fuel processing for fuel cells

Gunther Kolb
Fuel Processing
Further Reading
K. Sundmacher, A. Kienle, H. J. Pesch, J. F. Berndt, G. Huppmann (Eds.)
Molten Carbonate Fuel Cells
Modeling, Analysis, Simulation, and Control
2007
ISBN: 978-3-527-31474-4
W. Vielstich, A. Lamm, H. Gasteiger (Eds.)
Handbook of Fuel Cells - Fundamentals, Technology, Applications
4 volume set
2003
ISBN: 978-0-471-49926-8
B. Elvers (Ed.)
Handbook of Fuels
Energy Sources for Transportation
2007
ISBN: 978-3-527-30740-1
A. Züttel, A. Borgschulte, L. Schlapbach (Eds.)
Hydrogen as Future Energy Carrier
2008
ISBN: 978-3-527-30817-0
H W. Häring (Ed.)
Industrial Gases Processing
2008
ISBN: 978-3-527-31685-4
Gunther Kolb
Fuel Processing
for Fuel Cells
The Author

Dr. Gunther Kolb
IMM - Institut für Mikrotechnik Mainz GmbH
Carl-Zeiss-Str. 18 - 20
55129 Mainz
Germany
Cover Illustration:
Photograph courtesy of Nuvera.
The APU model was developed
by Tenneco within the European
project Hytran (‘‘ Hydrogen and
Fuel Cell Technologies for Road Transport’’ ),
contract no. TIP3-CT-2003-502577
co-ordinated by Volvo Technology
Corporation.
All books published by Wiley-VCH are carefully
produced. Nevertheless, authors, editors, and
publisher do not warrant the information contained
in these books, including this book, to be free of
errors. Readers are advised to keep in mind that
statements, data, illustrations, procedural details or
other items may inadvertently be inaccurate.
Library of Congress Card No.: applied for
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the
British Library.
Bibliographic information published by
the Deutsche Nationalbibliothek
Die Deutsche Nationalbibliothek lists this
publication in the Deutsche Nationalbibliografie;
detailed bibliographic data are available on the

Internet at <>.
# 2008 WILEY-VCH Verlag GmbH & Co. KGaA,
Weinheim
All rights reserved (including those of translation into
other languages). No part of this book may be
reproduced in any form – by photoprinting,
microfilm, or any other means – nor transmitted or
translated into a machine language without written
permission from the publishers. Registered names,
trademarks, etc. used in this book, even when not
specifically marked as such, are not to be considered
unprotected by law.
Typesetting Thomson Digital, Noida, India
Printing Strauss GmbH, Mörlenbach
Binding Litges & Dopf GmbH, Heppenheim
Cover Grafik-Design Schulz, Fußgönheim
Printed in the Federal Republic of Germany
Printed on acid-free paper
ISBN: 978-3-527-31581-9
Contents
Acknowledgement IX
1 Introduction and Outline 1
2 Fundamentals 3
2.1 Common Fossil Fuels 3
2.2 Basic Definitions, Calculations and Legislation 6
2.3 The Various Types of Fuel Cells and the Requirements of the
Fuel Processor 12
2.3.1 PEM Fuel Cells 12
2.3.2 High Temperature Fuel Cells 15
3 The Chemistry of Fuel Processing 17

3.1 Steam Reforming 17
3.2 Partial Oxidation 22
3.3 Oxidative Steam Reforming or Autothermal Reforming 29
3.4 Catalytic Cracking of Hydrocarbons 38
3.5 Pre-Reforming of Higher Hydrocarbons 39
3.6 Homogeneous Plasma Reforming of Higher Hydrocarbons 43
3.7 Aqueous Reforming of Bio-Fuels 44
3.8 Processing of Alternative Fuels 44
3.8.1 Dimethyl Ether 44
3.8.2 Methylcyclohexane 45
3.8.3 Sodium Borohydride 45
3.8.4 Ammonia 46
3.9 Desulfurisation 46
3.10 Carbon Monoxide Clean-Up 48
3.10.1 Water–Gas Shift 48
3.10.2 Preferential Oxidation of Carbon Monoxide 49
3.10.3 Methanation 51
Fuel Processing for Fuel Cells. Gunther Kolb
Copyright Ó 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
ISBN: 978-3-527-31581-9
V
3.11 Catalytic Combustion 52
3.12 Coke Formation on Metal Surfaces 52
4 Catalyst Technology for Distributed Fuel Processing Applications 57
4.1 A Brief Introduction to Catalyst Technology and Evaluation 57
4.1.1 Catalyst Activity 58
4.1.2 Catalyst Stability 60
4.1.3 Catalyst Coating Techniques 61
4.1.4 Specific Features Required for Fuel Processing Catalysts
in Smaller Scale Applications 68

4.2 Reforming Catalysts 69
4.2.1 Catalysts for Methanol Reforming 71
4.2.2 Catalysts for Ethanol Reforming 77
4.2.3 Overview of Catalysts for Hydrocarbon Reforming 80
4.2.4 Catalysts for Natural Gas/Methane Reforming 81
4.2.5 Catalysts for Reforming of LPG 84
4.2.6 Catalysts for Pre-Reforming of Hydrocarbons 86
4.2.7 Catalysts for Gasoline Reforming 88
4.2.8 Catalysts for Diesel and Kerosene Reforming 92
4.2.9 Cracking Catalysts 96
4.2.10 Deactivation of Reforming Catalysts by Sintering 98
4.2.11 Deactivation of Reforming Catalysts by Coke Formation 98
4.2.12 Deactivation of Reforming Catalysts by Sulfur Poisoning 101
4.3 Catalysts for Hydrogen Generation from Alternative Fuels 105
4.3.1 Dimethyl Ether 105
4.3.2 Methylcyclohexane 106
4.3.3 Sodium Borohydride 107
4.3.4 Ammonia 107
4.4 Desulfurisation Catalysts/Adsorbents 108
4.5 Carbon Monoxide Clean-Up Catalysts 111
4.5.1 Catalysts for Water–Gas Shift 111
4.5.2 Catalysts for the Preferential Oxidation of Carbon Monoxide 116
4.5.3 Methanation Catalysts 123
4.6 Combustion Catalysts 124
5 Fuel Processor Design Concepts 129
5.1 Design of the Reforming Process 129
5.1.1 Steam Reforming 129
5.1.2 Partial Oxidation 146
5.1.3 Autothermal Reforming 149
5.1.4 Catalytic Cracking 154

