. Tsai, J. Hazard. Mater. 175 (2010) 809-815. [3] F. Diehl, J. Barbier Jr., D. Duprez, I. Guibard, G. Mabilon, Appl. Catal. B 95 (2010) 217-227. [4] B. Puertolas, B. Solsona, S. Agouram, R. Murillo, A.M. Mastral, A. Aranda, S.H. Taylor, T. Garcia, Appl. Catal. B 93 (2010) 395-405. [5] A.M. Mastral, M.S. Callen, Environ. Sci. Technol. 34 (2000) 3051-3057. [6] X. Zhang, S. Shen, L.E. Yu, S. Kawi, K. Hidajat, K.Y. Simon Ng, Appl. Catal. A 250(2003) 341-352. [7] J.I. Park, J.K. Lee, J. Miyawaki, S.H. Yoon, I. Mochida, J. Ind. Eng. Chem. 17 (2011) 271-276. [8] E. N. Ndifor, T. Garcia, S. H. Taylor, Catal. Lett. 110 (2006) 125-128. [9] S.C. Kim, S.W. Nahm, W.G. Shim, J.W. Lee, H. Moon, J. Hazard. Mater 141 (2007) 305-314. [10] E.N.Ndifor, S.H. Taylor, Top. Catal. 52 (2009) 528-541. [11] C.C. Chang, C.Y. Chiu, C.Y. Chang, C.F. Chang, Y.H. Chen, D.R. Ji, Y.H. Yu, P.C. Chiang, J. Hazard. Mater 161 (2009) 287-293. [12] L. Yu, X. Li, X. Tu, Y. Wang, S. Lu, J. Yan, J. Phys. Chem. A 114 (2010) 360-368. [13] M.H. Yuan, Y.Y. Lin, C.Y. Chang, C.C. Chang, J.L. Shie, C.H. Wu, IEEE Transactions On Plasma Science 39 (2011) 1092-1098. [14] S.C. Marie-Rose, T. Belin, J. Mijoin, E. Fiani, M. Taralunga, F. Nicol, X. Chaucherie, P. Magnoux, Appl. Catal. B 90 (2009) 489-496. [15] A. Bampenrat, V. Meeyoo, B. Kitiyanan, P. Rangsunvigit, T.Rirksomboon, Catal. Commun. 9 (2008) 2349-2352. [16] T. Garcia, B. Solsona, D. Cazorla-Amorós, Á. Linares-Solano, S.H. Taylor, Appl. Catal. B 62 (2006) 66-76. [17] T. Garcia, B. Solsona, S.H. Taylor, Appl. Catal. B 66 (2006) 92-99. [18] A. K. Neyestanaki, L.E Lindfors, T. Ollonqvist, J. Väyrynen, Appl. Catal. A 196 (2000) 233-246. [19] M. Moreno-Mañas, R. Pleixats, Acc. Chem. Res. 36 (2003) 638-643. [20] T. Teranishi, M. Miyake, Chem. Mater. 10 (1998) 594-600.


160

[21] S.Domínguez-Domínguez, Á.Berenguer-Murcia, D.Cazorla-Amorós,Á.Linares-Solano, J. Catal. 243 (2006) 74-81. [22] S. Domínguez-Domínguez, Á.Berenguer-Murcia, B.K. Pradhan, Á.Linares-Solano, D.Cazorla-Amorós, J. Phys. Chem. C 112 (2008) 3827-3834. [23] S.Domínguez-Domínguez, Á.Berenguer-Murcia, Á.Linares-Solano, D.Cazorla-Amorós, J. Catal. 257 (2008) 87-95. [24] I. Miguel-García, Á. Berenguer-Murcia, D. Cazorla-Amorós, Appl. Catal. B 98 (2010) 161-170. [25] J.M. López, M.V. Navarro, T. García, R. Murillo, A.M. Mastral, F.J. Varela-Gandía, D. Lozano-Castelló, A. Bueno-López, D. Cazorla-Amorós, Microporous Mesoporous Mater. 130 (2010) 239-247. [26] J.M. Campelo, F. Lafont, J.M. Marinas, J. Catal. 156 (1995) 11-18. [27] D. Lozano-Castelló, F. Suárez-García, D. Cazorla-Amorós, Á. Linares-Solano, in: F. Beguin, E. Frackowiak (Eds), Carbon Materials for Electrochemical Energy Storage Systems, Ed. CRC Press (2010) pp. 115-162 [28] K. S. W. Sing, D. H. Everett, R. A. W. Haul, L. Moscou, R. A. Pierotti, J. Rouquerol, T. Siemieniewska, Pure&App. Chem. 57 (1985) 603-619. [29] Boudart, Kinetics of Heterogeneous Catalytic Reactions, Ed. Princeton Univ. Press, (1984). [30]E. Ntainjua, A.F. Carley, S.H. Taylor, Catal. Today 137 (2008) 362-366. [31]Y. Borodko, S.E. Habas, M.Koebel, P. Yang, H. Frei, G.A. Somorjai, J. Phys. Chem. B 110(2006) 23052-23059. [32] Y. Borodko, S.M. Humphrey, T.D. Tilley, H. Frei, G.A. Somorjai, J. Phys. Chem. C 111 (2007) 6288-6295. [33] J.K. Navin, M.E. Grass, G.A. Somorjai, A.L. Marsh, Anal. Chem. 81 (2009) 6295-6299. [34] J. Shie, C. Chang, J. Chen, W. Tsai, Y. Chen, C. Chiou, C. Chang, Appl. Catal. B 58 (2005) 289-297. [35] J. Park, J. Lee, J. Miyawaki, W. Pang, S. Yoon, I. Mochida, Catal. Commun. 11 (2010) 1068-1071.

Chapter 7. Total oxidation of

naphthalene at low temperatures using

palladium nanoparticles supported on inorganic oxide-coated cordierite honeycomb monoliths.

Chapter 7

161

7.1. Introduction.

Polycyclic aromatic hydrocarbons (PHAs), such as naphthalene, are environmentally

hazardous compounds produced as a result of the incomplete combustion or

pyrolysis of organic material [1-3]. Naphthalene is considered the least toxic and

simplest molecule and thus, it is used as a model compound for this group of

pollutants. As mentioned in Chapter 6, catalytic combustion is the most promising

for the removal of PAHs from polluted air streams [2]. Based on our previous

Chapter, the use of polymer protected Pd nanoparticles supported on zeolites (BETA

and ZSM-5) is the most promising for the total combustion of naphthalene due to the

low oxidation temperature reached (165-180 ºC) and the high stability of the

prepared catalyst.

However, for practical applications the use of structured catalysts is preferred and

thus, powder catalysts are less attractive than monolithic substrates (see Chapter 1).

Concerning the deposition method, different strategies can be performed for coating

the monolith walls with a support material like alumina, silica, zeolites… followed

by the impregnation of the active phase [4]. In the case of zeolites or SAPO

materials, there are two ways to coat the monoliths: hydrothermal synthesis using

either direct synthesis or seeded growth (we will call this method as in-situ synthesis)

and dip-coating or wash-coating, which consists in the deposition from a slurry of

zeolite particles followed by a stabilizing thermal treatment [5].

The objectives of the present study are (i) to prepare catalysts based on cordierite

monoliths coated with BETA and ZSM-5 zeolites, SAPO-5 molecular sieve and γ-

Al2O3, and subsequently impregnated with Pd-based nanoparticles, and (ii) to study

of the total oxidation of naphthalene, paying attention to the stability and

recyclability of the structured catalyst in long time-on-stream experiments.


162

7.2. Experimental.

7.2.1. Coating of BETA, ZSM-5 and SAPO-5.

The growth of BETA and ZSM-5 zeolites and a silicoaluminophophate molecular

sieve (SAPO-5) was carried out by in-situ synthesis onto the cordierite monolith

avoiding the use of binders. The general experimental procedure (six steps) for the

synthesis of the zeolites and the silicoaluminophosphate onto the cordierite monolith

was the same as exposed in Chapter 5.

In this case, two consecutive synthesis steps were performed. Thus, after finishing

the whole described procedure for the first time, the samples were placed again in an

autoclave for a second synthesis run using a fresh synthesis solution (i.e. points (ii) to

(vii) were repeated, see Chapter 5). After preparing the coated monoliths using two

consecutive syntheses, the removal of the template was carried out by calcination at

the desired temperature for each material in a muffle furnace under static air

conditions (heating rate 1 ºC/min). Finally, the total weight increase for both the first

coating step performed and the whole coating process was measured. The amount of

weight gained is given as amount loaded (wt %).

Coating of the monoliths with zeolite BETA was carried out following the

methodology described in Chapter 5. Two consecutive synthesis steps with a

crystallization temperature of 132 ºC for 48 hours were performed. The BETA

coated monoliths, named MBETA, were submitted to a calcination step to remove

the template at 500 ºC for 6 hours.

ZSM-5 monoliths, which will be called as MZSM-5 in this work, were obtained after

two consecutive synthesis steps. As it was mentioned previously by Ulla et al. [6],

the coating of high aluminium content zeolites is hindered in Al-rich supports such as

cordierite and therefore it is necessary to combine two synthesis protocols using

materials with different Si/Al ratios. For that reason, the first-crystallization step was

Chapter 7

163

the preparation of a silicalite-1 coating (which is isostructural to ZSM-5 zeolite but

with no aluminium content in its framework) in order to ensure the crystallization

and anchoring of the ZSM-5 zeolite on the monolith. Thus, the synthesis was

performed adapting the methodology for the preparation of silicalite-1/carbon

membranes reported by Berenguer-Murcia et al. [7]. The silicalite-1 solution was

obtained by adding 3.420 g of TEOS (tetraethoxysilane) dropwise in a 1M

Tetrapropylammonium hydroxide TPA-OH solution (4.920 g TPA-OH solution, in

31.570 g of distilled water). The resulting solution was aged for 90 min and the

monoliths were coated with a layer of silicalite-1 after a crystallization time of 6

hours at 180 ºC. The second crystallization step consisted in the preparation of a

ZSM-5 zeolite coating. The details of the experimental procedure used for the

preparation of ZSM-5 zeolite are given elsewhere [8,9]. According to reference [8],

the crystallization conditions are 190ºC during 10 hours. Template removal was

carried out by calcination in static air in a furnace at 550 ºC during 4 hours.

SAPO-5 monoliths, named MSAPO-5, were prepared by adapting the procedure

described by Campelo et al. [10] for the preparation of powdered SAPO-5. The

crystallization was performed at 200 ºC and crystallization times between 18 and 24

hours were used for studying the coating process. The calcination process for

MSAPO-5 monoliths was performed at 600ºC during 6 hours.

7.2.2 . Coating of γ-Al2O3.

Coating of γ-Al2O3 on honeycomb cordierite monolith was performed by dip-coating

and the samples were labeled as M-γ-Al2O3. Prior to the dip-coating step, the

cordierite monoliths were wrapped with Teflon tape to avoid deposition of γ-Al2O3

on their outer surface. M-γ-Al2O3 monoliths were prepared following the procedure

described by Villegas et al. [11]. The γ-Al2O3 powder was dispersed by HNO3 (γ-

Al2O3/H2O = 25 wt. %, HNO3/ γ-Al2O3 = 2 mmol/g) using a high-shear mixer


164

(ULTRA TURRAX T25, IKA Labortechnik). Initially, the required amount of γ-

Al2O3 powder was added to the acid solution for 10 min at 7000 rpm. After vigorous

stirring (14000 rpm) for 5 min at room temperature, the stable suspension was used

for dip-coating. Series of 8 monoliths were prepared with each dip-coating

suspension, in order to check the reproducibility of the coated monoliths. The

monoliths were immersed vertically into the suspension for 2 minutes and the excess

suspension was blown using compressed air. Finally, the monoliths were dried at

room temperature for 24 hours while rotated horizontally to ensure an optimum

solids distribution. The procedure was performed twice with a fresh suspension. M-γ-

Al2O3 monoliths were calcined in static air at 600ºC for 4 hours (heating rate

1ºC/min).


Palladium nanoparticles protected by polyvinylpirrolidone (PVP) were synthesized

by the reduction-by-solvent method as reported by Miguel-García et al. [12]. The

catalysts used in this work were prepared by the impregnation method, as reported in

the precious chapter. The aforementioned coated monoliths were wrapped with

Teflon tape in order to fix them vertically in 10 mL Teflon liners and to prevent

deposition of Pd nanoparticles on the outer surface of the cordierite monolith. The

necessary amount of Pd nanoparticles suspension in methanol was added bearing in

mind the total amount of coated BETA, ZSM-5, SAPO -5 or γ-Al2O3 in the monolith

to yield a nominal 1wt.% of metallic loading with respect to the zeolite, SAPO or

Al2O3 loadings. To ensure the complete immersion of the coated monoliths in the

suspension additional methanol was added for a final volume of 5 ml. Then, Teflon

liners filled with the monolith and Pd suspension were introduced in a thermostated

bath, stirred at 60 rpm at room temperature for two days. Later, they were transferred

to an oven, where they were left at 60 ºC in order to evaporate the solvent and obtain

Chapter 7

165

the final structured honeycomb monolith, labeled Pd/MBETA, Pd/MZSM-5,

Pd/MSAPO-5 and Pd/Mγ-Al2O3.

