fuel processing for fuel cells

Acknowledgement IX 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 Hi

Trang 3

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

A Züttel, A Borgschulte, L Schlapbach (Eds.)

Hydrogen as Future Energy Carrier

Trang 4

Fuel Processingfor Fuel Cells

Trang 5

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.

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 Nationalbibliogra ﬁe; detailed bibliographic data are available on the Internet at <http://dnb.d-nb.de>.

# 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 speci ﬁcally 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 Gra ﬁk-Design Schulz, Fußgönheim Printed in the Federal Republic of Germany Printed on acid-free paper

ISBN: 978-3-527-31581-9

Trang 6

Acknowledgement IX

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

3.2 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

Fuel Processing for Fuel Cells Gunther Kolb

Copyright Ó 2008 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim

Trang 7

3.11 Catalytic Combustion 52

3.12 Coke Formation on Metal Surfaces 52

4 Catalyst Technology for Distributed Fuel Processing Applications 574.1 A Brief Introduction to Catalyst Technology and Evaluation 574.1.1 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 984.2.12 Deactivation of Reforming Catalysts by Sulfur Poisoning 1014.3 Catalysts for Hydrogen Generation from Alternative Fuels 1054.3.1 Dimethyl Ether 105

4.3.2 Methylcyclohexane 106

4.3.3 Sodium Borohydride 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 1164.5.3 Methanation Catalysts 123

4.6 Combustion Catalysts 124

5 Fuel Processor Design Concepts 129

5.1 Design of the Reforming Process 129

Trang 8

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

7.1.1 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

Trang 9

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 35510.2 Production Techniques for Fuel Processors 359

10.2.1 Fabrication of Ceramic and Metallic Monoliths 35910.2.2 Fabrication of Plate Heat-Exchangers/Reactors 36110.2.3 Fabrication of Microchannels 365

10.2.4 Fabrication of Chip-Like Microreactors 367

10.2.5 Fabrication of Membranes for Hydrogen Separation 36910.2.6 Automated Catalyst Coating 370

References 373

Index 409

Trang 10

I would like to cordially thank my colleagues at IMM, in particular Dr Karl-PeterSchelhaas for fruitful discussions and input in thefields of calculations and materialproperties, Dr Hermann Ehwald for input in thefield of desulfurization catalysts,Tobias Hang for dealing with thefigures, Carola Mohrmann and Christina Miesch-Schmidt for dealing with the tables, Dr Athanassios Ziogas and Martin O’Connellfor dealing with the literature ordering and Sibylle for dealing with me when I was

‘‘hacking’’ through weekends and nights

Gunther Kolb

Trang 12

Introduction and Outline

Mankinds energy demand is increasing exponentially Between 1900 and 1997, theworlds population more than tripled and the average energy demand per humanbeing has also more than tripled, resulting in greater than thirteen times higheroverall global emissions [1] Thus the carbon dioxide concentration rose from 295parts per million in 1900 to 364 parts per million in 1997 [1] In 1997 almost allEuropean countries committed to reducing greenhouse gas emissions to an amount8% 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 thepotential to lower these emissions, owing to a potentially superior efﬁciencycompared with combustion engines Fuel cells require hydrogen for their operationand consequently numerous technologies are under investigation worldwide for thestorage of hydrogen, aimed at distribution, and mobile and portable applications.The lack of a hydrogen infrastructure in the short term, along with the highlyattractive energy density of liquid fossil and regenerative fuels, has created wide-spread research efforts in theﬁeld of distribution and on-board hydrogen generationfrom various fuels This complex chemical process, generally termed fuel processing,

is the subject of this book

The electrical power output equivalent of the fuel processors that are currentlyunder development world wide covers a wide range, from less than a watt to severalmegawatts Portable and small scale mobile fuel cell systems promise to be theﬁrstcommercial market for fuel cells, according to a market study of Fuel Cell Today in July

2003 [2] According to the same report, the number of systems built has increaseddramatically to up to more than 3000 in 2003 To date, most of these systems haveused Proton Exchange Membrane (PEM) fuel cells

Low power fuel processors (1–250 W) compete with both conventional storageequipment, such as batteries, and simpler fuel cell systems, such as Direct MethanolFuel Cells (DMFC)

Fuel cell systems for residential applications are typically developed for thegeneration 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

Trang 13

For mobile applications, systems designed to move a vehicle need to be guished from the Auxiliary Power Unit (APU), which either creates extra energyfor the vehicle (e.g., the air conditioning and refrigerator system of a truck) or works

distin-as a stand alone system for the electrical power supply

This book provides a general overview of theﬁeld of fuel processing for fuel cellapplications Its focus is on mobile, portable and residential applications, but thetechnology required for the smaller stationary scale is also discussed

In the second chapter fundamental deﬁnitions and the basic knowledge of fuelcell technology are provided, as far as is required to gain an insight into the interplaybetween the fuel cell and its hydrogen supply unit– the fuel processor

The third chapter deals with the reforming chemistry of conventional andalternative 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 ofChapter 5 Design concepts and numerical simulations presented in the openliterature are discussed for reforming, catalytic carbon monoxide clean-up andphysical clean-up strategies, such as membrane separation and pressure swingadsorption In addition, fuel processor concepts are then presented and the interplaybetween the various fuel processor components is explained Details of the basicengineering of fuel processors and dynamic simulations are discussed, coveringstart-up and control strategies Some tips and the basic knowledge required toperform such calculations are provided