5.1.5 Pre-Reforming 155
5.2 Design of the Carbon Monoxide Clean-Up Devices 155
5.2.1 Water–Gas Shift 155
5.2.2 Preferential Oxidation of Carbon Monoxide 161
VI Contents
5.2.3 Selective Methanation of Carbon Monoxide 164
5.2.4 Membrane Separation 164
5.2.5 Pressure Swing Adsorption 174
5.3 Aspects of Catalytic Combustion 176
5.4 Design of the Overall Fuel Processor 181
5.4.1 Overall Heat Balance of the Fuel Processor 181
5.4.2 Interplay of the Different Fuel Processor or Components 188
5.4.3 Overall Water Balance of the Fuel Processor 190
5.4.4 Overall Basic Engineering of the Fuel Processor 192
5.4.5 Dynamic Simulation of the Fuel Processor 205
5.4.6 Control Strategies for Fuel Processors 213
5.5 Comparison with Conventional Energy Supply Systems 215
6 Types of Fuel Processing Reactors 217
6.1 Fixed-Bed Reactors 217
6.2 Monolithic Reactors 217
6.3 Plate Heat-Exchanger Reactors 221
6.3.1 Conventional Plate Heat-Exchanger Reactors 223
6.3.2 Microstructured Plate Heat-Exchanger Reactors 225
7 Application of Fuel Processing Reactors 227
7.1 Reforming Reactors 227
7.1.1 Reforming in Fixed-Bed Reactors 227
7.1.2 Reforming in Monolithic Reactors 230
7.1.3 Reforming in Plate Heat-Exchanger Reactors 240
7.1.4 Reforming in Membrane Reactors 254
7.1.5 Reforming in Chip-Like Microreactors 260

7.1.6 Plasmatron Reformers 264
7.2 Water–Gas Shift Reactors 269
7.2.1 Water–Gas Shift in Monolithic Reactors 269
7.2.2 Water–Gas Shift in Plate Heat-Exchanger Reactors 270
7.2.3 Water–Gas Shift in Membrane Reactors 272
7.3 Catalytic Carbon Monoxide Fine Clean-Up 272
7.3.1 Carbon Monoxide Fine Clean-Up in Fixed-Bed Reactors 272
7.3.2 Carbon Monoxide Fine Clean-Up in Monolithic Reactors 273
7.3.3 Carbon Monoxide Fine Clean-Up in Plate Heat-Exchanger
Reactors 275
7.3.4 Carbon Monoxide Fine Clean-Up in Membrane Reactors 282
7.4 Membrane Separation Devices 283
7.5 Catalytic Burners 285
8 Balance-of-Plant Components 289
8.1 Heat-Exchangers 289
8.2 Liquid Pumps 290
8.3 Blowers and Compressors 290
Contents VII
8.4 Feed Injection System 292
8.5 Insulation Materials 293
9 Complete Fuel Processor Systems 295
9.1 Methanol Fuel Processors 295
9.2 Ethanol Fuel Processors 316
9.3 Natural Gas Fuel Processors 317
9.4 Fuel Processors for LPG 327
9.5 Gasoline Fuel Processors 332
9.6 Diesel and Kerosine Fuel Processors 344
9.7 Multi-Fuel Processors 348
9.8 Fuel Processors Based on Alternative Fuels 350
10 Introduction of Fuel Processors Into the Market Place – Cost

and Production Issues 355
10.1 Factors Affecting the Cost of Fuel Processors 355
10.2 Production Techniques for Fuel Processors 359
10.2.1 Fabrication of Ceramic and Metallic Monoliths 359
10.2.2 Fabrication of Plate Heat-Exchangers/Reactors 361
10.2.3 Fabrication of Microchannels 365
10.2.4 Fabrication of Chip-Like Microreactors 367
10.2.5 Fabrication of Membranes for Hydrogen Separation 369
10.2.6 Automated Catalyst Coating 370
References 373
Index 409
VIII Contents
Acknowledgement
I would like to cordially thank my colleagues at IMM, in particular Dr. Karl-Peter
Schelhaas for fruitful discussions and input in the fields of calculations and material
properties, Dr. Hermann Ehwald for input in the field of desulfurization catalysts,
Tobias Hang for dealing with the figures, Carola Mohrmann and Christina Miesch-
Schmidt for dealing with the tables, Dr. Athanassios Ziogas and Martin O’Connell
for dealing with the literature ordering and Sibylle for dealing with me when I was
‘‘ hacking’’ through weekends and nights.
Gunther Kolb
IX
Fuel Processing for Fuel Cells. Gunther Kolb
Copyright Ó 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
ISBN: 978-3-527-31581-9