7.2.4. Characterization of the catalysts.

ZSM-5 and BETA zeolites and SAPO-5 silicoaluminophosphate molecular sieve

supported on cordierite monolith were characterised by XRD, using a 2002 Seifert

powder diffractometer. The scanning rate was 2º/min and Cu-Kα radiation was used.

The coated monoliths were characterized by SEM in a JEOL microscope (model

JSM-80). The monoliths were carefully cut parallel to the monolith axis. This

working procedure allowed the analysis of the surface of the monolith channels and

the thickness and homogeneity of the coated layer. In the case of the prepared

catalysts, they were cut in the same manner and analyzed by SEM in a HITACHI S-

3000 microscope. Energy-dispersive X-ray spectrometry (EDX) was used to

ascertain the Pd concentration along the channels and the different cuts were

performed for each of the prepared catalysts.

Textural characterization of the coated monoliths and the catalysts was carried out by

means of adsorption of N2 at -196 ºC and CO2 at 0 ºC (Autosorb 6, Quantachrome).

Prior to the adsorption measurements, the samples were outgassed under vacuum

(10-2 mbar) at 250 ºC for 4 h to remove any adsorbed impurities. Surface area was

calculated from nitrogen adsorption isotherms using the BET equation (SBET). Total

micropore volume (VDR (N2)) was calculated applying the Dubinin-Radushkevich

(DR) equation to the N2 adsorption data at -196 ºC [13].


Catalytic activity tests for naphthalene oxidation were carried out in a fixed bed

reactor (D=1.6 cm). The feed stream consisted, in all cases, of 100 ppmv naphthalene


166

in a mixture of 20% O2 and 80% He. The total flow was set to 50 ml/min (GHSV

1220 h-1). Analysis of reactants and reaction products was performed by on-line gas

chromatography using thermal conductivity and flame ionization detectors. The

catalytic activity was measured over the temperature range 100-200ºC in incremental

steps and temperatures were measured by a thermocouple placed in the catalyst bed

connected to a PID controller. Data were collected at each temperature after a

stabilization time of 20 mins. Three analyses were made at each temperature.

Oxidation activity is expressed as a yield of carbon dioxide. Furthermore, time-on-

line experiments for long-term use of these catalysts were performed with the aim to

determine their life time. For this purpose, the catalysts were submitted to reaction

conditions at 250 ºC for 48h.


7.3.1. Coated monoliths characterization.

Table 7.1 shows the percentage of weight increase on coated monoliths after the first

and second coating steps. As discussed in Chapter 5, in order to improve zeolite

loading and deposit a thin and homogeneous zeolite layer, autoclave rotation during

the hydrothermal treatment and a second synthesis step are necessary. Furthermore,

all the coated zeolites and SAPO materials are highly stable towards sonication,

indicating a good adherence of the zeolite on the monoliths surface. Therefore, the

error in the estimation of the weight increase after two coating steps is reasonably

low, considering the complex process under study (including, crystallization,

cleaning or calcinations steps). M-γ-Al2O3 monoliths, prepared in batches of 8

monoliths, possess a high reproducibility with a weight increase and standard

deviation similar to Villegas et al. [11].

Chapter 7

167

Table 7.1. Monolith weight increase (wt %) for single and two steps coatings. Coated monolith First coating step (wt %) Second coating step (wt %) MBETA 3.6±0.4 17.7±2.0 MZSM-5 4.0±0.4 15.8±1.8 MSAPO-5 4.9±0.9 16.1±1.7 Mγ-Al2O3 9.4±1.4 15.0±1.7

X-ray diffraction was used to confirm the phase purity and crystallinity of the zeolite

or SAPO layers coated on the cordierite monoliths after two synthesis steps. Figure

7.1 compiles the diffractograms of uncoated cordierite and coated monoliths

(MBETA, MZSM-5 and MSAPO-5). Powder zeolites diffractograms have been

added to confirm the crystallinity of the coating layer; the diffractogram show the

characteristic peaks of the powder zeolites [9]. Therefore, it is possible to observe

that the coated monolith shows the characteristic peaks of cordierite and those of the

coating zeolite, confirming the crystallinity of the supported zeolite on the coated

samples.

Figure 7.1. XRD diffractograms of cordierite, coated monoliths and powder zeolites.

5 10 15 20 25 30 35 40 45 502θ

Cordierite

MBETAPowder Beta

MZSM-5Powder ZSM-5

MSAPO-5

Powder SAPO-5

●

●●

●●●●

●


168

In the case of MSAPO-5 monoliths, a synthesis time between 18-24 h was used.

From the analysis of the XRD of the samples prepared after 18, 20, 22 and 24 h

(results not shown), it was confirmed that only the MSAPO-5 prepared after 18h

(XRD included in Figure 7.1) showed the characteristic peaks of the SAPO-5

silicoaluminophosphate (AFI structure) without the appearance of other crystalline

phases. For that reason, all the monoliths synthesized for this work were prepared

with a synthesis time of 18 h.

Nitrogen adsorption was performed to analyze the porous texture of the as-prepared

materials. From the nitrogen isotherms (see Figure 7.2) it is possible to confirm that

cordierite exhibits negligible porosity; therefore, it can be assumed that there is no

contribution of the cordierite support to the adsorption properties of the coated

monoliths. Table 7.2 summarizes the surface areas (SBET) and the total micropore

volume (VDR(N2)) calculated per weight of the monolith.

Figure 7.2. Nitrogen adsorption/desorption isotherms at -196 °C.

0

10

20

30

40

50

0 0.2 0.4 0.6 0.8 1

Am

ount

ads

orbe

d (S

TP)

(c

m3 /g

)

P/PoCordierite MBETA MZSM-5 MSAPO-5 Mγ-Al2O3Mγ-Al2O3

Chapter 7

169

In the case of Mγ-Al2O3 it presents similar nitrogen adsorption isotherm shape than

the powder γ-Al2O3 used to coat the monolith, which indicates that the dip-coating

procedure used to prepare this sample does not modify significantly the porosity of

the γ-Al2O3.

In reference to zeolite and SAPO coated materials, MBETA, MZSM-5 and MSAPO-

5, the nitrogen adsorption isotherms obtained have also the same shape (including the

hysteresis loop) than the powdered samples prepared in Chapter 6. To compare the

specific surface area (SBET) area and micropore volume of the coating zeolites with

the powder zeolites, estimations for the monoliths have been done considering the

weight of the zeolite coatings. These values have been excluded in this Chapter for

brevity, but it can be said that both BETA (see Chapter 5) and SAPO-5 layers have

very similar porosity than the powder materials, thus being practically unaffected by

the in-situ coating process.

Table 7.2. Porous texture characterization results of the coated monoliths Sample SBET (m2/g) VDR (N2) (cm3/g) Cordierite 1 0.00 MBETA 100 0.05 MZSM-5 35 0.02 MSAPO-5 35 0.02 Mγ-Al2O3 8 —(*)

(*) not possible to determine

On the other hand, the ZSM-5 layer, confirmed by X-ray diffraction, shows a lower

porosity than the one expected from the comparison with powder samples. The main

reason is due to the coating procedure used for this sample. As we have mentioned

previously, the coating of high aluminium content zeolites is hindered in Al-rich

supports such as cordierite. According to Ulla et al. [6] the first step in the

hydrothermal synthesis of zeolite films is the formation of a precursor gel layer onto

the substrate, which serves as the primary source of nuclei. However, the addition of

aluminum to the solution accelerates gel layer formation but delays the nucleation


170

and zeolitization processes within it. Therefore, the use of Al-rich synthesis gels has

a direct influence on the kinetics of the process and, as a consequence, high loadings

of well-crystallized and anchored zeolite films are difficult to obtain. In a first

synthesis, Silicalite-1 is used to provide a suitable environment for the growth of a

high aluminium content zeolite upon the second synthesis step. This first layer is

partially dissolved during the second synthesis, helping the nucleation and growth of

the second layer. Thus, this procedure can block part of the first zeolite coating

producing a reduction in the adsorption properties of the zeolite layer.

Figure 7.3 shows the top view of the coated monoliths with ZSM-5, SAPO5 and γ-

Al2O3, respectively. In general, after a first synthesis step (See Figure 7.3, left),

MZSM-5 and MSAPO-5 monoliths have crystals which do not cover all the surface

of the monoliths and, therefore, do not form a continuous layer. These individual

crystals, however will act as seeds for the second in-situ synthesis step performed.

The individual crystals formed onto the cordierite monoliths present the typical

morphology for each type of material, ZSM-5 [7,9] and SAPO-5 [14,15]. In

reference to the Mγ-Al2O3 the slurry for dip-coating partially covered the surface of

the monoliths filling the external openings of the cordierite monoliths.

When a second synthesis was performed (See Figure 7.3 right), it is possible to cover

all the surface of the monolith with a complete and homogeneous layer of intergrown

crystals. MZSM-5 crystals maintained their respective shape. It must be pointed out

that MSAPO-5 monolith presents a bimodal particle size distribution. The crystals

appearing after the first synthesis have increased their size up to 12 µm

approximately and small SAPO-5 crystals have grown (1 µm approximately) filling

the gaps of the larger crystals, giving rise to a continuous layer. Finally, Mγ-Al2O3

monolith prepared by the “dip-coating” method is formed by a homogeneous and

continuous layer of Mγ-Al2O3 along the entire monolith wall. In spite of the

deposited monolith layer being homogeneous, a significant quantity of γ-Al2O3 was

Chapter 7

171

deposited on the corners preferentially towards the centre of the channel and small

cracks are observed after the calcination step.

Figure 7.3. SEM images of top view for the four prepared monoliths after the first

coating step (left) and the second coating step (right).

With the aim to analyze the thickness of the coated layer onto the monolith walls,

SEM images of the cross sectional view are useful for this purpose (see Figure 7.4).

All samples in this study prepared by the in-situ synthesis method showed multiple

First Step Second Step

1 µm

MBETA1 µm

MZSM-510 µm10 µm

MSAPO-510 µm10 µm

Mγ-Al2O3100 µm100 µm


172

layers of zeolite or SAPO crystals. There were no significant differences between the

coating zeolite or SAPO layer in the centre and in the channel open ends, as well as

along the channels, indicating that the movement of the autoclaves during synthesis

permits to obtain a homogeneous growth along the channel length.

A detailed analysis of SEM images has allowed us to conclude that MBETA

monolith is formed by a layer of intergrown zeolite crystals with an average

thickness of 3 µm (see Figure 7.4). In the case of MZSM-5 monoliths, the coating

layer consists of intergrown crystals with an average thickness of 6µm. It is worth

mentioning that the previous Silicalite-1 crystals deposited during the first synthesis

cannot be observed. As it was explained previously, these crystals act as seeds for the

deposition of a stable ZSM-5 layer and are partially dissolved during hydrothermal

synthesis. MSAPO-5 monoliths are formed by a coated layer of intergrown crystals

with a bimodal particle size distribution with a layer thickness of around 12 µm. To

conclude with the monoliths prepared by in-situ synthesis, it is important to note that

MZSM-5 and MSAPO-5 only exhibit the crystallization of individual zeolite crystals

in the internal openings without filling the porosity in the monolith walls, contrary to

the behaviour observed for the MBETA monolith (see Chapter 5).

Analyzing the γ-Al2O3 layer on Mγ-Al2O3 monoliths (See Figure 7.4) a variation in

thickness has been found. The thickness of the alumina wash-coating varies from 3-6

µm on the left side of the monolith to 16 µm on the other end-opening, being about

12 µm in the middle of the channel as determined by SEM. This behaviour can be

attributed to the immersion/extraction rate of the monolith from the slurry and the

final step of blowing with air that can alter the coated layer in one of the ends.

Chapter 7

173

Figure 7.4. SEM images of cross sectional for the four prepared monoliths after the

second coating step.

