C473etukansi.kesken.fm Page 1 Tuesday, October 22, 2013 12:07 PM
C 473
OULU 2013
UNIV ER S IT Y OF OULU P. O. B R[ 00 FI-90014 UNIVERSITY OF OULU FINLAND
U N I V E R S I TAT I S
S E R I E S
SCIENTIAE RERUM NATURALIUM
Professor Esa Hohtola
HUMANIORA
University Lecturer Santeri Palviainen
TECHNICA
Postdoctoral research fellow Sanna Taskila
MEDICA
Professor Olli Vuolteenaho
ACTA
UN
NIIVVEERRSSIITTAT
ATIISS O
OU
ULLU
UEEN
NSSIISS
U
Prem Kumar Seelam
E D I T O R S
Prem Kumar Seelam
A
B
C
D
E
F
G
O U L U E N S I S
ACTA
A C TA
C 473
HYDROGEN PRODUCTION
BY STEAM REFORMING OF
BIO-ALCOHOLS
THE USE OF CONVENTIONAL AND MEMBRANEASSISTED CATALYTIC REACTORS
SCIENTIAE RERUM SOCIALIUM
University Lecturer Hannu Heikkinen
SCRIPTA ACADEMICA
Director Sinikka Eskelinen
OECONOMICA
Professor Jari Juga
EDITOR IN CHIEF
Professor Olli Vuolteenaho
PUBLICATIONS EDITOR
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)
UNIVERSITY OF OULU GRADUATE SCHOOL;
UNIVERSITY OF OULU,
FACULTY OF TECHNOLOGY,
DEPARTMENT OF PROCESS AND ENVIRONMENTAL ENGINEERING ,
MASS AND HEAT TRANSFER PROCESS LABORATORY
C
TECHNICA
TECHNICA
ACTA UNIVERSITATIS OULUENSIS
C Te c h n i c a 4 7 3
PREM KUMAR SEELAM
HYDROGEN PRODUCTION BY STEAM
REFORMING OF BIO-ALCOHOLS
The use of conventional and membrane-assisted
catalytic reactors
Academic dissertation to be presented with the assent of
the Doctoral Training Committee of Technology and
Natural Sciences of the University of Oulu for public
defence in OP-Pohjola-sali (Auditorium L6), Linnanmaa,
on 4 December 2013, at 12 noon
U N I VE R S I T Y O F O U L U , O U L U 2 0 1 3
Copyright © 2013
Acta Univ. Oul. C 473, 2013
Supervised by
Professor Riitta L. Keiski
Docent Mika Huuhtanen
Reviewed by
Professor Jordi Llorca
Professor Lars J. Pettersson
Opponent
Associate Professor Yohannes Kiros
ISBN 978-952-62-0276-1 (Paperback)
ISBN 978-952-62-0277-8 (PDF)
ISSN 0355-3213 (Printed)
ISSN 1796-2226 (Online)
Cover Design
Raimo Ahonen
JUVENES PRINT
TAMPERE 2013
Seelam, Prem Kumar, Hydrogen production by steam reforming of bio-alcohols.
The 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
Acta Univ. Oul. C 473, 2013
University 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 growth
and urbanization. There is a need for alternative and non-conventional energy sources, which are
CO2-neutral, and a need to produce less or no environmental pollutants and to have high energy
efficiency. 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 a
potential fuel and clean energy carrier being at the same time a carbon-free element. Moreover, H2
is utilized in many processes in chemical, food, metallurgical, and pharmaceutical industry and it
is also a valuable chemical in many reactions (e.g. refineries). Non-renewable resources have been
the major feedstock for H2 production for many years. At present, ~50% of H2 is produced via
catalytic steam reforming of natural gas followed by various down-stream purification steps to
produce ~99.99% H2, the process being highly energy intensive. Henceforth, bio-fuels like
biomass derived alcohols (e.g. bio-ethanol and bio-glycerol), can be viable raw materials for the
H2 production. In a membrane based reactor, the reaction and selective separation of H2 occur
simultaneously in one unit, thus improving the overall reactor efficiency. The main motivation of
this work is to produce H2 more efficiently and in an environmentally friendly way from bioalcohols with a high H2 selectivity, purity and yield.