There are three basic types of fuel processing reactors, namely ﬁxed catalystbeds, monoliths and plate heat-exchangers, which are explained in Chapter 6.Chapter 7 then shows the practical applications of such reactors, as published inthe literature

In Chapter 8 some important aspects of balance-of-plant components are cussed, and Chapter 9 presents complete fuel processors for all types of fuels,while cost and production issues are the subject of Chapter 10

Trang 14

Fundamentals

This chapter provides information about common fossil fuels, necessary deﬁnitions

in thefield of fuel processing and the basic knowledge from the wide field of fuel celltechnology It is by no means comprehensive and is not a substitute for the dedicatedliterature in thesefields Rather, it provides a brief summary for readers who wish togain an overview of the topic of fuel processing without the need to use too muchadditional literature

2.1

Common Fossil Fuels

Fuels are solid, liquid or gaseous energy carriers To date, practically all of the fuelsavailable on the market are based upon fossil sources and thus contain hydrocarbons

of varying composition However, alternative fuels such as alcohols and hydrides mayserve as future energy carriers Table 2.1 provides an overview of the conventionalfuels and of the most important alternative fuels, which may act as future hydrogensource for fuel cells along with their key properties

A comparison of the gravimetric and volumetric density of various hydrogencarriers shows that liquid hydrocarbons have– apart from borohydrides – by far thebest combined properties (see Figure 2.1)

Table 2.2 shows the maximum amount of work that can be converted intoelectricity from various fuels, in theory Compared with the gravimetric and volu-metric energy density of 1 MJ kg1or<2 MJ L1of lithium-ion and zinc-air batteries,these values are considerably higher

The composition of fossil hydrocarbon fuels may vary widely depending on thesource of the crude oil that is processed in the reﬁnery

The composition of natural gas is predominantly methane, and also containsseveral percent ethane and propane In addition, minor amounts of butane andhigher hydrocarbons are present, plus carbon dioxide and nitrogen

Table 2.3 shows the composition of natural gas from various sources [5] Naturalgas also contains sulfur compounds at the ppm-level, such as hydrogen sulﬁde and

Trang 16

diethyl sulﬁde, and mercaptanes, such as ethyl mercaptane [(C2H5)CHS] and tertiarybutyl mercaptane [(CH3)3CHS].

Amongst all the fossil fuels, propane contains the highest amount of hydrogen on agravimetric basis, which even exceeds liqueﬁed hydrogen, when the weight of thestorage tanks is taken into consideration [6] Propane is usually marketed as liqueﬁedpetroleum gas, which is a mixture of propane and butane in various ratios

For gasoline, only approximate characterization parameters are provided, such asthe octane number, the boiling point distribution, and the saturated hydrocarbons(alkanes), unsaturated hydrocarbons (oleﬁns) and aromatics content The content ofcontaminants, such as sulfur, is important

Volumic density (g H

Lantanides

Hydrides base(TI et/ou Zr)Nanotubes

Compress

700bars

350bars

Liquids H

2

Liquids hydrocarbon

FullereneHydr ides MBH4

D.O.Ex

Glass micro-sphere

Hydr ides base Mg

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 H2 via reforming

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.

Trang 17

Regular gasoline, at least according to German standards, is well represented bythe overall formula C7H12[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 widelyused in the world are Jet fuel A and A1 [10] with a boiling range between 150 and 300C.Diesel fuels contain mainly iso-paraffins, but also n-paraffins, mono-, di-, tri-,tetra-cycloparaffins, alkylbenzenes, naphthalenes and phenanthrenes and evenpyrenes [11]

2.2

Basic Definitions, Calculations and Legislation

Fuel processing is the conversion of hydrocarbons, alcohol fuels and other alternativeenergy carriers into hydrogen containing gas mixtures The chemical conversion isachieved in most instances in the gaseous phase, normally heterogeneously catalysed

in the presence of a solid catalyst and less frequently homogeneously at hightemperature without a catalyst

Theﬁrst step of the conversion procedure is generally termed reforming, and hasbeen well established in large scale industrial processes for many decades Theindustrial 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 carbonmonoxide and hydrogen, which is then used for numerous processes in large scalechemical production, which are not subject of this book

Rather, the focus of this book is the technology that provides a hydrogen containinggas mixture, termed the reformate, which is suitable for feeding into a fuel cell Thefuel cell then converts hydrogen into electrical energy Carbon monoxide may also beconverted, 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 CxHyOzis easily determined by thefollowing formula [13]:

Table 2.3 Composition of natural gas from various sources [5].