7.3.2. Monolithic catalysts characterization.

After impregnation, the monolithic catalysts were analyzed by XRD and no

appreciable changes were observed in the prepared catalysts (results not shown). To

analyze the distribution and loading of Pd nanoparticles on the monolith, Energy-

dispersive X-ray spectrometry (EDX) coupled to the SEM was used to measure the

quantity of Pd along all the channels of the monolith. For that purpose, several

coated monoliths impregnated with Pd nanoparticles were cut perpendicularly to the

monolith channels, and analyzed by EDX. From the individual analysis performed in

the four catalyst prepared, Pd/MBETA, Pd/MZSM-5, Pd/SAPO-5 and Pd/Mγ-Al2O3,

MBETA MZSM-5

MSAPO-5 Mγ-Al2O3

A B

C D

10 µm 10 µm

10 µm 20 µm


174

the general conclusion drawn was that the amount of Pd in all the monoliths was

around 1wt% of Pd along all the analyzed channels with respect to the coating layer.

Nitrogen adsorption was performed on the monolithic catalysts in the same manner

as the coated monoliths. Table 7.3 includes the BET surface areas (SBET) and the total

micropore volume (VDR(N2)). Comparing the results with those obtained for the

coated monolith (see Table 7.2), it can be established that the catalyst preparation

method has affected the adsorption properties of the coated layer on the Pd/MBETA

and Pd/MZSM-5 and Pd/Mγ-Al2O3 to a minor extent. Only in the case of

Pd/MSAPO-5 catalyst the sample has significantly reduced its N2 adsorption

capacity compared to the MSAPO-5 monolith before impregnation with Pd

nanoparticles. From the N2 adsorption isotherms (results not shown), it is possible to

confirm that, after impregnation with the Pd-based catalyst, the shape of the

isotherms is the same as for the coated monoliths.

Table 3. Porous texture characterization results of the monolithic catalysts. Sample SBET (m2/g) VDR (N2) (cm3/g) Pd/MBETA 95 0.05 Pd/MZSM-5 35 0.02 Pd/MSAPO-5 10 —(*) Pd/Mγ-Al2O3 7 —(*)

(*) not possible to determine

Comparing these results with powder catalysts prepared previously (see Chapter 6),

Pd/BETA and Pd/ZSM-5, the decrease in SBET is lower in the coated monoliths. This

behavior could be tentatively assigned to the Pd nanoparticles deposition. When a

powder catalyst is prepared, the Pd nanoparticles are located on all the zeolite

crystals homogeneously. On the contrary, Pd nanoparticles in the catalytic monolith

are predominantly distributed on the surface of the coating layer due to the highly

dense and compacted layers formed, as it was corroborated by EDX. Similar to

Chapter 7

175

Pd/SAPO-5 powder catalyst (see Chapter 6), the catalytic monolith show the same

decrease in apparent surface area.

As mentioned in the previous Chapter, the synthesized Pd colloid shows very low

polydispersity with a particle size of 1.8 ± 0.3 nm. In the case of the prepared

monoliths it is not possible to observe Pd dispersion, but deposition of polymer

protected Pd nanoparticles results in a noticeable change in Pd particle size according

to our previous Chapter, so it is expected that the particle size increases when the

polymer protected Pd nanoparticles are impregnated on the coated monolith

7.3.3. Catalytic performance in the oxidation of naphthalene.

The catalytic activity of naphthalene oxidation was measured for the four monolithic

catalysts. Three reaction cycles were performed with each catalyst in order to ensure

the stability of the sample by increasing the reaction temperature to the desired set

point and then cooling down to room temperature in order to start the next catalytic

cycle. Additionally, it is remarkable to note that no CO was detected as a reaction

product in any of the experiments performed. Previous to testing the monolithic

catalysts, the coated monoliths (MBETA, MZSM-5, MSAPO-5 and Mγ-Al2O3) were

tested in the oxidation of naphthalene and, in general, none of the coated monoliths

was active at the temperature range studied (i.e., between 100 and 200ºC). In this

way, MBETA monolith was active at temperatures above 300ºC reaching a 45%

conversion of naphthalene to CO2 at 350ºC.

Figure 7.5 shows the variation of the catalytic activity (expressed as conversion

towards CO2) for naphthalene oxidation, as a function of the reaction temperature for

the four tested monolithic catalysts. After the first cycle, outstanding catalytic

performance has been observed for the Pd/Mγ-Al2O3 and Pd/MSAPO-5, whilst

Pd/MZSM-5 and Pd/MBETA were able to reach full naphthalene conversion at

somewhat higher temperatures. It is worth mentioning that the catalytic activity


176

remained unchanged after three activity cycles and maintained its high naphthalene

oxidation activity to CO2 without evidences of catalyst deactivation.

Figure 7.5. Variation of the catalytic activity for naphthalene total oxidation

(expressed as yield to CO2) as a function of reaction temperature over the four catalysts. (A) Pd/MBETA; (B) Pd/MZSM-5; (C) Pd/MSAPO-5; (D) Pd/Mγ-Al2O3

Table 7.4 shows the temperature required for total naphthalene oxidation to CO2 for

each monolithic catalyst over the three cycles studied. Comparing these results with

our previous Chapter, a small increase in the temperature needed to reach total

oxidation is observed for the zeolite-coated monolith (BETA and ZSM-5). This can

be due to the dense thin layer of the zeolite that may have an influence in

naphthalene adsorption/desorption, thus having some effect in the kinetic of the

process. Nevertheless, these are minor differences that show that the catalyst based

on Pd nanoparticles supported on the coated-monoliths have a similar activity to the

powder catalyst (see Chapter 6).

0.0

20.0

40.0

60.0

80.0

100.0

50 100 150 200

Con

vers

ion

to C

O2

(%)

Temperature (ºC)

Pd/MBETA

CYCLE 1CYCLE 2CYCLE 3

A

0.0

20.0

40.0

60.0

80.0

100.0

50 100 150 200C

onve

rsio

n to

CO

2(%

)Temperature (ºC)

Pd/MZSM-5


B

0.0

20.0

40.0

60.0

80.0

100.0

50 100 150 200

Con

vers

ion

to C

O2

(%)

Temperature (ºC)

Pd/MSAPO-5


C

0.0

20.0

40.0

60.0

80.0

100.0

50 100 150 200

Con

vers

ion

to C

O2

(%)

Temperature (ºC)

Pd/Mγ-Al2O3


D

Chapter 7

177

Table 7.4. Temperature required for naphthalene total oxidation for the different catalysts for the three oxidation cycles tested

Catalyst Temperature (ºC) for total conversion to CO2

Cycle 1 Cycle 2 Cycle 3 Pd/MBETA 195 180 180 Pd/MZSM-5 195 180 180 Pd/MSAPO-5 180 165 165 Pd/Mγ-Al2O3 165 165 165

A useful tool to identify the most effective monolith-supported metal catalyst for

naphthalene total oxidation is the evaluation of the catalyst life time. An additional

accelerated experiment to test the stability against ageing was conducted at 250 ºC

for 48h. Figure 7.6 shows the time-on-line data for the four monolithic catalysts. It

was evident that the Pd/MBETA and Pd/ZSM-5 samples were stable for at least 48 h

and there is no evidence of catalyst deactivation over the testing period. The yield of

CO2 at the beginning of the experiment was 100% and it was not diminished after

48h. Contrary to these two catalysts, Pd/MSAPO-5 and Pd/Mγ-Al2O3 exhibit a 100%

of conversion during the first hours, but their conversion to CO2 dropped to 92% and

90% for Pd/MSAPO-5 and Pd/Mγ-Al2O3, respectively, indicating a possible

deactivation effect, probably induced by the agglomeration of the Pd nanoparticles.

Therefore, Pd/MBETA and Pd/MZSM-5 are the most suitable catalyst for the

naphthalene removal. The rugosity of the prepared zeolite layer prevents the

agglomeration of the Pd nanoparticles. These observations are in agreement with the

powder catalyst (see Chapter 6).

In a first approximation, these results could assess the viability of the prepared

monolithic catalysts in naphthalene removal by catalytic oxidation but a critical

comparison with monolithic catalysts from the literature must be added. Neyestenaki

et al. [16] deposited ZSM-5 on cordierite monolith substrates by means of a ball-

milled zeolite-water slurry (400cpsi, d=21.8, h=20 mm) by dip-coating several times

(several dipping-drying steps) until reaching an amount of wash-coated loading


178

between 13.5 and 30.8% and Pd, loaded by classical methods, was employed as

catalyst. All the monolithic catalysts were tested for the removal of the emissions of

pollutants from the combustion of biofuels in simulated mixture based on

naphthalene (50 ppm), methane, CO, CO2 and O2, balance in N2 (GHSV 20000 h-1).

As it was expected, an increase in wash-coating loading resulted in decreased light-

off temperatures of all the pollutants. Focusing on the monolith with similar

characteristics (0.95%Pd/NaZSM-5) to our monolithic catalyst, the T50% was 246 ºC

and 243 ºC for coated monoliths with an amount of ZSM-5 of 14.6%wt and

24.3%wt, respectively.

Figure 7.6. Conversion of naphthalene to CO2 as a function of time-on-line at 250ºC

for the four monolithic catalysts.

Ferrandon et al. [17] studied the total oxidation of mixtures with different

concentrations of the same aforementioned pollutants, including 50 ppm of

naphthalene (GHSV 23000 h-1, catalyst volume 6.5 ml). The monoliths used were

composed of cordierite honeycomb monoliths (400cps) wash-coated with γ-Al2O3

and used as supports of noble metal catalysts (Pd or Pt, 0.1% molar based), metal

oxides (MnOxCuOx,10% molar basis) and combinations of noble metals with metals

0.00

20.00

40.00

60.00

80.00

100.00

0 5 10 15 20 25 30 35 40 45 50

Con

vers

ion

to C

O2

(%)

Time (h)

Pd/MBETAPd/MZSM-5Pd/MSAPO-5Pd/M-Al2O3Pd/M-Al2O3

Chapter 7

179

oxides. Analyzing the T50%parameter for the fresh monolithic catalyst, the Pd based

catalyst was the most active (T50% = 200 ºC) while the metal oxides alone, MnOx and

CuOx, were able to reach a conversion of 50% at 365 and 420 ºC, respectively. It was

pointed out that impregnation of said oxides with Pd or Pt significantly decreases the

T50% values for the resulting catalysts, despite still being higher than those obtained

in this study at least by 50ºC. To conclude, all our monolithic catalysts show T50%

values around 150-160 ºC after 3 cycles, indicating the benefits of the use of PVP

polymer protected nanoparticles with controllable properties. Furthermore, total

oxidation of naphthalene is reached at lower temperatures.

Concerning the zeolite coated monoliths a synergetic effect between zeolites (BETA

or ZSM-5) prepared by an in-situ methodology with PVP-Pd nanoparticles has been

found. In the first place, the catalytic monoliths are able to produce the total

oxidation of naphthalene at low temperatures together with its removal from the gas

stream by adsorption at low temperature where the nanoparticles are not yet active.

Secondly, the time-on-line experiments have clearly pointed out that the catalysts are

highly stable and thus should be regarded and suitable for the total oxidation of

naphthalene in a practical application.


180

7.4. Conclusions

Zeolite BETA and ZSM-5, silicoaluminophosphate molecular sieve, SAPO-5 have

been coated on cordierite honeycomb monoliths of 400cpsi by in-situ synthesis. Two

synthesis steps were required in order to coat the cordierite monoliths. The amount of

zeolite introduced during the first single-step synthesis cannot cover the monolith

surface completely but it is used as seeds for the second synthesis step which

produces homogeneous and thin zeolite films. Furthermore, a BETA coating layer is

formed by a compact three-dimensional zeolite network grown in all the openings

conferring a stronger zeolite anchorage to the cordierite structure. In general, the

reproducibility of the prepared coated materials was very high.

All the monoliths prepared were impregnated with Pd-PVP-protected nanoparticles

and were used as catalysts for the total oxidation of naphthalene. All the catalysts

produced a high activity for naphthalene conversion to CO2, with total conversion

taking place at 165 ºC after three cycles for Pd/MSAPO-5 and Pd/Mγ-Al2O3 and

180ºC after three cycles for Pd/MBETA and MSAPO-5 monoliths. All the catalysts

possess high stability because their properties remain largely unchanged after testing

for several oxidation cycles.

Time-on-line experiments have been used to test the stability of the catalysts. The

Pd/MBETA and Pd/MZSM-5 samples were stable after accelerated ageing time-on-

line for 48 h at 250 ºC. However, despite being the most active after three cycles,

Pd/SAPO-5 and Pd/Mγ-Al2O3, decreased their catalytic activity for the conversion of

naphthalene to 95% possibly due to a possible deactivation after ageing.

As a result it has been established that the use of Pd-based nanoparticles supported

on BETA and ZSM-5 zeolites supported on honeycomb monoliths are very

interesting options for the abatement of PAHs by catalytic oxidation.