1
Introduction and Outline
Mankinds energy demand is increasing exponentially. Between 1900 and 1997, the
worlds population more than tripled and the average energy demand per human

being has also more than tripled, resulting in greater than thirteen times higher
overall global emissions [1]. Thus the carbon dioxide concentration rose from 295
parts per million in 1900 to 364 parts per million in 1997 [1]. In 1997 almost all
European countries committed to reducing greenhouse gas emissions to an amount
8% below the emissions of 1990 in the period from 2008 to 2012. With this scenario,
fuel cell technology is attracting increasing attention nowadays, because it offers the
potential to lower these emissions, owing to a potentially superior efficiency
compared with combustion engines. Fuel cells require hydrogen for their operation
and consequently numerous technologies are under investigation worldwide for the
storage of hydrogen, aimed at distribution, and mobile and portable applications.
The lack of a hydrogen infrastructure in the short term, along with the highly
attractive energy density of liquid fossil and regenerative fuels, has created wide-
spread research efforts in the field of distribution and on-board hydrogen generation
from various fuels. This complex chemicalprocess,generallytermed fuel processing,
is the subject of this book.
The electrical power output equivalent of the fuel processors that are currently
under development world wide covers a wide range, from less than a watt to several
megawatts. Portable and small scale mobile fuel cell systems promise to be the first
commercial market forfuel cells, according to amarketstudy of Fuel Cell Today in July
2003 [2]. According to the same report, the number of systems built has increased
dramatically to up to more than 3000 in 2003. To date, most of these systems have
used Proton Exchange Membrane (PEM) fuel cells.
Low power fuel processors (1–250 W) compete with both conventional storage
equipment, such as batteries, and simpler fuel cell systems, such as Direct Methanol
Fuel Cells (DMFC).
Fuel cell systems for residential applications are typically developed for the
generation of power and heat, which increases their overall efficiency considerably,
because even low temperature off-heat may be utilised for hot water generation,
which reduces energy losses considerably.
Fuel Processing for Fuel Cells. Gunther Kolb

Copyright Ó 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
ISBN: 978-3-527-31581-9
j
1
For mobile applications, systems designed to move a vehicle need to be distin-
guished from the Auxiliary Power Unit (APU), which either creates extra energy
for the vehicle (e.g., the air conditioning and refrigerator system of a truck) or works
as a stand alone system for the electrical power supply.
This book provides a general overview of the field of fuel processing for fuel cell
applications. Its focus is on mobile, portable and residential applications, but the
technology required for the smaller stationary scale is also discussed.
In the second chapter fundamental definitions and the basic knowledge of fuel
cell technology are provided, as far as is required to gain an insight into the interplay
between the fuel cell and its hydrogen supply unit – the fuel processor.
The third chapter deals with the reforming chemistry of conventional and
alternative fuels, and with the chemistry of catalytic carbon monoxide clean-up,
sulfur removal and catalytic combustion.
An overview of catalyst technology for fuel processing applications is provided
in Chapter 4, covering all the processes described in Chapter 3.
The design of the individual components of the fuel processor is the subject of
Chapter 5. Design concepts and numerical simulations presented in the open
literature are discussed for reforming, catalytic carbon monoxide clean-up and
physical clean-up strategies, such as membrane separation and pressure swing
adsorption. In addition, fuel processor concepts are then presented and the interplay
between the various fuel processor components is explained. Details of the basic
engineering of fuel processors and dynamic simulations are discussed, covering
start-up and control strategies. Some tips and the basic knowledge required to
perform such calculations are provided.
There are three basic types of fuel processing reactors, namely fixed catalyst
beds, monoliths and plate heat-exchangers, which are explained in Chapter 6.

Chapter 7 then shows the practical applications of such reactors, as published in
the literature.
In Chapter 8 some important aspects of balance-of-plant components are dis-
cussed, and Chapter 9 presents complete fuel processors for all types of fuels,
while cost and production issues are the subject of Chapter 10.
2
j
1 Introduction and Outline
2
Fundamentals
This chapter provides information about common fossil fuels, necessary definitions
in the field of fuel processing and the basic knowledge from the wide field of fuel cell
technology. It is by no means comprehensive and is not a substitute for the dedicated
literature in these fields. Rather, it provides a brief summary for readers who wish to
gain an overview of the topic of fuel processing without the need to use too much
additional literature.
2.1
Common Fossil Fuels
Fuels are solid, liquid or gaseous energy carriers. To date, practically all of the fuels
available on the market are based upon fossil sources and thus contain hydrocarbons
of varying composition. However, alternative fuels such as alcoholsand hydrides may
serve as future energy carriers. Table 2.1 provides an overview of the conventional
fuels and of the most important alternative fuels, which may act as future hydrogen
source for fuel cells along with their key properties.
A comparison of the gravimetric and volumetric density of various hydrogen
carriers shows that liquid hydrocarbons have – apart from borohydrides – by far the
best combined properties (see Figure 2.1).
Table 2.2 shows the maximum amount of work that can be converted into
electricity from various fuels, in theory. Compared with the gravimetric and volu-
metric energy density of 1 MJ kg

À1
or <2MJL
À1
of lithium-ion and zinc-air batteries,
these values are considerably higher.
The composition of fossil hydrocarbon fuels may vary widely depending on the
source of the crude oil that is processed in the refinery.
The composition of natural gas is predominantly methane, and also contains
several percent ethane and propane. In addition, minor amounts of butane and
higher hydrocarbons are present, plus carbon dioxide and nitrogen.
Table 2.3 shows the composition of natural gas from various sources [5]. Natural
gas also contains sulfur compounds at the ppm-level, such as hydrogen sulfide and
Fuel Processing for Fuel Cells. Gunther Kolb
Copyright Ó 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
ISBN: 978-3-527-31581-9
j
3
Table 2.1 Overview of important fuels for fuel processing.
Fuel Formula
Sulphur content [wt. ppm]
[only commercially
available fuels]
Lower heating
value [kJ/mol]
Flammability
limits lower,
higher [Vol.%]
Density
[kg/m
3

]
Boiling point or
boiling range [

C]
Heat of
vaporization
[kJ/mol]
Heat capacity
[J/(mol K)]
At 20

C
Hydrogen H
2
– 240 4.1, 74 0.09 À252.7 0.92 28.6
Methanol CH
3
OH – 643 7.3, 36 794 (l) 64.6 35.2 49.0 (g)
Ethanol C
2
H
5
OH – 1240 4.3, 19 790 (l) 78.3 38.9 77.3
Methane CH
4
– 802 0.72 À161.4 8.2 35.5
Natural gas C
1.07
H

