ABCDEFG

UNIVERSITY OF OULU P .O. B 00 F I -90014 UNIVERSITY OF OULU FINLAND

A C T A U N I V E R S I T A T I S O U L U E N S I S

S E R I E S E D I T O R S

SCIENTIAE RERUM NATURALIUM

HUMANIORA

TECHNICA

MEDICA

SCIENTIAE RERUM SOCIALIUM

SCRIPTA ACADEMICA

OECONOMICA

EDITOR IN CHIEF

PUBLICATIONS EDITOR

Professor Esa Hohtola

University Lecturer Santeri Palviainen

Postdoctoral research fellow Sanna Taskila

Professor Olli Vuolteenaho

University Lecturer Hannu Heikkinen

Director Sinikka Eskelinen

Professor Jari Juga

Professor Olli Vuolteenaho

Publications Editor Kirsti Nurkkala

ISBN 978-952-62-0276-1 (Paperback)ISBN 978-952-62-0277-8 (PDF)ISSN 0355-3213 (Print)ISSN 1796-2226 (Online)

U N I V E R S I TAT I S O U L U E N S I SACTAC

TECHNICA

U N I V E R S I TAT I S O U L U E N S I SACTAC

TECHNICA

OULU 2013

C 473

Prem Kumar Seelam

HYDROGEN PRODUCTION BY STEAM REFORMING OF BIO-ALCOHOLSTHE USE OF CONVENTIONAL AND MEMBRANE-ASSISTED CATALYTIC REACTORS

UNIVERSITY OF OULU GRADUATE SCHOOL;UNIVERSITY OF OULU,FACULTY OF TECHNOLOGY,DEPARTMENT OF PROCESS AND ENVIRONMENTAL ENGINEERING ,MASS AND HEAT TRANSFER PROCESS LABORATORY

C 473

ACTA

Prem K

umar Seelam

C473etukansi.kesken.fm Page 1 Tuesday, October 22, 2013 12:07 PM

A C T A U N I V E R S I T A T I S O U L U E N S I SC Te c h n i c a 4 7 3

PREM KUMAR SEELAM

HYDROGEN PRODUCTION BY STEAM REFORMING OF BIO-ALCOHOLSThe use of conventional and membrane-assisted catalytic reactors

Academic dissertation to be presented with the assent ofthe Doctoral Training Committee of Technology andNatural Sciences of the University of Oulu for publicdefence in OP-Pohjola-sali (Auditorium L6), Linnanmaa,on 4 December 2013, at 12 noon

UNIVERSITY OF OULU, OULU 2013

Copyright © 2013Acta Univ. Oul. C 473, 2013

Supervised byProfessor Riitta L. KeiskiDocent Mika Huuhtanen

Reviewed byProfessor Jordi LlorcaProfessor Lars J. Pettersson

ISBN 978-952-62-0276-1 (Paperback)ISBN 978-952-62-0277-8 (PDF)

ISSN 0355-3213 (Printed)ISSN 1796-2226 (Online)

Cover DesignRaimo Ahonen

JUVENES PRINTTAMPERE 2013

OpponentAssociate Professor Yohannes Kiros

Seelam, Prem Kumar, Hydrogen production by steam reforming of bio-alcohols.The use of conventional and membrane-assisted catalytic reactorsUniversity of Oulu Graduate School; University of Oulu, Faculty of Technology, Department ofProcess and Environmental Engineering, Mass and Heat Transfer Process LaboratoryActa Univ. Oul. C 473, 2013University of Oulu, P.O. Box 8000, FI-90014 University of Oulu, Finland

Abstract

The energy consumption around the globe is on the rise due to the exponential population growthand urbanization. There is a need for alternative and non-conventional energy sources, which areCO2-neutral, and a need to produce less or no environmental pollutants and to have high energyefficiency. One of the alternative approaches is hydrogen economy with the fuel cell (FC)technology which is forecasted to lead to a sustainable society. Hydrogen (H2) is recognized as apotential fuel and clean energy carrier being at the same time a carbon-free element. Moreover, H2is utilized in many processes in chemical, food, metallurgical, and pharmaceutical industry and itis also a valuable chemical in many reactions (e.g. refineries). Non-renewable resources have beenthe major feedstock for H2 production for many years. At present, ~50% of H2 is produced viacatalytic steam reforming of natural gas followed by various down-stream purification steps toproduce ~99.99% H2, the process being highly energy intensive. Henceforth, bio-fuels likebiomass derived alcohols (e.g. bio-ethanol and bio-glycerol), can be viable raw materials for theH2 production. In a membrane based reactor, the reaction and selective separation of H2 occursimultaneously in one unit, thus improving the overall reactor efficiency. The main motivation ofthis work is to produce H2 more efficiently and in an environmentally friendly way from bio-alcohols with a high H2 selectivity, purity and yield.

In this thesis, the work was divided into two research areas, the first being the catalytic studiesusing metal decorated carbon nanotube (CNT) based catalysts in steam reforming of ethanol(SRE) at low temperatures (<450 °C). The second part was the study of steam reforming (SR) andthe water-gas-shift (WGS) reactions in a membrane reactor (MR) using dense and composite Pd-based membranes to produce high purity H2. CNTs were found to be promising support materialsfor the low temperature reforming compared to conventional catalyst supports, e.g. Al2O3. Themetal/metal oxide decorated CNTs presented active particles with narrow size distribution andsmall size (~2–5 nm). The ZnO promoted Ni/CNT based catalysts showed the highest H2selectivity of ~76% with very low CO selectivity <1%. Ethanol was shown to be a more suitableand viable source for H2 than glycerol. The dense Pd-Ag membrane had higher selectivity but alower permeating flux than the composite membrane. The MR performance is also dependent onthe active catalyst materials and thus, both the catalyst and membrane play an important role.Overall, the membrane–assisted reformer outperforms the conventional reformer and it is apotential technology in pure H2 production. The high purity of H2 gas with a CO-free reformatefor fuel cell applications can be gained using the MR system.

Keywords: bio-alcohols, bio-ethanol, carbon nanotube, catalysts, hydrogen production,membrane reactor, nanomaterials, palladium, steam reforming, water-gas shift

Seelam, Prem Kumar, Vedyn tuottaminen bio-alkoholeista höyryreformoimalla.Perinteiset ja membraaniavusteiset katalyyttiset reaktoritOulun yliopiston tutkijakoulu; Oulun yliopisto, Teknillinen tiedekunta, Prosessi- jaympäristötekniikan osasto, Lämpö- ja diffuusiotekniikan laboratorioActa Univ. Oul. C 473, 2013Oulun yliopisto, PL 8000, 90014 Oulun yliopisto

Tiivistelmä

Maailman energiankulutus on kasvussa räjähdysmäisen väestönkasvun ja voimakkaan kaupun-gistumisen myötä. Tällä hetkellä energian tuottamisen aiheuttamat ympäristöongelmat ja talou-dellinen epävarmuus ovat seikkoja, joiden ratkaisemiseksi tarvitaan vaihtoehtoisia ja ei-perintei-siä energialähteitä, joilla on korkea energiasisältö ja jotka tuottavat vähän hiilidioksidipäästöjä.Eräs vaihtoehtoisista lähestymistavoista on vetytalous yhdistettynä polttokennotekniikkaan, min-kä on esitetty helpottavan siirtymistä kestävään yhteiskuntaan. Vety on puhdas ja hiilivapaa polt-toaine ja energian kantaja. Lisäksi vetyä käytetään monissa prosesseissa kemian-, elintarvike-,metalli- ja lääketeollisuudessa ja se on arvokas kemikaali monissa prosesseissa (mm. öljynjalos-tamoissa). Uusiutumattomat luonnonvarat ovat olleet tähän saakka merkittävin vedyn tuotannonraaka-aine. Tällä hetkellä noin 50 % vedystä tuotetaan maakaasun katalyyttisellä höyryreformoi-nilla. Puhtaan (yli 99,99 %) vedyn tuotanto vaatii kuitenkin useita puhdistusvaiheita, jotka ovaterittäin energiaintensiivisiä. Integroimalla reaktio- ja puhdistusvaihe samaan yksikköön (memb-raanireaktori) saavutetaan huomattavia kustannussäästöjä. Biopolttoaineet, kuten biomassapoh-jaiset alkoholit (bioetanoli ja bioglyseroli), ovat vaihtoehtoisia lähtöaineita vedyn valmistukses-sa. Tämän työn tavoitteena on tuottaa vetyä bioalkoholeista tehokkaasti (korkea selektiivisyys jasaanto) ja ympäristöystävällisesti.

Tutkimus on jaettu kahteen osaan, joista ensimmäisessä tutkittiin etanolin katalyyttistä höy-ryreformointia matalissa lämpötiloissa (<450 °C) hyödyntämällä metallipinnoitettuja hiilinano-putkia. Työn toisessa osassa höyryreformointia ja vesikaasun siirtoreaktioa tutkittiin membraani-reaktorissa käyttämällä vedyn tuotantoon tiheitä palladiumpohjaisia kalvoja sekä huokoisia pal-ladiumkomposiittikalvoja. Hiilinanoputket (CNT) havaittiin lupaaviksi katalyyttien tukimateri-aaleiksi verrattuna tavanomaisesti valmistettuihin tukiaineisiin, kuten Al2O3. CNT-tukiaineellepinnoitetuilla aktiivisilla aineilla (metalli-/metallioksidit) todettiin olevan pieni partikkelikoko(~2–5 nm) ja kapea partikkelikokojakauma. Sinkkioksidin (ZnO) lisäyksellä Ni/CNT-katalyyt-teihin saavutettiin korkea vetyselektiivisyys (~76 %) ja erittäin alhainen hiilimoksidiselektiivi-syys (<1 %). Etanolin todettiin olevan parempi vedyn raaka-aine kuin glyserolin. Tiheillä Pd-Ag-kalvoilla havaittiin olevan vedyn suhteen korkeampi selektiivisyys mutta matalampi vuoverrattuna palladiumkomposiittikalvoihin. Membraanireaktorin suorituskyky oli riippuvainenmyös katalyytin aktiivisuudesta, joten sekä kalvolla että katalyyttimateriaalilla oli merkittävärooli kyseisessä reaktorirakenteessa. Yhteenvetona voidaan todeta, että membraanierotukseenperustuva reformointiyksikkö on huomattavasti perinteistä reformeriyksikköä suorituskykyisem-pi mahdollistaen tehokkaan teknologian puhtaan vedyn tuottamiseksi. Membraanitekniikallatuotettua puhdasta vetyä voidaan hyödyntää mm. polttokennojen polttoaineena.

Asiasanat: bioalkoholit, bioetanoli, hiilinanoputki, höyryreformointi, katalyytti,membraanireaktori, nanomateriaali, palladium, vedyn tuotanto, vesikaasun siirtoreaktio

‘I have no special talent. I am only passionately curious’ - Albert Einstein

Thesis is dedicated to my brother Naveen (Kali)

8

9

Acknowledgements

The thesis work was done in the Mass and Heat Transfer Process Laboratory,

Department of Process and Environmental Engineering, University of Oulu

(2008–2013). The membrane reactor studies were carried out at ITM-CNR,

Rende, Italy in 2009 (two months) and 2011 (three months). I dreamed to do

scientific research in my under-graduation days, but things went very hard. After

moving to Finland, finally, my dream comes true. During the last nine years in

Finland, it was a great journey and learning period in research, education and

well-being. Finally, the journey of PhD is finished and I hope I have achieved the

goals of my thesis objectives.

In order to achieve anything, there are a few people who always help and

support you to reach that goal. First of all and foremost, my sincere and deepest

gratitude to my principal supervisor Professor Riitta Keiski for the support and

guidance. Thanks to her for selecting and providing me this great opportunity. She

has given me flexibility to work, providing valuable information and inputs. I

immensely appreciate her patience and versatility. I must also express special

thanks and gratitude to my thesis supervisor Docent Mika Huuhtanen. He helped

me in every aspect related to the experimental part, manuscript writing and

practical matters. We had many good discussions and arguments; I learned many

valuable things from his experience. I knew very little about the membranes in

gas phase separation, until I met Dr. Angelo Basile, who introduced to me so-

called membrane reactors in hydrogen production. I would like to express my

gratitude to Dr. Angelo Basile for giving me the opportunity to work at ITM-CNR

lab, Italy. I must also thank Dr. Adolfo Iulianelli, who helped me in the MR set-up

and permeation tests. I would like to share my sincere gratitude to Dr. Simona

Liguori for her hospitality and sharing the knowledge during my stay in Italy and

also teaching me the MR system. I must say special thanks to Simona and her

family members.

My sincere gratitude to Dr. Krisztian Kordas and his group, especially to

Anne-Riikka Rautio, for providing the catalyst materials and characterization

results. We had good discussions about the material synthesis and results. I also

got valuable feedback from Dr. Krisztian Kordas, which improved my writing and

analytical thinking. I want also say thanks to Esa Turpeinen for helping me in the

reforming experiments and analyzing the products. We had very funny moments

and jokes in the coffee room. I must also thank Dr. Esa Muurinen for his help

10

related to work computer and other practical issues. I have to say thanks to my

roommate Timo Kulju for his friendship and nice chat during the breaks.

I acknowledge the entire laboratory colleagues who shared the time and good

discussions in the kahvihuone. I would like to thank Anna Valtanen for the

valuable and good discussions regarding the work and other things. I had very

nice discussions with Minna Pirilä, Auli Turkki, Rauli Koskinen, Paula

Saavalainen and Tanja Kolli in Finnish (suomeksi). Thanks for the good moments

we had during the coffee breaks (kahvitaukoja). I say thanks to Satu Pitkaäho for

helping me in the summary writings and other issues related to the dissertation.

Thanks to Auli Turkki (again), with whom I had discussions and a nice chat

regarding Indian projects and other matters. It was a really international

environment in the laboratory, I had good discussions with Christian Hirschman

and friends from Morocco. Thanks to Dr. Junkal Landaburu, with whom I usually

went for lunch and had very good discussions. I am thankful to Jorma Penttinen,

Erkki, Outi L, Heidi Österholm and Marja T (Neste Oil).

Without the financial support, this work would be impossible. I must

acknowledge the financial support from Fortum Foundation, ESF COST Action-

543 (STSM to Italy), EST Graduate School, Tauno Tönning Foundation, Finnish

Cultural Foundation, Tekniikan Edistämissäätiö and the Academy of Finland.

I am grateful to my parents for how they taught me and made me a person

with hardworking and good qualities. I acknowledge my elder brothers Praveen,

Naveen, Mahesh and Deepika (vaddina) who always were there for me,

supporting me during my studies and bad times. I would like to say thanks to my

Wesleyan and Panbazar friends. Thanks to my friends from Denmark, Italy,

Netherlands and Norway (Amar, Ram, Pathi) and friends from Helsinki: Sundeep

(thanks for invaluable discussions), Sandy, Vijender (Tampere). Thanks to my

friends and teachers from Loyola Academy 98CT batch. I would like to express

my acknowledgements to the list of groups in the past and present actively

involved: O-India (Oulu), Oulu cricket club, badminton (Oulu), floor ball (Oulu)

Desifinns (Turku), Fintia (Helsinki), and people in those groups made my stay in

Finland a wonderful experience.

The last person whom I want to acknowledge from the bottom of my heart

and soul to my sweet wife Harinya. I think there are no words to describe her

indispensable support, guidance, understanding, supporting morally and patiently

throughout my PhD work and she was instrumental in this very achievement.

Prem Kumar Seelam Oulu, 10.10.2013

11

List of acronyms and symbols

AC Activated carbon

AFC Alkaline fuel cell

ATR Autothermal reforming

BESR Bio-ethanol steam reforming

BET Brunauer-Emmett-Teller

BGSR Bio-glycerol stream reforming

BJH Barrett-Joyner-Halenda

CCS Carbon capture and storage

CDW Cold diffusion welding

CHP Combined heat and power

CMR Catalytic membrane reactor

CNT Carbon nanotube

CR Conventional reactor (or reformer)

CWM Catalytic wall microreactor

DWCNT Double walled carbon nanotube

EDX Energy dispersive X-ray spectroscopy

ELP Electroless plating

EU European Union

EW Ethanol-water mixture

EWAG Ethanol-water-acetic acid-glycerol mixture

FBR Fixed bed reactor

FC Fuel cell

FTIR Fourier Transform Infrared Spectroscopy

GC Gas chromatography

GCB Graphitic carbon black

GHG Greenhouse gases

GHSV Gas-hourly-space-velocity (h-1)

HC Hydrocarbon

HPLC High-performance liquid chromatography

HPP Hydrogen permeate purity

HRF Hydrogen recovery factor

HT High temperature

HTS High temperature shift

I.D. Inner diameter (µm or mm or nm)

ICE Internal combustion engine

12

IGCC Integrated gasification combined cycle

IMR Inert membrane reactor

IMRCF Inert membrane reactor with a catalyst on the feed side

IR Infra-red

LHV Lower heating value (MJ/kg)

LT Low temperature (°C)

LT-FC Low temperature-fuel cell

LTS Low temperature shift

MEMS Microelectromechanical systems

MF Microfabrication

MM Micromembrane

MMR Membrane microreactor

MR Membrane reactor (or) reformer

MS Microstructure

MWCNT Multi-walled carbon nanotube

NG Natural gas

NP Nanoparticle

O.D. Outer diameter (µm or mm or nm)

PBMR Packed bed membrane reactor

PBR Packed bed reactor

PEM Proton exchange membrane

PEMFC Proton exchange membrane fuel cell

PI Process intensification

PMO Prime Minster Office (Finland)

POX Partial oxidation

PP Partial pressure (bar or kPa)

PSA Pressure swing adsorption

PSS Porous stainless steel

RT Room temperature (°C)

SOFC Solid oxide fuel cell

SR Steam reforming

SRE Steam reforming of ethanol

SRM Steam reforming of methane

SS Stainless steel

STP Standard temperature pressure

SWCNT Single walled carbon nanotube

TCD Thermal conductivity detector

13

TEM Transmission electron micrograph

TOS Time-on-stream (min)

TR Traditional reactor (or reformer)

WGS Water-gas-shift

WHSV Weight hourly space velocity (h-1)

XRD X-ray powder diffraction method

Symbols

abs. Absolute pressure (bar or kPa)

As Membrane cross section area (m2)

Cd Concentration of H2 in metal layer at desorption (mol m-3)

CH2 Concentration of H2 in membrane sub layers (mol m-3)

Cs Concentration of H2 in metal layer at surface adsorption (mol m-3)

DH2 Diffusion coefficient of hydrogen (mol m-2)

dTEM Average nanoparticle size measured by TEM (nm)

dXRD Average nanoparticle size measured by XRD (nm)

Ea Activation energy (kJ mol-1)

Fglycerol, in Glycerol inlet feed flow rate (mol min-1) (for BGSR, in MR)

F´acetic acid, in Acetic acid inlet feed flow rate (mol min-1) (in MR)

F´ethanol, in Ethanol inlet feed flow rate (mol min-1) (in MR)

F´glycerol, in Glycerol inlet feed flow rate (mol min-1) (for BESR in MR)

FCH4, p Molar flow rate of CH4 in permeate (mol min-1)

FCH4, p+r Sum of CH4 molar flow rates at both permeate+retentate (mol min-1)

FCO, in Inlet flow rate of CO (mol min-1)

FCO, out Outlet flow rate of CO (mol min-1)

FCO, p Molar flow rate of CO in permeate (mol min-1)

FCO, p+r Sum of the molar flow rates in both permeate and retentate CO (mol

min-1)

FCO2, p Molar flow rate of CO2 in permeate (mol min-1)

FCO2, p+r Sum of molar flow rates of CO2 on permeate and retentate sides

(mol min-1)

Fethanol, in Ethanol inlet feed flow rate (mol min-1)

Fethanol, out Ethanol outlet feed flow rate (mol min-1)

FH2, p Hydrogen molar flow rate at permeate side (mol min-1)

FH2, r Hydrogen molar flow rate at retentate side (mol min-1)

Fi Flow rate of product ‘i’ (i = H2,CO,CO2,CH4,CH3CHO) (mol min-1)

14

Fi-total Total gas flow rate ‘i’ (retentate + permeate flow, i = H2, CO2, CO,

CH4) (mol min-1)

Fsweep-gas Sweep-gas N2 flow rate (mol min-1) (for MR-only)

JH2 Hydrogen permeating flux through the membrane (mol m2 s-1)

k Hydrogen concentration, corresponds to n´ = 1 (mol m-3)

ka Adsorption rate constant (mol m-2 s-1 kPa-1)

kd Desorption rate constant (mol m-2 s-1)

KH Henry’s law constant

ki Hydrogen dissolution rate constant into the membrane (mol m-2 s-1)

ko Hydrogen dissolution rate constant out of membrane (mol m-2 s-1)

L Length (cm or mm)

LHVH2 Lower heating value of hydrogen (MJ kg-1)

M’feed Mass flow rate of the feed (g h-1)

Mcat Mass of the catalyst pellet (g)

n Pressure exponential factor

pav Average pressure: pretentate+ ppermeate/2 (bar)

pd Hydrogen partial pressure at desorption (Pa)

Pe,H2 Hydrogen permeability (mol m-1 s-1 Pa-0.5)

Pe0 Pre-exponential factor (mol m-1 s-1 Pa-0.5)

Ple Permeance (mol m-2 s-1 Pa-1)

ppermeate Hydrogen partial pressure in the permeate side (bar)

preaction Reaction pressure (bar)

pretentate Hydrogen partial pressure in the retentate side (bar)

ps Hydrogen partial pressure at surface adsorption (Pa)

pTotal Total pressure: retentate + permeate (bar)

R Gas constant (kJ K-1 mol-1)

R2 Regression coefficient (-)

S/E Steam-to-ethanol molar ratio

S’i Selectivity of product gases i = H2, CO2, CO and CH4 (-)

SBET Specific surface area (m2g-1)

SH Solubility coefficient of hydrogen (ml atm mol-1)

Sp Selectivity of products (-)

T Reaction temperature (°C)

Vcat Volume of the catalyst pellet (cm3)

V´feed Volumetric feed flow rate (cm3 h-1)

YH2 Yield of hydrogen (-)

ΔH° Standard enthalpy of the reaction at 298 K (kJ mol-1)

15

Δp Pressure difference (kPa or bar)

Greek letters

αH2/gas_i Ideal selectivity of H2 with respect to other gases ‘i’ (N2, He)

αPd Crystalline phase of Pd-H hydride at low H atomic concentration

βPd Crystalline phase of Pd-H hydride at high H atomic concentration

δ Thickness of Pd layer (µm)

θH Surface coverage of Pd site by H-atom

ρb Catalyst bulk density (g cm-3)

16

17

List of original publications related to thesis

I Seelam PK, Huuhtanen M, Sápi A, Szabó M, Kordás K, Turpeinen E, Tóth G & Keiski RL (2010) CNT-based catalysts for H2 production by ethanol reforming. International J Hydrogen Energy 35: 12588–12595.

II Seelam PK, Rautio AR, Huuhtanen M, Turpeinen E, Kordás K & Keiski RL (2013) Low temperature steam reforming of ethanol over advanced carbon nanotube based catalysts. Manuscript.

III Seelam PK, Liguori S, Iulianelli A, Pinacci P, Calabrò V, Huuhtanen M, Keiski R, Piemonte V, Tosti S, De Falco M & Basile A (2012) Hydrogen production from bioethanol steam reforming reaction in a Pd/PSS membrane reactor. Catalysis Today 193: 42–48.

IV Iulianelli A, Seelam PK, Liguori S, Longo T, Keiski R, Calabrò V & Basile A (2011) Hydrogen production for PEM fuel cell by gas phase reforming of glycerol as byproduct of bio-diesel. The use of a Pd-Ag membrane reactor at middle reaction temperature. International J Hydrogen Energy 36: 3827–3834.

V Liguori S, Pinacci P, Seelam PK, Keiski R, Drago F, Calabrò V, Basile A & Iulianelli A (2012) Performance of a Pd/PSS membrane reactor to produce high purity hydrogen via WGS reaction. Catalysis Today 193: 87–94.

VI Seelam PK, Huuhtanen M & Keiski RL (2013) Chapter 5. Microreactors and membrane microreactors: fabrication and applications, In: Basile A (ed) Handbook of membrane reactors, Volume 2: Reactor types and industrial applications. Woodhead Publishing Series in Energy No. 56. Cambridge UK, Woodhead Publishing: 188–235.

VII Huuhtanen M, Seelam PK, Kolli T, Turpeinen E & Keiski RL (2013) Chapter 11.Advances in catalysts for membrane reactors, In: Basile A (ed) Handbook of membrane reactors, Volume 1: Fundamental materials science, design and optimization. Woodhead Publishing Series in Energy No. 55. Cambridge UK, Woodhead Publishing: 401–432.

The author addressed the research gap and problem identification in this thesis by

articles compilation. In publications 1–3, the author’s main contribution was in

conducting experiments, data analyses and writing the manuscripts. In

publications 4 and 5, he was the co-author, performed experiments, data analysis

and wrote the manuscript in close collaboration with the first author. Publications

6 and 7 are book chapters (review-articles) which are supporting the information

on the future research trends.

18

19

Contents

Abstract

Tiivistelmä

Acknowledgements 9 List of acronyms and symbols 11 List of original publications related to thesis 17 Contents 19 1 Introduction 21

1.1 Hydrogen economy and fuel cell technology .......................................... 21 1.2 Thesis scope and its objectives ............................................................... 24

2 Bio-alcohols for hydrogen production 29 2.1 Overall energy efficiency: bio-alcohols vs. hydrogen ............................. 31 2.2 Catalytic steam reforming in a conventional reformer ............................ 33

2.2.1 Catalysis in steam reforming of bio-ethanol ................................. 33 2.2.2 SRE reaction using CNT supported catalysts ............................... 37

2.3 Catalytic steam reforming in a membrane reformer ............................... 38 2.3.1 Hydrogen production in a membrane assisted reactor .................. 39 2.3.2 Hydrogen permeation through a Pd-metal membrane .................. 42

3 Future trends in membrane-assisted reactors: a short summary 47 3.1 Microreactors and membrane microreactors: fabrication and

applications ............................................................................................. 47 3.2 Advances in catalysts for membrane reactors ......................................... 49

4 Materials and their preparation 51 4.1 CNT supported catalysts ......................................................................... 51

4.1.1 Catalyst preparation ...................................................................... 51 4.1.2 Catalyst characterization .............................................................. 52

4.2 Catalysts and membrane materials in a membrane reformer .................. 53 4.2.1 Catalyst materials in a membrane reformer .................................. 53 4.2.2 Pd-Ag self-supported membrane .................................................. 54 4.2.3 Pd/PSS supported composite membrane ...................................... 54

5 Experimental methods and procedures 55 5.1 Conventional reformer ............................................................................ 55

5.1.1 Reactor set-up ............................................................................... 55 5.1.2 Catalyst activity tests .................................................................... 56

5.2 Membrane reformer ................................................................................ 57 5.2.1 Membrane reactor set-up .............................................................. 57

20

5.2.2 Permeation and performance tests ................................................ 59 6 Results and discussion 63

6.1 Bio-ethanol steam reforming in a conventional reformer ....................... 63 6.1.1 Catalyst characterization ............................................................... 63 6.1.2 Monometallic carbon-based catalysts ........................................... 70 6.1.3 Pt and ZnO promoted Ni/CNT-based catalysts ............................ 77

6.2 Bio-ethanol steam reforming in a membrane reformer ........................... 84 6.2.1 Hydrogen permeation test and factor ’n’ for Pd/PSS MR ............. 84 6.2.2 Effect of reaction pressure ............................................................ 86 6.2.3 Ni/ZrO2 and Co/Al2O3 catalysts in MR ........................................ 89

6.3 Bio-glycerol steam reforming in a membrane reformer .......................... 93 6.3.1 Hydrogen permeation test and factor ‘n’ for Pd-Ag MR .............. 93 6.3.2 Comparison of a membrane reformer and a traditional

reformer ........................................................................................ 94 6.4 Water-gas-shift reaction in a membrane reformer ................................... 98

6.4.1 Effect of He, CO, CO2 and H2O additions on H2

permeation .................................................................................... 99 6.4.2 WGS reaction test: Effect of CO/H2O ratio ................................ 100 6.4.3 WGS reaction test using syngas compositions ........................... 103

7 Summary and concluding remarks 107 7.1 Summary ............................................................................................... 107 7.2 Concluding remarks .............................................................................. 110

References 113 Original papers 123

21

1 Introduction

1.1 Hydrogen economy and fuel cell technology

The world’s energy demand is increasing rapidly due to urbanization, and despite

various regulations, the energy resources are diminishing. The global population

is on the rise and by 2025 it will reach over eight billion (Dorokhina & Tessema

2011, ExxonMobil 2013). Moreover, the main problems that the human race is

facing today are energy security, climate change, and air pollution. For many

decades, crude oil has been a very important fuel and energy source by which the

whole world was driven, and the economy was fuelled. There is a need for clean

energy which can reduce the environmental problems like air pollution and global

warming caused by an increase in greenhouse gases (GHGs) and can also

improve energy efficiency. Many governmental energy departments, institutes,

and industry are working on alternative clean energy resources in order to

improve energy security, and to reduce GHGs (Halman & Steinberg 1999,

Timilsina et al. 2006, Kaltschmitt & Streicher 2007). In Finland, a long-term

energy and climate policy was initiated by the government in order to reduce 80%

of carbon emissions by 2050 to achieve “Low Carbon Society” (PMO 2009).