Chapter 7

181

References

[1] T. Garcia, B. Solsona, D. Cazorla-Amorós, Á. Linares-Solano, S.H. Taylor, Appl. Catal. B 62 (2006) 66-76. [2] B. Puertolas, B. Solsona, S. Agouram, R. Murillo, A.M. Mastral, A. Aranda, S.H. Taylor, T. Garcia, Appl. Catal. B 93 (2010) 395-405. [3] A.M. Mastral, M.S. Callen, Environ. Sci. Technol. 34 (2000) 3051-3057. [4] T.A. Nijhuis, A.E.W. Beers, T. Vergunst, I.Hoek, F. Kapteijn, J.A. Moulijn, Catal. Rev. 43 (2001) 345380. [5] J.M. Zamaro, M.A. Ulla, E.E. Miró, Chemical Engineering Journal 106 (2005) 25-33. [6] M.A. Ulla, R. Mallada, J. Coronas, L. Gutierrez, E. Miró, S. Santamaría, Appl. Catal. A 253 (2003) 257-269. [7] Á Berenguer-Murcia, J Garcı ́a-Martı ́nez, D Cazorla-Amorós, Á Linares-Solano, A.B Fuertes, Microporous Mesoporous Mater. 59 (2003) 147-159. [8] H. Lechert, R. Kleinwort, H. Robson (Eds.), Verified Synthesis of Zeolitic Materials, second edition,.Ed. Elsevier (2001) pp. 198-200. [9] J.M. López, M.V. Navarro, T. García, R. Murillo, A.M. Mastral, F.J. Varela-Gandía, D. Lozano-Castelló, A. Bueno-López, D. Cazorla-Amorós, Microporous Mesoporous Mater. 130 (2010) 239-247. [10] J.M. Campelo, F. Lafont, J.M. Marinas, J. Catal. 156 (1995) 11-18. [11] L. Villegas, F. Masset, N. Guilhaume, Appl. Catal. A: General 320 (2007) 43-55. [12] I. Miguel-García, Á. Berenguer-Murcia, D. Cazorla-Amorós, Appl. Catal. B 98 (2010) 161-170. [13] D. Lozano-Castelló, F. Suárez-García, D. Cazorla-Amorós, Á. Linares-Solano, in: F. Beguin, E. Frackowiak (Eds),Carbon Materials for Electrochemical Energy Storage Systems, Ed. CRC Press, (2010) p.p. 115-162. [14] J.M. Campelo, F. Lafont, J.M. Marinas, M. Ojeda, Appl. Catal. A 192 (2000) 85-96. [15] J.M. Campelo, F. Lafont, J.M. Marinas, Zeolites 15 (1995) 97-103 [16] A. K. Neyestanaki, L.-E Lindfors, T. Ollonqvist, J. Väyryen, Appl. Catal. A 196 (2000) 233-246. [17] M. Ferrandon, J. Carnö, S. Järås, E. Björnbom, Appl. Catal. A 180 (1999) 153-161.

Chapter 8. Preferential oxidation of

CO catalyzed by palladium nanoparticles supported on

inorganic oxides and on inorganic oxide-coated

cordierite honeycomb monoliths.

Chapter 8

183

8.1. Introduction.

The growing concern of society to climate change issues associated with the adverse

consequences of emissions of greenhouse gases has encouraged the development of

technologies that allow power generation from renewable sources. Thus, H2 use as a

suitable energy carrier for replacing gasoline and other fossil fuel has been widely

discussed as a way to implement hydrogen economy [1].

As described in Chapter 1, in mobile applications, hydrogen fuel cells are an

attractive candidate to replace conventional devices, but the main issue is to find a

way to produce hydrogen with the appropriate purity [2]. Nowadays, hydrocarbon

reforming is the most prominent industrial process to produce hydrogen. However,

the CO concentration exiting the reformer is between 1000 and 10000 ppm.

Therefore, it is necessary to purify the H2 flow before being supplied to the fuel cell

by reducing the CO concentration below 10 ppm [3]. One possible solution is the

catalyzed preferential oxidation of CO (PrOx-CO).

This chapter reports the use of monolithic catalysts for the PrOx-CO reaction. These

monolithic catalysts are based on PVP polymer protected Pd nanoparticles supported

on BETA and ZSM-5 zeolites and SAPO-5 molecular sieve thin films deposited by

synthesis in-situ onto cordierite honeycomb monoliths.

8.2. Experimental.

8.2.1. Catalyst preparation and characterization.

The powder catalysts, Pd/BETA, Pd/ZSM-5 and Pd/SAPO-5 and the monolithic

catalysts (Pd/MBETA, Pd/MZSM-5 and Pd/MSAPO-5) used in this study, were

prepared following the synthetic procedure exposed in Chapters 6 and 7,

respectively.

Hydrogen purification by PrOx-CO

184

Both powder catalysts and monolithic catalysts were characterized by XRD, SEM,

EDX, N2 adsorption (-196 ºC) and CO2 adsorption (0 ºC) and the results have been

presented in previous chapters (Chapters 6 and 7).


The powder and monolithic catalysts were tested in the preferential oxidation of CO.

Prior to the test reaction, the samples were submitted to a reduction treatment. Thus,

each sample was reduced at 200 ºC for 2 h (heating rate 5 ºC/min) in 10% H2 in He

(100 ml/min). The reaction tests were performed with a reactant composition of 2%

CO, 2% O2, 30% H2, balance He. Non-isothermal experiments were done up to

200ºC (heating rate 2 ºC/min) and a total flow rate of 100 ml/min was used. Two

cycles were performed on each sample. An U-shape quartz reactor (16 mm of inner

diameter) was used for powder catalysts (using 0.150 g of sample) and a horizontal

reactor (16 mm of inner diameter) for the monolithic catalysts (using around 200-250

mg of sample depending on the coated zeolite or SAPO layer). The exhaust gases

leaving the reactor were analyzed by gas-chromatography (Agilent Technologies

6890N, equipped with a CTRI column operating at 80 ºC and a TCD detector) and a

mass spectrometer (Balzer, Thermostar GSD 301T). The resulting space velocities

(h-1) were 12000 and 2500 h-1 for powder and monolith catalysts, respectively. The

turnover frequency defined as moles of CO reacted per mol of surface metal per

second was calculated for the three powder catalysts.

Conversion and selectivity in the reactions were calculated as follows:

. %–

100 (8.1)

% 50 (8.2)

Chapter 8

185


Powder catalysts were tested with the aim to analyze the potential application of

these catalysts for the PrOx reaction. The catalytic activity of the three catalysts

prepared is shown in Figure 8.1.

In the PrOx reaction, H2 oxidation occurs as a side reaction to a certain extent,

generating H2O in the reactor, which could be an important drawback from both a

technological and economical point of view [4]. Therefore, the selectivity towards

CO oxidation as a function of the temperature is an important parameter in the PrOx

reaction. This information is included in Figure 8.1.

Additionally, Table 8.1 includes the maximum selectivity, the maximum conversion,

the turnover frequency (TOF) and the selectivity at 25% of CO conversion.

Figure 8.1. Variation of the catalytic activity for CO oxidation and the selectivity towards CO oxidation as function of reaction temperature over the three powder

catalysts.

In general, the conversions of CO for the prepared powder catalysts are very high,

reaching values close to 90% in all the cases. The maximum selectivity reached, in

0102030405060708090100

0102030405060708090

100

0 50 100 150 200

Sele

ctiv

ity (

%)

CO

conv

ersi

on (

%)

Temperature (ºC)

Pd/BETAPd/ZSM-5Pd/SAPO-5Pd/BETAPd/ZSM-5Pd/SAPO-5

CO

NV

ERSI

ON

SELE

CTI

VIT

Y


186

all the cases, is above 60%. The ideal catalyst for this reaction should have a

selectivity values towards CO oxidation over 50%, which means that CO oxidation is

preferential over H2 oxidation and that, theoretically, all CO could be removed from

the H2 flow. The selectivity values obtained for 25% of CO conversion is between 60

and 80% for the three samples. Comparing these results with those obtained in our

group using Pd/Alumina catalyst [4], where a 55% of selectivity was reached for a

25% of CO conversion, it can be said that the materials of the present study are quite

interesting for PrOx-CO reaction. Comparing with other results reported in the

literature, Manasilp and Gulari [5] used a catalyst based on 2% Pt/γ-Al2O3 and tested

in a H2 containing flow (1% CO, 1% O2, 60% H2, balance He), obtaining a 80% of

maximum CO conversion and selectivity of 50% approximately at 170ºC. Iwasa et

al. [6] showed that the Cs-modified Pd/ZnO catalyst had a high CO conversion of

82% with a selectivity of 60% indicating the enhanced effect of using Cs as a catalyst

promoter.

Table 8.1 also includes the turnover frequency (TOF) for the three catalysts during

the PrOx reaction calculated at 150 ºC. Similarly to naphthalene oxidation results

(see Chapter 6), despite using the same Pd-PVP nanoparticles in all the catalysts

tested, the TOF values are different depending on the support, demonstrating the

influence of the support in the final properties of the catalyst. In this sense, the TOF

value for the Pd/BETA catalyst at 150ºC is double than the TOF value for Pd/SAPO-

5 catalyst. Furthermore, comparing these results with the literature, Miguel-Garcia et

al. [4] published that the Pd/Alumina catalyst reached a TOF number of 0.10 s-1at

150 ºC, whilst the use of BETA zeolite as support gives a slightly higher TOF

number, indicating the potential use of BETA zeolite as support of the Pd

nanoparticles for the PrOx reaction.

Chapter 8

187

Table 8.1. Catalytic results in PrOx reaction. Maximum selectivity, maximum conversion and selectivity at 25% of CO conversion are specified for each sample. TOF (s-1) calculated at 150ºC is also included. Sample Maximum

selectivity (%)

Maximum conversion

(%)

Selectivity (%) at 25% of CO

conversion

TOF (s-1) (150ºC)

Pd/BETA 68 (151 ºC) 91 68 (151 ºC) 0.14 Pd/ZSM-5 81 (155 ºC) 88 81(155 ºC) 0.09 Pd/SAPO-5 70 (155 ºC) 83 60(162 ºC) 0.07

The reason for the different behavior of the catalysts studied in the PrOx reaction lies

in the effect that produces the support on the Pd nanoparticles. The promising

properties of the Pd/BETA catalyst in contrast to the other powder catalysts could be

tentatively related to the different interaction between the capping agent (PVP), the

Pd nanoparticles and the support. As reported by our group [4,7], the carbonyl

functional groups of the protecting polymer interacts with the metallic palladium,

withdrawing electron density from its surface. This effect corresponds to the

appearance of the two binding energies in the Pd XPS results (See Table 6.7 in

Chapter 6), which can be assigned to reduced and oxidized Pd. This effect produces

enhanced properties in the PrOx oxidation. Furthermore, Miguel-Garcia et al. [7]

reported that this effect is reduced with the time the Pd nanoparticles are in MeOH

suspension, where they are stored before being impregnated in the support, pointing

out that, as the PVP is removed from the surface by the solvent, the catalytic

properties are drastically modified.

As discussed in Chapter 6, the interactions support-PVP and PVP-Pd nanoparticle

are different depending on the support used. Bearing in mind the XPS results (See

Table 6.7 in Chapter 6) for the three powder catalysts, the Pd/BETA is the catalyst

with the highest amount of oxidized Pd, which indicates that the interaction PVP-Pd

is higher in the case of Pd/BETA powder catalyst compared with the rest of powder

catalysts. These differences in the Pd species may explain the different behavior in

the catalytic properties for CO oxidation of the three powder catalysts.


188

From a practical point of view, the use of structured catalyst is preferred, and thus

powder catalysts are less attractive than monolithic substrates. Therefore, structured

catalysts prepared as reported in Chapter 7, based on Pd nanoparticles supported on

zeolite thin films (BETA and ZSM-5) or SAPO-5, in-situ grown onto cordierite

honeycomb monoliths were used for the PrOx reaction. The catalysts prepared were

named as indicated in the experimental section of this Chapter: Pd/MBETA,

Pd/MZSM-5 and Pd/MSAPO-5.

Figure 8.2 shows the variation of the catalytic activity (expressed as CO conversion)

for PrOx reaction, as a function of the reaction temperature for the three tested

monolithic catalysts and for the first cycle. Table 8.2 summarizes the catalytic

properties of the three monolithic catalysts after the first and second CO oxidation

cycle. Analogously to the powder catalysts, the general trend is that all the

monolithic catalysts have a high conversion.

Figure 8.2. Variation of the catalytic activity for CO oxidation and the selectivity

towards CO oxidation as function of reaction temperature over the three monolithic catalysts for the first reaction cycle.