4.1
7–25 797 5.3, 15 0.77 – 8.2 –
Propane C
3
H
8
– 2015 2.2, 9.5 1.96 À42 73.5 73.5
Liquified Petroleum
Gas [LPG]
C
3
H
8
/C
4
H
10
50–200 2024 1.5, 11 540 (l)
(at 8 bar)
À42–À0.5 ––
Iso-octane C
8
H
18
– 4731 ––99.3 34.8 –
Gasoline C
7.1
H
14.3
150 4,720 0.8, 8 720–770 30–200 33.5 180

50 (European
Regulation 2005)
Dodecane C
12
H
26
– 7,392 – 750 216.4 40.6 270.2
Hexadecane C
16
H
34
– 9,792 – 770 286.8 51.3 386.5
Diesel C
13.6
H
27.1
50 (European regulation 2005)
Heating oil 2,000
8,080 1, 6 0.856 120–430 47.0 340
Bio-diesel C
18.7
H
34.5
O
2
<3,000 10,800 ––– 67.8 –
4
j
2 Fundamentals
diethyl sulfide, and mercaptanes, such as ethyl mercaptane [(C

2
H
5
)CHS] and tertiary
butyl mercaptane [(CH
3
)
3
CHS].
Amongst all the fossil fuels,propanecontains the highest amount of hydrogen ona
gravimetric basis, which even exceeds liquefied hydrogen, when the weight of the
storage tanks is taken into consideration [6]. Propane is usually marketed as liquefied
petroleum gas, which is a mixture of propane and butane in various ratios.
For gasoline, only approximate characterization parameters are provided, such as
the octane number, the boiling point distribution, and the saturated hydrocarbons
(alkanes), unsaturated hydrocarbons (olefins) and aromatics content. The content of
contaminants, such as sulfur, is important.
Volumic density (g H
2
/litre)
Lantanides
Hydrides base
(TI et/ou Zr)
Nanotubes
Compress
700
bars
350
bars
Liquids H

2
Liquids hydrocarbon
Fullerene
Hydrides MBH4
D.O.E
x
H
2
Glass micro-sphere
Hydrides base Mg
Activ carbon
Gravimetric density (%)
10 20 50 200
1
5
10
20
Figure 2.1 Comparison of gravimetric and volumetric storage
densities as provided by Heurtaux et al. [3].
Table 2.2 Energy density of various fuels related to different properties [4].
Maximum amount of work
Fuel MJ/Mol fuel MJ/Kg fuel MJ/L fuel MJ/Mol C in fuel
MJ/Mol H
2
via reforming
Methanol À0.69 À22 À17
a
À0.69 À0.23
Ethanol À1.31 À28 À22
a

À0.65 À0.22
n-Octane À5.23 À46 À32
a
À0.65 À0.21
Ammonia À0.33 À19 À10
a
À0.22
Methane À0.80 À50 À3.9
b
À0.80 À0.20
Hydrogen À0.23 À113 À0.89
b
À0.23
a
density of the liquid fuels calculated at 298K and 1 bar, for ammonia at 10 bar.
b
density of the gaseous fuels calculated at 298K and 100 bar.
2.1 Common Fossil Fuels
j
5
Regular gasoline, at least according to German standards, is well represented by
the overall formula C
7
H
12
[7].
A standard jet fuel that is frequently cited is the American JP-8 fuel. It contains about
1 000 ppm sulfur and up to 1.5 vol.% non-volatile hydrocarbons [8, 9]. Jet fuels widely
used inthe worldareJet fuelAandA1[10] with a boiling rangebetween 150and300


C.
Diesel fuels contain mainly iso-paraffins, but also n-paraffins, mono-, di-, tri-,
tetra-cycloparaffins, alkylbenzenes, naphthalenes and phenanthrenes and even
pyrenes [11].
2.2
Basic Definitions, Calculations and Legislation
Fuel processing is the conversion of hydrocarbons, alcohol fuels and other alternative
energy carriers into hydrogen containing gas mixtures. The chemical conversion is
achieved in mostinstances in the gaseous phase, normally heterogeneously catalysed
in the presence of a solid catalyst and less frequently homogeneously at high
temperature without a catalyst.
The first step of the conversion procedure is generally termed reforming, and has
been well established in large scale industrial processes for many decades. The
industrial applications most commonly (about 76% [12]) use natural gas as feedstock.
The purpose of this process is the production of synthesis gas, a mixture of carbon
monoxide and hydrogen, which is then used for numerous processes in large scale
chemical production, which are not subject of this book.
Rather, the focusof this book is the technologythat provides a hydrogen containing
gas mixture, termed the reformate, which is suitable for feeding into a fuel cell. The
fuel cell then converts hydrogen into electrical energy. Carbon monoxide may also be
converted, which depends on the fuel cell type (see Section 2.3.2).
The lower heating value of a chemical substance is defined as its standard enthalpy
of formation. The lower heating value of any fuel C
x
H
y
O
z
is easily determined by the
following formula [13]:

LHV kJ mol
À 1
ÂÃ
¼
y
2
þ 2x À z

198:8 þ 25:4 ð2:1Þ
Table 2.3 Composition of natural gas from various sources [5].
Component North Sea Qatar Netherlands Pakistan Ekofisk
CH
4
(Vol.%) 94.86 76.6 81.4 93.48 85.5
C
2
H
6
(Vol.%) 3.90 12.59 2.9 0.24 8.36
C
3
H
8
(Vol.%) 2.38 0.4 0.24 2.85
i-C
4
H
10
(Vol.%) 0.15 0.11 0.04 0.86
n-C