In this thesis, the work was divided into two research areas, the first being the catalytic studies
using 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) and
the water-gas-shift (WGS) reactions in a membrane reactor (MR) using dense and composite Pdbased membranes to produce high purity H2. CNTs were found to be promising support materials
for the low temperature reforming compared to conventional catalyst supports, e.g. Al2O3. The
metal/metal oxide decorated CNTs presented active particles with narrow size distribution and
small size (~2–5 nm). The ZnO promoted Ni/CNT based catalysts showed the highest H2
selectivity of ~76% with very low CO selectivity <1%. Ethanol was shown to be a more suitable
and viable source for H2 than glycerol. The dense Pd-Ag membrane had higher selectivity but a
lower permeating flux than the composite membrane. The MR performance is also dependent on
the 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 a
potential technology in pure H2 production. The high purity of H2 gas with a CO-free reformate
for 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 reaktorit
Oulun yliopiston tutkijakoulu; Oulun yliopisto, Teknillinen tiedekunta, Prosessi- ja
ympäristötekniikan osasto, Lämpö- ja diffuusiotekniikan laboratorio
Acta Univ. Oul. C 473, 2013
Oulun yliopisto, PL 8000, 90014 Oulun yliopisto
Tiivistelmä
Maailman energiankulutus on kasvussa räjähdysmäisen väestönkasvun ja voimakkaan kaupungistumisen myötä. Tällä hetkellä energian tuottamisen aiheuttamat ympäristöongelmat ja taloudellinen epävarmuus ovat seikkoja, joiden ratkaisemiseksi tarvitaan vaihtoehtoisia ja ei-perinteisiä energialähteitä, joilla on korkea energiasisältö ja jotka tuottavat vähän hiilidioksidipäästöjä.
Eräs vaihtoehtoisista lähestymistavoista on vetytalous yhdistettynä polttokennotekniikkaan, minkä on esitetty helpottavan siirtymistä kestävään yhteiskuntaan. Vety on puhdas ja hiilivapaa polttoaine ja energian kantaja. Lisäksi vetyä käytetään monissa prosesseissa kemian-, elintarvike-,
metalli- ja lääketeollisuudessa ja se on arvokas kemikaali monissa prosesseissa (mm. öljynjalostamoissa). Uusiutumattomat luonnonvarat ovat olleet tähän saakka merkittävin vedyn tuotannon
raaka-aine. Tällä hetkellä noin 50 % vedystä tuotetaan maakaasun katalyyttisellä höyryreformoinilla. Puhtaan (yli 99,99 %) vedyn tuotanto vaatii kuitenkin useita puhdistusvaiheita, jotka ovat
erittäin energiaintensiivisiä. Integroimalla reaktio- ja puhdistusvaihe samaan yksikköön (membraanireaktori) saavutetaan huomattavia kustannussäästöjä. Biopolttoaineet, kuten biomassapohjaiset alkoholit (bioetanoli ja bioglyseroli), ovat vaihtoehtoisia lähtöaineita vedyn valmistuksessa. Tämän työn tavoitteena on tuottaa vetyä bioalkoholeista tehokkaasti (korkea selektiivisyys ja
saanto) ja ympäristöystävällisesti.
Tutkimus on jaettu kahteen osaan, joista ensimmäisessä tutkittiin etanolin katalyyttistä höyryreformointia matalissa lämpötiloissa (<450 °C) hyödyntämällä metallipinnoitettuja hiilinanoputkia. Työn toisessa osassa höyryreformointia ja vesikaasun siirtoreaktioa tutkittiin membraanireaktorissa käyttämällä vedyn tuotantoon tiheitä palladiumpohjaisia kalvoja sekä huokoisia palladiumkomposiittikalvoja. Hiilinanoputket (CNT) havaittiin lupaaviksi katalyyttien tukimateriaaleiksi verrattuna tavanomaisesti valmistettuihin tukiaineisiin, kuten Al2O3. CNT-tukiaineelle
pinnoitetuilla aktiivisilla aineilla (metalli-/metallioksidit) todettiin olevan pieni partikkelikoko
(~2–5 nm) ja kapea partikkelikokojakauma. Sinkkioksidin (ZnO) lisäyksellä Ni/CNT-katalyytteihin saavutettiin korkea vetyselektiivisyys (~76 %) ja erittäin alhainen hiilimoksidiselektiivisyys (<1 %). Etanolin todettiin olevan parempi vedyn raaka-aine kuin glyserolin. Tiheillä PdAg-kalvoilla havaittiin olevan vedyn suhteen korkeampi selektiivisyys mutta matalampi vuo
verrattuna palladiumkomposiittikalvoihin. Membraanireaktorin suorituskyky oli riippuvainen
myös katalyytin aktiivisuudesta, joten sekä kalvolla että katalyyttimateriaalilla oli merkittävä
rooli kyseisessä reaktorirakenteessa. Yhteenvetona voidaan todeta, että membraanierotukseen
perustuva reformointiyksikkö on huomattavasti perinteistä reformeriyksikköä suorituskykyisempi mahdollistaen tehokkaan teknologian puhtaan vedyn tuottamiseksi. Membraanitekniikalla
tuotettua 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
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 socalled 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
9
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 Action543 (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
10
Oulu, 10.10.2013
List of acronyms and symbols
AC
AFC
ATR
BESR
BET
BGSR
BJH
CCS
CDW
CHP
CMR
CNT
CR
CWM
DWCNT
EDX
ELP
EU
EW
EWAG
FBR
FC
FTIR
GC
GCB
GHG
GHSV
HC
HPLC
HPP
HRF
HT
HTS
I.D.