0102030405060708090100

0102030405060708090

100

0 50 100 150 200 250

Sele

ctiv

ity (

%)

CO

conv

ersi

on (

%)

Temperature (ºC)

Pd/MBETAPd/MZSM-5Pd/MSAPO-5Pd/MBETAPd/MZSM-5Pd/MSAPO-5

CO

NV

ERSI

ON

SELE

CTI

VIT

Y

Chapter 8

189

Regarding the selectivity, the results in Table 8.2 show that the Pd/MBETA catalyst

has the highest selectivity (higher than 70%, which is very similar to the results

obtained with Pd/BETA powder catalyst).

Table 8.2. Catalytic results in PrOx reaction. Maximum selectivity and maximum conversion and selectivity at 25% of CO conversion for each monolithic catalyst after the first and second cycle. Sample Maximum

selectivity (%)Maximum

conversion (%) Selectivity (%) at 25% of CO

conversion

Pd/MBETA Cycle 1 78 (138 ºC) 88 77 (139 ºC) Cycle 2 72 (140 ºC) 86 72 (140 ºC)

Pd/MZSM-5 Cycle 1 59 (170 ºC) 85 59 (170 ºC) Cycle 2 42 (168 ºC) 76 42 (168 ºC)

Pd/MSAPO-5 Cycle 1 53 (175 ºC) 92 46 (160 ºC) Cycle 2 45 (160 ºC) 77 42 (162 ºC)

Comparing the PrOx results obtained with monoliths and the powder catalysts, it is

observed that: (i) Pd/MBETA catalyst has very similar catalytic properties

performance in the PrOx reaction than the analogous powder catalyst (Pd/BETA).

This can be explained considering the results presented in Chapter 5, where it was

shown that the properties (in terms of porosity and crystal morphology) of the BETA

thin film deposited on honeycomb monoliths were very similar to the powder BETA

zeolite, and (ii) the oxidation properties of the Pd/MZSM-5 and Pd/MSAPO-5

catalysts have diminished with the change in configuration (i.e. from powder to

monolith). As mentioned in Chapter 7, for the synthesis of ZSM-5, it was necessary

to coat first the cordierite monoliths with silicalite-1 crystals, which were used as

seeds for preparing the ZSM-5 layer. In the case of SAPO-5, the coated layer has a

different morphology when a second synthesis was performed. Then, these

differences in crystal morphology, porosity and so on in the ZSM-5 and SAPO-5 thin

films, with respect to powders, affect to the final catalytic properties of these two

monolithic catalysts.


190

Several studies have been performed with structured catalyst in a monolithic shape

[8,9]. Neri et al. [8] studied the PrOx reaction (5% CO, 1.5% O2, balance H2) with

honeycomb (size: 0.5x5cm, 225x103 h-1) monoliths based on Pt supported on a wash-

coated honeycomb with a zeolitized-pimenous material. From the PrOx oxidation

results, the maximum conversion reached was 60% at 100ºC with selectivity close to

60%. However, the authors pointed out that at temperatures higher than 150 ºC the

CO conversion and selectivity decreased due to the competitive H2 oxidation

reaction. Roberts et al. [9] prepared a monolithic catalyst based on Pt (5 wt. %)

promoted with Fe (1.0 wt%) supported on cordierite honeycomb monoliths (400cps)

dip-coated with alumina for the PrOx oxidation (T = 100 ºC, O2/CO = 1, 1% CO,

30000 h-1). The use of a promoter produced a beneficial effect in the final catalyst

properties increasing the CO conversion from 15% to 80% approximately with a

selectivity of 40% due to the fact that iron oxide was located adjacent to Pt particle,

providing an alternative site for O2 adsorption because this is not “blocked” by the

adsorption of CO. Comparing those results found in the literature for structured

catalysts with those obtained in the present study, it can be said that the Pd/MBETA

catalyst is a high performance monolithic catalyst for PrOx reaction.

To analyze the stability of the monolithic catalysts, two consecutive reaction cycles

were tested for the three catalysts and the results have been included in Table 8.2.

The Pd/MBETA is the most stable catalyst tested reaching approximately the same

values of catalytic properties after two cycles. In the case of Pd/MZSM-5 catalyst

and, and more pronounced for the Pd/MSAPO-5, their catalytic properties are

reduced. In the case of the Pd/MZSM-5 the maximum selectivity has been reduced

down to 40% while the Pd/MSAPO-5 monolithic catalyst has drastically reduced

both the selectivity and the CO conversion.

Chapter 8

191

8.4. Conclusions

Novel monolithic catalysts based on Pd nanoparticles supported on a zeolite thin film

prepared by in-situ synthesis onto a cordierite honeycomb monoliths have been

synthesized and tested for PrOx-CO reaction. The preliminary results show that

Pd/MBETA monolith catalyst is a high performance catalyst for PrOx reaction, being

the most promising for H2 purification towards its use in PEM fuel cells.


192

References

[1] M. Fan, C.P. Huang, A. Bland, Z. Wang, R. Slimane, I. Wright, Environanotechnology, Ed. Elsevier (2010) pp. 221-222. [2] R. Masel, Nature 442 (2006) 521-522. [3] R. Farrauto, S. Hwang, L. Shore, W. Ruettinger, J. Lampert, T. Giroux, Y. Liu, O. Ilinich, Annu. Rev. Mater. Res. 33 (2003) 1-27. [4] I. Miguel-Garcia, A. Berenguer-Murcia, D. Cazorla-Amoros, Appl. Catal. B 98 (2010) 191-170. [5] A. Manasilp, E. Gulari, Appl. Catal. B 37 (2002) 17-25. [6] N. Iwasa, S. Arai, M. Arai, Appl. Catal. B 79 (2008) 132-141. [7] I. Miguel-Garcia, Á. Berenguer-Murcia, D. Cazorla-Amorós Catal. Today 187 (2012) 161-170. [8] G. Neri, G. Rizzo, F. Corigliano, I. Arrigo, M. Caprì, L. De Luca, V. Modafferi, A. Donato, Catal. Today 147 (2009) S210-S214. [9] G. W. Roberts, P. Chin, X. Sun, J.J. Spivey, App. Catal. B 46 (2003) 601-611.

Chapter 9. General conclusions.

Chapter 9

193

In this PhD Thesis, zeolite thin films have been prepared and characterized for gas

purification in different applications: (i) hydrogen purification for its use in PEM fuel

cells; (ii) hydrocarbon emission control during the cold start in combustion engines;

and (iii) total abatement of polycyclic aromatic emissions. From the work presented

in this PhD Thesis, the following general conclusions have been obtained:

In relation to the zeolite membranes supported on macroporous carbon discs the

main conclusions are:

• Zeolite layers based on zeolite LTA deposited on porous carbon discs have

been successfully prepared following an established approach. Subsequent

ion-exchange carried out by a simple and reproducible methodology allowed

to modify the as-grown Na-LTA zeolite into K-, Rb-, Cs-forms. This permits

to tailor the porosity of the final material.

• Concerning the permeation tests performed with binary mixtures (50% H2,

1.25% CO in He), our as-prepared Na-LTA/carbon membrane is unsuitable

for H2 purification. The best result obtained is with the Cs-LTA membrane.

In this case, the H2 permeation flow obtained is similar to the Na-LTA/carbon

membrane but CO permeation is negligible. Thus, a H2 rich flow with a CO

concentration lower than 10 ppm, which is mandatory in order to use it in

PEM fuel cell, can be reached with the Cs-LTA/carbon membrane.

• The effect of CO2 and H2O in the membrane permeation properties of Na-

LTA/carbon and Cs-LTA/carbon membranes has been analyzed. It has been

observed that the as prepared Na-LTA/carbon membrane shows no CO

permeation at room temperature due to the competitive adsorption produced

by the presence of CO2 in the feed stream. However, this phenomenon

becomes less pronounced as the temperature increases. Furthermore, the H2

General conclusions

194

permeation values are not affected by the presence of CO2. Regarding the Cs-

LTA/carbon membrane, CO2 neither affects the CO nor H2 permeations.

• In the case of a simulated reformer mixture, it is demonstrated that the Cs-

LTA/carbon membrane is a high quality and stable membrane (no

performance loss at working times over 60 h has been observed). The

permeations of CO and H2 are not affected by the presence of water, making

these membranes very promising candidates for their prospective use in H2

purification devices.

Regarding the zeolite thin films supported on cordierite honeycomb monoliths and

the structured catalysts the main conclusions reached are:

• Cordierite honeycomb monoliths were coated with BETA zeolite by in-situ

crystallization with different zeolite loadings after one, two or three

consecutive synthesis steps. Depending on the number of synthesis the

properties of the final zeolite coated material are different. The optimum

conditions found are after three consecutive synthesis steps, reaching a high

zeolite loading and a stable and compact zeolite layer. This high zeolite

loading and the effective intergrowth between crystals has a positive effect on

the supported zeolite performance as HC trap in cold start conditions.

Propene diffusion is affected by the thickness of the supported BETA zeolite

layer, and the coated monolith prepared after three consecutive synthesis

steps, MBETA3, has the best performance.

• A highly effective HC trap for the abatement of cold start emissions has been

synthesized and tested in the laboratory. This material shows 100% of toluene

retention, and accomplishes the requested performance as a HC trap,

desorbing propene at temperatures close to 300 ºC, and being stable after

cycling.

Chapter 9

195

• Zeolites BETA and ZSM-5, silicoaluminophosphate molecular sieve SAPO-5

and γ-alumina impregnated, with Pd nanoparticles have been used as catalysts

for the total oxidation of naphthalene. All the catalysts produced have high

activity for naphthalene conversion to CO2, with total conversion taking place

between 165 and 180 ºC. Furthermore, all the catalysts have high stability

because their properties remain largely unchanged after testing for several

oxidation cycles. The Pd/ZSM-5 and Pd/BETA samples were stable after

accelerated ageing time-on-line for 48 at 250 ºC. Therefore, the use of Pd-

based nanoparticles supported on zeolites (BETA or ZSM-5) could be

potential options for the abatement of PAHs by catalytic oxidation.

• Cordierite honeycomb monoliths have been coated with zeolites BETA and

ZSM-5 and silicoaluminophosphate molecular sieve, SAPO-5 by in-situ

synthesis and two synthesis steps were required in order to properly coat the

monoliths. The amount of zeolite introduced after the first syntehsis cannot

cover the monolith surface completely but it functions as seeds for the second

synthesis step which produces homogeneous and thin zeolite films.

• All the monoliths prepared were impregnated with Pd-PVP-protected

nanoparticles and were used as catalysts for the total oxidation of

naphthalene. All the catalysts produced a high activity for naphthalene

conversion to CO2, with total conversion taking place at 165 ºC after three

cycles for Pd/MSAPO-5 and Pd/Mγ-Al2O3 and 180 ºC after three cycles for

Pd/MBETA and Pd/MZSM-5 monoliths. All the catalysts possess high

stability because their properties remain largely unchanged after testing for

several oxidation cycles.

• Time-on-line experiments have been used to test the stability of the catalysts.

As a result it has been established that the use of Pd-based nanoparticles

supported on BETA and ZSM-5 zeolites supported on honeycomb monoliths

are very interesting options for the abatement of PAHs by catalytic oxidation.

General conclusions

196

• The monolithic catalysts based on Pd nanoparticles were also tested for PrOx-

CO reaction. The preliminary results show that Pd/MBETA catalyst is a high

performance monolithic catalyst for PrOx reaction, being the most promising

for H2 purification.

Resumen en Castellano

Resumen de la Tesis Doctoral

197

1. Introducción general1.

1.1. Zeolitas, membranas de zeolita y películas delgadas de zeolita.

Las zeolitas son una familia de aluminosilicatos cristalinos con una estructura

tridimensional que presenta un sistema regular de poros bien definidos formados por

canales y cavidades de dimensiones moleculares. Las zeolitas están formadas por

tetraedros, TO4 (T= Si, Al), unidos entre sí compartiendo todos sus oxígenos.

Además de las zeolitas clásicas, para la síntesis de zeolitas también se pueden

emplear metales de transición u otros elementos P, B, Ga, Fe, Cr, Ti, V, Mn, Co, Zn,

Be ó Ca. A todos estos materiales cristalinos, se les denomina zeotipos. De todos

ellos, se pueden destacar los tamices moleculares de aluminofosfato (AlPO4),

silicoaluminofostatos (SAPO) o los metaloaluminofosfatos (MeAPO).

La variedad en composición química, así como la modificación del tamaño del poro

o la geometría de la estructura porosa son los rasgos más característicos de las

zeolitas. En general, las zeolitas existentes se pueden clasificar en base al tamaño de

poro, delimitado por el número de tetraedros que forman su entrada que varía entre 8,

10, 12 o más de 12 unidades de TO4. Por ello, las zeolitas presentan un elevado

número de aplicaciones en distintas áreas como son el intercambio iónico, la catálisis

o la separación de compuestos químicos.