Trang 18

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 thehydrogen and carbon monoxide that are produced to the LHV of the fuelconsumed:

hFuel processor¼LHVðH2Þ nH 2þ LHVðCOÞ nCO

up to 120% efﬁciency according to the Eqs (2.2) and (2.3)

The carbon monoxide content of the reformate obviously needs to be minimisedfor low temperature proton exchange membrane fuel cells, but other fuel cells maywell utilize it as a fuel (see Section 2.3.2) The same applies for methane in certain fuelcells Therefore, the heating value of the hydrogen alone does not provide theappropriate number for the calculation of efﬁciency in this instance

A modified definition of the fuel processor efficiency provides a more realisticvalue than Eqs (2.2) and (2.3) [14]:

However, for PEM fuel cells methane and carbon monoxide could be excludedfrom efﬁciency calculations, because they are not converted in the fuel cell

The following definition of efficiency was proposed by Feitelberg [15] It wasmodified to also take methane and carbon monoxide fed to the fuel cell intoconsideration as discussed above:

hFuel processor¼ LHVðH2ÞnH 2þ LHVðCOÞnCOþ LHVðCH4ÞnCH 4

LHVðFuelÞnFuelþ ½LHVðH2Þ nH 2þ LHVðCOÞnCO

þ LHVðCH4Þ nCH

ð2:5Þ

Trang 19

This deﬁnition seems to be the most realistic, because it takes all products intoconsideration in the numerator and all feed entering the fuel processor in thedenominator, which is in agreement with the rules for energy balancing.

Hagh [13] has derived a general formula for the fuel processor efﬁciency Thus, bydescribing the fuel processing reactions with the following general and simpliﬁedformula (by-products such as methane are not taken into consideration, the sameapplies for unconverted methane in case of methane fuel processing):

CxHyOzþ a O2þ b H2O! d CO þ e CO2þ f H2 ð2:6Þdeﬁning the stoichiometric ratio (SR) as the ratio of oxygen to oxygen required forcomplete combustion:

e þ1

2

0B

1C

Ax zy

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 isneither consumed by the water–gas shift reaction downstream of the reformer (seeSection 3.10.1), nor required to prevent dry-out of the membrane of low temperaturePEM fuel cells (see Section 2.3.1), nor to prevent coking in high temperature fuel cells(see Section 2.3.2), the water should be recovered to avoid a negative water-balance ofthe system However, the heat of condensation is difﬁcult to recover and commonlylost to the cooling air Thus, excess water should be minimised in fuel processor/fuelcell systems Condensers might be integrated upstream or downstream of thefuel cell to ensure water recovery and net positive water balancing of the wholesystem Such condensers may also recover water produced by the fuel cell itself

Trang 20

Low temperature heat losses are usually mandatory in all fuel cell/fuel processorsystems because the efﬁciency of heat-exchangers is limited and the system needs towork with cooling air, which might even have elevated temperatures in the summertime The major portion of the heat losses does not originate from the fuel processorbut rather from the fuel cell itself for low temperature proton exchange (PEM) fuelcells Thus the efﬁciency of the fuel processor is usually high provided heat losses areneglected, because internal heat-exchangers can keep the high temperature heatwithin 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 fuelprocessor require an elevated operating temperature This temperature will notdecrease at partial load Therefore heat losses remain constant 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 thusproposed as a modification to the definition provided by Schmid and W€unning [16]and others:

hFuel processor; operation¼ 1 Q_cond;FPþ Q_losses

Trang 21

and the molar heat capacity is according to:

To date, no emission regulations exist for fuel cell systems European legislationhas directives for heating systems based on natural gas or Liquiﬁed Petroleum Gas(LPG) They limit nitrous oxides (NOx) 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 NOxto 80 ppm and carbon monoxide to

60 ppm

In instances where homogeneous combustion is applied in fuel processors, theformation of NOxis inevitable and may lead to emissions exceeding the limitationsset 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%, NOxemission to 10% and volatile organic compounds (without methane) to10% using a fuel processor/fuel cell system compared with an internal combustionengine, 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 generated1.8 mg km1 carbon monoxide, 0.3 mg km1 NOx and 3.2 mg km1 unconvertedhydrocarbons The Super Ultra Low Emissions vehicle regulation allowed muchhigher values, namely 625 mg km1 carbon monoxide, 12 mg km1 NOx and

6 mg km1unconverted hydrocarbons

Trang 22

Table 2.4 Key chemical properties of gases most relevant for fuel

processing (source: IMM, Institut f€ur Mikrotechnik Mainz).

c

105Pa s

k

W m1k1CH3OH(g) 32.043 g mol1 25 43.71 201167 675990 1.293 0.959 0.0157

Trang 23

The Various Types of Fuel Cells and the Requirements of the Fuel Processor

The principle of the fuel cell was discovered more than 100 years ago by the frequentlycited Sir William Grove but also by Christian 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 remainingtechnical problems, such as durability issues

Only a few aspects of the complex theory of electrical power generation by fuel cellswill be discussed brieﬂy below, in order to highlight the consequences of these basicrules 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 totransport protons, the Proton Exchange Membrane (PEM), and of a catalyst, such asplatinum, positioned on both sides of the membrane on conducting material thatserves as the electrode This arrangement is termed the Membrane ElectrodeAssembly (MEA) Naﬁon membranes, aﬂuorocarbon polymer of sulfuric aciddeveloped by DuPont, are the most frequently used membrane materials Wherehydrogen is fed to one side of the MEA and oxygen to the other, hydrogen is oxidisedinto 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 Toachieve a voltage higher than 1 V, several MEAs need to be switched in series, forming

a fuel cell stack Because the hydrogen and oxygen need to be distributed to each MEAand over its entire area, gas distribution layers are also required, which must bemanufactured from electrically conductive material, in most instances this isgraphite or metal

The side of the MEA, which catalyses the hydrogen dissociation:

is the anode of the MEA, whereas the opposite side, which converts oxygen into water:

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 isconverted and that nitrogen and unconverted oxygen need to be removed from theMEA In practical systems a surplus of oxygen is fed to the cathode to avoid extremelylow concentrations at the exit Frequently a two-fold or higher surplus of thestoichiometric ratiol:

Trang 24

For the anode, however, it is not typically the stoichiometric ratio but rather theamount of hydrogen (or hydrogen and carbon monoxide, depending on the fuel celltype) converted in the fuel cell as a percentage of the feed that is specified Thisamount is termed the hydrogen utilisation For practical PEM fuel cell systemsrunning on reformate, 80% hydrogen utilisation may be assumed This value,however, is by no meansfixed When decreasing the electrical power withdrawalfrom the fuel cell, while keeping the reformateflow constant, hydrogen utilisationmay drop to lower values.