4
H
10
(Vol.%) 0.21 0.1 0.06
C
5
þ
(Vol.%) 0.02 0.41 0.22
N
2
(Vol.%) 0.79 0.24 14.2 4.02 0.43
S (ppm) 4 1.02 1 N/A 30
6
j
2 Fundamentals
The performance of a fuel processor is measured by its overall efficiency, which
is commonly defined as the ratio between the Lower Heating Value (LHV) of the
hydrogen and carbon monoxide that are produced to the LHV of the fuel
consumed:
h
Fuel processor
¼
LHVðH
2
Þ n
H
2
þ LHVðCOÞ n
CO
LHVðFuelÞ n

Fuel
ð2:2Þ
n are the molar flows and the lower heating value is in units of kJ mol
À1
. The
efficiency of the reformer may be calculated by a simplified version of Eq. (2.2):
h
Reformer
¼
LHVðH
2
Þn
H
2
LHVðFuelÞn
Fuel
ð2:3Þ
A certain portion of the hydrogen produced by the fuel processor is frequently fed
back to it, because it is not completely consumed by the fuel cell (see Section 2.3). The
curious situation maythenarise where the fuel processor efficiency exceeds 100%. In
particular, this is the situation for steam reforming, where substantially more heat is
required to run the process compared with partial oxidation and autothermal
reforming (see Section 3). A fuel processor running on steam reforming may reach
up to 120% efficiency according to the Eqs. (2.2) and (2.3).
The carbon monoxide content of the reformate obviously needs to be minimised
for low temperature proton exchange membrane fuel cells, but other fuel cells may
well utilize it as afuel(see Section 2.3.2). The same applies for methane in certainfuel
cells. Therefore, the heating value of the hydrogen alone does not provide the
appropriate number for the calculation of efficiency in this instance.
A modified definition of the fuel processor efficiency provides a more realistic

value than Eqs. (2.2) and (2.3) [14]:
h
Fuel processor
¼
LHVðH
2
Þn
H
2
þ LHVðCOÞn
CO
þ LHVðCH
4
Þn
CH
4
À LHVðH
2
Þn
H
2
þ LHVðCOÞn
CO
þ LHVðCH
4
Þ n
CH
4
½
recirculated

LHVðFuelÞn
Fuel
ð2:4Þ
In addition to the formula provided by Lutz et al. [14], it takes intoconsideration the
release of unconverted methane and the formation of methane by the reforming
process (see Section 3). Unconverted methane is commonly re-circulated to the fuel
processor, along with unconverted carbon monoxide, in particular for high temper-
ature fuel cells.
However, for PEM fuel cells methane and carbon monoxide could be excluded
from efficiency calculations, because they are not converted in the fuel cell.
The following definition of efficiency was proposed by Feitelberg [15]. It was
modified to also take methane and carbon monoxide fed to the fuel cell into
consideration as discussed above:
h
Fuel processor
¼
LHVðH
2
Þn
H
2
þ LHVðCOÞn
CO
þ LHVðCH
4
Þn
CH
4
LHVðFuelÞn
Fuel

þ½LHVðH
2
Þ n
H
2
þ LHVðCOÞn
CO
þ LHVðCH
4
Þ n
CH
4

recirculated
ð2:5Þ
2.2 Basic Definitions, Calculations and Legislation
j
7
This definition seems to be the most realistic, because it takes all products into
consideration in the numerator and all feed entering the fuel processor in the
denominator, which is in agreement with the rules for energy balancing.
Hagh [13] has derived a general formula for the fuel processor efficiency. Thus, by
describing the fuel processing reactions with the following general and simplified
formula (by-products such as methane are not taken into consideration, the same
applies for unconverted methane in case of methane fuel processing):
C
x
H
y
O

z
þ a O
2
þ b H
2
O ! d COþ e CO
2
þ f H
2
ð2:6Þ
defining the stoichiometric ratio (SR) as the ratio of oxygen to oxygen required for
complete combustion:
SR ¼
2a
y
2
þ 2x À z
ð2:7Þ
and by further defining the hydrogen ratio (HR) as the ratio of moles of hydrogen
produced to the moles of fuel reformed:
HR ¼
y
2
þ
3
d
e
þ 4
d
e

þ 1
À 2
0
B
@
1
C
A
x À z
y
2
þ 2x À z
À SR ð2:8Þ
The fuel processor efficiency is expressed by the simple term:
h
Fuel processor
¼
HR
1 À SR
ð2:9Þ
However, these calculations still assume ideal conditions, namely complete
conversion of the fuel, absence of methane formation and negligible heat losses
and they do not consider the water excess fed to practical systems (see Section 3.1).
Unconverted fuel is generally not desirable for most fuel processors, because the
fuel molecules might well contaminate the gas purification devices and the fuel cell
itself, which is especially critical for higher hydrocarbons and alcohols. However, for
methane fuel processors an incomplete conversion is feasible, which is discussed in
Section 2.3.2.
Excess water fed to the reformer decreases efficiency, so this water needs to
be removed from the system downstream. In other words, if this excess water is