Aunque la zeolitas se obtienen principalmente en forma de polvo, éstas se puede

preparar también como películas sobre una gran variedad de soportes tanto

inorgánicos (vidrio, cerámicas, metálicos, grafitos porosos) como orgánicos

(plásticos, celulosa o madera). Una de las ventajas que presenta de este tipo de

configuración es que permite transferir las propiedades de una zeolita (adsorción,

1NOTA: En este resumen no se incluyen referencias bibliográficas. Se puede encontrar una revisión bibliográfica exhaustiva durante la Tesis Doctoral.


198

catálisis, reconocimiento molecular y difusión) a una estructura bidimensional con

aplicaciones en diversas áreas como reactores químicos, separaciones moleculares o

análisis químico. Por lo tanto, resulta una opción interesante para muchas

aplicaciones potenciales, la posibilidad de tener las propiedades de una zeolita en una

configuración tipo membrana.

Una membrana se define como una barrera semipermeable selectiva entre dos o más

compuestos que actúa como barrera al transporte de materia entre dos fases

adyacentes. Hoy en día, existen diferentes tipos de membranas y se ordenan en tres

categorías: (i) poliméricas, (ii) metálicas e (iii) inorgánicas como las membranas de

zeolita. En primer lugar, las membranas poliméricas tienen varias ventajas como son

un bajo coste y no presentar caídas de presión significativas. Sin embargo, sufren

problemas mecánicos y poseen una elevada sensibilidad al hinchamiento y

compactación, limitando su utilidad. El segundo tipo, las membranas metálicas

presentan una permeancia de H2 elevada pero sufren debilitamiento mecánico por

hidrógeno a bajas temperaturas. Este aspecto se elimina mediante la formación de

aleaciones con metales nobles, pero aumenta el coste de las membranas. El ultimo

tipo, las membranas de zeolita, combinan las ventajas de las membranas inorgánicas,

como la estabilidad térmica y química, con las de las membranas poliméricas, es

decir, compuestas por una capa delgada y homogénea.

Durante los últimos 25 años se ha producido un gran avance dentro de la

investigación en membranas de zeolita. Sin embargo, aunque existen alrededor de

201 estructuras de zeolita de acuerdo con la Asociación Internacional de Zeolitas

(IZA, acrónimo del inglés de International Zeolite Association), sólo se han

empleado para la preparación de membranas alrededor de 20 estructuras. Entre estas,

se pueden destacar las membranas basadas en las estructuras SOD, CHA, LTA,

DDR, FER, MEL, MFI, ERO-OFF, MOR, FAU o BEA. El notable progreso

realizado en el desarrollo de membranas de zeolita se puede resumir en varios hitos.


199

Las membranas basadas en zeolitas fueron dadas a conocer por Kulprathinja cuando

preparó membranas compuestas basadas en cristales de zeolita ZSM-5 embebidas en

un polímero. Sin embargo, no fue hasta 1987 cuando se introduzco el concepto y la

preparación de películas continuas basadas en cristales de zeolita interconectados,

intensificando de esta forma la preparación de nuevos tipos de membranas de zeolita.

No fue hasta 1989 cuando se preparó la primera membrana basada en zeolita A con

propiedades de separación y Matsukata y col. prepararon el primer reactor basado en

membranas. Finalmente, uno de los avances más importantes dentro de la

preparación de membranas de zeolita, se dio cuando se aplicó el método de

crecimiento secundario para la preparación de membranas de zeolita A por Kita y

col. Actualmente, las membranas de zeolita han llegado a escala industrial. Así,

“Mitsui Engineering” ha preparado membranas de zeolita para procesos de

pervaporación de etanol/agua y “Smart Co.” Ofrece membranas de zeolita A para la

eliminación de aguas de disolventes, principalmente a escala de laboratorio.

Destacar que las membranas de zeolita combinan las ventajas de las membranas

inorgánicas, con una selectividad a la forma perfecta. Comparando las membranas de

zeolita con las membranas orgánicas, éstas poseen ciertas ventajas y desventajas.

Así, las membranas de zeolita presentan ciertas ventajas como que son estables a alta

temperatura, resistentes a medios agresivos o altas presiones, inertes a la degradación

por microorganismos o fáciles de limpiar después de su uso. Todas estas ventajas le

otorgan unas propiedades interesantes para solventar ciertos inconvenientes de cara a

su implementación a escala industrial como son un coste elevado, el escalado, la baja

permeabilidad de membranas muy selectivas (membranas densas) o problemas de

sellado de los módulos de membrana.

Con el objetivo de que el transporte sólo ocurra a través de los poros de la

membrana, se necesita que durante la síntesis de una membrana de zeolita se

produzca el desarrollo de una capa continua casi bidimensional y libre de defectos.


200

El procedimiento estándar para la obtención de una membrana de zeolita consiste en

depositar cristales de zeolita en un soporte, que le va a conferir las propiedades

mecánicas necesarias y permite, a su vez, el desarrollo de una estructura más extensa.

En referencia a los métodos de síntesis de membranas, el primer método descrito fue

la síntesis directa sobre un soporte. En general, este tipo de procedimiento se divide

en dos pasos: (i) se prepara el gel de síntesis y el soporte se introduce en este y, (ii)

se somete a síntesis hidrotérmica según las condiciones de síntesis de la membrana.

Un gran número de investigaciones se han centrado en este procedimiento a pesar de

las desventajas que presenta como son la necesidad de largos tiempos de síntesis y la

realización de varias etapas de síntesis hasta la obtención de la película continua de

membrana de zeolita. Recientemente, se ha desarrollado un nuevo método de

preparación de membranas de zeolita denominado como crecimiento secundario o

por sembrado. Las principales ventajas que presenta este método en comparación con

el anterior son: (i) permite la preparación de películas delgadas y orientadas, (ii)

mejora el crecimiento de la zeolita en diferentes soportes y (iii) solo requiere de la

realización de una etapa de síntesis.

De forma general, la síntesis de zeolita se divide en dos etapas: nucleación y

crecimiento de cristal. La etapa limitante durante la síntesis es la nucleación, y va a

determinar el crecimiento cristalino. Teniendo en cuenta los dos métodos descritos

anteriormente para la preparación de membranas de zeolita (directo o por sembrado),

en el caso de síntesis directa, hay una limitación en el intercrecimiento entre los

cristales de zeolita que van a formar la membrana debido al tiempo de inducción

necesario en la etapa de nucleación. Sin embargo, en el método de sembrado, esta

etapa se elimina, porque los cristales de zeolita sembrados actúan como núcleos de

cristalización. Además, permite obtener cristales de mayor calidad y un mejor

crecimiento intercristalino reduciendo los tiempos de síntesis. Por tanto, separar el

crecimiento de los cristales de la etapa de nucleación permite una gran flexibilidad


201

para poder modificar la microestructura o escalar la membrana de zeolita. En

relación al método de sembrado, en los últimos años se han desarrollado diferentes

métodos de sembrado como son la modificación superficial de soportes metálicos

mediante polímeros orgánicos, inmersión, por fricción, succión, ultrasonidos,

sembrado por giro (del inglés, spin-coating) o depósito electroforético, entre otros.

En cuanto a los soportes empleados para la síntesis de membranas de zeolita,

destacar los tubos o platos inorgánicos de alúmina, acero o basados en materiales

carbonosos.

Otro tipo de materiales basados en zeolitas son las películas delgadas soportadas en

materiales estructurados. En los últimos años, se ha obtenido gran avance en la

preparación de películas delgadas de zeolita en cualquier tipo de soporte, influyendo

en el control de su morfología, orientación o, incluso, grado de orientación de los

cristales interconectados. Por tanto, las características de las películas de zeolita

preparadas, así como el procedimiento experimental se pueden variar en función

aplicación final.

Para la preparación de materiales estructurados, se han empleado distintos soportes

(o estructuras abiertas), como son los monolitos cerámicos (cordierita, sílice), mallas

metálicas enrolladas, fibras metálicas sinterizadas, microreactores de acero, láminas

de Mo, láminas de Si, microfibras de cuarzo o materiales carbonosos (fibras, telas,

etc.).

Desde un punto de vista industrial, la principal razón por la cual se necesita usar una

película delgada de zeolita soportada en vez de materiales en polvo directamente es

que, debido al pequeño tamaño de las partículas de zeolita, éstas puede causar

problemas como elevadas caídas de presión en corrientes de gases y que,

normalmente, su manejo y separación es muy complicado. En primer lugar, los

problemas mencionados se pueden solventar mediante la preparación de pellets de

zeolita mediante su extrusión con ayuda de un agente aglomerante. Sin embargo, la


202

principal desventaja es que el empelo de un aglomerante puede alterar la porosidad

de la zeolita, provocando la aparición de problemas difusionales y la disminución de

su actividad catalítica. De esta forma, si se quiere tener una configuración en forma

de monolito, las principales condiciones que ha de cumplir este soporte son que evite

las caídas de presión, que permita una distribución uniforme del flujo y una buena

tolerancia a la presencia de partículas que puedan bloquearlo.

En primer lugar, para la aplicación de zeolitas soportadas sobre monolitos, se tiene

que determinar cuáles son los requerimientos para el soporte. Así, los monolitos

cerámicos, son los soportes estándar más empleados para las aplicaciones en fase gas

por su carácter no catalítico y su resistencia térmica. La cordierita, con una

composición química basada en óxidos de magnesio, silicio y aluminio

(2MgO·2Al2O3·5SiO2), es uno de los monolitos cerámicos más empleados y muestra

unas propiedades excepcionales como son bajas caídas de presión en tubos de

escape, buena resistencia térmica, su carácter refractario, buena adherencia de

suspensiones y/o compatibilidad con distintas suspensiones de catalizadores.

En cuanto al método de depósito de zeolitas, se puede hacer de dos formas

diferentes. El primer método consiste en la inmersión (del inglés, Dip-coating) en

una suspensión que contiene la zeolita seguido de un tratamiento térmico para la

estabilización de la capa depositada. El segundo método consiste en la síntesis

hidrotérmica de forma directa (también llamado síntesis in-situ) o mediante

sembrado y una síntesis posterior.

En referencia a la aplicaciones, las zeolitas soportadas sobre monolitos se pueden

aplicar a procesos de adsorción o catalíticos. En cuanto a aplicaciones basadas en

adsorción, se han empleado como adsorbentes de compuestos orgánicos volátiles.

Como catalizadores monolíticos, las zeolitas intercambiadas con metales nobles son

muy eficientes en sistemas de fase gas, como la eliminación de NOx de motores

diesel.


203


Los gases generados por la combustión del petróleo y carbón mineral constituyen un

problema medioambiental debido a que poseen una elevada contribución al efecto

invernadero. Este aspecto medioambiental unido a la disminución de las reservas de

combustibles fósiles, ha llevado al estudio del uso del hidrógeno como nuevo vector

energético. Por ello, recientemente se ha discutido la posibilidad de emplear el H2,

como un nuevo portador de energía que pueda reemplazar tanto a la gasolina como a

otro tipo de combustibles fósiles, en lo que se llama la “Economía del hidrógeno”.

De forma específica en los últimos años se ha planteado el uso del H2 en sistemas

portátiles, en aplicaciones tales como la industria del automóvil o para sistemas

portátiles electrónicos como ordenadores portátiles o teléfonos móviles.

Centrando la atención en las aplicaciones móviles, las pilas de combustible

alimentadas por H2 podrían reemplazar a los motores convencionales. De esta forma,

las pilas de combustible basadas en membranas de electrolito polimérico (PEM) son

una alternativa interesante debido a que son capaces de transformar la energía

química en energía eléctrica, evitando las limitaciones termodinámicas y los

requerimientos mecánicos de los motores convencionales. A pesar de que resulta un

sistema muy prometedor, uno de los problemas con que se encuentran las pilas tipo

PEM es la dificultad de encontrar una forma de producir H2 de la pureza adecuada

para un sistema portátil. Hoy en día, el reformado de hidrocarburos es el proceso

industrial más importante para la producción de elevadas cantidades H2. Las ventajas

más importantes del reformado de hidrocarburos (metanol o etanol) con vapor de

agua son que requiere bajas temperaturas, no necesita el aporte de O2 y se obtiene

una buena relación H2/CO (≈3/1) durante la producción de H2. Pero, el principal

escollo del reformado de hidrocarburos, es que la concentración de CO que sale del

reformador, y que después del proceso de desulfuración y su reducción mediante la

reacción de desplazamiento del gas de agua (del inglés Water-gas-shift, WGS), se


204

encuentra entre 1000 y 10000 ppm. La presencia de una pequeña cantidad de CO

envenena el electrocatalizador de Pt del ánodo, siendo éste uno de los principales

obstáculos para la implantación del reformado en las pilas de combustible. Así, se ha

establecido un máximo de 10 ppm de CO como concentración máxima para evitar la

degradación de la pila de combustible (PEM). De este modo, se necesita una etapa de

purificación de la corriente de H2 previa a la entrada de los gases a la pila de

combustible.