Hydrogen utilization of less than 80% might be the preferred option when theenergy of unconverted hydrogen is required to supply processes downstream of theanode (see Section 4.2)

The electrical power output of the fuel cell refers to about 50% of its energygeneration, the remaining energy is released as heat The overall efficiency is evenlower 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, acomparison with the overall efficiency of conventional passenger cars driven byinternal combustion engines reveals even lower values, 12% for gasoline-poweredcars 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 beremoved by water cooling, which in turn makes the system more complex andexpensive Cooling by the anode and cathode gasﬂows are simpler alternatives Theybecome more attractive the smaller the fuel cell system is, because cost issues aremore stringent in such instances

Dilution of hydrogen with inert gases such as carbon dioxide and nitrogen, as is thesituation for reformate from fuel processors, should not impair the fuel cell powergeneration by more than 10%, even if fuel utilisation is high and the hydrogencontent is as low as 40 vol.% [22]

Conventional low temperature PEM fuel cell anodes are sensitive to carbonmonoxide, which poisons the catalyst In other words, the carbon monoxide ispreferentially adsorbed at the catalyst and thus the desired reactions can no longertake place However, the poisoning is partially reversible [23] The dilution of thehydrogen 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 certainalloys of platinum with other metals, amongst them, most commonly, is ruthe-nium, but also iron, cobalt, molybdenum and tungsten This beneﬁcial effectoriginates from the fact that platinum adsorbs carbon monoxide preferentially,but water is adsorbed at the second metal, which makes the oxidation into carbondioxide feasible [25] The long term carbon monoxide tolerance of PEM fuel cellsmay 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 moretolerant towards carbon dioxide [26], which is of course essential when reformate

is applied as the fuel Increasing the fuel cell operating temperature decreasesthe poisoning effect, which also applies to high temperature PEM fuel cells

Trang 25

(see Section 2.3.2) [24] The negative effect of carbon dioxide on fuel cell stabilityoriginates from the reaction with the hydrogen adsorbed at the platinum to formcarbon monoxide, which means, in other words, that a reverse water–gas shiftreaction [see Eq (3.4), Section 3.1] is taking place [27] However, this effect is, yetagain, less signiﬁcant over platinum/ruthenium catalysts.

Another measure to reduce the detrimental effect of carbon monoxide on theanode performance is the addition of a small amount of air during normaloperation, 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 thepreferential oxidation of carbon monoxide in a dedicated clean-up reactor (seeSection 3.10.2), addition of air to the hydrogen containing reformate generatessafety issues

The poisoning effect of formaldehyde on PEM fuel cells is less dramatic comparedwith carbon monoxide The tolerance of PEM fuel cells to small amounts of formicacid is approximately ten times higher [23] However, Amphlett et al judge formicacid to be a severe poison for PEM fuel cells [28] Formic acid causes irreversibleperformance 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 fuelprocessor, can be tolerated in concentrations up to 0.5 vol.% according to Amphlett

et al [29] Kawatsu, from Toyota, suggested that methanol forms formaldehyde inthe fuel cell anode and also shows cross-over through the fuel cell membrane [30].Application of a platinum/ruthenium anode catalyst reduced these detrimentaleffects [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 lessthan 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 comparedwith carbon monoxide It originates from preferential adsorption; 1 ppm leads tosigniﬁcant performance losses [24]

Metallic ions such as copper, iron and sodium, which might be released from a fuelprocessor or from fuel processing catalysts, impair the fuel cell performance.Hydrogen peroxide might be formed from iron and copper ions, which attacks themembrane [24]

Even if the reformate is puriﬁed by catalytic carbon monoxide clean-up to wellbelow 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 2 kW PEMfuel cell stack with reformate produced from liquid hydrocarbons

Direct methanol and direct ethanol fuel cells are alternatives to PEM fuel cellsoperated with methanol reformers However, these types of fuel cells are not withinthe scope of this book, and hence will not be discussed

Trang 26

High Temperature Fuel Cells

A major drawback to PEM fuel cells, apart from their sensitivity to carbon monoxidepoisoning, is the osmotic drag of the Naﬁonmembrane, which causes water migra-tion from the side of the anode to that of the cathode This causes performance lossesunless the feed is humidiﬁed [32] Polybenzimidazole membrane material dopedwith phosphoric acid is an alternative, which allows for higher operating temperature

of the PEM fuel cell of between 150 and 170C and requires no humidiﬁcation [32].Additionally, it has signiﬁcantly higher tolerance towards carbon monoxide in con-centrations exceeding 1 vol.% This membrane material was developed by the formerHoechst Company; the technology is now owned by the BASF Company