neither consumed by the water–gas shift reaction downstream of the reformer (see
Section 3.10.1), nor required to prevent dry-out of the membrane of low temperature
PEM fuel cells (seeSection 2.3.1), nor to prevent coking inhigh temperature fuel cells
(see Section 2.3.2), the water should be recovered to avoid a negative water-balance of
the system. However, the heat of condensation is difficult to recover and commonly
lost to the cooling air. Thus, excess water should be minimised in fuel processor/fuel
cell systems. Condensers might be integrated upstream or downstream of the
fuel cell to ensure water recovery and net positive water balancing of the whole
system. Such condensers may also recover water produced by the fuel cell itself.
8
j
2 Fundamentals
Low temperature heat losses are usually mandatory in all fuel cell/fuel processor
systems because the efficiency of heat-exchangers is limited and the system needs to
work with cooling air, which might even have elevated temperatures in the summer
time. The major portion of the heat losses does not originate from the fuel processor
but rather from the fuel cell itself for low temperature proton exchange (PEM) fuel
cells. Thus the efficiency of the fuel processor is usually high provided heat losses are
neglected, because internal heat-exchangers can keep the high temperature heat
within the fuel processor.
Heat losses affect the fuel processor efficiency, especially at partial system load.
This becomes obvious when taking into consideration that the reactors of the fuel
processor require an elevated operating temperature. This temperature will not
decrease at partialload. Therefore heat losses remainconstant and become dominant
at partial load of the system. Smaller system size has similar effects.
To judge fuel processor efficiency, the following operational efficiency factor is thus
proposed as a modification to the definition provided by Schmid and W
€
unning [16]
and others:

h
Fuel processor; operation
¼ 1 À
Q
_
cond;FP
þ Q
_
losses
LHVðfuelÞn
Fuel
ð2:10Þ
Q
cond,FP
is the heat generated bycondensation of steam, which is not requiredforfuel
cell membrane dry-out. Thus the total fuel processor efficiency is defined as follows:
h
Fuel processor; total
¼ h
Fuel processor
h
Fuel processor; operation
ð2:11Þ
The fuel cell efficiency is defined as the ratio of the electrical power output of the fuel
cell P
Fuel cell
to the lower heating value of the fuel converted by the electrochemical
reactions:
h
FC

¼
P
Fuel cell
LHVðH
2
Þn
H
2
þ LHVðCOÞn
CO
ð2:12Þ
The overall efficiency of a fuel processor/fuel cell system is commonly defined as
the ratio of the electrical fuel cell stack power output P
Fuel cell
to the LHV of the fuel:
h
system
¼
P
Fuel cell
LHVðFuelÞn
Fuel
ð2:13Þ
Some basic thermodynamic definitions and calculations of enthalpy differences
when heating and evaporating fluids and for the reaction enthalpy are provided
below. The heat capacity c
p
is the derivative of enthalpy by temperature:
c
p

¼
dh
dT
ð2:14Þ
The enthalpy difference between the reference temperature T
0
and temperature T
is calculated as follows:
ð
T
T
0
dh ¼ hð TÞÀhðT
0
Þ¼
ð
T
T
0
c
p
ðTÞdT ð2:15Þ
2.2 Basic Definitions, Calculations and Legislation
j
9
and the molar heat capacity is according to:
c
pm
ðTÞ¼
1

T À T
0
ð
T
T
o
c
p
ðTÞdT ¼
hðTÞÀhðT
0
Þ
T À T
0
ð2:16Þ
The enthalpy of the fluid at temperature T is then:
hðTÞ¼hðT
0
Þþc
pm
ðTÞÁðT À T
0
Þð2:17Þ
To heat a substance from temperature T
1
to T
2
requires the enthalpy difference
Dh
1

:
Dh
1
¼ hðT
2
ÞÀhðT
1
Þ¼c
pm
ðT
2
ÞÁðT
2
À T
0
ÞÀc
pm
ðT
1
ÞÁðT
1
À T
0
Þð2:18Þ
The phase transition of a substance from phase A to phase B requires the enthalpy
difference Dh
2
:
Dh
2

¼ h
B
ðTÞÀh
A
ðTÞ¼h
B
ðT
0
ÞÀh
A
ðT
0
Þþc
pm
B
ðTÞÁðT À T
0
ÞÀc
pm
A
ðTÞ
ÁðT À T
0
Þð2:19Þ
While the reaction enthalpy Dh
R
of the chemical reaction
A þ B ! C ð2:20Þ
is calculated as follows:
Dh ¼ h

C
ðTÞÀh
A
ðTÞÀh
B
ðTÞ¼h
C
ðT
0
ÞÀh
A
ðT
0
ÞÀh
B
ðT
0
Þþc
pm
C
ðTÞ
ÁðT À T
0
ÞÀc
pm
A
ðTÞÁðT À T
0
ÞÀc
pm

B
ðTÞÁðT À T
0
Þð2:21Þ
Table 2.4 provides the most important physical and chemical properties of
substances present in the reformate.
To date, no emission regulations exist for fuel cell systems. European legislation
has directives for heating systems based on natural gas or Liquified Petroleum Gas
(LPG). They limit nitrous oxides (NO
x
) to 200 ppm and carbon monoxide to 100 ppm.
However, the legislation in some EU member countries are well below these values,
German emission control regulations limit NO
x
to 80 ppm and carbon monoxide to
60 ppm.
In instances where homogeneous combustion is applied in fuel processors, the
formation of NO
x
is inevitable and may lead to emissions exceeding the limitations
set by legislation [17]. This is not expected for catalytic combustion.
Based on their experimental results with a methanol fuel processor, Emonts et al.
calculated that for a light duty vehicle, carbon monoxide emissions could be reduced
to 1%, NO
x
emission to 10% and volatile organic compounds (without methane) to
10% using a fuel processor/fuel cell system compared with an internal combustion
engine, fulfilling the EU standards of 2005 for the new European driving cycle [18].
The catalytic afterburner was the only source of emissions in the system. It generated
1.8 mg km

À1
carbon monoxide, 0.3 mgkm
À1
NO
x
and 3.2 mg km
À1
unconverted
hydrocarbons. The Super Ultra Low Emissions vehicle regulation allowed much
higher values, namely 625 mg km
À1
carbon monoxide, 12 mg km
À1
NO
x
and
6mgkm
À1
unconverted hydrocarbons.
10
j
2 Fundamentals
Table 2.4 Key chemical properties of gases most relevant for fuel
processing (source: IMM, Institut f
€
ur Mikrotechnik Mainz).
t