Se han desarrollado varias alternativas para la purificación de H2, siendo las más

estudiadas la separación mediante membranas o bien la oxidación selectiva de CO

usando un catalizador adecuado (reacción denominada “Preferential Oxidation”,

PrOx). En cuanto al empleo de membranas selectivas, las principales ventajas que

ofrecen son su fácil preparación, bajo consumo energético y bajos costes para flujos

de gas bajos. Entre los diferentes tipos de membranas, las membranas de zeolita, son

muy interesantes por sus propiedades de separación y su posibilidad de emplearlas

como reactores de membrana catalíticos.

La otra tecnología interesante para la purificación de H2 es el empleo de reacción de

oxidación selectiva de CO catalizada (PrOx-CO). En esta reacción se emplea O2

como oxidante y posee un carácter exotérmico. Sin embargo, durante la reacción de

PrOx-CO, se produce como reacción paralela la oxidación del H2 a agua. Esta

reacción paralela se tiene que evitar en la medida de lo posible para asegurar una

eficiencia elevada de la pila de combustible. Por ello, se necesitan catalizadores muy

activos y selectivos a la oxidación de CO a CO2 que permitan una pérdida mínima de

H2, teniendo en cuenta la baja concentración de CO comparada con la de H2 con una

relación entre 1/30 y 1/60.

Para la oxidación selectiva de CO se han empleado diferentes metales como Pt, Au,

Pd, Ru o Cu, pero, de entre todos, los más estudiados han sido los catalizadores

basados en metales nobles. En general, los métodos de preparación de catalizadores


205

heterogéneos más empleados han sido la impregnación o el intercambio iónico. Sin

embargo, en las últimas décadas se ha desarrollado un nuevo método de preparación

de catalizadores basados en la síntesis de nanopartículas metálicas mediante métodos

coloidales. Las mayores ventajas que presentan la preparación de nanopartículas es

que permiten controlar el tamaño, la forma o la composición de éstas que va a afectar

tanto a su actividad catalítica como a su selectividad. De esta forma, dentro de la

oxidación selectiva de CO se han empleado catalizadores basados en nanopartículas

de Pt, Rh o Pd. Además, como soporte de catalizadores, las zeolitas presentan unas

propiedades interesantes como es promocionar la oxidación selectiva de CO debido

al efecto de tamiz molecular.

1.3. Reducción de las emisiones de hidrocarburos durante el

arranque en frío en vehículos.

En los últimos años, se han regulado más estrictamente las emisiones de CO,

hidrocarburos y NOx de los gases emitidos por los vehículos. Así, se han desarrollado

diferentes métodos para poder tratar todos estos contaminantes. Estas nuevas

regulaciones implican mejorar, por un lado la eficiencia del combustible consumido

por el vehículo y, por otro lado, purificar lo máximo posible los gases emitidos por

éste.

Una de las preocupaciones en el control de las emisiones de un vehículo es el control

de los hidrocarburos emitidos durante el arranque en frío de los motores de

combustión interna. Los vehículos se encuentran equipados con un catalizador de

tres vías. Este catalizador está formado por metales preciosos soportados en un

monolito recubierto de una película de un soporte. En general, los catalizadores de

tres vías se utilizan para controlar las emisiones de hidrocarburos de un vehículo. Sin

embargo, la temperatura a partir de la cual son catalíticamente activos es de 170ºC

para un catalizador fresco y alrededor de 200-225ºC para un catalizador envejecido.


206

Por ello, estos catalizadores necesitan un tiempo para adquirir su temperatura de

funcionamiento (entre 60 y 120 segundos) y que sean eficaces en la eliminación de

estos compuestos. Durante este tiempo, que se produce durante el arranque en frío

del motor, se producen las mayores emisiones de hidrocarburos (alrededor de un

80%). Por tanto, el control de las emisiones de hidrocarburos durante el arranque en

frío es esencial para poder reducir el impacto ambiental de los motores de gasolina.

Hasta la fecha, las formas utilizadas para abordar este problema han sido muy

diversas, presentando cada una sus ventajas e inconvenientes. Todas las soluciones

planteadas se pueden dividir en dos grupos. En el primero se encuentran las técnicas

basadas en llevar al catalizador a la temperatura de trabajo de forma inmediata, como

son: (i) acercamiento del catalizador a la salida de los gases del motor, (ii)

catalizadores calentados eléctricamente, (iii) aumentar la relación aire/combustible

para producir una mayor combustión ó (iv) calentamiento químico de los

catalizadores. En el segundo grupo se encuentra la preparación de una trampa que

sea capaz de adsorber los hidrocarburos mientras que el catalizador es inactivo y

emitirlos una vez el catalizador ha alcanzado su temperatura de trabajo para que sean

catalíticamente oxidados por el catalizador de tres vías.

De entre todas ellas, el uso de una trampa de HC situada delante del catalizador de

tres vías es la tecnología que presenta un mayor interés. Los factores críticos que

debe de cumplir dicho adsorbente son (i) una elevada capacidad de adsorción de

hidrocarburos a bajas temperaturas, (ii) que empiece a desorber a temperaturas

superiores de 200ºC, (iii) que el proceso sea reversible y (iv) que el material sea

resistente a elevadas temperaturas (750ºC). De esta forma, debido a las

características tan exigentes de este proceso, el tipo de sólido que puede cumplir

todos estos requisitos es muy limitado. Por ello, las zeolitas se presentan como un

posible candidato como adsorbente para esta aplicación, teniendo en cuenta que,


207

además, presentan una estabilidad elevada bajo condiciones muy severas (térmica y

químicamente).

1.4. Reducción de las emisiones de compuestos aromáticos

policíclicos (PAH).

Uno de los tipos de compuestos orgánicos volátiles (VOC) son los hidrocarburos

aromáticos policíclicos (PAH) y se emiten principalmente durante la combustión de

la materia orgánica. Hoy en día, existe una gran variedad de fuentes de emisión de

PAH como la combustión incompleta de hidrocarburos, diesel o gasolina, motores de

combustión interna, plantas de transformación de asfalto o de producción de energía

a partir de carbón mineral. La Agencia de Protección Medioambiental (EPA) ha

establecido una lista con 16 PAH cuyas emisiones deben ser controladas. Los

compuestos que forman la familia de PAH están basados en compuestos aromáticos

de 2, 3, 4 o incluso, 5 anillos aromáticos. A modo de ejemplo, las emisiones de un

motor diesel contienen distintas concentraciones de PAH formados por estructuras de

2 a 5 anillos entre los que se encuentran compuestos tales como naftaleno,

fenantreno, pireno o fluoranteno. La necesidad de erradicar las emisiones de PAH se

centra en que ha sido identificado como un grupo de compuestos peligrosos para la

salud y el medio ambiente. De todos ellos, se suele escoger el naftaleno como

molécula modelo de este grupo de contaminantes.

En los últimos años, se ha incrementado el desarrollo de nuevas tecnologías para la

reducción de PAH. Por ello, para reducir el nivel de contaminación de PAH, se han

desarrollado distintas técnicas tales como biodegradación, ozonización, adsorción,

incineración térmica, irradiación con laser de alta energía y oxidación catalítica. A

pesar de la existencia de numerosas técnicas, la oxidación catalítica para la

producción de CO2 y H2O es la técnica más prometedora debido a que permite


208

trabajar a temperaturas moderadas y posee una elevada selectividad a la producción

de CO2.

En la bibliografía se encuentran numerosos estudios acerca de la eliminación de PAH

mediante oxidación catalítica, a partir del uso de metales nobles soportados sobre γ-

alúmina, óxidos metálicos basados en Co, Mn, Cu, Fe, Ce o Ti, óxidos de cerio

mesoporosos o zeolitas intercambiadas con metales nobles. A pesar de todos los

estudios previos realizados, el empleo de partículas metálicas nanoestructuradas no

ha sido estudiado para esta reacción. De forma general, las nanopartículas metálicas

son partículas de tamaños entre 1 y 50 nm que se tienen que aislar evitando su

aglomeración. Estas entidades nanométricas se tienen que estabilizar mediante el

empleo de moléculas orgánicas que actúan como ligandos. El metal rodeado del

polímero orgánico resultante se denomina coloide que si está disperso en agua recibe

el nombre de “hidrosol” y en disolvente orgánico, “organosol”. Durante los últimos

años se han dedicado grandes esfuerzos a la preparación de nanopartículas metálicas,

siendo su preparación, estructura y las diversas aplicaciones temas de gran interés.

En concreto, las nanopartículas de Pd se han empleado en catálisis en numerosas

reacciones orgánicas. Dentro de esta Tesis Doctoral se va a estudiar el efecto

producido por la combinación de las propiedades de las zeolitas con las propiedades

de las nanopartículas metálicas para la oxidación de los compuestos aromáticos

policíclicos.

2. Objetivos de la Tesis Doctoral.

El objetivo principal de la presente memoria de investigación es preparar zeolitas y

películas delgadas de zeolita para la purificación de corrientes de gases.

En cuanto la purificación de H2, los objetivos son:


209

• Preparar membranas de Na-LTA soportadas sobre discos porosos de grafito,

modificar su porosidad mediante intercambio iónico con cationes alcalinos

(K, Rb o Cs) y estudiar su comportamiento de separación de las mezclas H2 y

CO y de mezclas que simulen los gases emitidos por un reformador.

• Preparar catalizadores estructurados basados en nanopartículas de Pd

soportadas en películas delgadas de zeolita crecidas en monolitos de

cordierita. Estudiar la oxidación la oxidación selectiva de CO en corrientes

ricas en H2.

En referencia a la eliminación de las emisiones de hidrocarburos, los objetivos

específicos han sido:

• Preparar y caracterizar películas delgadas de zeolita BETA crecidas en

monolitos de cordierita de estructura celular.

• Estudiar sus propiedades de eliminación de hidrocarburos en condiciones que

simulan el arranque en frio de un motor de combustión interna.

En relación a la eliminación de PAHs, los objetivos concretos son:

• Preparar catalizadores en polvo basados en nanopartículas de Pd soportadas

sobre las zeolitas BETA y ZSM-5, el tamiz molecular SAPO-5 y γ-alúmina

para estudiar la oxidación selectiva de naftaleno.

• Preparar y caracterizar catalizadores estructurados basados en nanopartículas

de Pd depositadas en películas delgadas de zeolita BETA, ZSM-5, un tamiz

molecular SAPO-5 mediante síntesis in-situ y γ-alúmina, mediante el método

de inmersión, y estudiar la oxidación total de naftaleno.


210

3. Resultados y conclusiones de la Tesis Doctoral.

Capítulo 3: Purificación de hidrógeno para pilas de combustible tipo PEM

mediante membranas preparadas por intercambio iónico de membranas de Na-

LTA/Carbón.

Uno de los primeros objetivos de la presente Tesis Doctoral es la purificación de H2

mediante membranas de zeolita. En este capítulo se ha realizado un estudio de la

preparación de membranas de zeolita A (referida como Na-LTA) soportada sobre

discos porosos de grafito. Además se ha modificado el tamaño de la porosidad para

la obtención de una porosidad más estrecha. Para ello, estas membranas de zeolita, se

han modificado mediante un método simple y reproducible como es el intercambio

iónico con sales de metales alcalinos (K, Rb y Cs).

Todos las membranas preparadas se han empleado para analizar su viabilidad en la

purificación de H2 mediante el estudio de las propiedades de permeación de gases

individuales (H2 y CO) y de mezclas binarias (50% H2, 1.25% CO en He).

Las medidas de permeaciones de gases individuales (H2 y CO) en las diferentes

membranas preparadas revelan que la membrana Cs-LTA/carbón puede separar H2

de CO a 150 ºC, no observándose permeación de CO. En el caso de la membrana de

Rb-LTA/carbón, a todas las temperaturas estudiadas y de la membrana de Cs-

LTA/carbón, a temperatura ambiente y 125 ºC, presentan valores de permeación de

CO que pueden ser debidos a la presencia de defectos (espacios intercristalinos,

huecos micrométricos) en la membrana de zeolita.