Moreover, European Union (EU) directives on energy policy, enforcing to use

20% renewable energy sources and minimum 10% of biofuels in the transport

sector by 2020 in order to reduce GHGs (EU Directive 2009, EU Commission

2003). Globally, many ambitious projects and initiatives have been finalized to

demonstrate new and alternative energy technologies to shift to low carbon

societies. In Finland, the fuel cell and hydrogen research programme has been

ongoing (e.g. Polttokennot/Fuel Cells) and this programme was funded by the

Finnish Funding Agency for Technology and Innovation (Tekes) Finland.

Prototype demonstrations and commercialization of fuel cells (FCs) with

combined heat and power (CHP) units will be presented by 2013 (Fuel Cells and

Hydrogen in Finland 2012, Dunn 2002). To achieve zero or low carbon societies,

strong decisions have to be made to improve the existing technologies and to

introduce new alternative energy systems (Stambouli et al. 2002). Abatement of

CO2 and other GHGs in stationary and transportation sources is a critical issue; an

immediate plan is needed to shift to more environmentally sustainable society.

Henceforth, hydrogen economy with fuel cell technology is one of the emerging

technologies in order to solve the global problems. (Marbán et al. 2007, Sperling

22

& Cannon 2004, Raissi & Block 2004, Dunn 2002, Kleijn & Der Voet 2010)

There are many technological and political challenges to achieve complete

transition to hydrogen society. Hydrogen is an energy carrier which can be used in

fuel cells to generate heat and electricity for various applications (Figure 1).

Consequently, hydrogen fuel is a CO2 neutral energy carrier and it emits water as

a by-product during the combustion. Globally, there are no exact figures or

statistics about the hydrogen production capacity, but it was roughly 50–60

million metric tons or 400–500 billion cubic meters per year in 2010 (HyperSolar,

US DOE 2012, Zakkour and Cook 2010). Primarily, hydrogen is used in

ammonia production, chemical industry and also in refining, i.e. hydro-cracking

and hydro-treating reactions. Moreover, hydrogen has been used as a fuel

propellant in aerospace industry for many years. Hydrogen is a clean energy

carrier with no NOx, COx, SOx and particulate matter formed during its

combustion (Hoogers 2003, Borroni-Bird 1996). Integrating FC with CHP is an

energy efficient approach and it improves the overall system efficiency.

Fig. 1. Primary energy sources for hydrogen production and its technologies for fuel

cell applications in various sectors (© EU 2003).

Hydrogen production technologies are widely classified into four main processes,

i.e. thermochemical, electrochemical, photoelectrochemical, and photobiological

processes. The last two technologies are less energy intensive processes and under

research and development phase (Navarro et al. 2007, Holladay 2009). At present,

thermochemical and electrochemical processes are commercially applied

technologies. In thermochemical processes, steam reforming (SR), autothermal

23

reforming (ATR) and partial oxidation (POX) are the common and commercially

used routes to produce H2 (Holladay 2009, Raissi & Block 2004) (Table 1). Each

process has however its advantages and disadvantages. SR is still the dominant

technology to produce hydrogen with high efficiency (~70–85%) and high

hydrogen yield (Holladay 2009). Today, most of hydrogen is produced by using

fossil based feedstocks (e.g. natural gas and coal) via the steam reforming

reaction over Ni based catalysts. Hydrogen has the highest energy density per

mass (120.7 kJ/kg) compared to any other fuel and it exhibits three times more

energy content than gasoline. By using hydrogen as a fuel, the carbon footprint,

and emissions can be reduced and the overall energy efficiency can be improved

(Sperling & Cannon 2004). The demand of H2 is on the rise and in order to have

sustainable hydrogen production towards hydrogen economy and fuel cell

applications, the productivity levels should be increased by using non-fossil based

feedstocks in order to reduce greenhouse gas emissions and to enhance efficiency.

(Marbán and Valdés-Solís 2007, Navarro et al. 2007)

Table 1. Hydrogen production technologies and their efficiencies (Holladay et al.

2009).

Technology Feed stock Efficiency (%) # Maturity

Steam reforming Hydrocarbon (e.g. NG*) 70−85 Commercial

Partial oxidation Hydrocarbon 60−75 Commercial

Autothermal reforming Hydrocarbon 60−75 Near term

Plasma reforming Hydrocarbon 9−85 Long term

Aqueous phase reforming Carbohydrates 35−55 Medium term

Ammonia reforming Ammonia NA Near term

Biomass gasification Biomass 35−50 Commercial

Photolysis Sunlight + water <1 Long term

Dark fermentation Biomass 60−80 Long term

Photo fermentation Biomass + sunlight <<1 Long term

Microbial electrolysis cells Biomass + electricity 78 Long term

Alkaline electrolyzer H₂O + electricity 50−60 Commercial

PEM electrolyzer H₂O + electricity 55−70 Near term

Solid oxide electrolysis cells H₂O + electricity + heat 40−60 Medium term

Thermochemical water splitting H₂O + heat n.a. Long term

Photo electrochemical water splitting H₂O + sunlight 12 Long term

*NG = natural gas, n.a. = not applicable, # thermal efficiency based on the lower heating values

24

1.2 Thesis scope and its objectives

The main scope of this thesis is to study the hydrogen production from biomass-

derived alcohols with high purity and/or CO-free H2 for low temperature fuel cell

applications (e.g. proton exchange membrane fuel cell (PEMFC) and alkaline fuel

cell (AFC)). The thesis was divided into two main research areas, the first being

the catalytic studies and the second the membrane reformer studies, i.e. a

comparative study between conventional and membrane assisted reformers.

In the first phase of the thesis, the feasibility of carbon nanotubes (CNTs) as a

catalyst support for steam reforming of ethanol (SRE) at low temperatures (<

450 °C) was tested in a conventional reformer (CR) or fixed bed reactor (FBR).

The aim was to introduce novel nano-based catalyst materials by CNTs and to

improve the key properties of catalysts in the low temperature steam reforming

reaction. The screening and performance of CNT-based catalysts are reported in

this thesis. In the CR, it is difficult to produce COx-free hydrogen, as the

reformate stream consists of CO, CO2, CH4 along H2. Henceforth, the additional

purification and down-stream separation units are needed in order to produce

COx-free hydrogen for PEMFC.

The second phase of the thesis is devoted to the applicability of a membrane

assisted reformer in steam reforming (SR) and water-gas-shift (WGS) reactions.

The SR of bio-alcohols was performed in a tubular palladium based membrane

assisted reactor with a reforming catalyst bed to produce high purity and/or CO-

free H2. A comparative study between a traditional reformer (TR) (which is also

termed as a conventional reformer (CR)) and a membrane reformer (MR) was

studied. The H2 recovery factor and permeate purity levels were also discussed.

Further, the water-gas-shift (WGS) reaction was also performed in a Pd/PSS MR

in order to achieve a high purity and high yield H2 for feeding the PEMFC.

Additional supplementary papers, i.e. two book chapters, are the supporting

information and critical reviews on the future technologies on integration of

catalysts with membrane based reactor systems and membrane functionality in

microdevices. Finally, concluding remarks are made based on the results and

viability of bio-alcohols for hydrogen production, as well as suitability of CNTs

as support materials in SRE are discussed. The production of COx-free or high

purity H2 using a membrane assisted reactor for fuel cell applications was also

discussed in this thesis work.

In this work, bio-ethanol (Paper I, II and III) and bio-glycerol (Paper IV) are

considered as model compounds to produce hydrogen at low temperatures. The

25

WGS reaction was also studied using a syn-gas feed mixture (Paper V) in a

membrane-assisted catalytic reactor for further purification and enhancing the

hydrogen yield which is coming from the reformate stream.

The main objectives of the thesis were achieved by articles compilation

(Table 2 & 3) which is presented in Figure 2 and in more detail described as

follows.

– Production of hydrogen by the SRE reaction in a conventional reformer at

low temperatures (< 450 °C) over CNT-based catalysts, in order to have a

high hydrogen yield and selectivity with very low CO and CH4 formation.

– Comparison of a conventional reformer and a membrane reformer in

hydrogen production using a dense Pd-Ag MR.

– Production of high purity H2 or at least COx-free hydrogen gas for PEMFC

via SR reactions at ~400 °C in a Pd-based MR.

– Improvement of the H2 yield and purity level in the reformate stream via

water-gas-shift (WGS) reaction to reduce CO in a Pd/PSS MR.

Fig. 2. Thesis objectives are achieved based on the articles compilation (Papers I-V).

The overall aim is to provide new knowledge on the future technologies and

intensification approaches (Papers VI and VII) to build compact and energy efficient

systems, for example membrane-based microreactor systems to produce high purity

H2 for micro-fuel cell systems (e.g., portable electronics).

26

Table 2. The list of papers related to the thesis and their contribution to the thesis aim.

Original papers Contribution to the research aim

Paper I

CNT-based catalysts for hydrogen production

by ethanol reforming

Main motivation to work on hydrogen production

via steam reforming of ethanol (SRE) at low

temperatures (< 450 °C). The suitability of CNTs

as a catalyst support and influence of

metal/metal oxide decorated CNTs in SRE. The

influence of reaction temperature and

metal/metal oxide type decorated CNTs.

Paper II

Low temperature steam reforming of ethanol

over advanced carbon nanotube based

catalysts

SRE at low temperatures (< 450 °C) over Pt and

ZnO promoted Ni/CNT catalysts. Influence of Pt

and zinc oxide on Ni/CNT. The stability of

monometallic and promoted catalysts. Catalysts

performance evaluation for CO reduction and

enhancing the H2 production.

Paper III

Hydrogen production from bio-ethanol steam

reforming reaction in a Pd/PSS membrane

reactor

A study of simulated crude ethanol/water

mixtures with impurities in Pd/PSS MR at 400 °C.

The influence of crude ethanol impurities over the

MR performance. A comparison of the behavior

of two commercial catalysts in MR and their

performance evaluation. Effect of GHSV over a

cobalt catalyst in a Pd/PSS MR.

Paper IV

Hydrogen production for PEM fuel cells by gas

phase reforming of glycerol as a byproduct in

bio-diesel production. The use of a Pd-Ag

membrane reactor at middle reaction

temperatures

Viability of producing hydrogen from glycerol in a

Pd-Ag MR at 400 °C. A comparison of a

traditional versus a membrane assisted reformer.

The effect of WHSV and reaction pressure over

MR performance to produce pure or CO-free H2

gas with > 99.99% purity.

Paper V

Performance of a Pd/PSS membrane reactor to

produce high purity hydrogen via WGS reaction

Permeation analysis of pure and gaseous

mixtures in a Pd/PSS membrane. The potentiality

of composite Pd/PSS in WGS reaction. The

effect of CO-to-steam ratio, syngas feed mixtures

and membrane stability in WGS reaction in a

Pd/PSS MR.

27

Table 3. The list of supplementary papers – critical reviews on the future membrane-

assisted reactors applications.

Supplementary papers

(review chapters)

Critical reviews on future trends

(supporting information)

Paper VI

Microreactors and membrane microreactors:

fabrication and applications

The review article provides an overview on the

membrane integration with microreactor devices.

Providing different approaches and methods to

integrate and fabricate MMRs. State of the art on

MMR devices in catalytic reactions.

Paper VII

Advances in catalysts for membrane reactors This study reviews the catalyst and membrane

functionalities in membrane assisted catalytic

reactors. Phenomena integration and information

on new materials for catalytic membranes.

28

29

2 Bio-alcohols for hydrogen production

Bio-ethanol production

Bio-ethanol (simply ethanol) is produced from wide sources of raw materials

from corn-to-cellulosic materials. Generally, ethanol is produced mainly by two

main routes, i.e. acidic and enzymatic hydrolysis, followed by fermentation and

separation processes. It is not sustainable to produce bio-fuels from edible

materials like corn, sugar crops, etc. The second generation biofuels are

interesting raw materials to produce bio-ethanol, especially from ligno-cellulosic

waste materials (e.g. wood, straw, and cellulosic wastes). (Sun et al. 2002, Kim et al. 2004, Balata et al. 2008) Around ~40–50% of biomass on the earth are

cellulosic materials, and in that, ligno-cellulosic material comprises the major

amount (Sun et al. 2002). Ethanol of the second generation produced from, i.e. organic bio-wastes like agro-, industrial, wood-, forest-, and food wastes (e.g. cheese industry by-products wastes), can be an alternative and viable resources

(Sansonetti et al. 2009, Silveira et al. 2009, Rass-Hansen et al. 2008). One of the

aims in this work was to investigate SR of a simulated crude ethanol mixture (i.e. a mixture with impurities like glycerol and acetic acid and water) which can in

similar composition be produced by fermentation of cheese industries by-

products. The effect of impurities over the performance of two different catalysts

in a MR was studied in Paper III.

Hydrogen from biomass-derived fuels

Hydrogen is not available in nature in a free form; it must be produced from

primary sources. Moreover, hydrogen can be produced from a wide variety of

materials available on the earth. Today, more than 90% of hydrogen is produced

from fossil based feedstocks by SR. In order to produce hydrogen more

efficiently and in an environmentally friendly way, renewable materials should be

considered as H2-sources. Biomass is a versatile raw material that can be used for

sustainable hydrogen production. Bio-derived fuels are efficient candidates for

hydrogen production. (Huber et al. 2003, Lymberopoulos 2005, Tanksale et al. 2010, Rass-hansen et al. 2008) Furthermore, biofuels are renewable, especially

bio-alcohols (e.g. ethanol, butanol), polyols (e.g. glycerol), and organic acids (e.g. acetic acid), derived from a wide variety of wastes i.e. agro-, industrial, food-

30

(e.g. milk- or cheese by-products) and cellulosic wastes (Qinglan et al. 2010,

Valant et al. 2010, and Sansonetti et al. 2009) (Figure 3). They can be good

sources for hydrogen and/or syn-gas production to make further useful energy and

fuels (Brouwer 2010). Interestingly, these materials can significantly improve the

hydrogen production without any major breakthroughs in the technology. The

existing technology can be utilized to produce high purity hydrogen gas from bio-

derived alcohols due to their easy availability, non-toxicity and neutral carbon

cycle (Fig. 3). Moreover, alcohols can be transported easily and stored safely.

This can solve the hydrogen-infrastructure issues like storage and liquefaction

related problems.

Fig. 3. Renewable hydrogen production from biomass-derived alcohols and the

working principle under the closed carbon cycle.

Currently, SR of natural gas is a mature and highly efficient technology in

comparison with other hydrogen production technologies (Table 1). However, as

aforementioned, SR is a highly energy intensive process with high air emissions

(Holladay et al. 2009). Bio-alcohols are easy to reform at low temperatures with

high hydrogen selectivity which is advantageous (Busca et al. 2009) and

improves energy economics and environmental performance (Huber 2003, Byrd

et al. 2008, Yang et al. 2007). A small scale decentralized unit can be established

to produce bio-alcohols and reform them further to hydrogen fuel (on-board or

on-site). The on-board reforming is a feasible option to produce H2 for feeding the

31

FC systems. Thus, one can avoid the hydrogen storage and infrastructure issues

by liquid biofuel (alcohols) distributing units (Holladay et al. 2004).

2.1 Overall energy efficiency: bio-alcohols vs. hydrogen

Globally, there are different types of biofuels produced via thermal and

biochemical conversion processes. Among them, ethanol, glycerol, butanol,

methanol, and methane are interesting sources for hydrogen production due to

their high lower heating value (LHV) per mass and their availability (Table 4).

Moreover, bio-alcohols are sulphur- and nitrogen- free oxygenates which are

potential to replace the conventional fuels. Bio-ethanol is the most commonly

produced biofuel (in volumes) and it is currently used as a gasoline blending

additive (i.e. 5, 10, 15 and 85% of ethanol) (Akande et al. 2005, Rass-Hansen et al. 2008). In some countries, e.g. in Brazil, 100% of ethanol is used in gasoline

engine vehicles. Fermentation broth contains usually 5–12 vol.% of bio-ethanol.

Fermentation is followed by expensive fractional distillation and separation steps

(Rass-Hansen et al. 2008). Thus, downstream separation units are highly energy

intensive and expensive processes to produce fuel grade ethanol. In fact, bio-

ethanol is less energy efficient when used directly in internal combustion engines

(ICE) due to its slow engine start-up and low combustion efficiency. Instead,

directly from the fermentation broth, aqueous ethanol can be reformed to a

hydrogen rich stream to feed low temperature fuel cells (LT-FC) like in PEMFC

or alkaline fuel cells (AFC).

It would be more beneficial to produce bio-alcohols in decentralized units and

transporting the liquid (e.g. aq. ethanol) to reforming to produce hydrogen fuel for

electricity generation, especially in rural areas. For PEMFC, high purity of H2

stream is needed with CO concentration less than 20 ppm (Zhang et al. 2011).

Otherwise, CO can damage and deactivate the FC anode electrode. Types of

biofuels derived from biomass materials and their feasibility to onsite hydrogen

production are summarized in Table 4 (US DOE, Tanksale et al. 2010, Li et al. 2011).

32

Table 4. Types of biofuels derived from biomass materials and their feasibility to

onsite hydrogen production (US DOE, Tanksale et al. 2010, Li et al. 2011).

Biofuels Molecular

formula

H/C Moles

H2

LHV*

(MJ/kg)

Availability, source,

toxicity, etc.

Comments

Ethanol C2H5OH 3 3 27 Abundant, biomass,

non-toxic

Viable for onboard, onsite

H2

Glycerol C3H8O3 2.7 4 16 Abundant, biodiesel

by-product, biomass

Viable, highly viscous than

EtOH

Methanol CH3OH 4 2 20 Abundant, syn-gas,

biomass, toxic

Viable, widely studied

Butanol C4H9OH 2.5 5 34 Unclear, biomass,

toxic

Viable, 85% direct use in

ICE, alternative to gasoline

Methane CH4 4 2 47 Limited NG reserves,

bio-gas

Viable, commercially

compressed NG

Hydrogen H2 - 1 120 Abundant, not in a free

form, highly

combustible

Clean energy carrier,

combust to produce clean

electric power

*LHV = lower heating value (MJ/Kg), NG = natural gas

The CO2 and water management issues are also important in a fuel cell system.

For solid oxide fuel cells (SOFC), purity of H2 is not as critical as for LT-FC. The

main applications of LT-FC are in stationary small scale auxiliary power units,

transportation vehicles, and portable power electronics. It is important to know

the energy economics in order to evaluate the benefits to use ethanol or glycerol

in hydrogen production.

By definition, energy efficiency is the ratio of LHVH2 of the produced

hydrogen to the LHV of energy inputs (i.e. feedstock + transportation + electricity

+ downstream, etc.). Theoretically, the net energy efficiency is high in the case of

hydrogen from ethanol compared to pure ethanol as reported by many studies

(Morgenstern et al. 2005 and Akande et al. 2005). Moreover, in a membrane

based reforming processes, a high net energy efficiency is gained compared to the

traditional reforming or ethanol ICE (Roses et al. 2011, Manzolini et al. 2008,

Morgenstern et al. 2005, Mendes et al. 2010). As reported by Deluga et al. (2004), 1 mol of ethanol can generate 1 270 kJ of energy which is equal to

produce hydrogen fuel to generate electric power of 350 Wh. The efficiency of

ICE is around ~20% compared to ~40–60% for a FC. In order to produce fuel

grade ethanol, more input energy is required. Thus, aqueous ethanol could be

considered as a more efficient raw material to produce hydrogen than other non-

renewable sources. In a similar manner, alcohols like methanol and polyols like

33

ethylene glycol and glycerol are also promising liquid fuels which can be utilized

to produce hydrogen. Even though glycerol can produce more H2, due to more C

atoms, it can produce more COx molecules than ethanol (Busca et al. 2009, Zhang

et al. 2007). Glycerol is a highly viscous; thus, it requires more vaporization

energy than ethanol. At the moment, the market for the glycerol price will be

depending on the bio-diesel production. Henceforth, ethanol and glycerol are the

most favored bio-fuels for hydrogen production. The feedstock cost and

availability will be crucial for commercialization of hydrogen producing plants

for large, small and niche energy applications (Holladay 2004).

2.2 Catalytic steam reforming in a conventional reformer

2.2.1 Catalysis in steam reforming of bio-ethanol

Biomass-derived fuels (e.g. bio-ethanol and bio-glycerol) are potential raw

materials for hydrogen production for fuel cell applications. The catalytic steam

reforming of bio-ethanol (SRE) is a widely studied reaction and published in

many literature reviews, e.g. by Mattos et al. (2012), Ni et al. (2007), Vaidya et al. (2006), Haryanto et al. (2005), Cheekatamarla et al. (2006).

Thermodynamically, steam reforming of ethanol is an endothermic reaction and

requires heat to complete the reaction. A complete thermodynamic analysis of

SRE has also been reported by Rabenstein et al. (2008), and Sun et al. (2012),

and compared with other hydrogen producing reactions (i.e. ATR and POX).

Further, SRE is a thermodynamically limited reaction. The main products from

SRE at equilibrium compositions are H2, CO, CH4, CO2 and other minor

intermediate species like acetaldehyde, ethylene and coke (Fatsikostas et al. 2004,

Sun et al. 2012). It is possible that ethanol can be fully converted to gaseous

products at low temperatures (< 450 °C) (Chen et al. 2012, Sun et al. 2005, and

Panagiotopoulou et al. 2012). At low temperature (LT) (< 450 °C) SRE, it is

difficult to prevent the formation of CH4 (Roh et al. 2006, Ni et al. 2007, Chen et al. 2012). Generally, at high temperature (HT) reforming, i.e. above 500 °C, the

formation of CO significantly increases and methane decreases with temperature

(Vaidya et al. 2006, Bi et al. 2007). At LT-SRE, it is beneficial to control the

formation of CO at low concentrations. When the steam-to-ethanol (S/E) molar

ratio is higher than the stoichiometric 3, the coke formation can be avoided and

also a high theoretical hydrogen yield can be achieved (Ni et al. 2007). The water

34

utilization is very low during the reaction and it acts as an excess reactant. As

many authors reported, the reaction order with respect to ethanol is 1 and to water

it is zero or a negative order. Moreover, a high S/E ratio requires high

vaporization energy. The SRE reaction is a very complex reaction network and a

set of important reactions taking place during SRE at low temperature are

presented in Table 5.

Table 5. SRE reactions occur at low temperature (Bshish et al. 2011, Ni et al. 2007,

Panagiotopoulou et al. 2012).

No. Reaction Equation ∆H°298K (kJmol-1)

R1 Overall SRE CH3CH2OH+ 3H2O ⇌ 6H2+2CO2 256

R2 Ethanol dehydrogenation CH3CH2OH ⇌ CH3CHO +H2 68

R3 Acetaldehyde reforming CH3CHO+H2O ⇌ 3H2+2CO 187

R4 Ethanol decomposition CH3CH2OH → CH4+CO +H2 49

R5 Acetaldhyde decomposition CH3CHO → CH4+CO −19

R6 Ethanol dehydration CH3CH2OH ⇌C2H4+H2O 45

R7 WGS reaction CO+H2O ⇌ CO2+H2 −41

R8 Ethylene polymerization C2H4 → polymers →carbon −172

R9 Boudouard reaction 2CO ⇌ CO2+C −172

R10 Methane decomposition CH4 ⇌ 2H2+C 75

SRE involves many reaction intermediates and unstable products, especially at

low temperatures (Rabenstein et al. 2008). Catalysts play an important role to

activate and to rupture the C-C and C-H bonds in the ethanol molecule (Haryanto

et al. 2005, Ni et al. 2003, Mattos et al. 2012). Many noble and base (groups 8–

10) metal catalysts have been tested in SRE at different reaction conditions (e.g. Liguras et al. 2003, Breen et al. 2002, Seelam et al. 2010, Liberatori et al. 2007).

The noble metal catalysts are most promising due to their high coke resistant

tendency. Among them, rhodium (Chen et al. 2012), platinum (Ciambelli et al. 2010), and palladium (Casanovas et al. 2006) are found to be the active, selective

and stable catalytic materials for the SRE reaction. Further, Rh based catalysts are

the most active and selective and they are also reported to produce a CO-free H2

stream at low temperatures (Chen et al. 2012, Lee et al. 2012). Due to the cost,

noble metals are less preferred ones at the industrial scale. Consequently,

researchers are working more on non-noble metal catalysts like Ni and Co due to

their lower costs. Although non-noble metals are very active and selective, they

have serious deactivation problems (Profeti et al. 2009, Sanchez et al. 2007).

Nickel based catalysts have been applied industrially for many years. In SRE,

35

cobalt is also active, selective, and more coke resistant than Ni (Bichon et al. 2008). However, the catalyst support which is very critical to provide active phase

during the reaction plays an equally important role, especially at low

temperatures. Many catalyst support materials have been investigated in SRE.

Many conventional supports like Al2O3, SiO2, TiO2, ZnO, CeO2 and ZrO2

(Sun et al. 2005, Haga et al. 1997, Sánchez et al. 2007) have been tested and used

commercially. Alumina is the most widely studied and industrially applied

support material for, e.g. steam reforming of methane (SRM). Moreover, α- and

γ-Al2O3 are the most promising support materials due to their thermal and coke

resistance. Alumina is an acidic type support and exhibits strong metal-support

interactions which can lead to coke formation (especially at LT).

For LT fuel cells, the reformate stream with high temperature is not beneficial

and energy losses can occur. Thus, LT SRE reaction is needed in order to produce

H2, to reduce the heat losses and to feed H2 directly to PEMFC. The concentration

of CO should however be lower than 20 ppm (Zhang et al. 2013). Therefore, a

selective and active metal catalyst with a strong support function is required to

produce H2 more efficiently with low CO and CH4 formation.

Many studies have reported that ethanol is completely converted into gases at

temperatures above 500 °C and high hydrogen yield is gained along with the by-

products such as CO, CH4 and CO2 (Ni et al. 2007, Haryanto et al. 2005).

Moreover, when the SRE reaction is performed at temperatures 550–800 °C, coke

formation can be avoided (Rabenstein et al. 2008). In SRE, temperatures below

450 °C, the complete acetaldehyde decomposition (R5) takes place, while above

500 °C methanation, methane reforming and reverse WGS reactions are favoured

(Ni et al. 2007, Roh et al. 2006). As reported by Silva et al. (2009), carbon

deposition can be mitigated in low temperature SRE reactions. It is important to

avoid ethylene (R6) and acetaldehyde formation (R2) that may lead to carbon

formation. New materials are introduced, e.g. CNT and TiO2 with different

compositions and preparation methods, in order to improve the key catalytic

properties (Sánchez et al. 2007, Benito et al. 2007, Sun et al. 2005). Choosing

right and correct metal-support compositions is also a critical point in the

preparation of active catalysts (Ertl et al. 1997).