Classical phosphoric acid fuel cells use phosphoric acid as the electrolyte, which isimmobilized in a Teﬂon bonded silicon carbide matrix Phosphoric acid fuel cellsusually work at temperatures around 200C and are able to tolerate carbon monoxidelevels of up to 2 vol.% [1] Platinum/ruthenium as the anode catalyst may improve theperformance in presence of carbon monoxide, similar to PEM fuel cells [33]

Alkaline fuel cells use an aqueous solution of potassium hydroxide as theelectrolyte When the solution is concentrated (85%), the operating temperature ofthe alkaline fuel cell can be as high as 250C, lower concentrations result in loweroperating temperatures, below 120C Both carbon monoxide and carbon dioxideact as poisons, which means that this fuel cell type is not really suited forhydrocarbon reformate, unless the hydrogen is separated by membranes or pressureswing adsorption (see Section 5.2) Hydrogen produced from ammonia is aninteresting fuel for this type of fuel cell, because reformate obtained from ammoniacracking contains no carbon oxides [34] (see also Section 3.8)

Solid oxide fuel cells contain solid electrolytes, which are frequently based onzirconia stabilised by yttria On the cathode side strontium or calcium dopedlanthanum manganese oxide (LaMnO3) is most commonly used, while the anodeside comprises of yttria stabilized zirconia frequently doped with nickel to achieveelectrical conductivity The oxygen is reduced at the cathode and the oxygen anionsdiffuse through the electrolyte to the anode, where they oxidise the fuel Hydrogen

or carbon monoxide may serve as the fuel To achieve sufﬁcient mobility of theoxygen anions, high operating temperatures of between 800 and 1000C are usuallyrequired [35] This value may be lowered to 600C for thin electrolytes made fromcertain electrolyte materials, such as lanthanum gallate based perovskites [36].However, the nickel catalyst is subject to coke formation when carbon monoxide

is present in the reformate Addition of steam is one possible way to reduce cokeformation Through the addition of steam, internal reforming of light hydrocarbons,such as methane, becomes feasible within the solid oxide fuel cell [37] Internalreforming may either be performed at the catalyst positioned adjacent to the anode

or at the anode itself [37] However, the efﬁciency of the fuel cell is reduced due tothe dilution of anode feed by the steam [35] On the other hand, steam reformingconsumes energy, which helps to cool the fuel cell [38, 39] Various alternative anodematerials, such as copper-based systems, are under investigation, which reduce

Trang 27

carbon formation in the solid oxide fuel cell [35] These issues will not be discussedfurther in the present book.

Reforming of natural gas for solid oxide fuel cells is achieved either internally asdescribed above or externally by a pre-reformer reactor [40, 41] Further processing ofthe fuel is not required, because of the unlimited tolerance of the fuel cell to carbonmonoxide The re-circulation of anode off-gas to the pre-reformer [42] is an interest-ing option for solid oxide fuel cells Through these means, addition of water isomitted, which clearly decreases the complexity of the system and reduces cold startproblems (see also Section 3.5)

Molten carbonate fuel cells operate at temperatures around 650C and are tolerant

to unlimited amounts of carbon monoxide In most instances mixtures of lithiumcarbonate and potassium carbonate act as the electrolyte The electrolyte is suspended

in an insulating and chemically inert lithium aluminate ceramic Nickel or nickel–chromium alloys serve as the anode catalysts, while nickel oxide is used as the cathodecatalysts

Autothermal or steam reforming of methane was considered in thermodynamiccalculations by Cavallaro and Freni for reformers, which were integrated into amolten carbonate fuel cell [43] Direct or indirect internal reforming is possiblewithin the molten carbonate fuel cell The reforming may be performed either bythe anode itself or by a dedicated catalyst in the anode compartment in analogywith the solid oxide fuel cell, as has been explained above Direct reforming ofalcohol fuels is also possible in molten carbonate fuel cells [44], whereas processing

of liquid hydrocarbons requires a pre-reformer

Theoretical calculations to evaluate internal reforming of methanol, ethanol andmethane for molten carbonate fuel cells were performed by Maggio et al [45]

Trang 28

The Chemistry of Fuel Processing

The most important parameter of reforming as theﬁrst step in fuel processing is thefeed composition In this book the feed composition is provided by the followingterms:

The steam/carbon ratio S/C, which is the ratio of molar steamﬂow rate to the molarﬂow rate of the fuel CxHyOzmultiplied by the number of carbon atoms in the fuel, x:S

hydro-CxHyOzþ ðx zÞH2O! xCO þ

x z þy2

The product mixture of the reaction is known as the reformate The reaction isendothermic and thus requires a heat supply Besides hydrogen and carbon monox-ide, the reformate usually contains signiﬁcant amounts of unconverted steam, and to

a lesser extent some unconverted fuel and carbon dioxide, the latter being formed bythe consecutive water–gas shift reaction:

COþ H2O! CO2þ H2 DH0

Trang 29

The water–gas shift reaction increases the hydrogen concentration of the mate This reaction is usually fast enough at the elevated temperatures of hydrocar-bon reforming to achieve thermodynamic equilibrium Owing to its exothermiccharacter, higher reaction temperatures favour the reverse reaction (see Section 3.10).Methane is frequently formed in signiﬁcant amounts, up to several percent.Higher reaction temperatures suppresses methane formation, according to theequilibrium of the methanation reaction, which is of course the reverse reaction

refor-of methane steam reforming:

3H2þ CO $ H2Oþ CH4 DH0

298¼ 253:7 kJ mol 1 ð3:5ÞThe equilibrium conversion of methane steam reforming as calculated fordifferent S/C ratios and system pressure is shown in Figure 3.1

The carbon dioxide reforming reaction:

Pa) suffers from low methane conversion due tothe thermodynamic equilibrium [48] Such high pressure is required, for example, ifthe reforming process is combined with membrane separation using conventionalpalladium membranes (see Section 5.2.4) Industrial steam reformers work with

Figure 3.1 Equilibrium conversion of methane steam reforming

versus reaction temperature for various S/C ratios and system

pressures [46].