C
C

pm
J mol
À1
K
À1
h
J mol
À1
LHV
J mol
À1
q
g/L
c
10
5
Pa s
k
Wm
À1
k
À1
CH
3
OH(g) 32.043 g mol
À1
25 43.71 À201167 675990 1.293 0.959 0.0157
250 52.64 À189322 – 0.737 1.721 0.0413
500 61.90 À171766 – 0.498 2.489 0.0853
750 69.78 À150576 – 0.377 3.191 0.1443

CH
4
(g) 16.043 g mol
À1
25 35.34 À74873 802284 0.647 1.111 0.0335
250 41.76 À65478 – 0.369 1.742 0.0699
500 48.65 À51765 – 0.250 2.287 0.1187
750 54.94 À35042 – 0.189 2.743 0.1704
C
2
H
5
OH(g) 46.069 g mol
À1
25 65.18 À234810 1277678 1.858 0.891 1.1620
250 81.83 À216398 – 1.059 1.502 1.1891
500 98.89 À187836 – 0.717 2.133 1.2195
750 111.64 À153868 – 0.542 2.734 1.2497
C
3
H
8
(g) 46.096 g mol
À1
25 73.40 À103847 2043972 1.779 0.828 0.0183
250 96.56 À82121 – 1.014 1.362 0.0473
500 116.59 À48467 – 0.686 1.849 0.0902
750 133.12 À7336 – 0.518 2.303 0.1424
CO(g) 28.01 g mol
À1

25 29.14 À110541 282964 1.130 1.767 0.0249
250 29.48 À103908 – 0.644 2.660 0.0399
500 30.15 À96218 – 0.436 3.447 0.0530
750 31.03 À88045 – 0.329 4.102 0.0646
CO
2
(g) 44.01 g mol
À1
25 37.10 À393505 – 1.775 1.493 0.0166
250 41.40 À384191 – 1.012 2.496 0.0354
500 45.30 À371986 – 0.685 3.420 0.0547
750 48.07 À358651 – 0.517 4.193 0.0721
H
2
(g) 2.016 g mol
À1
25 28.98 0 241826 0.081 0.892 0.1780
250 28.79 6478 – 0.046 1.306 0.2747
500 29.02 13784 – 0.031 1.689 0.3686
750 29.39 21309 – 0.024 2.033 0.4548
H
2
O(g) 18.015 g mol
À1
25 33.63 À241826 – 0.727 0.908 0.0197
250 34.40 À234087 – 0.414 1.827 0.0383
500 35.69 À224872 – 0.280 2.854 0.0668
750 37.16 À214883 – 0.212 3.848 0.1004
N
2

(g) 28.013 g mol
À1
25 29.13 0 – 1.130 1.785 0.0260
250 29.31 6595 – 0.644 2.685 0.0396
500 29.91 14206 – 0.436 3.511 0.0537
750 30.65 22221 – 0.329 4.217 0.0670
O
2
(g) 31.999 g mol
À1
25 28.96 0 – 1.291 2.069 0.0260
250 30.87 6947 – 0.736 3.147 0.0417
500 31.96 15181 – 0.498 4.139 0.0575
750 32.72 23720 – 0.376 4.954 0.0717
2.2 Basic Definitions, Calculations and Legislation
j
11
2.3
The Various Types of Fuel Cells and the Requirements of the Fuel Processor
The principle of the fuel cell was discoveredmorethan100yearsago bythe frequently
cited Sir William Grove but also byChristian Friedrich Schoenbein [19, 20]. However,
despite its large potential for highly efficient power generation, it still lacks wide-
spread applications due mostly to the economic aspects and some remaining
technical problems, such as durability issues.
Only a few aspects of the complex theory of electrical power generation by fuel cells
will be discussed briefly below, in order to highlight the consequences of these basic
rules on the fuel processor and its design.
2.3.1
PEM Fuel Cells
The most commonly used fuel cell is composed of a membrane that is able to

transport protons, the Proton Exchange Membrane (PEM), and of a catalyst, such as
platinum, positioned on both sides of the membrane on conducting material that
serves as the electrode. This arrangement is termed the Membrane Electrode
Assembly (MEA). Nafion
Ò
membranes, a fluorocarbon polymer of sulfuric acid
developed by DuPont, are the most frequently used membrane materials. Where
hydrogen is fed to one side of the MEA and oxygen to the other, hydrogen is oxidised
into water in a controlled combustion. In parallel, an electric potential of about 0.9 V
is generated, which decreases when current is withdrawn from the arrangement. To
achieve a voltage higher than 1 V, several MEAs need tobeswitchedinseries, forming
a fuel cell stack.Because the hydrogen and oxygen needto be distributed to each MEA
and over its entire area, gas distribution layers are also required, which must be
manufactured from electrically conductive material, in most instances this is
graphite or metal.
The side of the MEA, which catalyses the hydrogen dissociation:
H
2
! 2H
þ
þ 2e
À
ð2:22Þ
is the anode of theMEA,whereas the opposite side, which converts oxygenintowater:
O
2
þ 4H
þ
þ 4e
À