SRE at low temperatures is very challenging due to thermodynamic

limitations and the coke forming region. In LT SRE, a strong support function

with optimal interactions of metal-support is needed (Roh et al. 2007). Moreover,

catalyst support plays an important role in selective production of hydrogen.

Consequently, the catalyst should produce hydrogen with low CO formation.

36

Coke formation is very critical at low temperature SRE reaction. In order to avoid

coke and its precursors at LT SRE, a better catalyst system with robust properties

is needed.

Fig. 4. Equilibrium composition (in moles) for steam reforming of ethanol as a function

of temperature (°C): (A) Molar ratio H2O: ethanol = 3:1 and (B) H2O: ethanol = 13:1.

100 200 300 400 500 6000

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

C

mol

Temperature

H2O(g)

H2(g)

CO2(g)CH4(g)

CO(g)

Temperature (°C)

B

Temperature (°C)

100 200 300 400 500 6000.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

C

mol

Temperature

H2O(g)H2(g)

CH4(g)

CO2(g)

CO(g)

C

A

Temperature (°C)

37

The equilibrium compositions for the SRE reaction is presented in Figure 4 for

the two molar feed ratios, i.e. steam to ethanol 3:1 and 13:1, which were used in

this work (Paper I, II and III). The higher molar ratio, i.e. steam-to-ethanol of 13,

is beneficial to be utilized, because of two main reasons. First, the ratio is the feed

composition from the ethanol fermentation broth, and second, at high S/E ratios

carbon and CO formation can be reduced. Moreover, at higher steam-to-ethanol

molar ratios, the hydrogen yield is enhanced.

2.2.2 SRE reaction using CNT supported catalysts

Catalysis by nanomaterials opens a new window of opportunities to make robust

and efficient catalytic systems for many chemical transformations (Somorjai et al. 2009, Xu et al. 2008). Nanomaterials in catalysis are not new but the new

approaches have broadened the horizons in catalysis research. Among them,

carbon based nanomaterials are of interest in applications in many fields. Carbon

materials are inert, neutral in nature, and exhibit high thermal and mechanical

stability. Carbon support nanomaterials are generally classified by their

morphology, for example carbon nanofibers, nanotubes, nanospheres and

nanowires (Serp et al. 2003, Halonen et al. 2010). The most interesting ones are

carbon nanotubes (CNTs) due to their excellent physico-chemical properties,

organized structures, tube confinement, site accessibility and avoidance of

diffusion limitations (Pan et al. 2008, Li et al. 2007).

CNTs are rolled graphene sheets in cylindrical forms and they exhibit

different sidewall layers like single wall (SWCNT), double (DWCNT) and multi-

walled carbon nanotubes (MWCNTs). MWCNTs are cheaper, easier to prepare

and more widely available than the others. (Serp et al. 2003) Moreover, in CNT

(or MWCNT), carbon surfaces are highly resistant to acidic and basic media, and

surfaces can be modified to control the polarity and hydrophobicity (Pan et al. 2008, Serp et al. 2003). MWCNTs are used as catalytic support materials in many

reactions due to their flexibility to disperse active metal phase (Zgolicz et al. 2012, López et al. 2012, Li et al. 2007). CNTs are widely studied for FCs as an

electro-catalytic material and also for many other applications like for energy and

fuel production (Oriňáková et al. 2011). Using CNTs as supports, a narrow and

uniform active metal distribution with good dispersion can be achieved with small

nanoparticle sizes (Hou et al. 2009, Seelam et al. 2013b). CNTs have been

already studied in methanol, methane, propane, and bio-oil steam reforming

reactions. For methanol SR (Yang et al. 2007), Cu/ZnO-CNTs have been prepared

38

by a chemical reduction method. The activity of Cu/ZnO-CNT was found to be

very promising, achieving ~100% methanol conversion with high H2 yield and

low CO formation. The high activity is due to the good Cu nano-particles (NPs)

dispersion over the surface of CNTs with the 10 nm particle size (Yang et al. 2007). A scheme is presented in Figure 5 which shows the SRE reaction over the

CNTs supported catalysts.

Fig. 5. Schematic for SRE reaction over a metal decorated CNT-based catalyst.

In a study reported by López et al. (2012), a Ni/MWCNT catalyst was tested in

propane SR reaction and Ni was deposited by the deposition-precipitation method

on the CNT support. Due to controlled metal dispersion over MWCNT, the

optimal metal content (20 wt.% Ni) and particle size facilitated higher activity

and selectivity compared to a commercial Ni/Al2O3 catalyst. Moreover, the active

metal phase and also promoters have optimal interactions with CNTs, which play

an important role in the catalytic reactions (Zgolicz et al. 2012).

2.3 Catalytic steam reforming in a membrane reformer

Membrane technology is the most promising and highly energy efficient method

that is studied widely in many industrial (e.g. wastewater treatment and

purification of gases), pilot and niche applications like energy production,

pharmaceuticals production, drinking water purification and O2 or N2 enriching,

(Baker 2004, Bruschke et al. 1995). Generally, membranes for gas separation are

classified as organic (e.g. polymers) and inorganic (zeolites, porous carbon,

ceramic, and metal) materials. Furthermore, the membranes are categorized into

39

porous and dense materials. In order to use membrane materials for a specific

function, important criteria are adsorption, diffusion and solubility factors of a gas

molecule. Thus, a membrane allows the permeation of a desired gas molecule

more selectively in comparison with other gas molecules. Each gas, ion or vapor

has affinity towards specific membrane materials. For example Pd exhibits a high

adsorption coefficient for H2, Ag for O2, polymers for H+ ion conductivity, etc.

(Baker 2004) Membranes are introduced into a reactor system with different

configurations, shapes and sizes. In membrane reactors or reformers (MRs), the

membrane can act in three types or functions, i.e. extractor, distributor and

contactor-type MR (Caro et al. 2010, Dittmeyer et al. 2008). In this study, the Pd-

membrane function is the extractor-type MR, in which H2 is continuously

removed from the reaction zone to the permeate side (Paper III, IV and V).

In MR, reforming reaction and selective hydrogen separation take place

simultaneously in one unit. Thus, the reactant conversion and yield enhancement

can be achieved. Thermodynamically limited reactions like SR and WGS can be

displaced through the shift effect, according to the Le Chatelier principle: “if one of the products is removed continuously, the reaction is forced to shift towards forward direction i.e. product side” (Bruijn et al. 2007, Dittmeyer et al. 2001).

2.3.1 Hydrogen production in a membrane assisted reactor

Figure 6 represents the traditional process for hydrogen production by SR of

natural gas in comparison with a membrane-assisted reformer. In MR, the

reaction, separation and purification is done simultaneously in one unit. A

conventional process in the hydrogen production plants consists of pre-treatment,

pre-reforming, and steam reforming sections, CO purification units (e.g. shift

reactors, POX), pressure swing adsorption (PSA) and compression units to obtain

high purity hydrogen, i.e. ~99.999% (Figure 6). In order to reduce CO, high

(HTS) and low temperature shift (LTS) reactors and partial oxidation (POX) were

utilized before the downstream separation steps. Henceforth, a traditional process

with a PSA or a cryogenic separation unit is highly energy and cost intensive. A

high purity COx-free H2 stream can be produced by using dense Pd-based MR

without further downstream separation units. Thus, a membrane based separation

process is highly energy and cost efficient. The conversion and yield enhancement

can be achieved in a membrane reformer due to the displacement of

thermodynamic equilibrium more to the product side through the shift effect.

40

Fig. 6. Schematic presentation of a traditional reforming process versus a membrane-

assisted reformer in hydrogen production.

An integrated fuel cell with a combined heat and power (FC-CHP) unit is the

most efficient way to recover the heat losses and improve the energy and system

efficiency. Coupling the membrane assisted reformer based fuel processor with

the FC-CHP unit is a viable solution to produce electricity and heat for the

portable and residential FC applications, as proposed in Figure 7.

Hydrogen separating membranes are widely studied in many chemical

reactions, especially in SR, WGS, oxidative SR, dehydrogenation reactions (Shu

et al. 1991 Iulianelli et al. 2009, Basile et al. 1996 & 2011, Liguori et al. 2012,

Dittmeyer et al. 2001) and also for carbon capture and storage (CCS) (Jansen et al. 2009). For example, in ammonia plant, the hydrogen from the recycle line is

purified by a membrane and recycled to the feed line (Rahimpour et al. 2009).

Choosing a membrane always depends on many factors like perm-selectivity (i.e. separation factor), permeability, thermal stability, resistance to impurities and

manufacturability (Baker et al. 2001, Adhikari & Sandun 2006). The membrane

function mainly depends on its operating temperature. For, e.g. low temperature

fuel cell (LT-FC) (< 100 °C) applications, polymers like Nafion polymer for

proton conducting in PEMFC are the most suitable materials. For high

temperature (200 °C − 500 °C) applications, inorganic membranes are the most

appropriate ones due to their high thermal stability in harsh conditions (Hsieh

1996). Inorganic membranes in hydrogen separation have been widely studied

41

and investigated for many years by several researchers (Kikuchi & Uemiya 1991,

Shu et al. 1991).

Fig. 7. Comparison of the number of stages for a conventional (A) and a membrane

reformer (B) for a fuel-processor and an integrated PEMFC with combined heat and

power (CHP) units.

The Pd and Pd-alloy membranes for hydrogen separation are well reported in the

literature for the SR of methane, methanol, ethanol, WGS and dehydrogenation

reactions (Shu et al. 1991, Basile et al. 2001, 1996, 2008, Dittmeyer et al. 2001,

Seelam et al. 2011 & 2012, Iulianelli et al. 2011, Liguori et al. 2012, Tosti et al. 2008). In the H2 separating membranes, Pd is the most desirable metal for its high

H2-adsorption coefficient and perm-selectivity compared to other metals. Many

researchers have reported the Pd applicability for selective hydrogen separation

and purification at operating temperatures of 350–500 °C (e.g. Shu et al. 1991,

Adhikari et al. 2006). Moreover, Pd-based membranes exhibit the H2

embrittlement factor at low temperatures (Yun et al. 2011). This may lead to

micro-cracks in the Pd layer. Absorption of hydrogen below the critical point, i.e. 298 °C at 2 MPa, produces two phases αPd and βPd. At low H/Pd ratios, the α-

phase is dominant at high temperatures. At LT below 300 °C, the co-existence of

both phases with maximum (αPd-phase) and minimum (βPd-phase) H/Pd ratios for

pure Pd leads to phase transition, i.e. αPd-to-βPd (Shu et al. 1991, Yun et al. 2011).

The strain and recrystallization may lead to defects and pin-holes in the Pd layer.

Another problem with pure Pd metal is the deactivation of Pd-surface by COx,

hydrocarbon species, steam, which may be adsorbed as poisons and interact with

Pd and reduce the overall flux (Li et al. 2000, Liguori et al. 2012). In order to

avoid all aforementioned problems related to pure Pd, alloying or introducing a

second metal to form a Pd-alloy, for example, Pd-Ag, Pd-Cu and Pd-Ni can be

used (Yun et al. 2011). Permeability can in that way be improved, phase

transitions are avoided and deactivation of the membrane surface will be

Pretreatment and

prereforming units

Steam reforming

WGS-HT reactor

WGS-LT reactor

Partial oxidation

PSA for separation

Compression -

99.999% H2

PEM Fuel cell

processor

Electricity and heat

(CHP)

Membrane reformer (reforming,

separation and purification of H2)

1st stage

Membrane reformer (WGS for retentate stream)

2nd stage (optional)

Compression 99.999% H2

PEM Fuel cell processor

Electricity and heat

(CHP)

B

A

42

diminished. Many scientists are working to prepare optimized materials which

exhibit high perm-selectivity with reasonable throughput fluxes. The cost of the

membrane materials can be reduced by lowering the thickness of the membrane,

which improves the overall flux (Lu et al. 2007). Hydrogen permeates through the

dense Pd-metal lattice following the solution-diffusion mechanism. Henceforth,

dense Pd-based membranes have an infinite or H2/N2 > 10 000 separation factor.

First of this kind of a Pd-membrane reformer was used in a commercial scale by

Tokyo Gas Company (Shirasaki et al. 2009).

2.3.2 Hydrogen permeation through a Pd-metal membrane

Generally, hydrogen permeation through metals takes place in a series of complex

steps. Hydrogen permeability is high in Pd-based membranes due to its high

solubility and diffusivity. Mainly four steps are involved, as reported by Shu et al. (1991); First, reversible dissociative chemisorption of hydrogen on the lattice

surface, i.e. on membrane surface; second, reversible dissolution of surface

atomic hydrogen into the bulk metal layer; third, diffusion of atomic hydrogen

(Pd-H) in the membrane. Finally, the recombination of hydrogen atoms takes

place to form H2-molecules which are desorbed. The rate-limiting step most often

is the bulk diffusion.

In Shu et al. (1991), the expression for the hydrogen flux with respect to the

difference in the partial pressure at the retentate (i.e. lumen) and permeate side

(i.e. shell) is derived as follows: At steady state, the H2 permeating flux (JH2)

follows the Fick’s law of diffusion for one dimension (Equation (1)) and the

overall hydrogen flux at the surface adsorption and desorption in metals is

described in Equation (2).

JH2 = -DH2CH2

δ (1)

JH2 = DH2

δ(Cs-Cd), (2)

where Cs is the concentration of H2 in the metal layer at surface adsorption (mol

cm3) and Cd is the concentration of H2 in the metal layer at desorption (mol cm3),

DH2 is the diffusion coefficient of hydrogen (cm2 s-1) and δ is the thickness of the

membrane (µm).

The diffusion of hydrogen at the outlet of the membrane is expressed as the

difference of desorption and adsorption rates (Equation (3)).

43

JH2 = kdθ 2 − kapH2' (1− θ )

2, (3)

where θH is the surface coverage of Pd site by a H-atom, p´H2 is the partial

pressure of hydrogen and k and k are the adsorption and desorption rate

constants.

Equation (3) represents the adsorption of hydrogen on two empty sites and

desorption of the recombined H-atoms in the adsorbed layer. At the steady state,

the hydrogen dissolution (or solubility factor) at the interface is denoted by

Equation (4).

JH2 = kon'(1− θ ) − kiθ (1− n'), (4)

where k0 and ki are the dissolution rate constants in and out of the membrane,

respectively. The atomic ratio of H/Pd is denoted by n´.

Assuming the diffusion step as the rate-limiting, Equations (2) and (3) are

solved resulting in the Sievert’s law expression under one hypothetic condition,

i.e. n <1. Therefore, diffusion of atomic hydrogen through the bulk metal lattice is

the rate determining step and the H2 partial pressure in the Pd-based metal lattice

is lower than the one (Sievert’s validity) (Caravella et al. 2010). This hypothesis

is valid when using the α-phase of the Pd-hydride. The concentration terms will

change to partial pressure of hydrogen and the Equation (5) follows a Sievert’s

expression.

JH2 = DH

δKs(ps

0.5 − pd0.5) (5)

where ps is the partial pressure of hydrogen at surface and pd is the partial

pressure of hydrogen at desorption.

The permeability of hydrogen is then calculated based on the following

expression (Equation (6)).

Pe,H2 = kDHko

ki

kd

ka

0.5, (6)

where k is the H2 concentration, corresponds to n´ = 1 (mol m-3). Ks is the

Sievert’s constant expressed by Equations (7) and (8).

Ks=ko

ki

kd

ka

0.5. (7)

The Henry’s law at constant temperature can be expressed by Equation (8).

KH= ∆CH2

∆pH2

, (8)

44

where KH is the Henry’s constant, pH2 is the partial pressure of hydrogen,

substitute the ΔCH2 = KHΔpH2, inserting Equation (8) into (1) leads to Equation

(9):

JH2 =-DH

δKH∆pH2

. (9)

By definition, permeability is the product of mobility (dissolution) and solubility

(sorption) factors, i.e. DH and KH (Equation (10)).

Pe,H2 = DHKH, (10)

where Pe,H2 is the hydrogen permeability (mol m-1 s-1 Pa-1). In Equation 10, KH is

the hydrogen solubility which is a temperature-dependent constant and termed

also KH = SH and DH is the diffusion coefficient which measures the mobility of

the H2 molecules within the membrane (m2 s-1) and SH is the solubility

coefficient; it measures the dissolution of the H2 gas molecules within the

membrane.

Permeance (Pel ) is defined as the ratio of permeability to the thickness of the

membrane layer (Equation (11)).

Pel =

Pe, H2

δ (11)

The separation factor (αH2/i ) or ideal selectivity of the hydrogen with respect to

other gases is defined as in Equation (12)

αH2/i = Pe,H2

Pe,i =

Pe,H2l

Pe,il , (12)

where i denotes the N2 or He species and Pe,H2 and Pe,i are the permeabilities of

hydrogen and specie i, Ple,H2 and Pl

e,i are the permeances of hydrogen and specie i.

The perm-selectivity is defined as the measure of the ability of the membrane

to selectively transport one specific gas. For ideal gases, the permeability is

described as the gas permeation rate (JH2) through the membrane with a surface

area (As) and the thickness of the membrane (δ) and the driving force for the

separation is the partial pressure difference across the membrane (Δp). In the

dense Pd-metal membranes, hydrogen follows the solution-diffusion mechanism

and obeys the Sievert’s law (Equation (13)).

JH2 = PeH2

δpS

0.5 − pd0.5 (13)

Equation 13 represents the Sievert’s law for dense Pd-metallic membranes and it

has infinite selectivity to hydrogen. At constant temperature, H2 permeation

45

through dense palladium membranes occurs via the solution/diffusion

mechanism. Generally, the hydrogen permeation through metal membranes

follows the Equation (14).

JH2 = PeH2

δ(pretentaten −ppermeate

n ), (14)

where pretentate is the pressure at retentate side (bar), ppermeate is the pressure at

permeate side (bar), δ is the thickness of the Pd layer (µm) and n is the pressure

exponential factor.

The factor ‘n’ determines the rate controlling step in the permeating H2 flux

through the Pd-membrane. The surface and internal diffusion of hydrogen through

the Pd-metal lattice plays a critical role (Caravella et al. 2010). The solubility

factor is an important feature in determining the permeation characteristics

through the metal. Henceforth it is important to calculate the factor ‘n’. If n = 0.5,

then Equation (15) obeys the Sieverts law, i.e. the hydrogen flux is proportional to

the square root of the pressure difference between retentate and permeate and

inversely proportional to the thickness of the permeable layer. The H2 flux

decreases as a function of the thickness of the membrane; therefore, thicker

membranes have lower permeability values compared to thinner ones. If n > 0.5,

then the series of mechanisms occurs during the hydrogen permeation, if n = 1,

i.e. mass transport through porous membrane controls from the series of steps, i.e. surface to desorption (Yun et al. 2011). Furthermore, for the value n = 0.5–1.0,

diffusion through a micro- or mesoporous membrane layer (due to holes or micro

defects), generally follows the Knudsen, activated or viscous flow regime. Thus,

the overall flux depends on the conjunction of the bulk permeation. (Yun et al. 2011)

The temperature dependency of hydrogen permeation follows the Arrhenius

kind of expression, i.e. Equation (15).

Pe,H2 = Pe0 exp (-Ea

RT), (15)

where Pe0 is the pre-exponential factor, Ea is the activation energy, R is the gas

constant, and T is the absolute temperature (K).

In the case of a dense Pd-Ag membrane the ideal perm-selectivity of H2 is

infinite and both the Sieverts' law and Arrhenius equation are followed. Thus, the

temperature dependence of the H2 permeability can be expressed by means of an

Arrhenius-like equation (Equation (15)), the hydrogen flux permeating through

the Pd–Ag membrane can be expressed by means of the Richardson equation

46

(Equation (16)), which combines the Sieverts' law and the Arrhenius equation.

The hydrogen flux can be evaluated to confirm the Richardson equation and no

changes on the permeation properties have been observed (Iulianelli et al. 2010).

JH2 = Pe0 exp (-

EaRT

)

δ(pretentate

0.5 − ppermeate0.5 ) (16)

The dense Pd-based membranes have a high separation factor (> 10 000) with

respect to other gases, i.e. (H2/other gas). The dense Pd-Ag membranes have a

few major drawbacks. If the thickness is above 25 µm, low permeability will

probably exist (Tong et al. 2004), since membrane thickness is inversely

proportional to the permeating flux. According to Adhikari et al. (2006), thickness

of the membrane is independent of permeability. The ideal selectivity and

permeability are the two trade-off factors of the membrane. Moreover, the cost of

Pd metal has been growing exponentially over the years. Thus, using a thick Pd

membrane commercially is not viable. In order to make Pd-based membranes for

industrial applications, reducing the Pd-thickness (few micrometers) or using non-

Pd metal membranes like Ni, Nb, Ti, and V (Ockwig & Nenoff 2007) should be

favored. Reduction of Pd thickness will affect the ideal selectivity or separation

factor. It is very challenging to prepare an optimal Pd-thickness to have high

permeability and reasonable ideal selectivity. Many authors have studied Pd

thicknesses less than 50 µm, and in the literature, a lot of discrepancies on the

factor ‘n’ and the Sievert’s law validity exist (Caravella et al. 2010, Yun et al. 2011). Some authors have also reported that the Pd layer with the thickness of 1

µm for H2 separation can be utilized (Seelam et al. 2013a). Moreover, thin

membranes which are prepared by conventional methods have a low thermal and

mechanical stability compared to those thin membranes prepared by

microfabrication methods (Gielens et al. 2006, Tong et al. 2004).

Water gas shift reaction has also been performed in a MR using different

membranes and catalysts. It has been studied extensively and a few literature

reviews are published (Mendes et al. 2010, Babita et al. 2011). In this work, a

simulated syn-gas feed with different molar feed ratios via WGS reaction was

tested in a Pd/PSS MR using a Fe-Cr based HT shift catalyst (Paper V).

47

3 Future trends in membrane-assisted reactors: a short summary

The main objectives of the thesis will be the basic and fundamental study

presented in Papers I-V (Section 1.2, Fig. 2). In this section, the two applied key

technologies in membrane-assisted reactors are shortly presented (Papers VI &

VII). This serves as the supplementary and background information related to the

thesis. These two papers provide knowledge on the novel technologies which are

potential and efficient methods to improve the overall process efficiency (e.g., in

high purity hydrogen production). One of the intensified technologies can be the

so-called membrane microreactor (MMR) (Papers VI). The other one is the

membrane-assisted reactor combining the catalyst and membrane phenomena into

one single unit, e.g. catalytic membrane-based reactor (CMR) (Paper VII). The

importance of catalyst materials for membrane-based reactor systems and

different methods to couple the catalyst function with membrane materials were

discussed as well in Paper VII. The two multi-functional reactors, i.e. MMRs and

CMRs, are shortly presented in this section. The two chapters explain new

methods and technologies that can be utilized in the future developments, for

example on-board hydrogen production in micro-reformers for micro-fuel cell

devices (e.g., portable electronics).

Intensification approaches (especially for MRs) are still at the research and

development stage. Many researcher groups are working in the field to improve

the overall performance of processes and also to reduce the environmental load

(Moulijn et al. 2006). All green technologies are part of process intensification

approaches, for example reducing waste and improving the atom economy. One

such intensified tool is the membrane-assisted reactor.

3.1 Microreactors and membrane microreactors: fabrication and applications

In Paper VI, a critical review and an insight into the state of the art on membrane

functionality integrated into microdevices are reported and discussed. Process and

reactor miniaturization is the fundamental concept in process intensification (PI).

Paper VI is restricted mainly to a microfabrication (MF) of zeolite and palladium

based micromembranes (MMs) and their applications in fine chemical synthesis

and hydrogen separation. In addition, the applicability of microfabricated Pd

MMs for hydrogen production was also discussed. The integration of MMs with

48

microstructured reactors can improve the overall system efficiency and reduce the

material costs. MMRs are multifunctional devices with a novel approach in

reaction and membrane engineering (e.g. Moulijn et al. 2006).

There are three main functions involved in MMRs, i.e. micromembrane with

selective separation, fast heat and mass transfer rates, and miniaturized channels

for the reaction with reactant or process flows. The synergistic effect of the

membrane with microdevices has not yet been explored in many fields. This is an

important research field, e.g. catalyzed reactions, where a chemical reaction

occurs on the catalytic wall of microchannels in a microreactor (Seelam et al. 2013a). Further, a selective separation of the desired component by

micromembrane layers integrated into micro-units can provide compact and

efficient systems for example for energy applications. Thus, reaction, separation

and purification of the desired component can be achieved in a micro-level. One

of these kinds of applications was reported by Wilhite et al. (2006) for hydrogen

production in a Pd-membrane microreactor. In Karnik et al. (2003), a compact

microreactor device with integrated Pd MM was fabricated and designed for a

micro-fuel cell processor, including SR and WGS reaction channels. Jong et al. (2006) reported the approaches to integrate micromembranes into microdevices

such as directly integrating commercial membranes into micro-devices, in-situ

preparation of membrane layers into microchannels and membranes as a part of

chip fabrication, etc. Moreover, when selecting the membrane materials for

microdevices, chemical, mechanical, and thermal stability, as well as

compatibility issues are critical issues to be taken into consideration in designing

and manufacturing of MMR units.

In the near future, new membranes and catalytic materials will be coupled

into integrated microstructured reactors and will be tested in many applications.

The effect of miniaturization, permeation properties of membranes, and

optimization will be investigated thoroughly, including catalytic studies.

Microfabrication technologies already exist and the technology is matured in

microelectromechanical systems (MEMS) and semiconductor industries (Jensen

et al. 2001, Mills et al. 2007, McMullen & Jensen 2010). Therefore, MF will be

important for the development of chemical microreactors and microprocess

engineering. Due to the commercial experience, MF techniques can be beneficial

to apply in the membrane technology field, in order to improve the material and

system efficiency. Operating conditions like temperature and pressure are also

taken into consideration.

49

3.2 Advances in catalysts for membrane reactors

A critical review on the catalyst integration into membrane-based reactors is

reported in Paper VII. A catalyst can be introduced into a reactor in different

forms, i.e. having various sizes and shapes. In the case of membrane reactors,

catalytic reaction and separation occur simultaneously in a single unit. In catalytic

membranes, the interactions between the catalyst and the membrane are coupled

and in the literature this topic is not well addressed. (Huuhtanen et al. 2013) In

the packed-bed membrane reactor (PBMR), the membrane can have three main

functions, i.e. it can be used as an extractor-type (removing desired product

selectively via membrane), a distributor type (dosing reactants through membrane

selectively, controlling the reaction), or as a contactor type (reactants are

contacted through membrane or flow through the membrane) function (Miachon

& Dalmon 2004). In PBMR, catalyst does not have any interactions with the

membrane layer (e.g., catalyst pellets are filled in the tubular membrane module).

The features of a catalyst and a membrane are combined to form a catalytic

membrane layer, i.e. a catalytic membrane reactor (CMR) where both the reaction

and separation happens on the same layer. This reduces the need of use of

materials and process units. One of its kind is reported by Li et al. (2011), i.e. a

bimodal catalytic membrane for H2 production. In CMR, the catalyst is

encapsulated into porous membrane material to provide reaction and the desired

product is permeated through the membrane.

In an extractor-type MR, the catalyst is packed in annulus or shell side and

the catalyst enhances the reaction rate and the membrane layer allows selective

permeation. For example, a tubular membrane is filled with the catalyst pellets

and inserted into a reactor module. Figure 8 represents different concepts used in

the membrane-assisted catalytic reactors, i.e. PBMR or MR and CMR. Paper VII

presents only the hydrogen production reactions, like SR and WSG in membrane

assisted reactors. In many studies (Shu et al. 1991, Babita et al. 2011, Basile et al. 2008b, Iulianelli et al. 2011, Seelam et al. 2012), commercial catalysts were used

in the reactions performed in PBMR.