Trang 30

pressures exceeding 20 bar because of the requirements of the puriﬁcation processes,such as pressure swing adsorption (see Section 5.2.5) and the processes downstream

of the reformer

Elevated pressure also decreases the equilibrium conversion of methanol steamreforming [49] At 21 bar pressure, which may be required to run a steam reformeralong with a membrane separation system, the equilibrium methanol conversiondecreases to 99.2% at a reaction temperature of 280C and to 98.1% at 260C with aS/C ratio 1.5 [50]

Figure 3.2 shows the equilibrium conversion of methanol steam reforming as afunction of the S/C ratio of the feed [25] It is obvious that the maximum hydrogenconcentration in the reformate is gained at S/C 1 However, to minimize the carbonmonoxide concentration in a practical system, a surplus of steam is required.Therefore, in practice, systems operate at S/C ratios of between 1.3 and 2.0 Elevatedpressure also decreases the selectivity towards carbon monoxide [49]

Further possible by-products of methanol steam reforming are formic acid andmethyl formate (CH3OCHO), which are both harmful, at least for PEM fuel cells (seeSection 2.3.1) These by-products are formed because methanol steam reformingprobably takes place with both of these species as intermediate products, as proposed

by Takahashi et al [51]:

Figure 3.2 Equilibrium gas composition of methanol steam

reform-ing versus S/C ratio of the feed; pressure 5 bar; temperature

200C [25].

Trang 31

The decomposition reaction of formic acid is usually fast [52] Another possible product of methanol steam reforming is dimethyl ether, which is generated bymethanol dehydration (see Section 4.2.1):

Dimethyl ether formation is favoured by decreasing the reaction temperature [53].Steam reforming of all fuels typically under investigation in the research commu-nity (except for methanol) is performed at temperatures of at least 400C or higher.Under these reaction conditions, light hydrocarbons may be formed, namelymethane, ethylene and propylene

The second important alcohol fuel frequently investigated is ethanol Ethanolsteam reforming:

C2H5OHþ H2O! 2CO þ 4H2 DH0

298 ¼ þ 256 kJ mol 1 ð3:11Þusually requires a higher reaction temperature than methanol steam reforming Theethanol decomposition reaction is one important side reaction, which generatescarbon monoxide and methane and is favoured by higher reaction temperatures [54]:

Hydrocarbon fuels require a higher reaction temperature compared with ethanol

As illustrated in Figure 3.3, the equilibrium dry gas composition of reformate gained

Trang 32

from steam reforming of n-heptane shows an increasing carbon monoxide contentand a decreasing methane content with increasing temperature, while the carbondioxide concentration remains almost constant.

The methane formation through the reforming process is not only dictated bythermodynamics but also by the properties of the catalyst that is applied

Reactions other than methanation may well be responsible for the formation ofmethane from hydrocarbon fuels [56]:

CxHy! CH4þ Cx 1Hy 4 ð3:17Þ

CxHyþ H2! CH4þ Cx 1Hy 2 ð3:18ÞHowever, methane can be tolerated by most fuel cell systems up to a concentration

of 5 vol.% without damage

Unconverted fuel and by-products are undesired because in most instances theycannot be converted by fuel cells and so they reduce the efﬁciency of the system Inaddition, they usually also poison the fuel cell catalyst and the catalyst of subsequentcarbon monoxide clean-up systems (see Sections 2.3 and 4.5)

In general, steam reforming of higher hydrocarbons is usually performed at S/Cratios exceeding the stoichiometry (S/C¼ 2) in order to suppress coke formation AnS/C ratio of 3 may be required for higher hydrocarbons in the absence of oxygen in thefeed as for autothermal reforming (see Section 3.3 below) However, excess steamreduces the overall efﬁciency of the system (see Sections 2.2 and 5.4.3)

Figure 3.3 Equilibrium composition of dry reformate from

n-heptane steam reforming versus temperature; pressure 30 bar;

S/C ratio 4 [55].

Trang 33

The reaction is signiﬁcantly faster than steam reforming and is usually performed

in the diffusion limited regime [57] The catalytic reaction of hydrocarbons such asoctanes over rhodium coated foams starts at relatively low temperatures of around

250C [58]

An obvious advantage of partial oxidation is that only an air feed is required, apartfrom the fuel This makes the system simpler because evaporation processes, asrequired for steam reforming, are avoided On the other hand, the amount of carbonmonoxide formed is considerably higher compared with steam reforming Thisputs an additional load onto the subsequent clean-up equipment, but only whereCO-sensitive fuel cells are connected to the fuel processor When fuels are converted

by partial oxidation, some total oxidation usually takes place as an undesired sidereaction [46] In practical applications, an excess of air is fed to the system andconsequently even more fuel is subject to total oxidation The water formed by thecombustion process in turn gives rise to some water–gas shift Another typical by-product of partial oxidation is methane, which is formed according to reaction (3.5).Coke formation is a critical issue (see Sections 4.1.1 and 4.2.11) Coke may be formed

by reaction of carbon monoxide with hydrogen:

All hydrocarbons can form coke by cracking reactions, as exempliﬁed by methane:

CH4! 2H2þ C DH0

Two reaction mechanisms do exist in the literature for partial oxidation One ofthese proposes that the reaction begins with catalytic combustion followed byreactions of lower rate, namely steam reforming, CO2reforming and water–gasshift [60] The other mechanism proposes direct partial oxidation at very shortresidence times [61]

Trang 34

Ioannides and Verykios coated only the inner surface of one tubular reactor withrhodium catalyst, while another reactor was coated on both sides Partial oxidation ofmethane was performed in such a manner that the reactants were fed to the tube andthen passed the tube outer surface counter-currently, in an annular gap between thetube and the reactor housing The reactor that was coated on both sides showed aneven temperature profile, while the reactor where only inside of the tube was coatedshowed hot spot formation The different performances were attributed to endother-mic reactions in the second annularflow path The endothermic reactions consumedthe energy that had been generated in thefirst flow path inside the tube by fast initialexothermic reactions [62].

Lyubovsky et al performed partial oxidation of methane at an O/C ratio of 1.2over Microlith metallic screens coated with ceramic catalyst carrier, which wereimpregnated with rhodium [63] They took gas samples over the short length ofthis catalyst bed As shown in Figure 3.4, the oxygen was completely consumed aftertheﬁrst 4 mm of the bed, while only a small amount of methane had been converted.The primary products were steam, carbon monoxide and hydrogen, carbon dioxidewas formed to a lesser extent The steam was then consumed downstream of thereactor, converting residual methane As expected from the thermodynamic equi-librium, conversion decreased with increasing pressure Figure 3.5 shows thedecreasing selectivity of the partial oxidation reaction towards steam with increasingconversion The dominant selectivity towards carbon monoxide remained almostunchanged when the conversion increased The hot spot in the reactor reached

1000C

Specchia et al performed partial oxidation of methane over rhodium/a-aluminaﬁxed catalyst beds for short contact times over a time range of between 10 and 40 ms[64] With increasing catalyst particle sizes, conversion decreased, which wasattributed to transport limitations Higher reactor temperature was observed forlarger particles and thus more exothermic reactions took place When increasing theparticle size and the weight hourly space velocity (see Section 4.1), the water content

in the product increased, while less carbon monoxide was found and carbon dioxideremained at an unchanged low concentration Similar to the results of Lyubovskidiscussed above, steam seemed to be a primary product of the reaction

Partial oxidation is highly exothermic, which makes heat removal a critical issue inorder to prevent damage to the catalyst structure (see Section 4.2)

Some important aspects of the partial oxidation reaction were highlighted byPanuccio et al., from the group working with L.D Schmidt [65] They investigated theconversion of octane isomers over rhodium coated ceramic foams made froma-alumina It could be demonstrated, experimentally, that branched hydrocarbonssuch as isooctane (2,2,4 trimethylpentane) show slightly lower selectivity towards thedesired products carbon monoxide and hydrogen, but generate higher amounts ofthe total oxidation products independently of the feed composition (O/C ratio, shown

as its inverse C/O in Figure 3.6) However, at O/C ratios exceeding unity (C/Os ratiosless than one) the differences were more pronounced Under these conditions, fullconversion of the fuel was always achieved but signiﬁcant amounts of light hydro-carbons such as ethylene, propylene and butylenes were formed Isooctane feed

Trang 35

generated signiﬁcantly less ethylene and more propylene and isobutene comparedwith n-octane, especially at O/C ratios of less than one.

The reaction temperature ranged between 900 (O/C ratio 1) and 1150C (O/C ratio1.25) in the adiabatic reactors [65]

Figure 3.4 Species concentration and temperature profile over a

metal screen catalyst bed coated with rhodium catalyst for

partial oxidation of methane; O/C ratio 1.2; pressure 2 bar (top),

4 bar (centre) and 8 bar (bottom) [63].

Trang 36

Furthermore, these workers investigated the effect of the different pore sizes ofthe foams on selectivity As shown in Figure 3.7, the larger pore size [45 ppi (poresper inch) or 470mm pore size] generated lower selectivity towards hydrogen andcarbon monoxide compared with the smaller pores (45 ppi or 250mm pore size) Notshown here is the selectivity towards light hydrocarbons, which was approximatelydouble in the large pores These effects were attributed to the higher contribution of

Figure 3.5 Product selectivity versus methane conversion for

partial oxidation of methane over rhodium catalyst; O/C ratio

1.2 [63].

Figure 3.6 Selectivities as determined for n-octane, isooctane and

a 1:1 mixture thereof over a rhodium coated ceramic foam [65].