! 2H
2
O ð2:23Þ
forms the cathode. It is obvious that it is much more convenient to provide air instead
of oxygen to the cathode. However, this implies that not all gas fed to the cathode is
converted and that nitrogen and unconverted oxygen need to be removed from the
MEA. In practical systems a surplus of oxygen is fed to the cathode to avoid extremely
low concentrations at the exit. Frequently a two-fold or higher surplus of the
stoichiometric ratio l:
l ¼
n
O
2
2 n
H
2
¼ 2 ð2:24Þ
is fed to the cathode.
12
j
2 Fundamentals
For the anode, however, it is not typically the stoichiometric ratio but rather the
amount of hydrogen (or hydrogen and carbon monoxide, depending on the fuel cell
type) converted in the fuel cell as a percentage of the feed that is specified. This
amount is termed the hydrogen utilisation. For practical PEM fuel cell systems
running on reformate, 80% hydrogen utilisation may be assumed. This value,
however, is by no means fixed. When decreasing the electrical power withdrawal
from the fuel cell, while keeping the reformate flow constant, hydrogen utilisation
may drop to lower values.
Hydrogen utilization of less than 80% might be the preferred option when the

energy of unconverted hydrogen is required to supply processes downstream of the
anode (see Section 4.2).
The electrical power output of the fuel cell refers to about 50% of its energy
generation, the remaining energy is released as heat. The overall efficiency is even
lower due to energy losses of the fuel processor and to balance-of-plant components.
Such a system might be regarded as having a low overall efficiency. However, a
comparison with the overall efficiency of conventional passenger cars driven by
internal combustion engines reveals even lower values, 12% for gasoline-powered
cars and 15% for diesel-powered vehicles [21].
The unconverted part of the energy is released as heat within the fuel cell stack,
which generates a heat removal problem for practical systems. The heat may be
removed by water cooling, which in turn makes the system more complex and
expensive. Cooling by the anode and cathode gas flows are simpler alternatives. They
become more attractive the smaller the fuel cell system is, because cost issues are
more stringent in such instances.
Dilution of hydrogen withinert gases such as carbon dioxideand nitrogen, as is the
situation for reformate from fuel processors, should not impair the fuel cell power
generation by more than 10%, even if fuel utilisation is high and the hydrogen
content is as low as 40 vol.% [22].
Conventional low temperature PEM fuel cell anodes are sensitive to carbon
monoxide, which poisons the catalyst. In other words, the carbon monoxide is
preferentially adsorbed at the catalyst and thus the desired reactions can no longer
take place. However, the poisoning is partially reversible [23]. The dilution of the
hydrogen in reformate from fuel processors amplifies the poisoning effect unfortu-
nately [24].
To a certain extent this poisoning is suppressed or at least reduced for certain
alloys of platinum with other metals, amongst them, most commonly, is rut he-
nium, but also iron, cobalt, molybdenum and tungsten. This beneficial effect
originates from the fac t that platinum adsorbs carbon monoxide preferentially,
but water is adsorbed at the second metal, which makes the oxidation into carbon

dioxide feasible [25]. The long term carbon monoxide tolerance of PEM fuel cells
may be increased from a few ppm to values between 50 and 100 ppm at the most
by these means. Bimetallic catalysts such as platinum/ruthenium are also more
tolerant towards carbon dioxide [26], which is of course essential when reformate
is applied as the fuel. Increasing the fuel c ell operating t emperature decreases
the poisoning effect, which also applies to high temperature PEM fuel cells
2.3 The Various Types of Fuel Cells and the Requirements of the Fuel Processor
j
13
(see Section 2.3.2) [24]. The negative e ffect of carbon dioxide on fuel cell stability
originates from the reaction with the hydrogen adsorbed at the platinum to form
carbon monoxide, which means, in other words, that a reverse water–gas shift
reaction [see Eq. (3.4), Section 3.1] is taking place [27]. However, thi s effect is, yet
again, less significant over platinum/rut henium catalysts.
Another measure to reduce the det rimental effect of carbon monoxide on the
anode performance is the addition of a small amount of air during normal
operation, which is commonly termed bleed air. It oxidises the carbon monox-
ide adsorbed on the active sites of a selective oxidation catalyst layer [26] at the anode
(see Section 4.1.2). However, similar to the oxygen addition performed for the
preferential oxidation of carbon monoxide in a dedicated clean-up reactor (see
Section 3.10.2), addition of air to the hydrogen containing reformate generates
safety issues.
The poisoning effect of formaldehyde on PEM fuel cells is less dramatic compared
with carbon monoxide. The tolerance of PEM fuel cells to small amounts of formic
acid is approximately ten times higher [23]. However, Amphlett et al. judge formic
acid to be a severe poison for PEM fuel cells [28]. Formic acid causes irreversible
performance losses at concentrations as low as 250 ppm.
The poisoning effect of methane is very small for conventional PEM fuel cells [23].
Up to 5 vol.% are known to have no detrimental effect on the performance.
Methanol, which may originate from incomplete conversion in a methanol fuel

processor, can be tolerated in concentrations up to 0.5vol.% according to Amphlett
et al. [29]. Kawatsu, from Toyota, suggested that methanol forms formaldehyde in
the fuel cell anode and also shows cross-over through the fuel cell membrane [30].
Application of a platinum/ruthenium anode catalyst reduced these detrimental
effects [30]. The tolerance of PEM fuel cells towards methyl formate is assumed
to be similar to that for methanol according to Amphlett et al. [28].
Ammonia impairs the proton conductivity of the electrode considerably at less
than 100 ppm [24], and sulfur containing compounds also affect the catalyst perfor-
mance [27].
Hydrogen sulfide has a more severe poisoning effect on PEM fuel cells compared
with carbon monoxide. It originates from preferential adsorption; 1 ppm leads to
significant performance losses [24].
Metallic ions such ascopper, iron and sodium,which might be released from a fuel
processor or from fuel processing catalysts, impair the fuel cell performance.
Hydrogen peroxide might be formed from iron and copper ions, which attacks the
membrane [24].
Even if the reformate is purified by catalytic carbon monoxide clean-up to well
below 50 ppm carbon monoxide and if other impurities are reduced to the ppb level,
performance losses are to be expected when running a fuel cell with reformate. A 7%
lower power production was observed by Shi et al. [31] when running a 2kW PEM
fuel cell stack with reformate produced from liquid hydrocarbons.
Direct methanol and direct ethanol fuel cells are alternatives to PEM fuel cells
operated with methanol reformers. However, these types of fuel cells are not within
the scope of this book, and hence will not be discussed.
14
j
2 Fundamentals