50

Fig. 8. Different concepts of integrating a catalyst and a membrane in an extractor-

type membrane-assisted reactor (A) PBMR = packed-bed membrane reactor or IMRCF

= inert membrane reactor with a catalyst packed in the feed side or IMR = inert

membrane reactor or MR = membrane reactor and (B) CMR = catalytic membrane

reactor, i.e. a catalyst coupling with a membrane (Mcleary et al. 2006 & Dittmeyer et al.

2001).

The hydrogen spill over effect is a crucial phenomenon in enhancing the

hydrogen permeation through the Pd-based membrane in PBMR. Moreover, the

catalyst bed pattern and the position are also important parameters to enhance the

hydrogen permeation in the packed catalyst bed in MR systems. (Rei et al. 2011)

The MR performance is affected by the activity, selectivity and stability of the

catalyst materials. Designing robust catalytic materials for the PBMR and CMR

are the important research areas in catalysis in membrane-assisted reactors. Most

importantly in MR, the catalyst function is critical to deactivation phenomena

which affects the membrane permeation. The carbon formation over the catalyst

during the reaction might adsorb over the membrane surface due to the operating

pressures. Henceforth, the catalytic material should be highly stable and coke

resistant in PBMR and CMR. New and novel nanostructured catalytic materials

like CNTs, carbon black and anodic alumina oxide are promising materials for the

catalytic membrane based systems (Job et al. 2009, Stair et al. 2006, Li et al. 2011).

51

4 Materials and their preparation

4.1 CNT supported catalysts

The metal decorated carbon nanotube materials (MWCNTs: L = 0.1–10 µm, O.D.

= 10–15 nm, I.D. = 2–6 nm) were purified, synthesized, characterized, and tested

in steam reforming of ethanol (SRE). In the membrane reformer, only commercial

catalysts were tested. In the conventional reformer, the reference catalysts were

also tested and compared with the CNT-based catalysts. All other carbon

supported (activated carbon (AC) and graphitic carbon black (GCB)) catalysts

were pre-treated and synthesized in a similar procedure as the CNT-based

catalysts. A reference catalyst Ni16.6/Al2O3 (HTC400, Crosfield extrudates) was

also tested in the SRE reaction (Paper I and II).

4.1.1 Catalyst preparation

Generally, CNTs are inert and hydrophobic and they also contain traces of

impurities like Fe, Al and amorphous carbon. First, carbon nanotubes (Sigma-

Aldrich) were pre-treated and functionalized with a liquid phase oxidation method

in order to remove impurities and to improve the hydrophilicity, i.e. interactions

between metal/CNT-support, by introducing carboxyl groups (Azadi et al. 2010,

Motchelaho 2011). Figure 9 represents the steps to prepare metal decorated CNTs.

Fig. 9. Preparation method for the CNT-based catalysts (Papers I-II).

The CNT support material was functionalized using the 70% HNO3 acid

treatment for 9 h (sometimes, the time varies based on the sample) under reflux

conditions to prepare carboxylated CNTs, denoted as CNT-COOH. After that,

carboxylated CNTs were washed with distilled water for several times and dried

for overnight at 100 °C. The metal decorated CNTs were prepared using a wet

52

impregnation method. Pre-treated MWCNTs and the desired amount of metal

precursors (e.g., Ni(acac)2) were sonicated for 3 h in 500 cm3 of toluene and

followed by stirring at RT. The solvent was evaporated under N2 flow and metal

decorated CNTs were dried overnight at 100 °C. The metal/metal oxide decorated

CNTs were subjected to step calcination under N2 and reduction under a H2 flow.

4.1.2 Catalyst characterization

Characterization of catalyst materials was done by using different analytical

methods as it is important to understand the physical, chemical and catalytic

properties of the materials. The CNT-based catalysts were characterized by X-ray

powder diffraction (XRD), transmission electron microscopy (TEM) with energy-

dispersive X-ray (EDX) and N2 adsorption method (Brunauer-Teller-Emmett

(BET) and Barrett-Joyner-Halenda (BJH)).

X-ray powder diffraction

The X-ray diffraction (XRD) technique was used to identify the phase

composition and volume averaged crystallite size of the powder form of CNT-

based catalysts. The XRD patterns were recorded using Siemens D5000 X-ray

diffractometer in the conventional Bragg-Brentano focusing geometry with fixed

detector slits and without a monochromator. A Ni-foil was used as a filter for

CuKα radiation from X-ray tube (Microelectronic Laboratory, University of

Oulu). Step scan diffraction data was collected over a diffraction angle range of

2θ = 20°-60°. Received data was analyzed by using Origin Pro software.

Broadening of the peaks in the diffraction patterns were evaluated by fitting

Lorentzian curves. The mean crystallite size was calculated using the Scherrer

equation.

Energy-filtered transmission electron microscopy

The particle sizes of the catalyst decorated CNTs were characterized by energy-

filtered TEM (EFTEM LEO 912 OMEGA, acceleration voltage 120 kV, LaB6

filament). Samples were prepared on sandwich TEM grid. More than 100

particles were measured from micrographs and the average particle size and

particle size frequencies were collected in histograms. Elemental concentrations

of the samples were determined with energy-dispersive X-ray spectroscopy (EDS,

53

installed on a FESEM, Zeiss Ultra plus facility) from six different sample

locations (Microelectronic Laboratory, University of Oulu).

N2 adsorption for textural properties

Adsorption-desorption isotherms were used to determine the pore size distribution

and average pore diameter. The Brunauer-Teller-Emmett (BET) method was used

to determine the specific surface area (SBET) of the catalyst samples. The net pore

volume and pore size distribution were calculated based on the Barrett-Joyner-

Halenda (BJH) method. The BET and BJH were measured by using N2 adsorption

isotherms at -196 °C (by ASAP 2020, Micromeritics). First, the catalyst sample

was heated with a heating rate of 10 °C min-1 in vacuum and followed by

evacuation at 90 °C for 30 min. After that, the catalyst samples were heated 10 °C

min-1 to 350 °C in vacuum and then evacuated at 350 °C for 3 h and finally

cooled down to room temperature (RT). The evacuation was performed for 30

min at RT to remove excess of gaseous N2. Based on the linear plot of N2-

adsorption isotherms, the SBET (m2 g-1) was determined. The net pore volume was

measured as the adsorbed amount of N2 at relative pressure p/p0 =0.99. Pore sizes

were determined by using the BJH method on the standard isotherm method and

cylindrical pore geometry. The physisorption measurements were performed in

the Mass and Heat Transfer Process Laboratory at the Department of Process and

Environmental Engineering, University of Oulu.

4.2 Catalysts and membrane materials in a membrane reformer

In MR studies, commercial catalyst materials in pellet shapes were used. The

palladium-based membranes were characterized and tested at the Institute for

Membrane Technology of the Italian National Research Council (ITM-CNR), c/o

University of Calabria, Italy. In this work, two different kinds of membrane

materials, i.e. dense Pd-Ag self-supported and composite Pd/PSS tubular

membranes, were utilized for the SR and WGS reactions.

4.2.1 Catalyst materials in a membrane reformer

The list of catalysts’ information used in different reactions tested in MR is

summarized in Table 6.

54

Table 6. List of catalyst materials used in MR studies.

Reaction Catalyst Metal loadings

(wt.%)

Manufacturer Membrane

BESR Ni/ZrO2 25 Catacol Pd/PSS

BESR Co/Al2O3 15 Johnson Matthey Pd/PSS

BGSR Ru/Al2O3 0.5 Johnson Matthey Pd-Ag

WGS Fe-Cr n.a. Johnson Matthey Pd/PSS

n.a. not applicable

4.2.2 Pd-Ag self-supported membrane

A pin hole- and defect-free tubular dense Pd-Ag (75 wt% Pd & 25 wt% Ag)

membrane was produced by a cold-rolling and diffusion welding (CDW) method

with a 50 µm thick layer (Pd layer cross section area of ~46 cm2). The CDW

method was modified and developed by Dr. Tosti’s group at ENEA Frascati, Italy

(Tosti et al. 2004 & 2008).

4.2.3 Pd/PSS supported composite membrane

The porous stainless steel (PSS) support was manufactured by Mott Metallurgical

Corporation (AISI 316L porous tube). The PSS has dimensions of 10 mm O.D.

and 0.1 µm nominal pore size with 95% rejection of particle sizes greater than 0.1

µm. The mean and maximum values of the pore size are larger than 2 and 5 µm,

which was determined by a mercury intrusion measurement (Mardilovich et al. 2002). The PSS support was welded with two non-porous AISI 316L stainless

steel tubes with one end of the tube being closed. The tubular membrane was

placed inside the membrane housing. The total length of the tube was 21 cm; Pd

active length was 7.7 cm (active surface area 23.2 cm2). Palladium was deposited

on the composite PSS support by an electroless plating (ELP) deposition

technique producing 20 µm Pd thickness layers over the supports, which were

manufactured at RSE laboratories, Milan, Italy. The procedure for the ELP

method is reported in the literature (Pinacci et al. 2012, Basile et al. 2011).

55

5 Experimental methods and procedures

In activity measurements, two experimental procedures and set-ups were utilized

in this work: First, the conventional reformer (Figure 10), where catalytic

materials were tested and screened for the activity in the SRE reaction; second,

the membrane reactor set-up used for the SR and WGS reactions. The detailed

experimental procedures and reactor set-ups are discussed in the following

sections.

5.1 Conventional reformer

5.1.1 Reactor set-up

Figure 10 shows the experimental set-up for the activity tests performed in a

quartz reactor (conventional reformer is also termed as a fixed-bed reactor (FBR)

or a traditional reactor (TR)). In this reactor set-up, three main devices were

utilized in the activity measurements, i.e. FTIR (Gasmet CR-4000/2000/1000)

spectroscopy for product gas analysis, H2 analyzer (XMTC) for the outlet H2

concentration, and the oxygen analyzer (ABB Optima).

Fig. 10. Reactor setup for the reforming experiments.

The FTIR spectrometer was used in catalytic activity studies in order to quantify

and identify the compounds from the gas flow during the reaction tests. The FTIR

unit contains an IR source, detector, interferometer, and optical mirrors. The main

working principal in IR spectroscopy is based on the measurements of vibration

of molecules. When the IR radiation is passed through a gaseous sample, the

56

molecules absorb certain wavelengths. The molecules start to vibrate or rotate

when they absorb energy. Different molecules have different vibrational or

rotational wavelengths of IR radiation. The vibration modes can have

symmetrical or asymmetrical stretching and bending as well as rotational modes,

and also all of them. The calibration of the peristaltic pump was performed before

each set of reactions.

5.1.2 Catalyst activity tests

Typically, 0.1 g of a catalyst in a powder form was packed between glass wool

layers, i.e. as zebra packing in the inner quartz reactor (I.D. 5 mm). The reactants

ethanol (96% v/v VRW CAS: 64-17-5) and water were fed using a peristaltic

pump and 0.09 cm3 min-1 feed flow rate along with the N2 carrier gas (total flow

of 600 cm3 min-1). The ethanol-water mixture was vaporized in the outer shell of

the reactor tube before passing through the catalyst bed. The gaseous reactant

mixture was fed along with the carrier gas N2 in the ratio of 1:4 and kept constant

throughout the experiments. The calibration of the Fourier Transform Infrared

Spectroscopy (FTIR, GasmetTM) analyzer was performed to analyze the unreacted

ethanol-water mixture and product gases CO, CO2, CH4, CH3CHO and C2H4

using the Calcmet program. The composition (in vol.%) of product gases and

unreacted ethanol were analyzed by the spectra and hydrogen was analyzed by a

thermal conductivity gas analyzer (XMTC) (dry gas flow).

First, the catalyst was pre-treated and reduced under 20 vol.% H2 in a N2 flow

at 350 °C for 30 min with a heating rate of 15 °C min-1. After that, the catalyst

was cooled down to room temperature (RT) under reducing atmosphere. For the

reforming test, the heating rate of 10 °C min-1 was set and the reaction was carried

out at the 150 °C − 450 °C range under atmospheric pressure. The reactant

mixture, i.e. ethanol-to-water molar ratio, was 1:3 (stoichiometric) and the GHSV

~38 000 h-1 was kept constant in all the experiments. The activity of each catalyst

was analyzed based on the data collected from the FTIR and H2 analyzer. The

performance of the catalysts in the SRE reaction was evaluated based on

Equations (17) - (19) considering the inlet and outlet flow rates.

Conversion of ethanol %) = Fethanol, in Fethanol, out

Fethanol, in (17)

Yield of hydrogen (YH2) = FH2, out

6Fethanol, in (18)

57

Selectivity of products S'p% =

Fi'

∑Fi' ×100 (19)

where Fethanol, in, and Fethanol, out are the feed flow rates of ethanol at inlet and outlet

(cm3 min-1); FH2, out is the hydrogen flow rate at outlet concentration (cm3 min-1),

and F`i is the flow rate of product ‘i’ (i = H2, CO, CO2, CH4, CH3CHO and C2H4).

5.2 Membrane reformer

5.2.1 Membrane reactor set-up

The membrane reactor set-up is shown in Figure 11. The MR set-up consists of

mass flow and temperature controllers, and an MR module housing containing a

tubular Pd-based membrane. The MR module is made up of steel and the

dimensions are length 280 mm and I.D. 20 mm. The direction of the hydrogen

permeation in the MR is shown with the arrows in Fig. 11. The reaction pressure

is controlled by a back-pressure control valve at the retentate side. There is a

slight variation based on the type of configuration used. The catalyst and glass

spheres are packed inside the annulus (or shell) side of the module. The MR set-

up includes a HPLC pump (Dionex) to feed the reactant mixture, i.e. ethanol or

glycerol and water, with a set molar ratio and flow rate. Before feeding, ethanol

and water are vaporized in the pre-heating zone at 400 °C and then the vaporized

stream was fed into the reaction side. The MR module with a tubular membrane

was heated up to the reaction temperature of ~400 °C under a N2 flow with a very

slow heating rate (1 °C min-1). The thermocouple is inserted into the shell side of

the module.

In bio-glycerol steam reforming (BGSR) tests, N2 was used as the sweep gas

in the permeate side and maintained at the constant pressure of 1 bar (abs.). No

sweep gas was utilized in bio-ethanol steam reforming (BESR) and water-gas-

shift (WGS) reactions which were performed in the Pd/PSS MR. The reaction

products from the retentate side were fed through a cold trap in order to condense

the unreacted water and glycerol (or ethanol). The retentate and permeate streams

were analyzed using a gas chromatograph (GC, HP6890) which was temperature

programmed with two thermal conductivity detectors (TCD) upto 250 °C. The

GC had three packed columns Carboxen 1000 (15 ft 1/8 in.) and Porapack

(R50/80 with 8 ft 1/8 in.) connected in series, and a molecular Sieve 5 Å (6 ft 1/8

in.). The tubular MR was packed with a commercial catalyst (3 g) in the annulus

58

side for BGSR and the shell side for BESR reaction. In this work, similar terms

were used, i.e. the membrane reformer is also termed as a membrane reactor

(MR) or a packed bed membrane reactor (PBMR) or an inert membrane with a

catalyst packed in the annulus.

Fig. 11. Experimental set-up for the membrane reactor system. (III published by

permission of Elsevier).

Two types of MR configurations were used, i.e. the first set-up (Figure 12A)

which was utilized for the BGSR and the second set-up (Figure 12B) for the

BESR reaction. A dense membrane having a 50 µm thick Pd-Ag self-supported

layer with a 14 cm active length (Paper IV) for the BGSR and a 20 µm Pd

59

supported on porous stainless steel composite membrane layer with 7.4 cm active

length for the BESR (Paper III).

Fig. 12. Two different MR configurations used in the experiments carried out for A)

BGSR and B) BESR reactions. (III, IV published by permission of Elsevier).

5.2.2 Permeation and performance tests

The permeation test is important before the reaction in order to verify the defects

in the Pd membrane and quantify the H2 flux and its perm-selectivity with respect

to other gases (e.g., N2). The permeation tests were also performed at each stage

of the experimental test, i.e. before and after the reaction test, and also after the

regeneration (with H2) process to verify the hydrogen flux behavior. First, the

permeation test was conducted with a pure H2 flow at reaction temperature

400 °C with the transmembrane pressure difference depending on the variation of

the retentate pressure and keeping permeate pressure constant at 1 bar. The

permeation of hydrogen through a Pd-based membrane is presented in Equation

(14).

60

The reaction test in MR was performed using 3 g of a commercial catalyst

with inert glass beads (2 mm diameter). For bio-ethanol steam reforming (BESR),

Ni/ZrO2 (Catacol) and Co/Al2O3 (Johnson Matthey), for bio-glycerol steam

reforming (BGSR) Ru/Al2O3 (Johnson Matthey), and for WGS Fe-Cr (Johnson

Matthey) catalytic materials (in a pellet shape) were used. Before the reaction, the

catalyst was reduced under a H2 flow for 2 h at 400 °C under atmospheric

pressure. The permeation test was conducted prior to the reaction test. The

reaction tests were performed in MR and, the flow rates at retentate and permeate

sides were measured by using a soap-bubble flow meter. One side of the tubular

membrane was closed and the reformed gases in the retentate and permeate sides

were analyzed by GC using two detectors (front and back). The experimental

results were evaluated until they reached steady state conditions during 90–125

min of testing. Each experimental point is an average value of 9–10 experiments

until the steady state was reached. The complete experimental test lasted around

150 min and after that the catalyst was subjected to regeneration with a pure H2

flow (18·10-4 mol min-1).

The performance of the reaction tests was evaluated based on the following

equations for each set of reaction.

The performance equations for the BESR reaction in MR

The conversion of bio-ethanol, i.e. simulated ethanol mixture with impurities such

as glycerol and acetic acid was calculated using Equation (20).

Conversion of bio-ethanol %) = FCO2, p+r

+ FCH4, p+r + FCOp+r

2F'ethanol, in +3F'

glycerol, in +2F'acetic acid, in

×100 (20)

where FCO2,p+r + FCH4,p+r + FCO,p+r denotes the sum of the molar flow rates (mol

min-1) at permeate and retentate sides (p+r) of CO2, CH4 and CO, F´ethanol,in,

F´glycerol,in and F´acetic acid,in represent the inlet molar flow rates of ethanol, glycerol

and acetic acid, respectively.

The hydrogen recovery factor (HRF) defined as the amount of hydrogen

recovered at the permeate side to the total hydrogen produced (permeate +

retentate sides) can be calculated based on Equation (21).

HRF %) = FH2, p

FH2, p+FH2, r

×100 (21)

where FH2, p and FH2, r are the hydrogen molar flow rates (mol min-1) at permeate

and retentate sides, respectively.

61

The hydrogen permeate purity (HPP) level can be calculated from Equation

(22) and the total hydrogen yield from Equation (23), respectively.

Hydrogen permeates purity (HPP%) = FH2, p

FH2, p+ FCO2, p+ FCOp+ FCH4, p

×100 (22)

Hydrogen yield YH2% = FH2, p+r

6F'ethanol, in+7F'glycerol, in+4F'acetic acid, in ×100 (23)

The selectivity of the products is denoted by the Equation (24).

Selectivity S'i%) = Fi-total

FH2- total

+ FCO2- total

+ FCO total+ FCH

4-total

×100 (24)

where Fi- total is the total molar flow rate of the compound ‘i’ (sum of retentate +

permeate flow rates) (mol min-1).

The performance equations for the BGSR reaction in MR

The conversion of bio-glycerol, yield of H2 and the weight hourly space velocity

were calculated by Equations (25), (26) and (27), respectively.

Conversion of bio-glycerol %) = FCOr+FCH4, r+FCO2,r

3Fglycerol, in ×100 (25)

Hydrogen yield YH2% = FH2, r+p

7F''glycerol, in ×100 (26)

Weight hourly space velocity WHSV) = M'feed

Mcat (27)

where Mcat is the mass of the catalyst pellet (g), M´feed is the mass flow rate of the

feed (g h-1).

The performance equations for the WGS reaction in MR

The ideal perm-selectivity or separation factor (Equation (28)) is defined as the

ratio of permeabilities (or permeance) of permeated gas molecules.

αH2/gas_i = PeH2

Pegas_i (28)

where gas_i is N2 or He, αH2/gas_i is the ideal selectity or perm-selectivity of H2

with respect to other gases, i.e. N2 or He and PeH2 is the permeance of H2 (mol m-2

s-1 Pa-1).

62

The conversion of CO is calculated based on Equation (29).

CO conversion %) = FCOin

- FCOout

FCOin

×100 (29)

where FCO,in and FCO,out are the inlet and outlet molar flow rates of carbon

monoxide (mol min-1). The gas hourly space velocity (GHSV) is expressed based

on Equation (30).

Gas hourly space velocity GHSV) = V´feed

Vcat (30)

where Vcat is the volume of the catalyst pellet (cm3) and V´feed is the volumetric

feed flow rate in the gas at STP (cm3 h-1).

The volume of the catalyst pellet in MR was calculated by Equation (31)

based on the bulk density of the catalyst (given by the manufacturer) and the mass

of the catalyst (g). The volume of the catalyst pellet (Vcat) is defined as:

Vcat = Mcat

ρb (31)

where ρb is the bulk density (g cm-3) and Mcat is the mass of the catalyst pellet (g).

63

6 Results and discussion

The results are discussed in two different parts. First, the results obtained from the

catalytic studies in the conventional reformer and second, the results from the

membrane reformer studies. A reference catalyst Ni/Al2O3 (16.6 wt.% Ni) was

also tested and compared with the novel CNT-based catalysts in a conventional

reformer.

6.1 Bio-ethanol steam reforming in a conventional reformer

6.1.1 Catalyst characterization

The TEM micrographs of three different carbon-supported fresh catalysts, i.e. Ni2/CNT, Ni2/GCB, and Ni2/AC catalysts (GCB = graphitic carbon black and AC

= activated carbon), are shown in Figure 13. The nature of the catalyst support

influences the uniformity of the particles and the active particle size (Zgolicz et al. 2012). The Ni nano-particles (NPs) are finely dispersed and decorated over the

surface of the CNTs (Fig. 13a). The average Ni size for Ni2/CNT is 2 nm which

was lower than that for Ni2/GCB (3.3 nm) and Ni2/AC (4.9 nm). Over the

Ni2/CNT catalyst, it was found that a narrow distribution of particle sizes was

achieved. The Ni NPs over the GCB (Fig. 13b) and AC (Fig. 13c) supports

displayed a broader distribution with bigger particles compared to Ni2/CNT. A

good metal dispersion was achieved over the CNT surface, as shown in the TEM

images and similar phenomena were reported by Li et al. (2007) and López et al. (2012).

64

Fig. 13. TEM images and Ni nano-particle size distributions on different carbon-

supported materials (a) Ni2/CNT, (b) Ni2/GCB, and (c) Ni2/AC catalysts. Note: CNT =

carbon nanotube, GCB = graphitic carbon black and AC = activated carbon.

The metal decorated CNT-supported catalysts (fresh) with different metal types

were characterized by TEM and XRD techniques. The TEM micrographs of metal

decorated CNTs are presented in Figure 14.

c.

65

Fig. 14. TEM images and corresponding X-ray diffraction patterns of (a) Ni/CNT, (b)

Co/CNT, (c) Pt/CNT and (d) Rh/CNT catalysts (before activation in H2 flow). The length

scale bar is 50 nm in each micrograph (I, published by permission of Elsevier).

Well dispersed NPs over the CNT surfaces were displayed (Figure 14 a-d).

Moreover, the metal NPs are highly dispersed with a narrow active metal

distribution. For NiO phases, 2θ = 37° and 43° is attributed to the planes (111)

(200), respectively and Co3O4 at (220) and (311) corresponds to 2θ = 32° and 38°,

respectively. In Fig. 14, the NPs decorated on the nanotube were identified as the

metallic phase (Pt0 and Rh0) and oxide phases (NiO and Co3O4) for respective

catalyst samples before the reduction step (Paper I). The order of reducibility of

the metal oxides on CNTs is as follows: CoOx/CNT > NiOx/CNT > PtOx/CNT

(Leino et al. 2013). Especially, in the case of Co/CNT, the metal/metal oxide

phase (Co/Co3O4) was detected and it is difficult to reduce the catalyst completely

at temperatures below 350 °C.

66

The average particle size of NPs measured from the broadening of the reflections

in XRD patterns (<dXRD> calculated by the Scherrer equation) give higher values

than those calculated from the TEM micrographs (<dTEM>) (Table 7). This can be

observed particularly for the Co/CNT and Pt/CNT catalyst samples which are

polydispersed and some large particles were observed in those catalysts by TEM.

Both Ni/CNT and Rh/CNT catalysts were found to be monodisperse. However, in

the case of Rh/CNT, the NPs are partially agglomerated to larger Rh clusters with

the size of ~20–40 nm (Fig. 14d, the biggest clusters). The NPs with various sizes

between 1 to 10 nm were detected and the average particle size (dTEM) of around 5

nm was measured by TEM (Table 7). The agglomerates of Rh metal particles

seem to be sintered to larger particles which are partially clustered to the size of

10–40 nm (Paper I).

Table 7. Average particle sizes of metal NPs over CNT supported catalysts measured

by TEM and XRD (I, published by permission of Elsevier).

Catalyst dTEM (nm) dXRD (nm) Remarks

Ni/CNT 4.8 5.0 monodisperse

Co/CNT 6.0 11.5 polydisperse

Pt/CNT 4.0 19.2 polydisperse

Rh/CNT 5.8 9.1 partially clustered, monodisperse

The XRD patterns and TEM micrographs of bimetallic Ni10Ptx (x = 0, 1, 1.5 and 2

wt.%) on CNT-supported catalysts are presented in Figure 15 a-d. The Ni NP

sizes range from 3.5 to 4.8 nm and the average particle size is around 4.2 nm.

From Fig. 15, it is difficult to distinguish between the Ni and Pt NPs, as the Ni

loading (10 wt.%) is much higher than the Pt loading (0–2 wt.%). Thus, most of

the NPs dispersed and detected over the CNT surface are most probably Ni.

Moreover, a few NPs (< 2 nm) might be present at the interior of the carbon

tubes, as the I.D. is around 2–6 nm.

67

Fig. 15. X-ray diffraction patterns and TEM images of (a) Ni10/CNT, (b) Ni10Pt1/CNT, (c)

Ni10Pt1.5/CNT and (d) Ni10Pt2/CNT catalysts (II, manuscript).

It was shown from the TEM images that there is no significant destruction in the

nanotubes structure due to pre-treatment and synthesis steps. In Ni10Ptx/CNT (x =

1–2 wt.%) catalysts, the Pt (111) metallic phase was observed with high intensity

at 2θ = 40° (Fig. 15), except for Ni10Pt1/CNT due to the small amount of Pt.

Moreover, as the Pt loading increases, the catalyst samples were more non-

uniform, as show in the TEM images. The Ni (200) peaks with the intensity at 37o

and 43o are assigned to the lattice phase of NiO.

Figure 16 represents the XRD pattern of the ZnO (101) peak at 2θ = 36°,

minor ZnO (100), (002) and (102) peaks were also found. The NiO (200) peak at

2θ = 43° and NiO (111) (phases) are shown. The oxide phases NiO and ZnO were

68

confirmed by the XRD patterns with narrow distributions (before reduction). The

ZnO is well dispersed in the bulk of the nanotube surface.

Fig. 16. XRD pattern and TEM micrograph of the (ZnO10)Ni10/CNT catalyst (II,

manuscript).

The bigger particles in Fig. 16 were found to be ZnO particles. Moreover, the

ZnO addition was incorporated after the Ni impregnation. Thus, ZnO forms

clustered particles which are bigger than the Ni counterparts and found had no

interactions with Ni.

