nuclear production of hydrogen · in addition, hydrogen had a long and successful career as a final...

INTERNATIONAL NUCLEARSOCIETIES COUNCIL

NUCLEAR PRODUCTION

OF HYDROGEN

— TECHNOLOGIES AND PERSPECTIVES FOR GLOBAL

DEPLOYMENT —

International Nuclear Societies Council(INSC)

Current Issues inNuclear Energy

Series

November 2004

INCSHydrogen-FM 12/8/04 8:13 AM Page i

Copyright 2004 International Nuclear Societies CouncilISBN: 0-89448-570-9

ANS Order Number: 690064All rights reserved.

Printed in the United States of AmericaPublished by the American Nuclear Society, Inc.

555 North Kensington AvenueLa Grange Park, Illinois 60526 USA

INCSHydrogen-FM 12/8/04 8:13 AM Page ii

MEMBERS OF THE INTERNATIONALNUCLEAR SOCIETIES COUNCIL

American Nuclear Society (ANS)Asociación Argentina de Tecnologia Nuclear (AATN)Associação Brasileira de Energia Nuclear (ABEN)

Atomic Energy Society of Japan (AESJ)Australian Nuclear Association (ANA)

Canadian Nuclear Society (CNS)Egyptian Society of Nuclear Science and Applications (ESNSA)

European Nuclear Society (ENS)Austrian Nuclear SocietyBelgian Nuclear Society

British Nuclear Energy SocietyBulgarian Nuclear SocietyCroatian Nuclear SocietyCzech Nuclear SocietyDanish Nuclear SocietyFinnish Nuclear SocietyFrench Nuclear Society

German Nuclear SocietyHungarian Nuclear Society

Israel Nuclear SocietyItalian Nuclear Society

Lithuanian Nuclear Energy AssociationNetherlands Nuclear Society

Nuclear Society of RussiaNuclear Society of Slovenia

Romanian Nuclear Energy AssociationSlovak Nuclear Society

Spanish Nuclear SocietySwedish Nuclear Society

Swiss Nuclear Society

Indian Nuclear Society (InNS)Israel Nuclear Society (IsNS)

Korean Nuclear Society (KNS)Latin American Section (LAS)

Nuclear Energy Society Taipei, China (NEST)Pakistan Nuclear Society (PNS)

Sociedad Nuclear Mexicana (SNM)

iii

INCSHydrogen-FM 12/8/04 8:13 AM Page iii

MEMBERS OF THE TASKGROUP ON NUCLEAR

ENERGY’S ROLE INTHE FUTURE

Chairman

Masao Hori AESJ, Japan

Members

Manuel Acero ENS, SpainJ. Stephen Herring ANS, USAAndrew C. Kadak ANS, USACesare Marchetti ENS, AustriaAlistair I. Miller CNS, CanadaJorge Spitalnik LAS/ANS, BrazilJ. M. Martinez Val ENS, SpainXavier Vitart ENS, FranceDavid C. Wade ANS, USABertram Wolfe ANS, USA

iv

INCSHydrogen-FM 12/8/04 8:13 AM Page iv

v

CONTRIBUTORS

Preface Cesare Marchetti (Austria)Chapter 1 Jorge Spitalnik (Brazil)Chapter 2 Masao Hori (Japan)Chapter 3 J. Stephen Herring, James E. O’Brien, Carl

M. Stoots, Paul A. Lessing, Raymond P.Anderson, Joseph J. Hartvigsen, and S. Elangovan (USA)

Chapter 4 Xavier Vitart (France)Chapter 5 J. M. Martinez Val, Jesús Talavera, and

Agustín Alonso (Spain)Chapter 6 Alistair I. Miller (Canada)Chapter 7 David C. Wade (USA)Chapter 8 Masao Hori (Japan)

EDITORS

Masao HoriJorge Spitalnik

INCSHydrogen-FM 12/8/04 8:13 AM Page v

INCSHydrogen-FM 12/8/04 8:13 AM Page vi

CONTENTS

Foreword, viii

Preface: The Making of a Strategy, x

1. Introduction, 1

2. Synergistic Production of Hydrogen by Fossil Fuelsand Nuclear Energy, 7

3. Hydrogen Production Through High-TemperatureElectrolysis, 23

4. Thermochemical Production of Hydrogen, 45

5. Safety Issues of Nuclear Hydrogen Production, 69

6. Hydrogen Economics for Automotive Use, 85

7. A Sustainable Nuclear Fission–Based HydrogenEnergy Supply Architecture, 107

8. Concluding Remarks, 125

References, 133

vii

INCSHydrogen-FM 12/8/04 8:13 AM Page vii

viii

FOREWORD

The perspective of the transition to a hydrogen econo-my, where electricity and hydrogen serve as complementa-ry energy carriers and where a sustainable energy supply isattained by eliminating carbon emission, is being currentlyconsidered of great importance in a world dependent onvolatile fossil fuel prices and climate change impacts.

In response to such considerations, the InternationalNuclear Societies Council (INSC) set up a Task Group todevelop this report, “Nuclear Production of Hydrogen:Technologies and Perspectives for Global Deployment,”reviewing technologies for the nuclear production of hydro-gen and perspectives for its global deployment. Membersof the Task Group were appointed by INSC nuclear soci-eties and agreement organizations. This report is beingpresented as a publication in the INSC series, CurrentIssues in Nuclear Energy.