ICE
Activated carbon
Alkaline fuel cell
Autothermal reforming
Bio-ethanol steam reforming
Brunauer-Emmett-Teller
Bio-glycerol stream reforming
Barrett-Joyner-Halenda
Carbon capture and storage
Cold diffusion welding
Combined heat and power
Catalytic membrane reactor
Carbon nanotube
Conventional reactor (or reformer)
Catalytic wall microreactor
Double walled carbon nanotube
Energy dispersive X-ray spectroscopy
Electroless plating
European Union
Ethanol-water mixture
Ethanol-water-acetic acid-glycerol mixture
Fixed bed reactor
Fuel cell
Fourier Transform Infrared Spectroscopy
Gas chromatography
Graphitic carbon black
Greenhouse gases
Gas-hourly-space-velocity (h-1)
Hydrocarbon
High-performance liquid chromatography
Hydrogen permeate purity
Hydrogen recovery factor
High temperature
High temperature shift
Inner diameter (µm or mm or nm)
Internal combustion engine
11
IGCC
IMR
IMRCF
IR
LHV
LT
LT-FC
LTS
MEMS
MF
MM
MMR
MR
MS
MWCNT
NG
NP
O.D.
PBMR
PBR
PEM
PEMFC
PI
PMO
POX
PP
PSA
PSS
RT
SOFC
SR
SRE
SRM
SS
STP
SWCNT
TCD
12
Integrated gasification combined cycle
Inert membrane reactor
Inert membrane reactor with a catalyst on the feed side
Infra-red
Lower heating value (MJ/kg)
Low temperature (°C)
Low temperature-fuel cell
Low temperature shift
Microelectromechanical systems
Microfabrication
Micromembrane
Membrane microreactor
Membrane reactor (or) reformer
Microstructure
Multi-walled carbon nanotube
Natural gas
Nanoparticle
Outer diameter (µm or mm or nm)
Packed bed membrane reactor
Packed bed reactor
Proton exchange membrane
Proton exchange membrane fuel cell
Process intensification
Prime Minster Office (Finland)
Partial oxidation
Partial pressure (bar or kPa)
Pressure swing adsorption
Porous stainless steel
Room temperature (°C)
Solid oxide fuel cell
Steam reforming
Steam reforming of ethanol
Steam reforming of methane
Stainless steel
Standard temperature pressure
Single walled carbon nanotube
Thermal conductivity detector
TEM
TOS
TR
WGS
WHSV
XRD
Transmission electron micrograph
Time-on-stream (min)
Traditional reactor (or reformer)
Water-gas-shift
Weight hourly space velocity (h-1)
X-ray powder diffraction method
Symbols
abs.