The average metal particle size measured by TEM and XRD is summarized

in Table 8. The average particle size calculated by XRD (<dXRD>) gives slightly

higher values due to broadening of the reflections compared to TEM

measurements (<dTEM>). The elemental composition of CNT-based catalysts from

EDX analyses are presented in Table 8. The carbon content in metal decorated

CNTs is reduced during the synthesis steps and is found to be lower than the

carbon amount in the CNT-COOH sample (i.e. 92 wt.% C). This has an influence

on the metal content values presented based on the EDX analysis (Table 8). Even

a small amount of trace elements detected in the catalyst samples can influence

the catalyst activity and these elements are existing in the pristine samples (Al

and Fe were used as catalyst materials for CNT production) (Paper II).

69

Table 8. The metal content measured by EDX and the average nano-particle size of

fresh CNT-supported catalysts measured by TEM and XRD (II, manuscript).

Catalyst Metal contents by EDX

(wt.%)*

dTEM

<nm>

dXRD

<nm>

Ni10/CNT Ni 10.7±9.3; C 79.9±8.4; O 7.5±2.6 3.5±2.6 NiO(200) 5.9±0.1

Ni10Pt1/CNT Ni 12.6±4.0; C 77.7±3.6; O 6.1±1.2;

Pt 1.4±0.2

3.9 ± 3.3 NiO(200) 5.2±0.2

Ni10Pt1.5/CNT Ni 10.9± 5.1; C 77.6±5.1; O 7.5±3.1;

Pt 2.2±1.3

4.1 ± 2.9 NiO(200) 6.3±0.1;

Pt(111) 6.5±0.7

Ni10Pt2/CNT Ni 22.9±10.9; C 63.6±12.1; O 5.3±2.0;

Pt 4.6±2.9

4.8 ± 3.7 NiO(200) 7.1±0.1;

Pt(111) 10.5±0.8

(ZnO10)Ni10/CNT Ni 6.9±1.9; C 65.4±11.8; O 20.4±10.6;

Zn 5.1±3.2

n.d. NiO(200) 6.4 ±0.3 ;

ZnO(101) 28.2±1.6

n.d. not determined, *Ni 10 wt.% nominal loading and traces of Al ad Fe are found

The textural properties of the prepared CNT supported catalysts are presented in

Table 9. The pristine multi-walled CNTs have a very low surface area compared

to the treated CNTs. During CNTs purification and liquid oxidation steps, the

creation of defects (e.g., vacancies or dislocations) and scavenging the impurities

were done, resulting in an increase in the surface area from 75 to 218 m2 g-1 and

the pore volume from 0.22 to 0.75 cm3g-1 for CNT-COOH (carboxylated-CNTs).

The specific surface area (SBET) of CNT-based catalysts varied between 209 to

269 m2 g-1. A similar SBET increment phenomenon was observed by Azadi et al. (2010) and Motchelaho (2011). Moreover, from the EDX analyses, few traces of

elements are detected, which might have partially blocked the tube openings in

the pristine sample. The SBET of the functionalized CNT support increased from

218 to 269 m2 g-1 after incorporation of Ni. On the other hand, the net pore

volume remained unchanged in the after-treatment steps. Commercial

Ni16.6/Al2O3 catalyst pellets were crushed and sieved to a powder form (< 250 µm

particle size).

70

Table 9. Textural properties of catalysts in SRE reaction.

Catalyst SBET (m2 g-1) Pore volume (cm3 g-1) Pore size (nm)

MWCNT-Pristine# 75 0.22 11.4

CNT-COOH 218 0.75 13.6

Ni2/CNT 284 0.77 10.2

Ni2/GCB 12 0.02 7.9

Ni2/AC 897 0.50 2.3

Ni10/CNT 269 0.66 9.4

Ni10Pt1/CNT 258 0.72 10.9

Ni10Pt1.5/CNT 251 0.70 10.8

Ni10Pt2/CNT 251 0.75 11.6

(ZnO10)Ni10/CNT 209 0.56 10.4

Ni10(ZnO10)/CNT 241 0.63 10.1

Ni16.6/Al2O3* 112 0.33 11.8

Subscript indicates the nominal metal loadings, n.d. not determined, # untreated CNTs,

*reference catalyst in TR

The functionalized and metal modified CNTs exhibit larger pore sizes due to the

acidic treatment. The carboxylated CNTs have a slightly larger pore size of 13.6

nm compared to the pristine (11.4 nm) and metal modified CNTs (9.4–11.6 nm).

The incorporation of Pt to Ni10/CNT does not induce any significant effect on the

textural properties in comparison with Ni10/CNT. For Ni10(ZnO10) and

(ZnO10)Ni10/CNT based catalysts there exhibits a larger difference in SBET and

pore characteristics. The SBET of Ni10/CNT was diminished by the addition of

ZnO from 269 to 209 m2 g-1. As presented in Figure 9, for the catalyst

(ZnO10)Ni10/CNT, the Ni precursor is impregnated prior to the Zn addition. In

(ZnO10)Ni10/CNT’s TEM image (Fig. 16), the ZnO clusters might be covering the

Ni nano particles and also the defects on the exterior of the tube (Paper II). Thus,

the (ZnO10)Ni10/CNT catalyst exhibits lower values of SBET and pore volume

compared to Ni10/CNT and Ni10(ZnO10)/CNT (Paper II).

6.1.2 Monometallic carbon-based catalysts

Influence of carbon support in Ni-based catalysts on SRE

Three different carbon-type supports with 2 wt.% Ni as the nominal loading were

prepared, i.e. Ni2/CNT, Ni2/GCB and Ni2/AC catalysts, and tested in the SRE

reaction. Figures 17 a-b clearly show that the ethanol conversion and hydrogen

71

production is high for Ni2/CNT in comparison with the other carbon supported

Ni-based catalysts. The normalized plots of ethanol converted and hydrogen

produced per catalyst surface area of Ni-supported carbon materials was

presented in Figures 17 c and 17 d.

Fig. 17. Effect of the type of carbon supported Ni-catalyst in SRE on (a) ethanol

conversion and (b) hydrogen production as a function of reaction temperature.

Normalised plots (c) mole of ethanol converted and (d) mole of hydrogen produced

per total surface area of the catalysts as a function of reaction temperature.

The pristine and functionalized CNTs were tested in the reaction and they

exhibited very low activity. The Ni2/CNT catalyst exhibits the highest activity and

selectivity in the SRE reaction. Due to the differences in the surface area values

of the catalysts, normalized plots for the activity in ethanol conversion and

hydrogen production per surface area was presented in Figure 17c and 17d. In

Ni2/CNT, active nanoparticles are highly dispersed over the surface of CNT, and

exhibit smaller particle sizes (2 nm) which are active in the reaction (Fig. 13).

Many authors have reported that the particle size has a great influence over the

activity. Moreover, the support structure influences the metal particle dispersion

72

and distribution (Chen et al. 2011, Yang et al. 2007). The results obtained from

the activity tests suggest that, due to smaller particles, a narrow particle size

distribution can be obtained over CNTs and these results in higher activity

compared to bigger particles. The incorporation of Ni on the CNTs resulted in

very high activity in the SRE reaction. The results gained are in good agreement

with the published data in Li et al. (2007) and Chen et al. (2012). The results

confirm that CNTs are promising carbon support materials in SRE, compared to

GC and AC, and also, that CNTs perform better compared to conventional

catalyst supports like Al2O3.

Influence of metal/metal oxide CNT supported catalysts on SRE

Metal/metal oxide decorated CNT-based catalysts were tested in a conventional

fixed bed reformer in the temperature range of 150–450 °C. The monometallic

supported CNT materials with different metal loadings (5–10 wt.%) were

prepared using a wet-impregnation method. The ethanol conversion increases

with temperature for all the tested catalysts (Figure 18). The reforming reaction is

an endothermic process and thus the temperature has a positive effect on

conversion.

In the SRE reaction, at very low temperatures between 150–180 °C, the

ethanol conversion starts to increase over all the catalysts, except Rh/CNT and

Pt/CNT catalysts. Already at 180 °C, the ethanol conversion of 25% is achieved

over the Ni-based catalysts (Ni/CNT & Ni16.6/Al2O3).

73

Fig. 18. Ethanol conversion as a function of reaction temperature with 1:3 ethanol-to-

steam molar ratio over Me/CNT-supported catalysts (where Me = Ni, Co, Pt and Rh) in

comparison with the reference Ni16.6/Al2O3 catalyst (I, published by permission of

Elsevier).

The reforming activity was enhanced by temperature. At 200 °C, the order of

activity in ethanol conversion was: Ni/CNT = Ni16.6//Al2O3 >> Co/CNT >>

Pt/CNT > Rh/CNT. The Ni-based catalysts exhibited the highest activity in

ethanol conversion between 150 and 200 °C. At 225–300 °C, over the Ni/CNT

catalyst, the ethanol conversion slightly decreased, until the temperature reached

300 °C probably due to the carbon formation via methane decomposition and also

CO adsorption resulting in carbon islands over the nanotube (Liberatori et al. 2007, Sun et al. 2005). Moreover, Ni-based catalysts are more prone to carbon

formation during the SRE reaction, as reported by Ni et al. (2007). At 300 °C, the

decreasing order of activity in ethanol conversion was: Co/CNT > Ni16.6//Al2O3 =

Ni/CNT >> Pt/CNT > Rh/CNT. Above 300 °C, Ni and Co achieve higher

conversions compared to the rest of the catalysts. Over Co/CNT, the ethanol

conversion is steadily increasing with temperature and reaching the complete

conversion at 450 °C. At 350 °C, Co/CNT and Ni/CNT exhibit as high as ~90%

conversion out-performing the commercial Ni16.6//Al2O3 catalyst. The Co/CNT

and Ni/CNT catalysts achieve over ~90% ethanol conversion at 400 °C (Figure

18). The Ni and Co decorated CNTs exhibit the best results which are in good

agreement with the published literature over inorganic supported catalysts (Homs

et al. 2006, Sun et al. 2005, Fatsikostas et al. 2004) even though the metal content

74

of Ni/CNT is almost ~60% lower than in the Ni16.6/Al2O3 catalyst (16.6 wt.% Ni).

Moreover, the amount of Ni16.6//Al2O3 catalyst used in the tests is two times

higher than that of all the Me/CNT-based catalysts (Me = Ni, Co, Pt and Rh) are

proved to be less active. Over the Ni/CNT and Co/CNT catalysts, 98% and 92%

ethanol conversions were achieved at 375 °C, respectively. Over Pt/CNT, the

ethanol conversion is ~95% at 450 °C and the H2 production steadily increases as

a function of temperature. The TEM analysis confirms the carbon formation over

the Ni/CNT catalyst after the reaction tests (figures not presented). Over Rh/CNT,

the ethanol conversion reaches ~20% at 450 °C and the hydrogen production of

around 1.5 vol.%, which is very low compared to all the other catalysts (Fig. 19).

The reason for the poor performance of the Rh catalysts may be due to the bigger

particle clusters (20–50 nm) (Paper I).

First, at low temperatures, ethanol dehydrogenates to acetaldehyde and

hydrogen. Then, acetaldehyde goes under a series of reactions. The products

composition is presented in Figure 19. Over the Ni/CNT and CO/CNT, the

acetaldehyde decomposition takes place resulting in methane and CO formation.

The acetaldehyde concentration is quite low in the product composition due to

faster dehydrogenation to H2 and subsequent decomposition to CO and methane.

The hydrogen production increases with temperature and achieves 10 vol.% H2

(max.) over Co/CNT and Ni/CNT. Over the Co/CNT and Ni/CNT catalysts, CO

starts decreasing at 400 °C and CO2 concentration increases due to the WGS

reaction.

75

Fig. 19. Composition of product gases over the CNT-based catalysts in SRE with 1:3

ethanol-to-steam molar ratio a) Ni/CNT, b) Co/CNT, c) Pt/CNT, d) Rh/CNT and e) Ni/Al2O3

(I, published by permission of Elsevier).

For Ni16.6/Al2O3, initially CO starts to decrease at 225 °C and then almost

stabilizes above 300 °C with the CO2 increasing trend. Methane was the dominant

product and it is favored over the Pt/CNT catalyst due to ethanol decomposition

reactions (Table 5).

The hydrogen yield increased with temperature and Co/CNT gained the

highest yield, i.e. 2.5 moles of H2 per mole of ethanol at 450 °C compared to all

76

the other catalysts. Over Ni/CNT a maximum of 2 moles of H2 per mole of

ethanol was achieved at 450 °C (figures not presented). For Co/CNT, the

concentrations of CO and methane are quite low compared to the rest of the

catalysts. Over the Pt/CNT and Rh/CNT catalysts, ethanol decomposition and

methanation reactions are predominant, whereas, over the Co/CNT and Ni/CNT

catalysts, ethanol dehydrogenation, SR, WGS and acetaldehyde decomposition

are favored. At 450 °C, the order of hydrogen yield was found to be as follows:

Co/CNT > Ni/CNT > Ni/Al2O3 > Pt/CNT >> Rh/CNT (Paper I).

The selectivity of hydrogen over the CNT-based catalysts is presented in

Figure 20. The Ni/CNT and Ni16.6/Al2O3 catalysts exhibit quite similar hydrogen

selectivity values at 200 °C. At 400 °C, Co/CNT exhibits the highest selectivity

compared to all the other catalysts. The temperature enhances the reforming

activity and also increases the selectivity of hydrogen over the studied catalysts

(Fig. 20). At low temperatures, i.e. at 200 °C and 300 °C, the selectivity of

hydrogen over Ni-supported on CNT and Al2O3 is high compared to the Co/CNT

catalyst. Overall, Co/CNT is the most active and selective catalyst in SRE and

Ni/CNT is also found to be active at lower temperatures but prone to more carbon

formation via methane and CO decomposition reactions.

Fig. 20. Hydrogen selectivity over metal supported CNT-based catalysts as a function

of reaction temperature in SRE reaction; the selectivity values have been calculated

based on the maximum conversion achieved in each temperature. (I, published by

permission of Elsevier).

77

6.1.3 Pt and ZnO promoted Ni/CNT-based catalysts

Influence of Pt addition on Ni/CNT

The reproducibility of the Ni10/CNT was found to be successful. Further tests

were conducted and the optimization of catalyst composition was done by adding

secondary base and noble metals to improve the performance of the Ni10/CNT

catalyst. From Figure 21 a-b, it can be seen that all NiPt containing catalysts have

a slight activity improvement in ethanol conversion and hydrogen production rate.

The Ni particle sizes are observed to increase by the Pt addition. A slight positive

influence on the ethanol conversion and hydrogen formation was found by adding

Pt to Ni10/CNT. The hydrogen production is slightly higher over the reference

Ni16.6/Al2O3 between 150–325 °C. At low temperatures, i.e. 150–275 °C, the

ethanol conversion over the Ni10Ptx supported CNT based catalysts is comparable

to that of the commercial catalyst. Above 325 °C, the hydrogen concentration

over the reference catalyst is comparable to that of Ni10/CNT and Ni10Ptx/CNT

catalysts. The decreasing order of H2 production at 350 °C is as follows:

Ni10Pt1.5/CNT > Ni16.6/Al2O3 > Ni10Pt2/CNT > Ni10Pt1/CNT = Ni10/CNT.

Fig. 21. Performance of Ni10Ptx/CNT (x = 0, 1, 1.5 and 2 wt.%) catalysts in SRE: a)

Ethanol conversion and b) Hydrogen production as a function of reaction temperature

(II, manuscript).

As reported by Panagiotopoulou et al. (2012) and Chen et al. (2012), CH3CHO

and ethylene are the main precursors for coke formation. Below 400 °C,

CH3CHO can undergo a complete decomposition (Liberatori et al. 2007).

According to Figure 22a, the concentration of CH3CHO was quite low, i.e. < 0.5%, at low temperature. It is probably due to the fact that CH3CHO undergoes

78

reforming and decomposition reactions. After dehydrogenation of ethanol at low

temperatures, a faster and subsequent CH3CHO decomposition can take place.

Over Ni10/CNT, the concentration of CH3CHO is highest at the temperature range

of 150–400 °C. In Figure 22b, the methane concentration steadily increases at T >

300 °C and onwards; this trend is similar for all the CNT-based catalysts.

Henceforth, at low temperatures, it is difficult to control methane formation and

only at high temperatures (above 500 °C) methane can be reformed

(Panagiotopoulou et al. 2012, Roh et al. 2007).

Fig. 22. The product concentrations (in vol.%) in SRE a) CH3CHO, b) CH4, c) CO and d)

CO2 concentrations (in vol.%) over Ni10Ptx/CNT (x = 0, 1, 1.5 and 2 wt.%) catalysts (II,

manuscript).

In Figures 22 c-d, it can be observed that the formation of CO and CO2 increases

with temperature. Over Ni16.6/Al2O3, the CO and CH4 formation is much higher at

the temperature range of 150–350 °C, due to ethanol decomposition reaction.

Moreover, over CNT-based catalysts, CO starts to increase until 400 °C and then

decreases to almost < 1 vol.% concentration. Over Ni10/CNT, the CO

concentration is around 1 vol.%, which is higher than over other NiPt-based

catalysts (< 0.6 vol.%). At 400 °C, the concentration of CO starts to decrease,

which might be explained by two reasons: due to the WGS and disproportionation

79

reactions. Furthermore, CO2 is formed in SR, and starts to increase rapidly due to

the SR + WGS reactions. At 400 °C, the decreasing order of CO formation is as

follows: Ni10/CNT > Ni16.6/Al2O3 > Ni10Pt1/CNT > Ni10Pt1.5/CNT > Ni10Pt2/CNT.

Over Ni10Pt2/CNT, the formation of CO2 is the highest at 400 °C compared to all

the other catalysts; this can be justified by the decreasing trend of CO via the

WGS reaction (Paper II).

At 450 °C, the decreasing order of CO2 concentration is as follows:

Ni10Pt2/CNT > Ni10Pt1.5/CNT = Ni10Pt1/CNT > Ni10/CNT. The addition of Pt (2

wt.%) was found to be slightly beneficial in CO and CH4 reduction. A similar

phenomenon was reported by Profeti et al. (2009). By adding secondary metals,

like Pt as promoters to the Ni10/CNT, might improve the reducibility of the

catalyst. This avoids particle agglomeration and affects the oxidation state (Profeti

et al. 2009, Furtado et al. 2009). Both Ni10Pt1/CNT and Ni10Pt1.5/CNT catalysts

behave similarly and exhibit similar activity values in product composition and

ethanol conversion; this being due to the similar catalytic properties (e.g. similar

particle size).

Influence of ZnO addition on Ni/CNT

ZnO was added into the Ni/CNT catalysts in two ways, i.e. ZnO was incorporated

after the Ni impregnation or vice versa. It can be seen from the textural properties

that both the catalysts exhibit slightly different surface characteristics. In the case

of (ZnO10)Ni10/CNT catalyst, the SBET value is lower than that of

Ni10(ZnO10)/CNT. In Figure 23, the ethanol conversion over Ni10(ZnO10)/CNT is

high compared to (ZnO10)Ni10/CNT in the studied temperature range. At 175 °C,

the Ni10(ZnO10)/CNT reaches ~30%, and (ZnO10)Ni10/CNT achieves ~10%

conversions. Ethanol conversion is higher than ~50% over Ni10(ZnO10)/CNT at

225 °C (T50 = 225 °C), and over (ZnO10)Ni10/CNT at 310 °C. A complete ethanol

conversion was achieved over the Ni10(ZnO10)/CNT catalyst at 350 °C. For

(ZnO10)Ni10/CNT, the complete ethanol conversion takes place at 425 °C. In the

case of Ni10(ZnO10)/CNT catalysts, the Ni NPs are probably well dispersed on

ZnO and CNTs. Thus, might have strong interactions between Ni and ZnO on the

CNTs surface, which can result in a higher ethanol conversion. On the contrary,

for the (ZnO10)Ni10/CNT; the bigger ZnO particles are segregated from the Ni

counterparts (Fig. 16). The hydrogen production over the NiZnO-based catalysts

is presented in Figure 23. Both NiZn-based catalysts achieve the highest H2

production compared to all the other catalysts, i.e. Ni and NiPt-based catalysts.

80

For both NiZnO-based catalysts, the hydrogen production steadily increases with

temperature.

Fig. 23. Effect of ZnO addition on the Ni10/CNT catalyst in SRE, i.e. ethanol conversion

(filled symbols) and hydrogen production (open symbols) as a function of reaction

temperature (II, manuscript).

The Ni10(ZnO10)/CNT produces a slightly higher amount of H2 between 150–

350 °C compared to (ZnO10)Ni10/CNT due to the enhanced steam reforming

activity. Both ZnO promoted catalysts activate almost similar H2 production rates

at the reaction temperature of 350 °C and above that. At 450 °C,

(ZnO10)Ni10/CNT produces more hydrogen, i.e. ~11 vol.% compared to ~9.5

vol.% over the Ni10(ZnO10)/CNT catalyst. By adding ZnO to Ni10/CNT, a

significant improvement in CO and CH4 reduction was gained. Over the

(ZnO10)Ni10/CNT catalyst, the undesirable by-products formation is reduced and

hydrogen production is enhanced. This might be caused by ZnO particles being

segregated and having weaker interactions with the Ni NPs which can probably

lead to the avoidance of decomposition and disproportionation reactions.

Moreover, the ZnO is known for its redox and basic properties, which play an

important role in SRE reaction (Yang et al. 2005, Llorca et al. 2003, Homs et al. 2006).

81

Fig. 24. Product concentration in SRE reaction at the reactor outlet stream a) CH3CHO,

b) CH4, c) CO and d) CO2 production (in vol.%) as a function of reaction temperature

over NiZnO-CNT based catalysts (II, manuscript).

Over the (ZnO10)Ni10/CNT catalyst, the CO concentration is almost zero in the

temperature range between 200−350 °C, and also the methane concentration is

quite low, i.e. < 0.2 vol.%, as shown in Figures 24b and 24c. For

Ni10(ZnO10)/CNT catalyst, the CO concentration increases until it reaches the

maximum at 350 °C and further above 400 °C it decreases to < 0.01 vol.%.

Methane is formed above 350 °C, and its amount starts to increase with

temperature, as shown in Fig. 24b. The CO concentration for Ni10(ZnO10)/CNT

increases with temperature until 400 °C and thereafter it decreases. Both catalysts

produce very low CO and CH4 concentrations compared to all the tested catalysts

in SRE. The CO2 concentration was found to be very high between 150−350 °C

(Fig. 24d) over the ZnO promoted catalysts. The addition of ZnO to the Ni/CNT

enhances the activity in the SR and WGS reactions, probably due to the surface

properties of ZnO (Llorca et al. 2001). The surface oxygen coverage is increased

by ZnO addition (Table 8), which might have an effect on the reduction in CO

formation. Thus, the synergistic effects of Ni, i.e. in C-C bond cleavage and in

82

reducing undesirable side reactions over segregated ZnO particles were observed.

Significant structural and surface properties occur over the Ni10(ZnO10)/CNT

catalyst due to the electronic transfer between ZnO and Ni and also with CNTs.

The selectivity of products over the NiZnO-based catalysts in SRE is

summarized in Table 10. The H2 selectivity over (ZnO10)Ni10/CNT steadily

increases from ~59% at 300 °C, to ~71% at 450 °C. For the Ni10(ZnO10)/CNT

catalyst, the SH2 increases slightly from 67% at 300 °C to 69% at 450 °C. The

steam reforming of ethanol activity is more pronounced over the

(ZnO10)Ni10/CNT catalyst confirmed by the CO2 selectivity which is high

between 300–450 °C. The selectivity towards H2 varies based on the mode of

ZnO addition due to the proximity and interactions between the Ni NPs and

bigger ZnO particles.

Table 10. The products selectivity (Si) over the CNT-based catalysts in SRE reaction

against reaction temperature over a) (ZnO10)Ni10/CNT and b) Ni10(ZnO10)/CNT catalysts

(II, manuscript).

T (°C) SH2 (%) SCH3CHO (%) SCO (%) SCH4 (%) SCO2 (%)

a. (ZnO10)Ni10/CNT catalyst

300 59.1 16.0 0.0 0.0 24.9

350 75.8 6.7 0.6 0.0 17.0

400 70.8 0.7 1.1 10.8 16.6

450 70.7 0.0 0.0 13.4 16.0

b. Ni10(ZnO10)/CNT catalyst

300 66.7 13.1 5.1 11.1 4.0

350 66.3 3.3 8.5 11.8 10.1

400 66.1 0 0.9 18.0 15.1

450 68.9 0 0.1 14.7 16.4

The selectivity to acetaldehyde (300–350 °C) is low over Ni10(ZnO10)/CNT, and

the CH3CHO decomposes to give methane and CO which are higher in

concentrations compared to the (ZnO10)Ni10/CNT catalysts. Therefore, the

acetaldehyde formation and decomposition are more pronounced over the

Ni10(ZnO10)/CNT catalysts, and also the WGS activity is lower compared to the

(ZnO10)Ni10/CNT catalysts. In the case of (ZnO10)Ni10/CNT catalyst, the

selectivity towards methane and CO is very low below 350 °C, which may

confirm the fact that the decomposition reactions are less pronounced over

(ZnO10)Ni10/CNT. In the ZnO promoted Ni/CNT catalyst, the bigger ZnO

particles on the CNT have greater role in CO reduction to form CO2 due to

83

surface oxidation by ZnO and facilitation of the shift reaction. At 350 °C, the CO2

selectivity is high over (ZnO10)Ni10/CNT compared to Ni10(ZnO10)/CNT. Between

350–450 °C, a reasonably high CO2 selectivity was gained compared to

Ni10(ZnO10)/CNT. The methane selectivity is also high over Ni10(ZnO10)/CNT

catalyst between 300–450 °C compared to (ZnO10)Ni10/CNT catalyst. Overall, the

ZnO promoted Ni10/CNT based catalysts were most active and H2 selective and

were having low CO formation (Paper II).

Deactivation of CNT-based catalysts

Stability tests over the Ni10/CNT and (ZnO10)Ni10/CNT catalysts were performed

at 400 °C as a function of time-on-stream (TOS) (Figure 25a). A stable ethanol

conversion was achieved over the Ni10/CNT (~95%) and (ZnO10)Ni10/CNT

(~90%) catalysts between 0–60 min TOS. The catalyst deactivation is probably

due to the formation of different types of carbon via the Boudouard and CH4

decomposition reactions. The hydrogen production is more stable over the ZnO

promoted catalyst than Ni10/CNT (Fig. 25a). The products composition and its

distribution over the (ZnO10)Ni10/CNT catalyst are presented in Figure 25b at

400 °C.

Fig. 25. Stability test of Ni10/CNT and (ZnO10)Ni10/CNT catalyst: a) Ethanol conversion

and hydrogen production, b) Product compositions (in vol.%) as a function of time-on-

stream (in min) over (ZnO10)Ni10/CNT catalysts in SRE reaction at 400 °C.

From Fig. 25b, over the (ZnO10)Ni10/CNT catalysts, it can be seen that hydrogen

is the major component produced, and at the reactor outlet, CO < 0.2 vol.% was

observed. Acetaldehyde is completely decomposed to CO and CH4 side products

at 400 °C. The H2 content is around ~8.5–9 vol.% and the CO2 content is stable

84

(2.2 vol.%) over the TOS. The methane concentration decreases with TOS and

probably decomposes to elemental carbon.

Table 11. Physical properties of used catalysts (SRE at 400 °C for 4 h).

Catalyst SBET

(m2 g-1)

Pore volume

(cm3 g-1)

Pore size

(nm)

Carbon

(wt.%)

Ni10CNT 283.9 0.44 6.20 87.7±4.0

(ZnO)10Ni10CNT 271.8 0.40 5.95 83.4±3.5

The catalyst deactivation takes place after 60 min TOS due to the formation of

significant amounts of amorphous and carbon nanofibers in addition to soot

formation observed by TEM images (Paper II, figures not presented). The

physical appearance of the used catalyst is like a gummy-carbon coated catalyst.