Trang 37

homogeneous reactions in the larger pores favouring the generation of light carbons rather than syngas (carbon monoxide and hydrogen) However, numericalsimulations revealed that homogeneous gas phase chemistry alone could not explainthe experimental results The catalyst also played a signiﬁcant role in productdistribution.

hydro-Some workers add steam to the feed of partial oxidation reactors (running at anO/C ratio higher than 1) and term this autothermal reforming This is, however,misleading, because oxygen will be consumed initially in the reactor Thus to a largeextent the fuel molecules are already converted and there is little feed left for steamreforming Such operating conditions decrease the system efﬁciency signiﬁcantly,

as shown in Section 3.3 This has the consequence that it is mainly water–gas shiftthat will take place in the reactor downstream of the inlet section, owing to thepresence of steam Because both partial oxidation and water–gas shift are exother-mic, this mode of operation does not help to reduce hot spot formation in the reactorsubstantially However, coke formation is certainly suppressed in the presence ofsteam

Running a fuel processor at an O/C ratio of 1.0 or higher in the presence of steamfeed should be termed steam supported partial oxidation

Figure 3.8 shows methane conversion, hydrogen yield and adiabatic temperaturerise of a methane partial oxidation reactor as a function of the O/C ratio (hereexpressed as the air ratio,l) [66] Owing to coke formation below O/C ¼ 1.2 (l ¼ 0.3)and it being more pronounced below O/C¼ 1 (l ¼ 0.25), the reactor temperaturerises only moderately until O/C reaches 1.2 Beyond this value the adiabatic

Figure 3.7 Selectivities as determined for n-octane and isooctane

over a rhodium coated ceramic foam of different pore sizes,

described as ppi (pores per inch) [65].

Trang 38

temperature increases rapidly because catalytic combustion becomes dominant,which also decreases the hydrogen yield considerably of course The adiabatictemperature of 743C calculated at O/C¼ 1.2 is relatively low for a partial oxidationreactor, which originates from the low pre-heating temperature of 200C assumedfor the feed Seo et al also demonstrated, by their calculations, that increasing thefeed temperature does not affect the O/C ratio required to prevent coke formation[66] However, the feed temperature increases the adiabatic temperature of thereactor At a feed temperature of 500C and an O/C ratio of 1.2, the adiabatic reactortemperature readily exceeded 900C.

Similar thermodynamic equilibrium calculations were performed by Docter andLamm for gasoline feed (C7H12) at various O/C ratios (again expressed as thel valuefor total combustion), which are shown for S/C ratios of 0 and 0.7 (partial oxidation)

in Figure 3.9 (steam supported partial oxidation) [7] This makes it clear that coke isonly stable up to a certain O/C ratio and coke formation decreases with an increasedconcentration of steam However, coke formation is not only dictated by thethermodynamic equilibrium but also affected by other factors such as catalystproperties (see Section 4.2) The decreasing content of methane in the reformatewith increasing reactor temperature becomes obvious in Figure 3.9, which origi-nates from the thermodynamic equilibrium In a practical system the extent ofmethane formation also depends on the performance of the catalyst The maximumhydrogen concentration was calculated for air ratios between 0.3 and 0.35, whichcorresponds to an O/C ratio of 1 This is equivalent to the stoichiometric composition

of partial oxidation It is important to realize that a much higher hydrogen content inthe reformate is achieved in the presence of steam, owing to the water–gas shiftreaction In the presence of steam, steam reforming may occur below l ¼ 0.3,

Figure 3.8 Effect of the O/C ratio [expressed as an air ratio,

l ¼ (O/C)/4] on the equilibrium values of conversion and

hydrogen yield, and on the adiabatic temperature rise of the

reactor for methane partial oxidation; pre-heating temperature

200C; pressure 1 bar [66].

Trang 39

because the oxygen is then insufﬁcient to convert the fuel completely through partialoxidation.

The suppression of methane formation with increasing reaction temperatureand the increasing formation of carbon monoxide according to thermodynamicswas also determined experimentally by Moon et al [67] as shown in Figure 3.10

Figure 3.9 Thermodynamic equilibrium gas composition and

reformer adiabatic temperature versus air ratio (l) for gasoline

reforming [7]; feed temperatures were 400C for air, 200C for

steam and 20C for the fuel; left, S/C ¼ 0; right, S/C ¼ 0.7.

Figure 3.10 Dry reformate composition during isooctane partial

oxidation in the presence of steam; S/C ¼ 3.0; O/C ¼ 1.0;

GHSV ¼ 8776 h 1 [67].

Trang 40

Oxidative Steam Reforming or Autothermal Reforming

Oxidative steam reforming is the general term for the operation of a steam reformer,

to which a certain amount of additional air is fed:

However, this is not the case in a practical system because heat losses need to becompensated for Usually an optimum atomic oxygen/carbon (O/C) ratio exists foreach fuel under thermally neutral conditions to achieve optimum efﬁciency Thisvalue amounts to O/C¼ 0.88 (or 0.44 for the O2/C ratio here named x) as shown formethane in Figure 3.11 [68]

The maximum efficiency at this optimum ratio amounts to 93.9% for methane,6.1% of the efficiency being lost mainly for the evaporation of water The generalformula for the optimum efficiency is:

Figure 3.11 Effect of O/C ratio on the lower heating value of

hydrogen and the efficiency of the methane reforming process as

calculated by Ahmed and Krumpelt [68].

Tiêu đề	Fuel Processing for Fuel Cells
Tác giả	Gunther Kolb
Trường học	Imm - Institut für Mikrotechnik Mainz GmbH
Chuyên ngành	Fuel Processing for Fuel Cells
Thể loại	Thesis
Năm xuất bản	2008
Thành phố	Mainz

Định dạng
Số trang	435
Dung lượng	9,38 MB