As
Cd
CH2
Cs
DH2
dTEM
dXRD
Ea
Fglycerol, in
F´acetic acid, in
F´ethanol, in
F´glycerol, in
FCH4, p
FCH4, p+r
FCO, in
FCO, out
FCO, p
FCO, p+r
FCO2, p
FCO2, p+r
Fethanol, in
Fethanol, out
FH2, p
FH2, r
Fi
Absolute pressure (bar or kPa)
Membrane cross section area (m2)
Concentration of H2 in metal layer at desorption (mol m-3)
Concentration of H2 in membrane sub layers (mol m-3)
Concentration of H2 in metal layer at surface adsorption (mol m-3)
Diffusion coefficient of hydrogen (mol m-2)
Average nanoparticle size measured by TEM (nm)
Average nanoparticle size measured by XRD (nm)
Activation energy (kJ mol-1)
Glycerol inlet feed flow rate (mol min-1) (for BGSR, in MR)
Acetic acid inlet feed flow rate (mol min-1) (in MR)
Ethanol inlet feed flow rate (mol min-1) (in MR)
Glycerol inlet feed flow rate (mol min-1) (for BESR in MR)
Molar flow rate of CH4 in permeate (mol min-1)
Sum of CH4 molar flow rates at both permeate+retentate (mol min-1)
Inlet flow rate of CO (mol min-1)
Outlet flow rate of CO (mol min-1)
Molar flow rate of CO in permeate (mol min-1)
Sum of the molar flow rates in both permeate and retentate CO (mol
min-1)
Molar flow rate of CO2 in permeate (mol min-1)
Sum of molar flow rates of CO2 on permeate and retentate sides
(mol min-1)
Ethanol inlet feed flow rate (mol min-1)
Ethanol outlet feed flow rate (mol min-1)
Hydrogen molar flow rate at permeate side (mol min-1)
Hydrogen molar flow rate at retentate side (mol min-1)
Flow rate of product ‘i’ (i = H2,CO,CO2,CH4,CH3CHO) (mol min-1)
13
Fi-total
Fsweep-gas
JH2
k
ka
kd
KH
ki
ko
L
LHVH2
M’feed
Mcat
n
pav
pd
Pe,H2
Pe0
Ple
ppermeate
preaction
pretentate
ps
pTotal
R
R2
S/E
S’i
SBET
SH
Sp
T
Vcat
V´feed
YH2
ΔH°
14
Total gas flow rate ‘i’ (retentate + permeate flow, i = H2, CO2, CO,
CH4) (mol min-1)
Sweep-gas N2 flow rate (mol min-1) (for MR-only)
Hydrogen permeating flux through the membrane (mol m2 s-1)
Hydrogen concentration, corresponds to n´ = 1 (mol m-3)
Adsorption rate constant (mol m-2 s-1 kPa-1)
Desorption rate constant (mol m-2 s-1)
Henry’s law constant
Hydrogen dissolution rate constant into the membrane (mol m-2 s-1)
Hydrogen dissolution rate constant out of membrane (mol m-2 s-1)
Length (cm or mm)
Lower heating value of hydrogen (MJ kg-1)
Mass flow rate of the feed (g h-1)
Mass of the catalyst pellet (g)
Pressure exponential factor
Average pressure: pretentate+ ppermeate/2 (bar)
Hydrogen partial pressure at desorption (Pa)
Hydrogen permeability (mol m-1 s-1 Pa-0.5)
Pre-exponential factor (mol m-1 s-1 Pa-0.5)
Permeance (mol m-2 s-1 Pa-1)
Hydrogen partial pressure in the permeate side (bar)
Reaction pressure (bar)
Hydrogen partial pressure in the retentate side (bar)
Hydrogen partial pressure at surface adsorption (Pa)
Total pressure: retentate + permeate (bar)
Gas constant (kJ K-1 mol-1)
Regression coefficient (-)
Steam-to-ethanol molar ratio
Selectivity of product gases i = H2, CO2, CO and CH4 (-)
Specific surface area (m2g-1)
Solubility coefficient of hydrogen (ml atm mol-1)
Selectivity of products (-)
Reaction temperature (°C)
Volume of the catalyst pellet (cm3)
Volumetric feed flow rate (cm3 h-1)
Yield of hydrogen (-)
Standard enthalpy of the reaction at 298 K (kJ mol-1)
Δp
Pressure difference (kPa or bar)
Greek letters
αH2/gas_i
αPd
βPd
δ
θH
ρb
Ideal selectivity of H2 with respect to other gases ‘i’ (N2, He)
Crystalline phase of Pd-H hydride at low H atomic concentration
Crystalline phase of Pd-H hydride at high H atomic concentration
Thickness of Pd layer (µm)
Surface coverage of Pd site by H-atom
Catalyst bulk density (g cm-3)
15
16
List of original publications related to thesis
I
II
III
IV
V
VI
VII
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.
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.
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.
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.
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.
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.
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.
17
18
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
19
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
20
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
21
& 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
22
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).
Efficiency (%) # Maturity
Technology
Feed stock
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
23
1.2
Thesis scope and its objectives
The main scope of this thesis is to study the hydrogen production from biomassderived 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 COfree 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
24
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).
25
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
Main motivation to work on hydrogen production
by ethanol reforming
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
SRE at low temperatures (< 450 °C) over Pt and
over
ZnO promoted Ni/CNT catalysts. Influence of Pt
advanced
carbon
nanotube
based
and zinc oxide on Ni/CNT. The stability of
catalysts
monometallic and promoted catalysts. Catalysts
performance evaluation for CO reduction and
enhancing the H2 production.