The mass of used catalysts was measured after the test and found to be 2.5 times

(for (ZnO10)Ni10/CNT)) and 4.6 times (for Ni10/CNT) increment that of fresh

catalyst. The average carbon formation rate over the (ZnO10)Ni10/CNT catalyst

was 0.043 gcoke h-1 and over Ni10/CNT catalyst was 0.077 gcoke h-1. The carbon

content by EDX analysis in the used catalysts increased from ~72 to ~88 wt.%

(Table 11) that confirms the carbon formation during the reactions. A slight

increase in SBET was observed due to the formation of amorphous carbon in the

proximity of catalyst particles (Paper II). The pore volume and pore size (Table

11) of CNT-based catalysts decreased probably due to the metal NPs and tube

openings of CNTs covered by soot and amorphous carbon deposition. The carbon

balance was calculated based on the inlet ethanol molar flow rate to the sum of

carbon containing compounds molar flow rates. The large deviation in the carbon

molar balance, i.e. the fact that it remains below 100% that indicates carbon

deposition on the metal catalyst surface during the reaction. The deactivation over

the ZnO promoted Ni10/CNT is much slower than that of Ni10/CNT.

6.2 Bio-ethanol steam reforming in a membrane reformer

6.2.1 Hydrogen permeation test and factor ’n’ for Pd/PSS MR

The permeation tests were conducted at each stage of the experimental work. The

following steps (1–3) were performed to verify the presence of any defects and

also to evaluate the permeating flux through the composite Pd/PSS membrane in

85

the pressure range of 2–10 bar. The permeation tests with pure and mixture gases

were performed as follows:

1. Permeation tests with pure single gases: He, N2, and H2.

2. Permeation tests with binary gas mixtures: He/N2, He/H2 and N2/H2 for the

calculation of ideal selectivities.

3. Permeation tests with gas mixtures: He/N2/H2.

First, the permeation tests were performed before and after the steam reforming

reaction with a pure H2 flow in order to measure and verify the flux behavior with

the transmembrane pressure difference of 2–4 bar (abs.). In Figure 26, the factor n

is 0.7 for the Pd/PSS membrane was presented with the highest coefficient of

determination value of R2 = 0.9997. In this case, the membrane was not dense and

does not obey the Sievert’s law; the rate limiting step is controlled by surface

diffusion combined with viscous and Knudsen flows.

Fig. 26. Hydrogen flux permeating through the Pd-Ag self-supported membrane at

390 °C as a function of H2 permeation, i.e. the pressures difference (∆kPan) at different

‘n’ values (V, published by permission of Elsevier).

The permeation tests were also conducted after the regeneration step. It was found

that the H2 flux declines after the reaction. In order to gain the original flux, the

regeneration process was performed during the experiments. After the permeation

test with a pure H2 flow, the test was also conducted with pure N2 and He gases.

The permeance and ideal selectivity (αH2/He and αH2/N2) values are summarized in

Table 11. The ideal selectivity deceases with increasing the pressure drop across

86

the membrane. The permeation behavior through the composite Pd/PSS layer is

attributed to the viscous or Knudsen flow through the defects of the membrane

layer. A complete flow mechanism was discussed by Mardilovich et al. (1998)

and Pinacci et al. (2012). The Pd/PSS membrane is highly perm-selective and

permeable to H2 gas with respect to other gases tested. The calculated H2

permeance value is 4.34·10-7 mol m-2 s-1 Pa-1 at Δp = 80 kPa (as reported in Table

12). The permeance of H2 slightly decreases, whereas for N2 and He, the

permeance increases with the driving force for the composite Pd/PSS. The partial

pressure difference of N2 and He have a positive influence on the permeance of

the studied composite membrane.

Table 12. The permeances of H2, N2 and He and ideal selectivity at different pressure

drops (i.e. driving force) through the Pd/PSS membrane at 390 °C (V, published by

permission of Elsevier).

∆p

(kPa)

Permeance H2

×10-7 (mol m-2 s-1 Pa-1)

Permeance N2

×10-9 (mol m-2 s-1 Pa-1)

Permeance He

×10-9 (mol m-2 s-1 Pa-1)

αH2/N2

(-)

αH2/He

(-)

80 4.34 1.43 2.36 303.4 183.4

100 4.33 1.51 2.36 286.4 183.3

170 4.24 1.69 2.84 249.8 149.2

200 4.19 1.73 3.03 241.3 138.3

260 4.04 1.93 3.60 208.7 112.2

300 3.95 2.03 3.68 194.3 107.2

360 3.81 2.19 3.74 173.5 102.0

The long term stability of the Pd/PSS membrane permeation was examined for

nine different thermal cycles at 390 °C as a function of average pressure (pav). The

total pressure drop across the membrane was kept constant (around 1.5 bar).

Initially the H2 flux was measured up to 1 000 h and then measured at the end of

the reaction tests. The stability and reproducibility of the membrane permeation

behavior was conducted to verify the thermal cycles. For the thermal cycles 4 and

5, the N2 permeance was measured, as well as after testing with different pav

values. Finally, it was concluded that the permeance values remained constant

throughout the testing period.

6.2.2 Effect of reaction pressure

Two different simulated crude bio-ethanol mixtures were tested in BESR. These

were the pure ethanol-water mixture (denoted as EW mix) with 1:13 molar ratio

87

and the ethanol, water, acetic acid and glycerol mixture (EWAG mix). The

simulated EWAG mixture consists of the following compounds (by vol.%): 76%

water and 19% ethanol with 4% glycerol and 1% acetic acid as impurities. This

simulated mixture is similar in composition to the one from the crude ethanol

fermentation broth which was prepared from cheese industry waste by-products

(Sansonetti et al. 2009). The influence of reaction pressure over the SR of pure

EW and EWAG mixtures with impurities was tested at 400 °C in Pd/PSS MR

with the pressure range of 8–12 bar (abs.) and the reactor packed with a Ni/ZrO2

reforming catalyst. As shown in Figure 27, the effect of impurities on the ethanol

conversion and hydrogen recovery factor (HRF) was studied. The results indicate

that impurities and steam have a negative effect on BESR. The bio-ethanol

conversion and HRF increase with reaction pressure for both mixtures. The

reaction pressure has two main effects on the MR system. The first one is from

the thermodynamic equilibrium, at high pressures, the bio-ethanol conversion

increased, i.e. the reaction proceeds towards the products side with an increasing

number of moles. The BESR reaction with EWAG mix consists of three

simultaneous reactions (including ethanol reforming). The parallel SR reactions

with the number of moles increased at the product side (Equations (32) and (33)).

C3H8O3 + 3H2O ⇌7H2 + 3CO2 ΔH°= +128kJ

mol (32)

CH3COOH + 2H2O ⇌4H2 + 2CO2 ΔH°= +134.8kJ

mol (33)

The second one is due to the pressure, which has a positive effect on the

equilibrium conversion; the higher the reaction pressure, the higher the hydrogen

permeation, i.e. the driving force. Therefore, the higher pressure at the retentate

side allows increasing the H2 partial pressure in order to get hydrogen stream to

the permeate side. The continuous removal of hydrogen from the reaction side to

the permeate side will shift the reaction towards the products side, i.e. in

Equations (32) – (33), the equilibrium will shift towards the right-hand side.

Thus, the shift effect caused by reaction pressure is more pronounced and

prevalent over the BESR reaction than the thermodynamic effect and an

increasing trend for conversion with pressure is shown in Figure 27.

88

Fig. 27. Bio-ethanol conversion and HRF as a function of reaction pressure in Pd/PSS

MR over the Ni/ZrO2 catalyst at 400 °C and GHSV = 3200 h-1. SR of EW and EWAG

mixtures (bio-ethanol conversion as a filled symbol and HRF as an open symbol) (III,

published by permission of Elsevier).

As higher reaction pressure results in a higher hydrogen permeating flux through

the membrane, a higher HRF can be achieved. The HRF values are very low,

probably due to carbon deposition and unreacted steam over the membrane

surface which affects the hydrogen permeation. The carbon formation was

confirmed during the regeneration step as methane formation by the reaction (34)

was detected in gas chromatography (GC) during the regeneration process.

C+ H2→ CH4 ΔH°= -75 kJ

mol (34)

The performance of BESR with the EW mixture produces higher conversion and

HRF values in comparison with the EWAG mixture. The coke formation over the

Ni/ZrO2 catalyst for the EWAG mixture is higher during the reaction. The

presence of impurities i.e. glycerol and acetic acid in the EWAG mixture, affects

negatively the reactor performance due to high amounts of carbon formed. This

degrades the membrane permeation, thus resulting in lower fluxes, i.e. low HRF.

Phenomena of this kind were confirmed by Le Valant et al. (2008) and (2010) and

they reported that the crude ethanol impurities will affect the conversion,

selectivity and stability of a catalyst. During the SR of the EWAG mixture, the

dehydration reactions are possible to form ethylene and/or propylene which are

the precursors for the carbon formation (Adhikari et al. 2007).

89

6.2.3 Ni/ZrO2 and Co/Al2O3 catalysts in MR

In this study, two commercial catalysts (Ni/ZrO2 and Co/Al2O3) were compared

and evaluated in BESR reaction in a Pd/PSS MR at 400 °C. The BESR reaction

was performed only with the EWAG mixture over the two different commercial

catalysts. As shown in Figure 28, the bio-ethanol conversion and HRF increases

with reaction pressure. Moreover, higher conversion and HRF were achieved over

the Co/Al2O3 catalyst compared to Ni/ZrO2. The Co catalyst in MR achieved 42%

and 82% conversions at 8 and 12 bar (abs), respectively, whereas for Ni 70% (at 8

bar) and 85% (at 12 bar), conversions were gained. Henceforth, Ni is more active

in bio-ethanol conversion than the Co catalyst. The Co/Al2O3 catalyst exhibits

however higher HRF values than Ni, even though lower conversion values are

achieved than over the Ni catalysts. As reported earlier (Ni et al. 2007), Co and Ni

are the most active catalysts, but Co is more tolerant towards coke forming

reactions than Ni.

Fig. 28. Bio-ethanol conversion (EWAG mix) and HRF versus the reaction pressure in

PSS/Pd MR over Ni and Co catalysts at 400 °C and GHSV = 3200 h-1 (III, published by

permission of Elsevier).

The selectivity to product gas molecules in BESR in Pd/PSS MR are summarized

in Table 13. As reported by Galetti et al. (2010) and in Paper I, the Ni based

catalysts are more prone to carbon and methane formation. Furthermore, nickel is

90

more active in dehydrogenation than cobalt; thus it produces more acetaldehyde

which decomposes to methane.

Table 13. Selectivity of product gases at different reaction pressures in BESR reaction

performed in a Pd/PSS MR at 400 °C, GHSV = 3200 h-1, over Ni/ZrO2 and Co/Al2O3

catalysts (III, published by permission of Elsevier).

Pressure (bar) SH2 (%) SCO (%) SCH4 (%) SCO2 (%)

a. Ni/ZrO2 catalyst

8 36.0 3.2 23.2 37.6

10 33.1 2.3 27.1 37.4

12 31.2 1.2 28.4 39.2

b. Co/Al2O3 catalyst

8 53.7 3.5 8.0 34.8

10 49.6 2.7 11.1 36.6

12 48.5 3.1 11.9 36.5

In Pd/PSS MR, Co/Al2O3 exhibits higher hydrogen and lower methane selectivity

compared to the Ni/ZrO2 catalyst. Over Co/Al2O3, the SH2 of ~54% and ~49% at 8

and 12 bar were achieved, respectively. The reaction pressure has a slightly

negative effect on the hydrogen selectivities. Both catalysts exhibit low CO

selectivity values < 5%.

The hydrogen yield over the two catalysts is presented in Figure 29a. The

cobalt based catalyst is very active in producing higher hydrogen yields compared

to the nickel catalyst. The catalyst support influences the reaction network. ZrO2

is shown to be less active than alumina in the SR reactions. Moreover, Ni/ZrO2 is

more prone to produce less reactive carbon than the alumina based catalyst as

reported by Ferreira-Aparicio et al. (2005). For Ni, the hydrogen yield is constant

over the reaction pressure range 8–12 bar, for cobalt, the H2 yield increases with

pressure and Co is able to gain 35% yield at 12 bar.

In Pd/PSS MR, the cobalt catalyst produces higher hydrogen permeate purity

(HPP) levels, e.g. around 95% at 12 bar, and for Ni 85%, as shown in Figure 29b.

In the case of Co/Al2O3 which is able to produce higher HRF, yield and selectivity

values along with higher HPP values are achieved compared to Ni. Over the Ni

catalysts, the reaction pressure does not have any effect on the HPP levels.

91

Fig. 29. Effect of the reaction pressure on (a) hydrogen yield and (b) hydrogen

permeate purity (HPP) at 400 °C over the Ni/ZrO2 and Co/Al2O3 catalysts, GHSV = 3200

h-1 in Pd/PSS MR (III, published by permission of Elsevier).

Effect of GHSV on Co/Al2O3 catalyst in Pd/PSS MR

The Co/Al2O3 catalyst was found to be the most active and selective catalyst over

the studied catalysts in BESR with the EWAG mixture at 400 °C. The effect of

GHSV over a cobalt based catalyst was studied. As the GHSV increases from 800

to 3200 h-1, the conversion, HRF and yield decreases (Figure 30). It is beneficial

to work at higher space velocities in order to achieve higher productivity levels.

The residence time becomes shorter as the GHSV increases, which means that

more reactant molecules will enter to the MR much faster, henceforth, a low

contact time between the catalyst and reactant molecules in provided. The GHSV

is inversely proportional to the conversion, yield and HRF.

The influence of GHSV over the Co based catalyst in MR was shown in

Figure 30 and the best results were obtained at 800 h-1 with ~94% conversion,

~40% HRF, ~40% yield, and ~95% HPP. The H2 permeation is enhanced at low

GHSVs due to higher contact times and finally results in higher yields and

conversions.

92

Fig. 30. Bio-ethanol conversion, hydrogen yield (YH2) and hydrogen recovery factor

(HRF) as a function of GHSV at 400 °C, preaction = 12 bar, over the Co/Al2O3 catalyst for

SR of EWAG mix in a Pd/PSS MR (III, published by permission of Elsevier).

The HPP value over Co remained constant at ~95% with GHSV. The molar

compositions of the products at the retentate and permeate sides are tabulated in

Table 14. At the permeate side, the molar composition and flow rate of hydrogen

are high, and this confirms the hydrogen purity level, i.e. ~95% at permeate

which was collected during the reaction test. This stream is suitable for feeding

the PEMFC system to produce clean energy. In the retentate stream, unrecovered

H2 and other product gas molecules which are not permeable were present. A very

low CO composition of ~0.14% was permeated, thus a high hydrogen rich stream

is collected and found to be useful for feeding PEMFC.

Table 14. Molar composition and flow rates of SR of the EWAG mix for permeate and

retentate streams. Operating conditions: preaction = 12 bar, T = 400 °C, GHSV = 800 h−1

over the Co/Al2O3 catalyst (III, published by permission of Elsevier).

Gas Retentate stream Permeate stream

Molar composition

(%)

Molar flow rate

(mol min-1)

Molar composition

(%)

Molar flow rate

(mol min-1)

H2 38 2.32·10-4 95 1.44·10-4

CO 2 9.96·10-6 0.1 2.13·10-7

CH4 13 8.02·10-5 0.7 9.86·10-7

CO2 48 2.93·10-4 4.7 7.16·10-6

93

6.3 Bio-glycerol steam reforming in a membrane reformer

6.3.1 Hydrogen permeation test and factor ‘n’ for Pd-Ag MR

First, the permeating flux and pressure exponential factor ‘n’ were calculated by

changing the transmembrane pressure difference. The self-supported Pd-Ag

membrane with a 50 µm thick Pd layer acts as a dense layer which is completely

selective towards hydrogen with respect to the other gases. Thus, at the permeate

side, only hydrogen is collected along with the sweep gas which is utilized to

remove H2. The factor n = 0.5 was calculated for the Pd-Ag membrane by

permeation tests with the highest coefficient of determination R2 = 0.9989 (Figure

31) and behavior of membrane obeys the Sievert’s law which follows the

solution-diffusion mechanism. Henceforth, this membrane is 100% selective to

hydrogen and no other gas will permeate through this dense Pd-Ag layer (Paper

IV).

Fig. 31. Hydrogen flux through the Pd-Ag self-supported membrane at 400 °C vs. H2

permeation driving force, i.e. pressures difference (∆kPan) at different ‘n’ values (IV,

published by permission of Elsevier).

94

The Sievert’s law states that, if the diffusion of atomic hydrogen through the

dense metal layer (i.e., Pd layer) is the rate-limiting step, then the hydrogen flux is

proportional to the square root of the hydrogen partial pressure difference

between retentate and permeates sides (Δp0.5). The temperature dependency of the

hydrogen permeability follows the Arrhenius-like expression, as reported in

Section 2.3.2. The activation energy (Ea) of 8.58 kJ mol-1 and the pre-exponential

factor (Pe0) of 1.14·10-6 mol m-1 s-1 kPa-0.5 were calculated based on the

permeation data using Equation (15). The values Ea and Pe0 are in good agreement

with other experimental data found in the literature for a similar kind of Pd-based

membrane (Basile et al. 2008b). The carbon formation was detected during the

reaction tests which are probably due to the ethylene and/or propylene formation

during glycerol reforming, which in their turn might be the responsible precursors

for the carbon formation (Adhikari et al. 2006). The carbon formation can be

reduced if the reaction temperature is higher than 730 °C and steam-to-glycerol

molar ratio greater than 6 (Lin et al. 2008, Slinn et al. 2008). In this work, at each

reaction test, a regeneration step is performed with a pure H2 flow in reaction

zone for 2 h in order to remove the carbon deposition over the membrane surface

(via Equation (34)). The regeneration step is performed in order to regain the

original flux which was obtained before the reaction.

6.3.2 Comparison of a membrane reformer and a traditional reformer

Effect of reaction pressure

The effect of reaction pressure was studied in bio-glycerol steam reforming

(BGSR) in a Pd-Ag MR over a 0.5 wt.% Ru/Al2O3 commercial catalyst at 400 °C.

The glycerol conversion increases in the membrane reactor (MR) and decreases in

the traditional reactor (TR) with reaction pressure. According to the

thermodynamic calculations in BGSR, the pressure has a negative effect on the

equilibrium conversion and yield in TRs (Adhikari et al. 2007). The reaction

pressure exhibits two main effects; one on the thermodynamics which is a

negative effect and the other is the shift effect which is positive. In MR, ~15%

glycerol conversion was achieved at 5 bar and for TR only ~5% conversion was

gained at 5 bar (Figure 32a). In MR, the pressure has a positive effect, i.e. by

increasing the pressure, H2 which is produced at the reaction side is permeated

through the Pd-Ag membrane. The H2 partial pressure decreases from retentate to

95

permeate. Henceforth, by removing H2 from the reaction side, the BGSR reaction

shifts towards the forward direction that will consume more glycerol to form

products. In this extractor type Pd-Ag MR, the enhancement of glycerol

conversion is achieved due to the shift effect. Thus, in Pd-Ag MR, the conversion

increases with the reaction pressure. Moreover, the permeating hydrogen flux

increases from the retentate to the permeate side with the reaction pressure from 1

to 5 bar. This results in a higher driving force to achieve higher HRF values

(follows the Fick’s-Sievert’s law). The HRF (i.e., CO-free H2) of ~9.5% and

~18% were obtained with 1 and 5 bar, respectively (Paper IV).

The low conversion values were obtained in MR and TR, probably due to the

carbon formation on the membrane surface, thus reducing the hydrogen

permeation. Moreover, the metal content of 0.5 wt.% Ru is relatively low (Hirai et al. 2005). In steam reforming reactions, acidic supports favor the dehydration

reaction which produces precursors for the carbon coke formation (Ni et al. 2007). The carbon formation has a strong effect on the hydrogen permeation and

thus exhibits low HRF values.

Fig. 32. Effect of pressure on (a) glycerol conversion and (b) hydrogen yield and HRF

(only for MR) for the dense Pd-Ag MR and TR (0.5 wt.% Ru/Al2O3 catalyst), conditions:

T = 400 °C, H2O/C3H8O3 = 6 (mol/mol), WHSV = 1.0 h-1, sweep-gas in the counter-current

flow configuration, ppermeate= 1.0 bar and Fsweep-gas/Fglycerol, in = 11.9 (IV, published by

permission of Elsevier).

At higher pressures, a higher conversion is achieved which can result in higher

hydrogen yields. In Pd-Ag MR, the pressure has a positive effect on the hydrogen

yield. From Fig. 32b, in the MR, the hydrogen yield increases with pressure. In

MR, the yield is around ~5.3% and ~7.35% at 1 and 5 bar, respectively. In the

case of TR, the yield decreases with reaction pressure and ~4.9% at 1 bar and

3.35% at 5 bar were achieved. On the contrary, the pressure has a negative effect

96

on the reaction equilibrium in TR. In MR, a CO-free HRF increases with the

pressure (Fig. 32b). As predicted, as the pressure increases, this results in higher

hydrogen permeation to the permeate side. Overall, the MR performs better than

TR, operating at the same conditions.

The effect of reaction pressure on the product gas selectivity in MR was also

calculated at WHSV of 1 h-1. The selectivity of CO decreases with the pressure,

i.e. the selectivity values of ~52% and 23% at 1 and 5 bar were achieved,

respectively (figures not presented). The selectivity of CO2 however increases

with pressure, i.e. from 46% at 1 bar to ~73% at 5 bar. The hydrogen permeating

driving force has a positive influence by shifting the WGS reaction towards the

forward direction, i.e. to the product side, by consuming more CO to produce CO2

and H2. The methane selectivity slightly increased with pressure, i.e. 2.1% at 1

bar to 4.4% at 5 bar, respectively.

Effect of WHSV

All the results presented in the previous section are with the WHSV value of 1 h-1.

The reduction of WHSV was performed in the counter-current flow configuration

in order to verify the effect of reactants residence time on the MR performance.

As the WHSV decreased from 1.0 to 0.1 h-1, the contact time between the catalyst

and the reactant molecules increased; thus, a longer residence leads to higher

conversions with higher HRF values (Figure 33).

Fig. 33. Glycerol conversion and HRF as a function of WHSV for the Pd–Ag MR, and

TR. Conditions: T = 400 °C, H2O/C3H8O3 = 6 (mol/mol), preaction = 5 bar (abs.), sweep-gas

in a counter-current flow configuration, ppermeate = 1.0 bar and Fsweep-gas/Fglycerol, in = 11.9

(IV, published by permission of Elsevier).

97

Lowering the WHSV in both systems, i.e. in MR and TR, has an increasing effect

on glycerol conversion. But in MR, a higher glycerol conversion is achieved than

in TR due to the shift effect, i.e. continuous removal of hydrogen through the

membrane (Paper IV). In MR, glycerol conversion of ~60% was achieved and for

TR, the conversion was 42% at WHSV of 0.1 h-1. Henceforth, at low WHSVs, a

higher hydrogen partial pressure was obtained at the reaction side favouring the

driving force and finally collecting a higher CO-free H2 recovery in the permeate

side. Overall, HRF of ~56% at WHSV of 0.1 h-1 was achieved and it drops to

17% at WHSV value of 1.0 h-1.

The selectivity to different product gas molecules is shown in Figure 34. The

selectivity to hydrogen remained constant with WHSV and gained ~52%. It was

observed that decreasing the WHSV has a positive effect on the CO selectivity. At

5 bar reaction pressure, the selectivity of CO decreases from 10% at 1 h-1 to 2% at

0.1 h-1. The CH4 selectivity decreases with increasing WHSV (Paper IV). In Fig.

34, the selectivity of CO increases and CO2 slightly decreases as a function of

space velocity. Moreover, the HRF drastically decreases with an increasing

WHSV value.

Fig. 34. CO-free HRF and selectivity of product gases against WHSV for the Pd–Ag

MR. Conditions: T = 400 °C, H2O/C3H8O3 = 6 (mol/mol), preaction = 5 bar (abs.), sweep-gas

in a counter-current flow configuration, ppermeate = 1.0 bar and Fsweep-gas/Fglycerol, in = 11.9.

From Figure 35, the highest hydrogen yield of ~28% was obtained at low WHSV,

i.e. 0.1 h-1 at 5 bar. Furthermore, TR obtained the highest yield of 16% at 0.1 h-1

and the lowest of 4% at 1 h-1. As predicted, low WHSVs exhibit higher yields in

98

both reactors, i.e. MR and TR. The space velocities have a significant effect on

the reactor performance and it is beneficial to operate at longer residence times to

achieve higher MR performance.

Fig. 35. Hydrogen yield as a function of WHSV for the Pd-Ag MR and TR. Conditions: T

= 400 °C, H2O/C3H8O3 = 6 (mol/mol), preaction = 5 bar (abs.), sweep-gas in a counter-

current flow configuration, ppermeate = 1.0 bar and Fsweep-gas/Fglycerol, in = 11.9 (IV, published

by permission of Elsevier).

6.4 Water-gas-shift reaction in a membrane reformer

The membrane stability test was performed with different gaseous mixtures to

verify the membrane permeation characteristics. The pressure exponential factor

‘n’ for the composite Pd/PSS membrane was discussed in Section 6.2.1.

The thermal cycles (1–5) are the permeation tests with pure gas and gas

mixtures (time length in parentheses): 1. He (268 h), 2. He/H2 (800 h), 3. He/H2

(1290 h), 4. N2/H2 (48 h) and 5. N2/He/H2 (776). Thermal cycles 6 and 7 for the

WGS reaction tests, i.e. H2O-to-CO molar ratios (1–7) of 3, 4 (740 h) and thermal

cycles 8 and 9, are for the WGS with syn-gas feed mixtures (168 h) and the

reproducibility tests (72 h) (Paper V).

99

6.4.1 Effect of He, CO, CO2 and H2O additions on H2 permeation

In Pd/PSS MR, the effect of H2O, He, CO, and CO2 additions over the membrane

permeation was investigated. The compositions with three different gaseous

mixtures (simulated syn-gas mixtures from an IGCC-plant) are presented in Table

15. The hydrogen permeating flux can be affected by many factors like dilution

with other gases (Li et al. 2000b) and vapors, competitive adsorption with other

gases (Wang et al. 2006b), hydrogen depletion, etc. (Augustine et al. 2011, Hou et al. 2002b, Amandusson et al. 2001). In thermal cycle V, i.e. He, N2 and H2 gas

mixture was fed to the MR and the influence of this mixture on H2 permeation

was studied. It was found that the gas mixture had no significant effect on the H2

permeation. The permeating flux for the binary mixture (Mix bin), i.e. H2/He, was

found to slightly reduce in comparison with the pure gas. This is probably due to

the dilution and depletion effects (Wang et al. 2006b).

Table 15. Composition of different syn-gas feed gaseous mixtures fed to Pd/PSS MR

at 390 °C for permeation tests.

Mixture Gas mixture Molar composition (%)

Mix Binary H2/He 45/55

Mix 1 H2/CO/CO2/H2O 45/8/31/16

Mix 2 H2/He/CO2/H2O 45/8/31/16

In Figure 36, it is clearly shown that the permeation for mix 1 and mix 2 is

declined over the pressure driving force compared to the pure and binary gases.

The addition of steam, He, CO and CO2 has a detrimental effect on the H2

permeation thorough the membrane. The flux declining is due to the presence of

CO and steam (Mix 1 and Mix 2) which has a strong effect on the hydrogen

permeation (Li et al. 2000b).

100

Fig. 36. Hydrogen permeating flux as a function of H2 partial pressure difference

(kPa0.7) at 390 °C with pure H2 and different gaseous mixtures (Table 15) Mix 1, Mix

2, Mix binary (Mix bin), and Pure H2 (V, published by permission of Elsevier).

Especially, over the Pd-based membranes, a competitive CO adsorption (i.e., COad) can hinder the H2 permeation. The steam addition can result in the H2O

decomposition and recombination mechanism. Initially, H2O forms adsorbed as

[OHad] and [Oad] nascent oxygen atoms which can poison the membrane surface

via forming PdO (Gao et al. 2004, Amandusson et al. 2000). The effect of steam

adsorption is a very complex over the Pd-surface and it is more deteriorate than

CO adsorption. Henceforth, the addition of steam and CO2 is more detrimental

over the membrane permeation than CO and He.