Las medidas de permeación llevadas a cabo con mezclas de H2/CO ponen de

manifiesto que las membranas preparadas de Na-LTA/carbón no son adecuadas para

la purificación de H2. La membrana de K-LTA/carbón se puede utilizar únicamente a

temperatura ambiente, pero el flujo de H2 obtenido es más bajo que en el caso de la

membrana de Na-LTA. Los mejores resultados se han obtenido con las membranas


211

de Rb-LTA/carbón y Cs-LTA/carbón puesto que pueden separar el H2 a todas las

temperaturas estudiadas. Sin embargo, los resultados más llamativos son los

correspondientes a la membrana de Cs-LTA/carbón. En este caso, la permeación de

H2 obtenida es similar a la membrana de Na-LTA/carbón, pero no observándose

permeación de CO. Por tanto, la membrana de Cs-LTA/carbón proporciona un flujo

rico en H2 con una concentración de CO menor de 10 ppm, que es el requisito para

que se pueda utilizar en las pilas de combustible.

Capítulo 4: Purificación de H2 mediante membranas de zeolita A/carbón en

mezclas que simulan las emisiones de un reformador.

En este capítulo se han utilizado la membrana original de Na-LTA/C y la mejor

membrana obtenida en el Capítulo anterior (Cs-LTA/C). Estas membranas se han

empleado para el estudio del efecto de la presencia de distintas concentraciones de

CO2 y de H2O sobre las propiedades de permeación de H2 y CO en mezclas que

simulan los gases emitidos por un reformador.

Del estudio del efecto del CO2 en las propiedades de permeación, se ha observado

que las membranas de Na-LTA/C no muestran permeación de CO a temperatura

ambiente debido a la adsorción competitiva del CO2 con respecto al CO. Sin

embargo, este mecanismo se encuentra menos favorecido conforme aumenta la

temperatura debido a que disminuye esta adsorción competitiva. En este sentido las

peores propiedades de permeación se alcanzan a 150 ºC y con un 20% de CO2 en la

corriente de gases, en donde la permeación de CO alcanza su máximo. Además, es de

destacar, que debido a la baja polarizabilidad y la elevada difusividad del H2 la

permeación de éste no se ve afectada por la presencia de CO2. En el caso de la

membrana de intercambiada Cs-LTA/C, la presencia de CO2 no afecta ni a la

permeación de CO ni de H2. En este caso, la permeación de CO2 se ha explicado

considerando que su mecanismo de permeación se encuentra gobernado por la

adsorción fuerte en la zeolita, seguida de una migración superficial.


212

Finalmente, en el caso de los ensayos con mezclas que simulan las emisiones de un

reformador (50% H2, 1.25% CO, 20% CO2 y 5% H2O), se ha determinado que la

membrana original, Na-LTA/C, permite el paso de CO a todas las temperaturas

estudiadas. Estos resultados son peores que los encontrados en base seca, indicando

que la presencia H2O puede producir una contracción de los cristales de zeolita

permitiendo la permeación de CO. Por otro lado, la membrana Cs-LTA/C presento

un elevado grado de estabilidad y las permeaciones de H2 y CO no se ven afectadas

por la presencia de elevadas concentraciones de CO2 y de H2O, obteniendo un

material óptimo para la purificación de H2.

Capítulo 5: Preparación y caracterización de zeolita BETA soportada en

monolitos de cordierita de estructura celular para la retención de hidrocarburos

en condiciones de arranque en frio.

En el presente capítulo se ha optimizado la síntesis para la preparación de zeolita

BETA soportada sobre monolitos de cordierita mediante el método de síntesis in-

situ, para su uso como trampa de hidrocarburos en condiciones de arranque en frío de

vehículos. Los hidrocarburos que componen estas emisiones se dividen en

hidrocarburos ligeros (compuesto modelo propeno) e hidrocarburos pesados

(compuesto modelo tolueno). La tendencia observada es que mientras que los

hidrocarburos pesados se encuentran atrapados adecuadamente, los hidrocarburos

ligeros se emiten a temperaturas bajas. Por tanto, se necesita optimizar una trampa de

hidrocarburos que retenga los hidrocarburos ligeros hasta temperaturas de al menos

200ºC.

Para ello, se ha llevado a cabo un estudio sistemático que ha permitido obtener las

condiciones de síntesis para una película delgada de zeolita BETA tras dos o tres

etapas consecutivas de síntesis. Se ha destacado el efecto de sembrado que produce

la primera etapa de síntesis que, si bien no genera un recubrimiento completo de la

superficie del monolito, los cristales depositados actúan como núcleos de


213

cristalización. La realización de dos etapas consecutivas de síntesis ha permitido

soportar un 17% en peso de zeolita BETA, mientras que con tres etapas de síntesis se

ha podido soportar hasta un 25% en peso de zeolita BETA.

Los monolitos recubiertos de zeolita BETA tras una, dos y tres etapas de síntesis, se

ha empleado para la realización de ensayos en condiciones de arranque en frío. Este

tipo de ensayos consiste en introducir la corriente de hidrocarburos (100ppmv

propeno, 87 ppmv tolueno, 1% O2, 10%H2O en Ar) en el sistema a la vez que se

aumenta la temperatura rápidamente, simulando el arranque de un motor.

De forma general, todos los monolitos recubiertos de zeolita BETA son capaces de

adsorber tanto propeno como tolueno. En el caso del tolueno, como era de esperar,

todos los monolitos recubiertos de zeolita BETA son capaces de retenerlo hasta

temperaturas elevadas. Por otro lado, para los hidrocarburos ligeros, los resultados

varían en función de la cantidad de zeolita BETA depositada. En el caso del monolito

recubierto tras una etapa de cristalización, se observa la emisión de propeno a

temperaturas bajas, mientras que los monolitos preparados tras dos etapas y tres

etapas de síntesis se comportan mejor como trampa de hidrocarburos debido a ese

mayor grosor de la capa depositada que permite la retención de hidrocarburos hasta

temperaturas más altas.

Así, el monolito preparado tras tres etapas de síntesis es el material que mejor se

comporta como trampa de hidrocarburos, al tener la mayor cantidad de zeolita y ser

compacta. Esta elevada cantidad de zeolita con el elevado grado de intercrecimiento

de los cristales obtenido mediante la síntesis in-situ, tiene un efecto positivo,

aumentando su temperatura de desorción hasta temperaturas de 290ºC, cumpliendo

con los requerimientos de la trampa de hidrocarburos. En el caso de los

hidrocarburos más pesados, estos no desorben hasta temperaturas a las que la zeolita

BETA es catalíticamente activa para su oxidación.


214

Capítulo 6: Oxidación total de naftaleno mediante nanopartículas de paladio

soportadas en BETA, ZSM-5, SAPO-5 y alúmina en polvo.

En el presente capítulo, se han preparado diferentes catalizadores basados en

nanopartículas de Pd soportadas (1% en peso carga nominal) en las zeolitas BETA,

ZSM-5, el tamiz molecular SAPO-5 (todos ellos preparados mediante síntesis

hidrotérmica) y γ-alúmina, para su empleo en la oxidación catalítica de naftaleno.

En primer lugar se estudiaron propiedades de los soportes y de los catalizadores

preparados tales como son cristalinidad, textura porosa, dispersión y tamaño de las

nanopartículas de Pd y relación de Pd (0)/Pd (II), de los catalizadores.

Los ensayos catalíticos para la oxidación de naftaleno se han llevado a cabo en un

reactor de lecho fijo. La mezcla de reacción empleada consiste en 20% O2, 80% He y

100 ppm de naftaleno. La conversión hacia CO2 se ha analizado mediante

cromatografía de gases. El intervalo de temperaturas empleado para medir la

actividad catalítica fue entre 100 y 200ºC. Con el fin de estudiar la viabilidad y el

tiempo de vida de los catalizadores preparados, estos se sometieron a experimentos

en continuo a 250ºC durante 48 h.

Del estudio realizado, se determinó que todos los catalizadores poseen una elevada

actividad catalítica después de cuatro ciclos de reacción. El catalizador de Pd/BETA

es el catalizador más efectivo para la oxidación naftaleno (T =165ºC), puesto que

posee una elevada estabilidad después de cuatro ciclos y tras el experimento en

continuo a 250ºC/48h. Por tanto, se trata de un catalizador muy prometedor para su

uso en la eliminación de compuestos aromáticos policíclicos.


215

Capítulo 7: Oxidación total de naftaleno a bajas temperaturas mediante

nanopartículas de paladio soportadas en monolitos de cordierita de estructura

celular recubiertos de óxidos inorgánicos.

En este capítulo se ha estudiado la preparación de los catalizadores monolíticos

recubiertos con películas de distintos sólidos cristalinos (BETA, ZSM-5, SAPO-5 y

γ-Al2O3). En primer lugar, se ha realizado un estudio detallado para la preparación de

películas delgadas de los sólidos anteriores con el fin de cubrir la superficie del

monolito.

En el caso de la preparación de una película delgada de zeolita BETA, se realizó

mediante dos etapas de síntesis, tal y como se describió en el Capítulo 5. En el caso

de la preparación de una película delgada de zeolita ZSM-5 o un tamiz molecular de

silicoaluminofosfato (SAPO-5), se realizó un estudio detallado para la preparación y

obtención de recubrimientos homogéneos de este tipo de materiales mediante síntesis

in-situ. De forma breve, se necesitaron dos etapas consecutivas de síntesis para la

obtención de películas delgadas de estos materiales. Durante la primera etapa de

síntesis se produce un depósito de zeolita o del tamiz molecular que no permite

recubrir por completo la superficie del monolito. Sin embargo, este recubrimiento, se

emplea como sembrado para una segunda etapa de síntesis que permite obtener una

película delgada y continua del material escogido. En el caso de la γ-alúmina, se

llevó a cabo mediante el método de inmersión, obteniendo un recubriendo

homogéneo después de dos etapas de inmersión. A continuación, se depositaron las

nanopartículas de Pd, previamente sintetizadas, sobre la película delgada de sólido

cristalino mediante impregnación (1% carga nominal).

Los ensayos catalíticos llevados a cabo fueron realizados de la misma forma que en

el apartado anterior. De estos ensayos se obtuvo que todos los catalizadores

estudiados permiten la oxidación completa de naftaleno entre 165-180ºC después de

3 ciclos de reacción. Además, los catalizadores basados en zeolitas poseen una


216

elevada estabilidad debido a que sus propiedades no se ven afectadas después de

varios ciclos de reacción. De los experimentos de estabilidad se determinó que los

catalizadores Pd/MBETA y Pd/MZSM-5 son estables y sus propiedades permanecen

invariables (oxidación total de naftaleno a CO2) después de 48h a 250ºC. Por lo

tanto, estos catalizadores constituyen una opción interesante para la reducción de los

PAH mediante oxidación catalítica

Capítulo 8: Oxidación selectiva de CO catalizada por nanopartículas de paladio

soportadas en óxidos inorgánicos en polvo y óxidos inorgánicos soportados

sobre monolitos de cordierita de estructura celular.

En este último capítulo se ha llevado a cabo el estudio de la purificación de H2

mediante la oxidación selectiva de CO. El estudio se ha llevado a cabo empleando

los catalizadores en polvo y monolíticos descritos en los capítulos 6 y 7.

Los ensayos catalíticos se han llevado a cabo en un reactor en U y en un reactor

horizontal, para catalizadores en polvo y monolíticos, respectivamente. Antes de

cada reacción, la muestra es sometida a un tratamiento de reducción previo,

empleando una corriente de 10% H2/He, calentando a 5ºC/min, hasta 200ºC y

durante 2h. La corriente de gases empleada para los ensayos catalíticos consiste en

2% CO, 2% O2, 30% H2 en He, y es similar a la composición de los gases de salida

de un reactor de Water-Gas Shift (WGS). Los ensayos se han llevado a cabo en

modo no isotermo, con una rampa de 2ºC/min hasta 200ºC. Se ha analizado la

conversión de CO y la selectividad de los catalizadores. En el caso de los

catalizadores monolíticos se han realizado dos ciclos de reacción.

De los resultados obtenidos de los catalizadores en polvo, se observó que todos eran

capaces de producir elevadas conversiones de CO con una selectividad elevada. Por

otro lado, los catalizadores soportados han demostrado tener un muy buen

comportamiento catalítico después del primer ciclo. Sin embargo, los catalizadores


217

Pd/MZSM-5 y Pd/MSAPO-5 pierden propiedades catalíticas después de dos ciclos

de reacción. En el caso del catalizador Pd/MBETA, no se observa pérdida de

actividad a lo largo de dos ciclos consecutivos, mostrando elevados valores de

conversión de CO con selectividades próximas al 75% a 140ºC.

Date post:	29-Nov-2018
Category:	Documents
Upload:	truongdiep
View:	214 times
Download:	0 times

Preparation of zeolite thin films for gas...

Documents