Paper III
Hydrogen production from bio-ethanol steam
A
reforming reaction in a Pd/PSS membrane
mixtures with impurities in Pd/PSS MR at 400 °C.
study
of
simulated
crude
ethanol/water
reactor
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
Viability of producing hydrogen from glycerol in a
phase reforming of glycerol as a byproduct in
Pd-Ag MR at 400 °C. A comparison of a
bio-diesel production. The use of a Pd-Ag
traditional versus a membrane assisted reformer.
membrane
The effect of WHSV and reaction pressure over
reactor
at
middle
reaction
temperatures
MR performance to produce pure or CO-free H2
gas with > 99.99% purity.
Paper V
Performance of a Pd/PSS membrane reactor to
Permeation
produce high purity hydrogen via WGS reaction
mixtures in a Pd/PSS membrane. The potentiality
analysis
of
pure
and
gaseous
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.
26
Table 3. The list of supplementary papers – critical reviews on the future membraneassisted reactors applications.
Supplementary papers
Critical reviews on future trends
(review chapters)
(supporting information)
Paper VI
Microreactors and membrane microreactors:
The review article provides an overview on the
fabrication and applications
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.
27
28
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 byproducts. 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, food29
(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 bioderived 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
30
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, bioethanol 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).
31
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
H/C
formula
Ethanol
Glycerol
Methanol
C2H5OH
C3H8O3
CH3OH
Moles
H2
3
2.7
4
3
4
2
LHV* Availability, source,
Comments
(MJ/kg) toxicity, etc.
27
16
20
Abundant, biomass,
Viable for onboard, onsite
non-toxic
H2
Abundant, biodiesel
Viable, highly viscous than
by-product, biomass
EtOH
Abundant, syn-gas,
Viable, widely studied
biomass, toxic
Butanol
Methane
Hydrogen
C4H9OH
CH4
H2
2.5
4
-
5
2
1
34
47
120
Unclear, biomass,
Viable, 85% direct use in
toxic
ICE, alternative to gasoline
Limited NG reserves,
Viable, commercially
bio-gas
compressed NG
Abundant, not in a free
Clean energy carrier,
form, highly
combust to produce clean
combustible
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 nonrenewable sources. In a similar manner, alcohols like methanol and polyols like
32
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
33
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).
∆H°298K (kJmol-1)
No.
Reaction
Equation
R1
Overall SRE
CH3 CH2 OH+ 3H2 O ⇌ 6H2 +2CO2
R2
Ethanol dehydrogenation
CH3 CH2 OH ⇌ CH3 CHO +H2
68
R3
Acetaldehyde reforming
CH3 CHO+H2 O ⇌ 3H2 +2CO
187
R4
Ethanol decomposition
CH3 CH2 OH → CH4 +CO +H2
R5
Acetaldhyde decomposition
CH3 CHO → CH4 +CO
R6
Ethanol dehydration
CH3 CH2 OH ⇌C2 H4 +H2 O
R7
WGS reaction
CO+H2 O ⇌ CO2 +H2
−41
R8
Ethylene polymerization
C2 H4 → polymers →carbon
−172
R9
Boudouard reaction
2CO ⇌ CO2 +C
−172
R10
Methane decomposition
CH4 ⇌ 2H2 +C
75
256
49
−19
45
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,
34
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 byproducts 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.
35
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.
mol
4.0
A
3.5
H2O(g)
3.0
H2(g)
2.5
2.0
CH4(g)
1.5
1.0
CO2(g)
C
0.5
CO(g)
0.0
100
200
300
400
500
600
Temperature
C
Temperature (°C)
mol
15
B
14
13
H2O(g)
12
11
10
9
8
7
6
5
H2(g)
4
3
2
CO2(g)
CH4(g)
1
0
100
CO(g)
200
300
400
500
600
Temperature
C
Temperature
(°C)
Temperature (°C)
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.
36
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 multiwalled 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
37
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
38
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 Pdmembrane 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.
39
Fig. 6. Schematic presentation of a traditional reforming process versus a membraneassisted 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
40
and investigated for many years by several researchers (Kikuchi & Uemiya 1991,
Shu et al. 1991).
A
Pretreatment
and
prereforming
units
Steam
reforming
WGS-HT
reactor
WGS-LT
reactor
Partial
oxidation
PSA for
separation
Compressi
on 99.999%
H2
PEM Fuel
cell
processor
Electricity
and heat
(CHP)
B
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)
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 αphase 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
41
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 = -DH2
JH2 =
DH2
δ
CH2
δ
(Cs -Cd ),
(1)
(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)).