6.4.2 WGS reaction test: Effect of CO/H2O ratio

The water-gas-shift (WGS) reaction was performed in the Pd/PSS MR packed

with 3 g of high temperature shift (HTS) catalyst Fe-Cr/ZnO with steam-to-CO

ratio of 1 (stoichiometric), operated at 390 °C and in 7–11 bar pressure range. The

CO conversion was constant and the HRF increased with increasing pressure

from 7 to 11 bar (Figure 37). As expected, the HRF is enhanced with pressure and

facilitates a higher driving force which is responsible to remove H2 continuously

from the retentate side to the permeate side. The CO conversion follows a

constant trend with the reaction pressure due to the positive effect of the H2

selective permeation. Moreover, the H2 permeation counterbalanced by the

101

negative effect due to the CO reactant loss as part of permeation through the

defects of the membrane layer. At Preaction of 11 bar, the CO conversion is ~80%

and HRF about ~60% (Paper V).

Fig. 37. CO conversion and HRF over the Fe-Cr shift catalyst in Pd/PSS MR versus

Preaction range of 7–11 bar. Conditions: T = 390 °C, GHSV= 5750 h-1, ratio CO/H2O = 1 and

ppermeate = 1 bar (V, published by permission of Elsevier).

The influence of GHSV was studied over the performance of Pd/PSS MR for CO

conversion and HRF at 390 °C. It is beneficial to operate the reactor at high space

velocities due to the time and economic constraints. The longer residence time is

always beneficial in order to have a better contact time.

Figure 38 illustrates the CO conversion and HRF as a function of GHSV with

two different H2O-to-CO molar ratios. The CO conversion and HRF are enhanced

by increasing the molar ratio from 3 to 4 at constant GHSV. By increasing GHSV,

the conversion and HRF decrease in all the cases due to the low contact times. At

3 450 h-1, the CO conversion is increased from 85% to 97% by increasing the

molar ratio of H2O-to-CO from, i.e. 3 to 4 (Fig. 38). This improvement is due to

the excess steam shifting the WGS reaction towards the product side, i.e. forward

direction according to the reaction R7 (Table 5) by consuming higher CO

amounts.

102

Fig. 38. CO conversion and HRF for the WGS reaction performed in a Pd/PSS MR with

two H2O/CO ratios, i.e. 3 and 4 against the GHSV at 390 °C (V, published by permission

of Elsevier).

At low GHSVs, it is always beneficial for the reactor performance to achieve

higher conversions and HRF due to longer contact times for catalyst and

reactants. The maximum hydrogen recovery of ~83% at 3450 h-1 was obtained

and it dropped to 31% at 14 000 h-1 GHSV. Moreover, a higher HRF value was

achieved at low GHSV, because higher conversion results in higher amounts of

hydrogen which were permeating continuously to the permeate side and a higher

H2 was recovered due to the shift effect in MR.

In Table 15, a HPP level of ~96% was gained at GHSV (5750 h-1) compared

to higher space velocity of 14 000 h-1 (i.e., ~90–94 % was achieved). Overall, a

higher HPP of above 90% was achieved. A higher H2O-to-CO ratio is found to be

beneficial than the stoichiometric. In particular, the increase in steam-to-CO ratio

more, as well as a decrease in GHSV probably shifts the WGS reaction towards

the product (H2 and CO2) side, thus minimizing the carbon formation due to the

Boudouard reaction. The coke formation was detected after each reaction test and

it is the main reason for the low H2 permeating flux. In Table 16, the coke

formation for the H2O-to-CO ratio of 4 is lower compared to the ratio of 3. The

deposition of carbon decreases by increasing the feed molar ratio of H2O/CO and

decreasing the GHSV value.

103

Table 16. Hydrogen permeate purity (HPP) and coke formation as a function of GHSV

for the WGS reaction performed in Pd/PSS MR with different H2O/CO molar ratios (i.e.,

3 and 4) (V, published by permission of Elsevier).

GHSV (h-1) HPP (%) Coke (g h-1)

H2O/CO = 3 H2O/CO = 4 H2O/CO = 3 H2O/CO = 4

3450 93.8 94.6 0.14 0.07

5750 95.9 94.7 0.26 0.10

14000 92.2 90.0 0.59 0.27

6.4.3 WGS reaction test using syngas compositions

In this section, syn-gas feed with three different compositions (Table 17) were

tested in Pd/PSS MR. The feed composition is similar to that of the reformate

stream which is coming out of the IGCC plant. Moreover, this type of a gaseous

mixture is usually found from the outlet of a gasifier or steam reformer. WGS

reaction was performed in MR in order to upgrade syn-gas feed to a H2 rich

stream to feed PEMFC.

Table 17. Syn-gas mixture compositions fed to the Pd/PSS MR and the H2 PP driving

force for each feed mixture during the WGS reaction (V, published by permission of

Elsevier).

Feed Composition (%) H2 partial pressure difference

∆pH2 (bar) CO H2 CO2 H2O H2O/CO

Feed 1 0.08 0.45 0.31 0.16 2 5.85

Feed 2 0.08 0.41 0.28 0.24 3 5.60

Feed 3 0.08 0.36 0.24 0.32 4 5.36

In Table 17, by changing the syn-gas feed composition from Feed 1 to Feed 3,

mainly corresponds to the increase in the feed molar ratio, i.e. H2O/CO, from 2 to

4 and the pressure decreased slightly in the drivign force (ΔpH2).

104

Fig. 39. CO conversion and HRF versus the three syngas feed compositions used in

the WGS reaction performed in Pd/PSS at 390 °C, GHSV = 3450 h-1, preaction = 11.0 bar,

and ppermeate = 1.0 bar (V, published by permission of Elsevier).

As the H2O/CO molar ratio increased from 2 to 4, the CO conversion slightly

increased from 67% for Feed 1 to 77% for Feed 3, as shown in Figure 39. The

HRF follows a decreasing trend, probably due to two effects, first being the

carbon deposition on the membrane surface, which inhibits the hydrogen

permating flux through the membrane. The second is the lowering of hydrogen

permeation driving force during the WGS reaction from 5.85 for feed 1 to 5.36

for Feed 3 (Table 17). In each reaction test, the HPP is higher than ~97%, which

confirms that the CO content in the permeate side is between ~0.2–0.3%. The

reproducibility of results was conducted by repeating the same tests with the same

three different feed compositions in order to confirm the results. Therefore, after

700 h time, the WGS reaction has been performed by varying the GHSV.

105

Table 18. Durability of Pd/PSS MR performance by comparing the results obtained

from original and repeated tests as a function of GHSV (V, published by permission of

Elsevier).

GHSV (h-1) CO conversion (%) HRF (%) HPP (%)

Original Repeated Original Repeated Original Repeated

3450 96.1 94.8 82.8 80.2 94.6 94.2

5750 90.2 89.7 75.5 73.6 94.7 93.5

14000 89.3 88.6 30.9 28.9 90.0 89.8

A comparison between the original and repeated reaction tests results is reported

in Table 18. As shown, there is no significant difference in the results obtained

and the reproducibility of the results is good. The pure gas permeation test was

performed at the end of the reaction test and it confirms that similar peremating

flux values were obtained.

106

107

7 Summary and concluding remarks

7.1 Summary

The application of biomass in distributed hydrogen production for fuel cells can

be a promising hydrogen energy route. In order to produce hydrogen in a more

environmental and efficient way, oxygenates like bio-alcohols can be potential

feedstocks for hydrogen which can reduce the carbon footprint and improve the

hydrogen capacity. Especially, the short chain alcohols like methanol, ethanol and

glycerol are viable hydrogen sources which can be reformed at low temperatures

and also can have a potential for on-board H2 production. There is a big challenge

in performing the steam reforming reaction (SRE) at low temperatures (< 450 °C)

using alcohols in order to control the CO formation and increase the hydrogen

production. Thus, catalyst supports and metal catalysts have a critical role to

produce H2 at low operating temperatures. In this work, two main objectives were

set, first studying the suitability of CNTs as support material in SRE and second

studying the applicability and potentiality of a membrane-assisted reformer in

producing high purity hydrogen. The conclusions are made based on the results

obtained from the catalyst activity tests and the membrane reformer studies.

In this work, pristine MWCNTs (or CNTs) are pre-treated and functionalized

by a liquid oxidation method in order to improve the solubility before anchoring

the active metal compounds. After the pre-treatment, the metal precursor was

impregnated into the carboxylated CNTs. There was no significant destruction of

nanotubes after the catalyst synthesis steps. The metal/metal oxide decorated

CNTs have a narrow particle distribution. The CNTs are mesoporous materials

with reasonably high SBET compared to commercial alumina-based catalysts.

Screening of different carbon supported Ni catalysts (Ni2/CNT, Ni2/GCB, and

Ni2/AC) in the SRE reaction was studied. Over the Ni2/CNT, small (~2 nm) and

narrow Ni, particle sizes were formed and found active in the reforming reaction.

At low metal loadings, the particle size had a greater influence on the catalyst

activity. The CNTs are promising supports in the SRE compared to conventional

supports, e.g., graphitic carbon black, activated carbon and alumina.

Four different types of metals were introduced over the CNTs, i.e. Co, Ni, Pt

and Rh on CNTs with 5–10 wt.% of nominal loadings. The Co/CNT and Ni/CNT

are found to be most active catalysts compared to Pt/CNT and Rh/CNT catalysts.

Over the Co/CNT, a high hydrogen yield with low CO and CH4 formation was

108

achieved. The reforming temperature had a significant effect on the catalyst

activity. The product formation and distribution were more dependent on

temperature over active metal/metal oxide type on CNTs.

Further tests were carried out with Pt and ZnO promoted Ni/CNT catalysts,

for enhancing the reforming activity and reducing the CO formation. There was a

significant changes in the textural properties of the pre-treated CNTs and

Me/CNT (Me = Ni, NiPt and NiZnO) due to the synthesis steps. The promotional

effect of Pt in Ni10Ptx/CNT (x = 1, 1.5 and 2 wt.%) catalysts was insignificant in

SRE. The ethylene formation over CNT based catalysts was found to be very low,

i.e. < 0.3 vol.%. The carbon formation over the CNT-based catalysts was detected

during the reforming experiments. The activity tests showed that the ZnO

promoted CNT-based catalysts were efficiently enhancing the hydrogen

production rate and inhibiting the carbon monoxide formation compared to all the

tested catalysts. The deactivation rate over the ZnO promoted Ni10/CNT was slow

compared to Ni10/CNT. Over the CNTs, the carbon deposition probably takes

place via the Boudouard and hydrocarbon decomposition reactions. A high H2

selectivity of ~76% with low CO selectivity (< 1%) was achieved over

(ZnO10)Ni10/CNT (at 350 °C). Moreover, a slow deactivation rate was found over

ZnO promoted Ni10/CNT catalysts. Overall, the ZnO promoted Ni10/CNT

catalysts outperforms the Ni10/CNT and NiPt/CNT-based catalysts. The catalyst

deactivation is attributed to the formation of different types of carbon on the

catalyst surface. Further work is needed to design and optimize the CNT-based

catalysts to have high durability for long term usability.

In the second stage of the work, an intensified approach was studied using a

Pd-based membrane-assisted catalytic reactor in the hydrogen production. Three

different reactions with two different Pd-based membranes were studied

separately, in order to gain new knowledge on the Pd-membrane applicability in

hydrogen producing reactions. Using an extractor type membrane reformer, the

reforming reaction and H2-selective separation through the metal membrane were

performed in a single device. The summary of Pd-based membrane and their

characteristics, including reforming and HTS catalysts, is presented in Table 19.

The pressure exponential factor ‘n’ provides the rate determining step which

controls the H2 diffusion through the palladium membranes. In this study, for the

dense Pd-Ag (50 µm thickness) membrane, the ‘n’ value is 0.5 and the permeation

follows the Sievert’s law with a solution-diffusion mechanism. Over the dense Pd,

the H2-permeability follows the Arrhenius-like temperature dependency

expression. For the Pd/PSS composite membrane, the factor ‘n’ is 0.7, which

109

deviates from the Sievert’s law validity. For the Pd/PSS membrane, the overall H2

flux follows the quasi solution-diffusion, and/or the Knudsen and viscous flow

mechanisms and the rate-controlling is no more the bulk diffusion. The dense Pd-

Ag membrane exhibits relatively low permeating fluxes in comparison to the

Pd/PSS. The Pd/PSS possesses high fluxes with the reasonable ideal selectivity

(i.e., H2/N2 = 303). Moreover, the Pd/PSS with Pd thickness of 20 µm was found

to be highly stable and durable for ~1 500 h. No sweep-gas was used in the case

of Pd/PSS MR, which could be an added advantage in the industrial scale. Two

different commercial catalysts, i.e. Ni/ZrO2 and Co/Al2O3 were tested in the

Pd/PSS MR. Not only the membrane properties, but also the active metal catalyst

and support materials are important for the overall performance of the MR.

Table 19. Membrane characteristics, preparation and catalysts used in the reactions

studied in this research work (Papers III, IV and V).

Reaction Membrane Thickness and conditions Catalysts Preparation Factor 'n'

BESR Pd/PSS δ = 20 µm (Pd)*; L = 7.4 cm;

p = 8–12 bar (abs.); T = 400 °C

Ni/ZrO2 &

Co/Al2O3

ELP 0.7

BGSR Dense Pd-Ag# δ = 50 µm (Pd-Ag); L =14 cm;

p = 1–4 bar (abs.); T = 400 °C

Ru/Al2O3 CDW 0.5

WGS Pd/PSS δ = 20 µm (Pd); L = 7.4 cm;

p = 7–11 bar (abs.); T = 390 °C

Fe-Cr/ZnO ELP 0.7

*Membrane thickness (δ), ELP = electroless plating, CDW = cold-rolling and diffusion welding, #self-

supported, abs. = absolute, L = active length

In BGSR, a dense Pd-Ag membrane reactor performs better than the traditional

reactor (or conventional reformer) at the same operating temperature. The effect

of operating variables, i.e. reaction pressure and space velocities, were found to

be significant on the MR performance. As predicted, an increase in the reaction

pressure will result in an increase in conversion, yield and HRF. Moreover, low

WHSVs or GHSVs will have a positive influence on the MR and TR performance

due to high contact time. In MR, the thermodynamic equilibrium is shifted

towards the product side due to the continuous removal of H2 from the retentate to

the permeate side.

The permeation through the Pd-based membrane is affected by the reactant

composition in comparison with single pure gas (i.e., H2). Steam and CO2 inhibit

the H2 permeation and thus affect negatively the MR performance. The steam-to-

carbon ratio has a positive effect on the MR performance. An increase in the

steam-to-carbon ratio causes a shift towards the product side and also reduces the

110

CO formation in WGS reaction in MR. The HPP levels for the Pd/PSS supported

membrane were reasonably high being ~95%, whereas for the dense Pd-Ag

membrane the HPP levels were greater than ~99.999%.

The conversion and yield enhancement by the extractor-type Pd based MR

had a positive effect on the thermodynamically limited SR and WGS reactions by

shifting the reactions more towards the product sides.

A small scale membrane reformer can be utilized in the production of

hydrogen with high purity from a wide variety of biomass sources for small

auxiliary power units and also for the niche markets. Moreover, the ability to

perform steam reforming reaction and separation in one unit reduces the number

of additional downstream separation units. Thus, a membrane assisted reformer is

more efficient compared to a conventional reformer.

7.2 Concluding remarks

The main results of the thesis can be concluded as follows. The main aims of this

thesis is to understand and gain the knowledge on catalyst development in ethanol

reforming at low temperatures by introducing CNT supported catalysts and

evaluating the conventional and membrane assisted rectors in hydrogen

production.

First, CNTs are shown suitable catalyst supports for the reforming reaction

due to its tubular structure which makes it ideal to disperse or anchor active

metal/metal oxide crystallites with small particle sizes. Among them Ni/NiO

supported CNTs were found to be the active and selective catalysts similar to that

of Cobalt based catalysts. However, the Ni based catalyst is more feasible due to

its low temperature reducibility and reproducibility. Further tests with ZnO-

promoted Ni10/CNT based catalysts showed them to be superior compared to Ni

catalysts. Incorporation of ZnO improves the electronic transfer between the

CNTs and Ni, and might improve the reducibility behavior of the catalyst. The

oxidative surface of the ZnO-promoted catalyst was responsible for the oxidation

of CO to CO2 and it also enhanced the WGS activity. In this thesis, the suitability

of CNTs in SRE is the fundamental issue and still a lot of research is needed to

improve the catalyst stability and durability for industrial applications. Future

work will be more on improving stability of CNTs by introducing and promoting

novel and new hybrid materials into CNTs.

Second, new information for the applicability of Pd-based MRs in SR and

WGS reactions is generated. In this thesis, the goal is to produce pure hydrogen

111

using bio-alcohols and syn-gas feed in MR for fuel cells. A comparison was made

between a conventional and a membrane based reformer. In all the cases, the MR

outperforms the conventional reactor. The tubular Pd/PSS is a potential

membrane material for commercial applications due to its excellent permeation

characteristics, operability, durability, and mechanical stability compared to the

dense Pd-Ag membrane. Based on the studied parameters, the membrane

reformer is an efficient technology to enhance the overall efficiency of SR and

WGS reaction processes. Understanding catalysis and membrane phenomena

separately is an important to improve the overall performance of catalytic

membrane based reactors. The prospective integration of a catalytic reaction and

selective downstream separation in a membrane assisted reactor can be foreseen

as a potential way to enhance pure hydrogen production.

112

113

References

Adhikari S & Sandun F (2006) Hydrogen membrane separation techniques. Ind Eng Chem Res 45: 875−881.

Adhikari S, Fernando S, Gwaltney SR, Filip SD, Bricka RM, Steele PH & Haryanto A (2007) A thermodynamic analysis of hydrogen production by steam reforming of glycerol. Int J Hydrogen Energy 32(14): 2875−2880.

Akande AJ, Idem RO & Dalai AK (2005) Synthesis, characterization and performance evaluation of Ni/Al2O3 catalysts for reforming of crude ethanol for hydrogen production. Appl Catal A 287: 159−175.

Amandusson H, Ekedahl LG & Dannetun H (2000) The effect of CO and O2 on hydrogen permeation through a palladium membrane. Appl Surf Sci 153: 259−267.

Amandusson H, Ekedahl LG & Dannetun H (2001) Hydrogen permeation through surface modified Pd and PdAg membranes. J Membrane Sci 193: 35−47.

Augustine AS, Ma YH & Kazantzis NK (2011) High pressure palladium membrane reactor for the high temperature water-gas shift reaction. Int J Hydrogen Energy 36: 5350−5360.

Azadi P, Farnood R & Meier E (2010) Preparation of multiwalled carbon nanotube-supported nickel catalysts using incipient wetness method. J Phys Chem A 114: 3962−3968.

Babita K, Sridhar S & Raghavan KV (2011) Review: Membrane reactors for fuel cell quality hydrogen though WGSR- review of their status, challenges and opportunities. Int J Hydrogen Energy 36: 6671−6688.

Baker RW (2002) Future Directions of Membrane Gas Separation Technology. Ind Eng Chem Res 41: 1393–1411.

Baker RW (2004) Membrane technology and applications. John Wiley & Sons. Balata M, Balata H & Öz C (2008) Progress in bioethanol processing. Prog Energy

Combust Sci 34: 551–573. Basile A, Chiappetta G, Tosti S & Violante V (2001) Experimental and simulation of both

Pd and Pd/Ag for a water gas shift membrane reactor. Sep Purif Technol 25: 549−571. Basile A, Drioli E, Santella F, Violante V, Capannelli G & Vitulli G (1996) A study on

catalytic membrane reactors for water gas shift reaction. Gas Sep Purif 10: 53−61. Basile A, Gallucci F, Iulianelli A & Tosti S (2008a) CO-free hydrogen production by

ethanol steam reforming in a Pd–Ag membrane reactor. Fuel Cells 8: 62–68. Basile A, Pinacci P, Iulianelli A, Broglia M, Drago F, Liguori S, Longo T & Calabrò V

(2011) Ethanol steam reforming reaction in a porous stainless steel supported palladium membrane reactor. Int J Hydrogen Energy 36: 2029−2037.

Basile A, Gallucci F & Tosti S (2008b) Chapter 8. Synthesis, characterization and applications of palladium membranes. In: Malada R & Menendez M (eds) Inorganic membranes: synthesis, characterization and applications. Membrane science and technology. Elsevier: 255−323.

114

Benito M, Padilla R, Rodríguez L, Sanz JL & Daza L (2007) Zirconia supported catalysts for bioethanol steam reforming: Effect of active phase and zirconia structure. J Power Sources 169: 167−176.

Bi JL, Hong YY, Lee CC, Yeh CT & Wang CB (2007) Novel zirconia-supported catalysts for low-temperature oxidative steam reforming of ethanol. Catal Today 129: 322−329.

Bichon P, Haugom G, Venvik HJ, Holmen A & Blekkan EA (2008) Steam Reforming of ethanol over supported Co and Ni catalysts. Top Catal 49: 38–45.

Bird CEB (1996) Fuel cell commercialization issues for light-duty vehicle applications. J Power Sour 61: 33−48.

Breen JP, Burch R & Coleman HM (2002) Metal-catalyzed steam reforming of ethanol in the production of hydrogen for fuel cell applications. Appl Catal B 39: 65−74.

Brouwer J (2010) On the role of fuel cells and hydrogen in a more sustainable and renewable energy future. Curr Appl Phys 10: S9−S17.

Bruijn FA, Rietveld B & Brink RW (2007) Hydrogen production and fuel cells as the bridging technologies towards a sustainable energy system. In: Centi G & Santen RA (eds) Catalysis for Renewables: From Feedstock to Energy Production. Weinheim, Wiley-VCH Verlag GmbH & Co. KGaA: 307−308.

Brunetti A, Bernardo P, Drioli E & Barbieri G (2010) Ch 14. Membrane Engineering: Progress, and Potentialities in Gas Separations. In: Yampolskii Y & Freeman B (eds) Membrane Gas Separation. Wiley UK.

Bruschke H (1995) Industrial application of membrane separation processes. Pure App Chem 67: 993−1002.

Bshish A, Yaako Z, Narayanan B, Ramakrishnan R & Ebshish A (2011) Steam-reforming of ethanol for hydrogen production. Chem Papers 65: 251–266.

Busca G, Tania Montanari T, Resini C, Ramis G & Costantino U (2009), Hydrogen from alcohols: IR and flow reactor studies. Catal Today 143 (1–2): 2−8.

Byrd AJ, Pant KK & Gupta RB (2008) Hydrogen production from glycerol by reforming in supercritical water over Ru/Al2O3 catalyst. Fuel 87: 2956–2960.

Caravella A, Scura F, Barbieri G & Drioli E (2010) Sieverts law empirical exponent for Pd-based membranes: Critical analysis in pure H2 permeation. J Phys Chem B 114: 6033−6047.

Caro J (2010) Basic aspects in membrane reactors. In: Drioli E & Giorno L (eds) Comprehensive Membrane Science and Engineering. Elsevier 3: 1−24.

Cheekatamarla PK & Finnerty CM (2006) Reforming catalysts for hydrogen generation in fuel cell applications. J Power Sources 160: 490−499.

Chen L, Choong CKS, Zhong Z, Huang L, Ang TP, Hong L & Lin J (2010) Carbon monoxide-free hydrogen production via low-temperature steam reforming of ethanol over iron-promoted Rh catalyst. J Catal 276: 197−200.

Chen Z, Guan Z, Li M, Yang Q & Li C (2011) Enhancement of the performance of a platinum nanocatalyst confined within carbon nanotubes for asymmetric hydrogenation. Angew Chem Int Ed 50: 4913−4917.

115

Ciambelli P, Palma V & Ruggiero A (2010) Low temperature catalytic steam reforming of ethanol. 1. The effect of the support on the activity and stability of Pt catalysts. Appl Catal B 96: 18–27.

Deluga GA, Salge JR, Schmidt LD & Verykios XE (2004) Renewable hydrogen from ethanol by autothermal reforming. Science 303: 993−997.

Dittmeyer R & Caro J (2008) Catalytic Membrane Reactors. In: Ertl G, Knözinger H, Schüth F & Weitkamp J (Eds.) Handbook of Heterogeneous Catalysis. Wiley-VCH Verlag GmbH & Co. KGaA: 2198–2248.

Dittmeyer R, Höllein V & Daub K (2001) Membrane reactors for hydrogenation and dehydrogenation processes based on supported palladium. J Mol Catal A Chem 173: 135–84.

Dorokhina A & Tessema D (2011) World Energy and Population: trends to 2050. Advanced Research Project. MBA International Business Management, Virginia International University: 627.

Dunn S (2002) Hydrogen futures: toward a sustainable energy system. Int J Hydrogen Energy 27: 235–264.

Ertl G, Knozinger H & Weitkamp J (1997) Handbook of Heterogeneous Catalysis. Vol 1. Wiley-VCH.

EU (2003) Hydrogen Energy and Fuel Cells A vision of our future. Directorate-General for Research, Luxembourg: EU commission, Office for Official Publications of the European Communities, EUR 20719.

EU (2009) DIRECTIVE 2009/28/EC OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL (20% biofuels). EU Parliament Council.

EU commission (2003) Hydrogen Energy and Fuel Cells – A vision of our Future. EUR 20719EN, Luxembourg: 11-12.

ExxonMobil (2013) The Outlook for Energy: A View to 2040. URI: www.exxonmobil.com/Corporate/files/news_pub_eo.pdf. Cited in June 2013.

Fatsikostas AN & Verykios XE (2004) Reaction network of steam reforming of ethanol over Ni-based catalysts. J Catal 225: 439–452.

Ferreira-Aparicio P, Benito M, Kouachi K & Menad S (2005) Catalysis in membrane reformers: a high-performance catalytic system for hydrogen production from methane. J Catal 231: 331–343.

Fuel Cells and Hydrogen in Finland (2012) Johnson Matthey PLC trading. Fuel Cell Today 13–25.

Furtado AC, Alonso CG, Cantão MP & Fernandes-Machado NRC (2009) Bimetallic catalysts performance during ethanol steam reforming: Influence of support materials. Int J Hydrogen Energy 34 (17): 7189–7196.

Galetti AE, Gomez MF, Arrúa AL & Abello AC (2010) Ni catalysts supported on modified ZnAl2O4 for ethanol steam reforming. App Catal A 380: 40–47.

Gao H, Lin YS & Zhang B (2004) Chemical stability and its improvement of palladium based metallic membranes. Ind Eng Chem Res 43: 6920–6930.

116

Gielens FC, Knibbeler RJJ, Duysinx PFJ, Tong HD, Vorstman MAG & Keurentjes JTF (2006) Influence of steam and carbon dioxide on the hydrogen flux through thin Pd/Ag and Pd membranes. J Membr Sci 279: 176–185.

Haga F, Nakajima T, Miya H & Mishima S (1997) Catalytic properties of supported cobalt catalysts for steam reforming of ethanol. Catal Lett 48: 223–227.

Halman MM & Steinberg M (1999) The science and the source of the greenhouse effect and global climate change. In: Halman MM & Steinberg M (eds) Greenhouse gas carbon dioxide mitigation Science and Technology. CRC Press: 1–13.

Halonen N, Rautio A, Leino AR, Kyllönen T, Tóth G, Lappalainen J, Kordás K, Huuhtanen M, Keiski RL, Sápi A, Szabó M, Kukovecz A, Kónya Z, Kiricsi I, Vajtai R & Ajayan PM (2010) Three dimensional carbon nanotube scaffolds as filters or catalyst support membranes. ACS Nano 4: 2003–2008.

Haryanto A, Fernando S & Murali N (2005) Current status of hydrogen production techniques by steam reforming of ethanol: a review. Energ Fuel 9: 2098–2106.

Hinsberger M & Hoogers G (2003) Stationary Power Generation. In: Hoogers G (ed) Fuel Cell Technology handbook. CRC Press LLC.

Hirai T, Ikenaga N, Miyake T & Suzuki T (2005) Production of hydrogen by steam reforming of glycerin on ruthenium catalyst. Energy Fuels 19: 1761–1762.

Holladay JD, Hu J, King DL & Wang Y (2009). An overview of hydrogen production technologies. Catal Today 139: 244–260.

Holladay JD, Wang Y & Jones E (2004) Review of developments in portable hydrogen production using micro reactor technology. Chem Rev 104: 4767–4790.

Homs N, Llorca J & Piscina RL(2006) Low-temperature steam-reforming of ethanol over ZnO-supported Ni and Cu catalysts: The effect of nickel and copper addition to ZnO-supported Cobalt-based catalysts. Catal Today 116: 361–366.

Hou K & Hughes R (2002) The effect of external mass transfer, competitive adsorption and coking on hydrogen permeation through thin Pd/Ag membranes. J Membrane Sci 206: 119–130.

Hou T, Yuan L, Ye T, Gong L, Tu J, Yamamoto M, Youshifumi T & Li Q (2009) Hydrogen production by low-temperature reforming of organic compounds in bio-oil over a CNT-promoting Ni catalyst. Int J Hydrogen Energy 34 (22): 9095–9107.

Hsieh HP (1996) Inorganic Membranes for Separation and Reaction. Elsevier Science BV. Huber GW, Shabaker JW & Dumesic JA (2003) Raney Ni–Sn catalyst for H2 production

from biomass-derived hydrocarbons. Science 30: 2075–2077. Huuhtanen M, Seelam PK, Turpeinen E, Kolli T & Keiski RL (2013) Chapter 11.

Advances in catalysts for membrane reactors. In: Basile A (ed) Handbook of Membrane reactors. Woodhead publisher: 401–432.

HyperSolar (2013) URI: http://www.hypersolar.com/market.php. Cited in April 2013. Iulianelli A & Basile A (2010) An experimental study on bio-ethanol steam reforming in a

catalytic membrane reactor. Part I: Temperature and sweep-gas flow configuration effects. Int J Hydrogen Energy 35: 3170–3177.

117

Iulianelli A, Longo T, Liguori S, Seelam PK, Keiski RL & Basile A (2009) Oxidative steam reforming of ethanol over Ru/Al2O3 catalyst in a dense Pd-Ag membrane reactor to produce hydrogen for PEM fuel cells. Int J Hydrogen Energy 34: 8558–8565.

Iulianelli A, Seelam PK, Liguori S, Longo T, Keiski RL, Calabrò V & Basile A (2011) Hydrogen production for PEM fuel cell by gas phase reforming of glycerol as byproduct of bio-diesel. The use of a Pd-Ag membrane reactor at middle reaction temperature. Int J Hydrogen Energy 36: 3827–3834.

Jansen D, Dijkstra JW, van den Brink RW, Peters TA, Stange M, Bredesen R, Goldbach A, Xu H, Gottschalk A & Doukelis A (2009) Hydrogen membrane reactors for CO2 capture. Energy Procedia 1: 253–260.

Jensen KF (2001) Microreaction engineering is small better?. Chem Eng Sci 56:293 –303. Job N, Berthon-Fabry S, Chatenet M, Marie J, Brigaudet M, Pirard J-P (2009)

Nanostructured carbons as platinum catalyst support for proton exchange membrane fuel cell electrode. Top Catal 52: 2117–2122.

Jong J D, Lammertink RGH & Wessling M (2006) Membranes and microfluidics: a review. Lab Chip 6: 1125–1139.

Kaltschmitt M & Streicher W (2007) Introduction. In: Kaltschmitt M & Streicher W (eds) Renewable Energy Technology, Economics, and Environment. Berlin, Springer: 1–19.

Kikuchi E & Uemiya S (1991) Preparation of supported thin palladium-silver alloy membranes and their characteristics for hydrogen separation. Gas Sep Pur 5: 261–266.

Kim S & Dale BE (2004) Global potential bioethanol production from wasted crops and crop residues. Biomass Bioenergy 26 (4): 361–375.

Kleijn R & Der Voet EV (2010) Resource constraints in a hydrogen economy based on renewable energy sources: An exploration. Renew Sust Energ Rev 14: 2784–2795.

Le Valant A, Can F, Bion N, Duprez D & Epron F (2010) Hydrogen production from raw bioethanol steam reforming: Optimization of catalyst composition with improved stability against various impurities. Int J Hydrogen Energ 35: 5015–5020.

Le Valant A, Garron, Bion N, Epron F & Duprez D (2008) Hydrogen production from raw bioethanol over Rh/MgAl2O4 catalyst: impact of impurities: heavy alcohol, aldehyde, ester, acid and amine. Catal Today 138: 169–174.

Lee JS, Han GB & Kang M (2012) Low temperature steam reforming of ethanol for carbon monoxide-free hydrogen production over mesoporous Sn-incorporated SBA-15 catalysts. Energy 44 (1): 248–256.

Leino A-R, Mohl M, Kukkola J, Mäki-Arvela P, Kokkonen T, Shchukarev A & Kordas K (2013) Low-temperature catalytic oxidation of multi-walled carbon nanotubes. Carbon 57: 99–107.

Li A, Liang W & Hughes R (2000) The effect of carbon monoxide and steam on the hydrogen permeability of a Pd/stainless steel membrane. J Membrane Sci 165: 135–141.

Li G, Kanezashi M, Tsuru T (2011) Highly enhanced ammonia decomposition in a bimodal catalytic membrane reactor for COx-free hydrogen production. Catal Comm 15: 60–63.

118

Li J, Yu H, Yang G, Peng F, Xie D, Wang H & Yang J (2011) Steam reforming of oxygenate fuels for hydrogen production: A thermodynamic study. Energy Fuels 25: 2643–2650.

Li Y, Lai G-H & Zhou R-X (2007) Carbon nanotubes supported Pt–Ni catalysts and their properties for the liquid phase hydrogenation of cinnamaldehyde to hydrocinnamaldehyde. App Sur Sci 253 (11): 4978–4984.

Liberatori JWC, Ribeiro RU, Zanchet D, Noronha FB & Bueno JMC (2007) Steam reforming of ethanol on supported nickel catalysts. Appl Catal A 327: 197–204.

Liguori S, Pinacci P, Seelam PK, Keiski RL, Drago F, Calabrò V, Basile A & Iulianelli A (2012) Performance of a Pd/PSS membrane reactor to produce high purity hydrogen via WGS reaction. Catal Today 193: 87–94.

Liguras DK, Kondarides DI & Verykios XE (2003) Production of hydrogen for fuel cells by steam reforming of ethanol over supported noble metal catalysts. Appl Catal B 43: 345–354.

Lin SY, Kim BH & Ha SY (2008) Hydrogen production from ethanol steam reforming over supported cobalt catalysts. Catal Lett 122: 295–301.

Llorca J, Homs N, Sales J & Piscina PR (2002) Efficient production of hydrogen over supported cobalt catalysts from ethanol steam reforming. J Catal 209: 306–317.

Llorca J, Piscina PR, Sales J & Homs N (2001) Direct production of hydrogen from ethanolic aqueous solutions over oxide catalysts. Chem Commun 641–642.

López E, Kim J, Shanmugharaj AM & Ryu SH (2012) Multiwalled carbon nanotubes-supported nickel catalysts for the steam reforming of propane. J Mater Sci 47: 2985–2994.

Lymberopoulos N (2005) Hydrogen production from renewables, EU commission DG-TREN, EESD Contract: NNE5-PTA-2002-003/1.

Manzolini G & Tosti S (2008) Hydrogen production from ethanol steam reforming: energy efficiency analysis of traditional and membrane processes. Int J Hydrogen Energy 33: 5571–5582.

Marbán G & Valdés-Solís T (2007) Towards the hydrogen economy?. Int J Hydrogen Energy 32: 1625–1637.

Mardilovich IP, She Y, Ma YH & Rei MH (1998) Defect-free palladium membranes on porous stainless-steel support. AIChE J 44: 310–322.

Mattos LV, Jacobs G, Davis BH & Noronha FB (2012) Production of hydrogen from ethanol: Review of reaction mechanism and catalyst deactivation. Chem Rev 112 (7): 4094–4123.

McLeary EE, Jansen JC & Kapteijn F (2006) Zeolite based films, membranes and membrane reactors: Progress and prospects. Micropor Mesopor Mat 90: 198–220.

McMullen JP & Jensen KF (2010) Integrated microreactors for reaction automation: New approaches to reaction development. Annu Rev Anal Chem 3: 19–42.

Mendes D, Mendes A, Madeira LM, Iulianelli A, Sousa JM & Basile A (2010) The water-gas shift reaction: from conventional catalytic systems to Pd-based membrane reactors – a review. Asia Pac J Chem Eng 5: 111–137.

119

Mendes D, Tosti S, Borgognoni F, Mendes A & Madeira LM (2010) Integrated analysis of a membrane-based process for hydrogen production from ethanol steam reforming. Catal Today 156: 107–117.

Miachon S. Dalmon J-A (2004) Catalysis in membrane reactors: what about the catalyst?. Top Catal 29: 59–65.

Mills PL, Quiram DJ & Ryley JF (2007) Microreactor technology and process miniaturization for catalytic reactions − A perspective on recent developments and emerging technologies. Chem Eng Sci 62: 6992−7010.

Morgenstern DA & Fornango JP (2005) Low-temperature reforming of ethanol over copper-plated raney nickel: A new route to sustainable hydrogen for transportation. Energ Fuels 19: 1708–1716.

Motchelaho AMM (2011) Iron and Cobalt catalysts supported on carbon nanotubes for use in the Fischer-Tropsch synthesis. PhD Thesis, University of Witwatersrand, Faculty of Engineering and the Built Environment, Johannesburg.

Moulijn JA, Stankiewicz A, Grievink J & Gorak A (2006) Process intensification and process system engineering: a friendly symbiosis. Computer Aided Chem Eng 21: 29–37.

Navarro RM, Pena MA & Fierro JL (2007) Hydrogen production reactions from carbon feedstocks: fossil fuels and biomass. Chem Rev 107: 3952–3991.

Ni M, Leung DYC & Leung MKH (2007) A review on reforming bio-ethanol for hydrogen production. Int J Hydrogen Energy 32: 3238–3247.

Ockwig NW & Nenoff TM (2007) Membranes for hydrogen separation. Chem Rev 107: 4078−4110.

Oriňáková R & Oriňák A (2011) Recent applications of carbon nanotubes in hydrogen production and storage. Fuel 90 (11): 3123–3140.

Pan X & Bao X (2008) Reactions over catalysts confined in carbon nanotubes. Chem. Commun 6271–6281.

Panagiotopoulou P& Verykios XE (2012) Mechanistic aspects of the low temperature steam reforming of ethanol over supported Pt catalysts. Int J Hydrogen Energy 37(21): 16333–16345.

Path (2011) Annual Report on World Progress in Hydrogen, A Report by the Partnership for Advancing the Transition to Hydrogen (PATH). URI: http://www.hpath.org/ReportBook.pdf. Cited in Jan 2012.

Pinacci P & Drago F (2012) Influence of the support on permeation of palladium composite membranes in presence of sweep gas. Catal Today 193 (1): 186–193.

PMO, Finland (2009) Government Foresight Report on Long-term Climate and Energy Policy: Towards a Low-carbon Finland, Prime Minister’s Office Publication 30/2009, Helsinki University Print. URI: http://vnk.fi/julkaisukansio/2009/j28-ilmasto-selonteko-j29-klimat-framtidsredogoerelse-j30-climate_/pdf/en.pdf. Cited in Oct 2012.

Profeti LPR, Ticianelli EA & Assaf EM (2009) Production of hydrogen via steam reforming of biofuels on Ni/CeO2–Al2O3 catalysts promoted by noble metals. Int J Hydrogen Energy 34 (12): 5049–5060.

120

Qinglan H, Chang W, Dingqiang L, Yao W, Dan L & Guiju L (2010) Production of hydrogen-rich gas from plant biomass by catalytic pyrolysis at low temperature. Int J Hydrogen Energy 35: 8884–8890.

Rabenstein G & Hacker V (2008) Hydrogen for fuel cells from ethanol by steam-reforming, partial-oxidation and combined auto-thermal reforming: a thermodynamic analysis. J Power Sources 185: 1293–1304.

Rahimpour MA & Asgari A (2009) Production of hydrogen from purge gases of ammonia plants in a catalytic hydrogen-permselective membrane reactor. Int J Hydrogen Energy 34 (14): 5795–5802.

Raissi AT & Block DL (2004) Hydrogen Automotive fuel of the future. IEEE Power Energy Mag: 41–45.

Rass-Hansen J, Johansson R, Møller M & Christensen CH (2008) Steam reforming of technical bioethanol for hydrogen production. Int J Hydrogen Energy 33: 4547–4554.

Rei M-H, Yeh G-T, Tsai Y-H, Kao Y-L, Shiau L-D (2011) Catalysis-spillover-membrane. III: The effect of hydrogen spillover on the palladium membrane reactor in the steam reforming reactions. J Membr Sci 369: 299–307.

Roh H-S, Wang Y, King DL, Platon A & Chin Y-H (2006) Low temperature and H2 selective catalysts for ethanol steam reforming. Catal Lett 108: 15–19.

Roses L, Gallucci F, Manzolini G, Campanari S & Annaland MVS (2011) Comparison between fixed bed and fluidized bed membrane reactor configurations for PEM based micro-cogeneration systems. Chem Eng J 171: 1415–1427.

Sanchez Marcano JG & Tsotsis ThT (Eds.) (2002). Catalytic membranes and membrane reactors. Wiley-VCH.

Sánchez MC, Navarro RM & Fierro JLG (2007) Ethanol steam reforming over Ni/MxOy–Al2O3 (M=Ce, La, Zr and Mg) catalysts: Influence of support on the hydrogen production. Int J Hydrogen Energy 32:1462–1471.

Sansonetti S, Curcio S, Calabrò V & Iorio G (2009) Bio-ethanol production by fermentation of ricotta cheese whey as an effective alternative non-vegetable source. Biomass and Bioenergy 33 (12): 1687–1692.

Seelam PK, Huuhtanen M & Keiski RL (2013a) Microreactors and membrane microreactors: fabrication and applications. In: Basile A (ed) Handbook of Membrane reactors. Woodhead Publishing Limited: 188–235.

Seelam PK, Huuhtanen M, Sápi A, Kordás K, Szabó M, Turpeinen E & Keiski RL (2010) CNT-based catalysts for hydrogen production by ethanol reforming. Int J Hydrogen Energy 35: 12588–12595.

Seelam PK, Leino A, Turpeinen E, Kordás K, Huuhtanen M & Keiski RL (2013b) Low temperature steam reforming of ethanol over advanced carbon nanotube based catalysts. Submitted to Int J Hydrogen Energy.

Seelam PK, Liguori S, Iulianelli A, Pinacci P, Calabrò V, Huuhtanen M, Keiski R, Piemonte V, Tosti S, Falco M & Basile A (2012) Hydrogen production from bio-ethanol steam reforming reaction in a Pd/PSS membrane reactor. Catal Today 193: 42–48.

121

Seelam PK, Liguori S, Iulianelli A, Pinacci P, Calabrò V, Huuhtanen M, Keiski R, Piemonte V, Tosti S, Falco M & Basile A (2011) Hydrogen production from bio-ethanol steam reforming reaction in a PSS supported palladium membrane reactor. PS-19, Book of abstracts, ICCMR10 Conference, St. Petersburg, Russia, 154–155.

Serp P, Corrias M & Kalck P (2003) Carbon nanotubes and nanofibers in catalysis. Appl Catal A 253: 337–358.

Shirasaki Y, Tsuneki T, Ota Y, Yasuda I, Tachibana S, Nakajima H & Kobayashi K (2009) Development of membrane reformer system for highly efficient hydrogen production from natural gas. Int J Hydrogen Energy 34 (10): 4482–4487.

Shu J, Grandjean BPA, Van Neste A & Kaliaguine S (1991) Catalytic palladium-based membrane reactors: A Review. Can J Chem Eng 69: 1036–1060.

Silva ALD, Malfatti CF & Muller IL (2009) Thermodynamic analysis of ethanol steam reforming using Gibbs energy minimization method: A detailed study of the conditions of carbon deposition. Int J Hydrogen Energy 34: 4321–4330.

Silveira JL, Braga LB, Souza ACC, Antunes JS & Zanzi R (2009) The benefits of ethanol use for hydrogen production in urban transportation. Renew Sust Energ Rev 13: 2525–2534.

Slinn M, Kendall K, Mallon C & Andrews J (2008) Steam reforming of biodiesel by-product to make renewable hydrogen. Bioresour Technol 99: 5851–5858.

Somorjai GA, Frei H & Park JY (2009) Advancing the frontiers in nanocatalysis, biointerfaces, and renewable energy conversion by innovations of surface techniques, J Am Chem Soc 131: 16589–16605.

Sperling D & Cannon JS (Eds.) (2004) The Hydrogen Energy Transition: Moving Toward the Post Petroleum age in Transportation. New York, Elsevier–Academic.

Stair PC, Marshall C, Xiong G, Feng H, Pellin MJ, Elam JW, Curtiss L, Iton L, Kung H, Kung M & Wang H-H (2006) Novel, uniform nanostructured catalytic membranes. Top Catal 39: 181–186.

Stambouli AB & Traversa E (2002) Fuel cells an alternative to standard sources of energy. Renew Sust Energ Rev 6: 295–304.

Sun J, Qiu XP, Wu F & Zhu WT (2005) H2 from steam reforming of ethanol at low temperature over Ni/Y2O3, Ni/La2O3 and Ni/Al2O3 catalysts for fuel-cell application. Int J Hydrogen Energy 30: 437–445.

Sun S, Yan W, Sun P & Chen J (2012) Thermodynamic analysis of ethanol reforming for hydrogen production. Energy 44 (1): 911–924.

Sun Y & Cheng J (2002) Hydrolysis of lignocellulosic materials for ethanol production: a review. Bioresource Technol 83: 1–11.

Tanksale A, Beltramini JN & Lu GM (2010) A review of catalytic hydrogen production processes from biomass. Renew Sust Energ Rev 14: 166–182.

The Outlook for Energy (2013) A View to 2040. URI: www.exxonmobil.com/Corporate/Files/news_pub_eo2013.pdf.

Timilsina GR & Shrestha RM (2006) General equilibrium effects of a supply side GHG mitigation option under the clean development mechanism, J Environ Manage 80: 327–341.

122

Tong HD (2004) Microfabricated palladium-based membranes for hydrogen separation. PhD Thesis. University of Twente, Transducers Science and Technology of the MESA+ Research Institute.

Tosti S & Bettinali L (2004) Diffusion bonding of Pd–Ag membranes. J Mater Sci 39: 3041–3046.

Tosti S, Basile A, Borgognoni F, Capaldo V, Cordiner S, Di Cave S, Gallucci F, Rizzello C, Santucci A & Traversa E (2008) Low temperature ethanol steam reforming in a Pd-Ag membrane reactor: Part 1: Ru-based catalyst. J Membr Sci 308: 250–257.

U.S. DOE (2012) Fuel Cell Technologies Program, Current U.S. Hydrogen Production. URI: http://www.hydrogen.energy.gov/pdfs/12014_current_us_hydrogen_production.pdf. Cited in June 2013.

U.S. DOE, Office of Energy Efficiency and Renewable Energy, Alternative Fuels Data Center. URI: http://www.afdc.energy.gov/fuels/properties.html. Cited in June 2013.

Vaidya PD & Rodrigues AE (2006) Insight into steam reforming of ethanol to produce hydrogen for fuel cells. Chem Eng J 117: 39–49.

Wang WP, Thomas S, Zhang XL, Pan XL, Yang WS & Xiong GX (2006) H2/N2 gaseous mixture separation in dense Pd/α-Al2O3 hollow fiber membranes: experimental and simulation studies. Sep Purif Technol 52: 177–185.

Wilhite BA, Weiss SE, Ying JY, Schmidt MA & Jensen KF (2006) High-purity hydrogen generation in a microfabricated 23 wt% Ag–Pd membrane device integrated with 8:1 LaNi0.95Co0.05O3/Al2O3 catalyst. Adv Mater 18: 1701–1704.

Xu W, Kong JS, Yeh Y-TE &, Chen P (2008) Single-molecule nanocatalysis reveals heterogeneous reaction pathways and catalytic dynamics. Nat Mater 7: 992–996.

Yang HM, Liao PH (2007) Preparation and activity of Cu/ZnO-CNTs nano-catalyst on steam reforming of methanol. Appl Catal A 317: 226–233.

Yang Y, Ma J & Wu F (2006) Production of hydrogen by steam reforming of ethanol over a Ni/ZnO catalyst. Int J Hydrogen Energ 31: 877–882.

Yun S & Oyama ST (2011) Correlations in palladium membranes for hydrogen separation: A review, J Membrane Sci 375: 28–45.

Zakkour P & Cook G (2010) CCS industry roadmap-high purity CO2 sources: final draft. Sectoral assessment carbon counts company (UK) Ltd. URI: http://www.globalccsinstitute.com/node/42391. Cited in June 2013.

Zgolicz PD, Stassi JP, Yañez MJ, Scelza OA & Miguel SRD (2012) Influence of the support and the preparation methods on the performance in citral hydrogenation of Pt-based catalysts supported on carbon nanotubes. J Catal 290: 37–54.

Zhang B, Tang X, Li Y, Xu Y & Shen W (2007) Hydrogen production from steam reforming of ethanol and glycerol over ceria-supported metal catalysts. Int J Hydrogen Energy 32: 2367–2373.

Zhang H, Yang D, Li B, Qian Z & Ma J (2013) Investigation of the temperature-related performance of proton exchange membrane fuel cell stacks in the presence of CO. Int J Energ Res. DOI: 10.1002/er.3028.

123

Original papers

I Seelam PK, Huuhtanen M, Sápi A, Szabó M, Kordás K, Turpeinen E, Tóth G & Keiski RL (2010) CNT-based catalysts for H2 production by ethanol reforming. International J Hydrogen Energy 35: 12588–12595.

II Seelam PK, Rautio AR, Huuhtanen M, Turpeinen E, Kordás K & Keiski RL (2013) Low temperature steam reforming of ethanol over advanced carbon nanotube based catalysts. Manuscript.

III Seelam PK, Liguori S, Iulianelli A, Pinacci P, Calabrò V, Huuhtanen M, Keiski R, Piemonte V, Tosti S, De Falco M & Basile A (2012) Hydrogen production from bioethanol steam reforming reaction in a Pd/PSS membrane reactor. Catalysis Today 193: 42–48.

IV Iulianelli A, Seelam PK, Liguori S, Longo T, Keiski R, Calabrò V & Basile A (2011) Hydrogen production for PEM fuel cell by gas phase reforming of glycerol as byproduct of bio-diesel. The use of a Pd-Ag membrane reactor at middle reaction temperature. International J Hydrogen Energy 36: 3827–3834.

V Liguori S, Pinacci P, Seelam PK, Keiski R, Drago F, Calabrò V, Basile A & Iulianelli A (2012) Performance of a Pd/PSS membrane reactor to produce high purity hydrogen via WGS reaction. Catalysis Today 193: 87–94.

VI Seelam PK, Huuhtanen M & Keiski RL (2013) Chapter 5. Microreactors and membrane microreactors: fabrication and applications, In: Basile A (ed) Handbook of membrane reactors, Volume 2: Reactor types and industrial applications. Woodhead Publishing Series in Energy No. 56. Cambridge UK, Woodhead Publishing: 188–235.

VII Huuhtanen M, Seelam PK, Kolli T, Turpeinen E & Keiski RL (2013) Chapter 11.Advances in catalysts for membrane reactors, In: Basile A (ed) Handbook of membrane reactors, Volume 1: Fundamental materials science, design and optimization. Woodhead Publishing Series in Energy No. 55. Cambridge UK, Woodhead Publishing: 401–432.

Reprinted with permission from Elsevier Limited (publications I, II, III, IV and

V) and Woodhead Publishing Limited for original publications VI and VII.

Original publications are not included in the electronic version of the dissertation.

124

A C T A U N I V E R S I T A T I S O U L U E N S I S

Book orders:Granum: Virtual book storehttp://granum.uta.fi/granum/

S E R I E S C T E C H N I C A

457. Nardelli, Pedro Henrique Juliano (2013) Analysis of the spatial throughput ininterference networks

458. Ferreira, Denzil (2013) AWARE: A mobile context instrumentation middlewareto collaboratively understand human behavior

459. Ruusunen, Mika (2013) Signal correlations in biomass combustion – aninformation theoretic analysis

460. Kotelba, Adrian (2013) Theory of rational decision-making and its applications toadaptive transmission

461. Lauri, Janne (2013) Doppler optical coherence tomography in determination ofsuspension viscosity

462. Kukkola, Jarmo (2013) Gas sensors based on nanostructured tungsten oxides

463. Reiman, Arto (2013) Holistic work system design and management : — aparticipatory development approach to delivery truck drivers’ work outside thecab

464. Tammela, Simo (2013) Enhancing migration and reproduction of salmonid fishes :method development and research using physical and numerical modelling

465. Yadav, Animesh (2013) Space-time constellation and precoder design underchannel estimation errors

466. Schaberreiter, Thomas (2013) A Bayesian network based on-line risk predictionframework for interdependent critical infrastructures

467. Halonen, Niina (2013) Synthesis and applications of macroscopic well-alignedmulti-walled carbon nanotube films

468. Remes, Jukka (2013) Method evaluations in spatial exploratory analyses ofresting-state functional magnetic resonance imaging data

469. Oravisjärvi, Kati (2013) Industry and traffic related particles and their role inhuman health

470. Czajkowski, Jakub (2013) Optical coherence tomography as a characterizationmethod in printed electronics

471. Haapalainen, Mikko (2013) Dielectrophoretic mobility of a spherical particle in2D hyperbolic quadrupole electrode geometry

472. Bene, József Gergely (2013) Pump schedule optimisation techniques for waterdistribution systems

C473etukansi.kesken.fm Page 2 Tuesday, October 22, 2013 12:07 PM

ABCDEFG

UNIVERSITY OF OULU P .O. B 00 F I -90014 UNIVERSITY OF OULU FINLAND

A C T A U N I V E R S I T A T I S O U L U E N S I S

S E R I E S E D I T O R S

SCIENTIAE RERUM NATURALIUM

HUMANIORA

TECHNICA

MEDICA

SCIENTIAE RERUM SOCIALIUM

SCRIPTA ACADEMICA

OECONOMICA

EDITOR IN CHIEF

PUBLICATIONS EDITOR

Professor Esa Hohtola

University Lecturer Santeri Palviainen

Postdoctoral research fellow Sanna Taskila

Professor Olli Vuolteenaho

University Lecturer Hannu Heikkinen

Director Sinikka Eskelinen

Professor Jari Juga

Professor Olli Vuolteenaho

Publications Editor Kirsti Nurkkala

ISBN 978-952-62-0276-1 (Paperback)ISBN 978-952-62-0277-8 (PDF)ISSN 0355-3213 (Print)ISSN 1796-2226 (Online)

U N I V E R S I TAT I S O U L U E N S I SACTAC

TECHNICA

U N I V E R S I TAT I S O U L U E N S I SACTAC

TECHNICA

OULU 2013

C 473

Prem Kumar Seelam

HYDROGEN PRODUCTION BY STEAM REFORMING OF BIO-ALCOHOLSTHE USE OF CONVENTIONAL AND MEMBRANE-ASSISTED CATALYTIC REACTORS

UNIVERSITY OF OULU GRADUATE SCHOOL;UNIVERSITY OF OULU,FACULTY OF TECHNOLOGY,DEPARTMENT OF PROCESS AND ENVIRONMENTAL ENGINEERING ,MASS AND HEAT TRANSFER PROCESS LABORATORY

C 473

ACTA

Prem K

umar Seelam

C473etukansi.kesken.fm Page 1 Tuesday, October 22, 2013 12:07 PM