42
JH2 = kd θ
2
2
− ka p'H2 (1 − θ ) ,
(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 = ko n' (1 − θ ) − ki θ (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
(p0.5
− p0.5
)
s
d
(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 = kDH
ko
kd 0.5
ki
ka
,
(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
kd 0.5
ki
ka
.
(7)
The Henry’s law at constant temperature can be expressed by Equation (8).
KH =
∆CH2
∆pH
,
(8)
2
43
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
δ
KH ∆pH .
2
(9)
By definition, permeability is the product of mobility (dissolution) and solubility
(sorption) factors, i.e. DH and KH (Equation (10)).
Pe,H2 = DH KH ,
(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 (Ple ) is defined as the ratio of permeability to the thickness of the
membrane layer (Equation (11)).
Ple =
Pe, H2
(11)
δ
The separation factor (αH2/ ) or ideal selectivity of the hydrogen with respect to
i
other gases is defined as in Equation (12)
αH2/i =
Pe,H2
Pe,i
=
Ple,H2
Ple,i
,
(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 Ple,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
δ
p0.5
− p0.5
S
d
(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
44
through dense palladium membranes occurs via the solution/diffusion
mechanism. Generally, the hydrogen permeation through metal membranes
follows the Equation (14).
JH2 =
PeH2
δ
(pnretentate − pnpermeate ),
(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
45
(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 =
E
Pe0 exp (- a )
RT
δ
(p0.5
− p0.5
)
retentate
permeate
(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 nonPd 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).
46
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
47
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.
48
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.
49
Fig. 8. Different concepts of integrating a catalyst and a membrane in an extractortype 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).
50
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 (SigmaAldrich) 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
51
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 energydispersive 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 CNTbased 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 energyfiltered 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,
52
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-JoynerHalenda (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 N2adsorption 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.
53
Table 6. List of catalyst materials used in MR studies.
Reaction
Catalyst
Metal loadings
Manufacturer
Membrane
(wt.%)
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).
54
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
55
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
Yield of hydrogen (YH2 ) =
56
Fethanol, out
Fethanol, in
FH2, out
6Fethanol, in
(17)
(18)
Selectivity of products S' p % =
F'i
∑F'i
×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 setup 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-gasshift (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
57
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
58
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).
59
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.
60
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%) =
Hydrogen yield YH2 % =
FH2, p
FH2, p + FCO2, p + FCOp + FCH4, p
×100
FH2, p+r
6F'ethanol, in +7F'glycerol, in +4F'acetic acid, in
(22)
×100 (23)
The selectivity of the products is denoted by the Equation (24).
Selectivity S'i %) =
Fi-total
FH
2- total
+ FCO
2- total
+ FCO total + FCH
×100
(24)
4-total
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 %) =
Hydrogen yield YH2 % =
FCOr +FCH4, r +FCO2,r
3Fglycerol, in
FH2, r+p
7F''glycerol, in
×100
Weight hourly space velocity WHSV) =
M'feed
Mcat
×100
(25)
(26)
(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).
61
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).
62
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).
63
c.
Fig. 13. TEM images and Ni nano-particle size distributions on different carbonsupported 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.
64
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.
65
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.
66
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 nonuniform, 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
67
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).
68
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
dTEM
dXRD
(wt.%)*
<nm>
<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;
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;
4.1 ± 2.9
NiO(200) 6.3±0.1;
4.8 ± 3.7
NiO(200) 7.1±0.1;
n.d.
NiO(200) 6.4 ±0.3 ;
Pt 1.4±0.2
Pt 2.2±1.3
Ni10Pt2/CNT
Ni 22.9±10.9; C 63.6±12.1; O 5.3±2.0;
(ZnO10)Ni10/CNT
Ni 6.9±1.9; C 65.4±11.8; O 20.4±10.6;
Pt(111) 6.5±0.7
Pt 4.6±2.9
Pt(111) 10.5±0.8
Zn 5.1±3.2
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).
69
Table 9. Textural properties of catalysts in SRE reaction.
SBET (m2 g-1)
Pore volume (cm3 g-1)
MWCNT-Pristine
75
0.22
11.4
CNT-COOH
218
0.75
13.6
10.2
Catalyst
#
Pore size (nm)
Ni2/CNT
284
0.77
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
70
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
71
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).
72
Fig. 18. Ethanol conversion as a function of reaction temperature with 1:3 ethanol-tosteam 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
73
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.
74
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
75
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).
76
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
77
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
78
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.
79
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).
80
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
81
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 (%)
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
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
a. (ZnO10)Ni10/CNT catalyst
b. Ni10(ZnO10)/CNT catalyst
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
82
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-onstream (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
83
(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
Pore volume
Pore size
(m2 g-1)
(cm3 g-1)
(nm)
(wt.%)
Ni10CNT
283.9
0.44
6.20
87.7±4.0
(ZnO)10Ni10CNT
271.8
0.40
5.95
83.4±3.5
Carbon
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
84
the pressure range of 2–10 bar. The permeation tests with pure and mixture gases
were performed as follows:
1.
2.
3.
Permeation tests with pure single gases: He, N2, and H2.
Permeation tests with binary gas mixtures: He/N2, He/H2 and N2/H2 for the
calculation of ideal selectivities.
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
n
390 °C as a function of H2 permeation, i.e. the pressures difference (∆kPa ) 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
85
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
Permeance H2
Permeance N2
Permeance He
αH2/N2
×10-7 (mol m-2 s-1 Pa-1)
×10-9 (mol m-2 s-1 Pa-1)
×10-9 (mol m-2 s-1 Pa-1)
(-)
(-)
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
(kPa)
αH2/He
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
86
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)).
C3 H8 O3 + 3H2 O ⇌7H2 + 3CO2
CH3 COOH + 2H2 O ⇌4H2 + 2CO2
ΔH° = +128
°
kJ
mol
kJ
ΔH = +134.8
mol
(32)
(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.
87
Fig. 27. Bio-ethanol conversion and HRF as a function of reaction pressure in Pd/PSS
-1
MR over the Ni/ZrO2 catalyst at 400 °C and GHSV = 3200 h . 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).
88
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
-1
PSS/Pd MR over Ni and Co catalysts at 400 °C and GHSV = 3200 h (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
89
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
-1
performed in a Pd/PSS MR at 400 °C, GHSV = 3200 h , over Ni/ZrO2 and Co/Al2O3
catalysts (III, published by permission of Elsevier).
Pressure (bar)
SH2 (%)
SCO (%)
SCH4 (%)
SCO2 (%)
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
a. Ni/ZrO2 catalyst
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.
90
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
-1
h 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.
91
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
over the Co/Al2O3 catalyst (III, published by permission of Elsevier).
Gas
H2
CO
CH4
CO2
92
Retentate stream
Permeate stream
Molar composition
Molar flow rate
Molar composition
Molar flow rate
(%)
(mol min-1)
(%)
(mol min-1)
2.32·10
-4
95
1.44·10-4
9.96·10
-6
0.1
2.13·10-7
8.02·10
-5
0.7
9.86·10-7
2.93·10
-4
4.7
7.16·10-6
38
2
13
48
−1
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
n
permeation driving force, i.e. pressures difference (∆kPa ) at different ‘n’ values (IV,
published by permission of Elsevier).
93
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
94
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:
-1
T = 400 °C, H2O/C3H8O3 = 6 (mol/mol), WHSV = 1.0 h , 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
95
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).
96
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
97
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 countercurrent 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).
98
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
Mix Binary
Gas mixture
Molar composition (%)
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).
99
Fig. 36. Hydrogen permeating flux as a function of H2 partial pressure difference
0.7
(kPa ) at 390 °C with pure H2 and different gaseous mixtures (Table 15)
2,
Mix binary (Mix bin), and
Mix 1,
Mix
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
100
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
-1
Preaction range of 7–11 bar. Conditions: T = 390 °C, GHSV= 5750 h , 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.
101
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.
102
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)
Coke (g h-1)
HPP (%)
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
CO
H2
CO2
H2O
H2O/CO
∆pH2 (bar)
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).
103
Fig. 39. CO conversion and HRF versus the three syngas feed compositions used in
-1
the WGS reaction performed in Pd/PSS at 390 °C, GHSV = 3450 h , 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.
104
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.
105
106
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
107
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
108
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 PdAg 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;
Ni/ZrO2 &
ELP
0.7
p = 8–12 bar (abs.); T = 400 °C
Co/Al2O3
δ = 50 µm (Pd-Ag); L =14 cm;
Ru/Al2O3
CDW
0.5
BGSR
Dense Pd-Ag#
p = 1–4 bar (abs.); T = 400 °C
WGS
Pd/PSS
δ = 20 µm (Pd); L = 7.4 cm;
Fe-Cr/ZnO ELP
0.7
p = 7–11 bar (abs.); T = 390 °C
*Membrane thickness (δ), ELP = electroless plating, CDW = cold-rolling and diffusion welding, #selfsupported, 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-tocarbon 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
109
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 ZnOpromoted 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
110
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.
111
112
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 nanotubesupported 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.
113
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.
114
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.
115
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 ZnOsupported 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.
116
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 SBA15 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.
117
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 nanotubessupported nickel catalysts for the steam reforming of propane. J Mater Sci 47: 2985–
2994.
Lymberopoulos N (2005) Hydrogen production from renewables, EU commission DGTREN, 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 watergas shift reaction: from conventional catalytic systems to Pd-based membrane reactors
– a review. Asia Pac J Chem Eng 5: 111–137.
118
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-ilmastoselonteko-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.
119
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 bioethanol steam reforming reaction in a Pd/PSS membrane reactor. Catal Today 193:
42–48.
120
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 bioethanol 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 byproduct 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.
121
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 PdAg 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 Ptbased 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.
122
Original papers
I
II
III
IV
V
VI
VII
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.
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.
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.
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.
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.
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.
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.
123
124
C473etukansi.kesken.fm Page 2 Tuesday, October 22, 2013 12:07 PM
ACTA UNIVERSITATIS OULUENSIS
SERIES C TECHNICA
457.
Nardelli, Pedro Henrique Juliano (2013) Analysis of the spatial throughput in
interference networks
458.
Ferreira, Denzil (2013) AWARE: A mobile context instrumentation middleware
to collaboratively understand human behavior
459.
Ruusunen, Mika (2013) Signal correlations in biomass combustion – an
information theoretic analysis
460.
Kotelba, Adrian (2013) Theory of rational decision-making and its applications to
adaptive transmission
461.
Lauri, Janne (2013) Doppler optical coherence tomography in determination of
suspension viscosity
462.
Kukkola, Jarmo (2013) Gas sensors based on nanostructured tungsten oxides
463.
Reiman, Arto (2013) Holistic work system design and management : — a
participatory development approach to delivery truck drivers’ work outside the
cab
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 under
channel estimation errors
466.
Schaberreiter, Thomas (2013) A Bayesian network based on-line risk prediction
framework for interdependent critical infrastructures
467.
Halonen, Niina (2013) Synthesis and applications of macroscopic well-aligned
multi-walled carbon nanotube films
468.
Remes, Jukka (2013) Method evaluations in spatial exploratory analyses of
resting-state functional magnetic resonance imaging data
469.
Oravisjärvi, Kati (2013) Industry and traffic related particles and their role in
human health
470.
Czajkowski, Jakub (2013) Optical coherence tomography as a characterization
method in printed electronics
471.
Haapalainen, Mikko (2013) Dielectrophoretic mobility of a spherical particle in
2D hyperbolic quadrupole electrode geometry
472.
Bene, József Gergely (2013) Pump schedule optimisation techniques for water
distribution systems
Book orders:
Granum: Virtual book store
http://granum.uta.fi/granum/
C473etukansi.kesken.fm Page 1 Tuesday, October 22, 2013 12:07 PM
C 473
OULU 2013
UNIV ER S IT Y OF OULU P. O. B R[ 00 FI-90014 UNIVERSITY OF OULU FINLAND
U N I V E R S I TAT I S
S E R I E S
SCIENTIAE RERUM NATURALIUM
Professor Esa Hohtola
HUMANIORA
University Lecturer Santeri Palviainen
TECHNICA
Postdoctoral research fellow Sanna Taskila
MEDICA
Professor Olli Vuolteenaho
ACTA
UN
NIIVVEERRSSIITTAT
ATIISS O
OU
ULLU
UEEN
NSSIISS
U
Prem Kumar Seelam
E D I T O R S
Prem Kumar Seelam
A
B
C
D
E
F
G
O U L U E N S I S
ACTA
A C TA
C 473
HYDROGEN PRODUCTION
BY STEAM REFORMING OF
BIO-ALCOHOLS
THE USE OF CONVENTIONAL AND MEMBRANEASSISTED CATALYTIC REACTORS
SCIENTIAE RERUM SOCIALIUM
University Lecturer Hannu Heikkinen
SCRIPTA ACADEMICA
Director Sinikka Eskelinen
OECONOMICA
Professor Jari Juga
EDITOR IN CHIEF
Professor Olli Vuolteenaho
PUBLICATIONS EDITOR
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)
UNIVERSITY OF OULU GRADUATE SCHOOL;
UNIVERSITY OF OULU,
FACULTY OF TECHNOLOGY,
DEPARTMENT OF PROCESS AND ENVIRONMENTAL ENGINEERING ,
MASS AND HEAT TRANSFER PROCESS LABORATORY
C
TECHNICA
TECHNICA