This report describes technologies currently existingand being developed for the production of hydrogen focus-ing on nuclear energy input and analyzes safety issuesassociated with such production. Since the implemen-tation of a hydrogen economy is greatly influenced by itsutilization in transportation, Hydrogen Economics forAutomotive Use has been included as a separate chapterin this study.

A reengineered world energy supply architecture,optimized for the combined production of nuclear energyand hydrogen and intended for global energy supply inmid–21st century market conditions, is portrayed as an

INCSHydrogen-FM 12/8/04 8:13 AM Page viii

ix

example of the development of future industrial and urbancomplexes.

INSC hopes this report will assist the nuclear commu-nity, policy and decision makers, and the public to be awareof the feasibility of a hydrogen economy based on nuclearenergy utilization for assuring sustainable developmentand mitigation of climate change effects.

Jorge SpitalnikChairman, International Nuclear Societies Council

FOREWORD

INCSHydrogen-FM 12/8/04 8:13 AM Page ix

PREFACE

The Making of a Strategy

One day in the mid-1960s, I was sitting in my office at theEuratom center in Ispra meditating on the long-term futureof nuclear energy. At that time, faith in the future of nuclearpower was untarnished, but I felt a strategic weakness in thecurrent view. Everybody was thinking in terms of electricityproduction even if the most optimistic projection would giveabout half of the primary energy electrified. Because in thelast couple of hundred years, energy consumption doubledevery 30 years or so, even if all electricity were nuclear, after30 years we would be back to square one in terms of fossilfuels consumption. Some words were said about an all-electric economy, which sounds a little eerie as the generat-ing system should be geared to the instant consumption ofenergy, on top of the fact that many kinds of vehicles are noteasily amenable to the use of electricity.

So I decided that nuclear power should be used toproduce some sort of fuel and the choice fell on hydrogen.It is transportable, flexible, nonpolluting, and starts fromwater and ends in water. I started a bibliographic searchon the use of hydrogen and discovered to my surprise thatall sorts of final uses were studied at the time—even fluo-rescent lamps where the phosphors are directly excited bythe oxidation of hydrogen over them. In addition, hydrogenhad a long and successful career as a final energy carrier.City gas is primarily hydrogen and was the energy back-bone of European and American cities until World War II.The first internal combustion engine built in Lucca by

x

INCSHydrogen-FM 12/8/04 8:13 AM Page x

Meucci and Barsanti ran on hydrogen, and experimentshad been done in the 1950s on hydrogen airplanes. Themost fascinating application for me was the production offood for the astronauts using microorganisms capable ofprocessing hydrogen. Therein lies the promise of liberat-ing man from the burden of agriculture. Incidentally,chlorophyll splits water into hydrogen and oxygen. Back tosquare one.

One of the problems insufficiently examined is whyman in the last couple of centuries passed from wood tocoal, and then to oil and gas. One trivial explanation is thatforests were overexploited, which may be locally true.However, world forests now shed something like 100 TW asbiomass while humanity consumes approximately 10 TW.No exhaustion in sight. One can say that forests are spreadaround the world. But so are oil fields. The real problem, inmy opinion, is that exploitations of forests have little or noeconomies of scale. To cut one tree per hour, one needsone lumberjack, and to cut two trees per hour, two lumber-jacks. Nor does the argument of facility of use have suffi-cient weight. It is true that it would be cumbersome to runa car stuffing it with wood, but in the 20th century, cities likeParis were consuming huge amounts of methyl alcohol pro-duced by wood distillation, an excellent fuel for cars. Afterall, no car runs on crude oil, and sophisticated refineriesare needed to produce the right type of fuel.

After carefully examining all the transitions, I came tothe conclusion that the winning parameter that leads to thetransition is composed of economies of scale. The drivingforce is the scale of the market, i.e., the level of consump-tion, and second is the spatial density of consumption. Thenext competitor wins on the basis of its economies of scalein the range dictated by consumption. The process is bet-ter seen inside a global enveloping technology, such as theelectric system. Here the capacity of a generator is dictat-ed by the spatial density of electricity consumption and itstransportability. So the power station “sees” a certain mar-ket via its distribution lines and the generator is sized to it.If the spatial density increases, the market seen by the

xiPREFACE

INCSHydrogen-FM 12/8/04 8:13 AM Page xi

power station increases even more quickly, because theelectricity transport system can use higher voltage, i.e.,higher-capacity lines that carry electricity over greater dis-tances. Examining the American electricity system, onesees in fact a doubling of consumption every 7 years anda doubling of the capacity of generators every 6 years. Inonly 100 years, they went from the 10 kW of the JumboDynamo of Edison to the 1 GW of today: an incrediblejump by a factor of 100 000.

Without such theatrical effects, the productivity (pas-sengers � km/hr) in commercial airplanes did increase bya factor of 100 in 50 years in tight response to demand.The same is true for oil tankers, where, at any time, thelargest ship carries a constant fraction of the total worldtraffic.

The essence of all this reasoning is that if we want toride a technology destined for success, we must create theconditions for the economies of scale to be part of its evo-lution. In the case of hydrogen as a vector of nuclear ener-gy, this will be implicit for the simple reason that hydrogenhas high transportability in pipelines and very high as LH2in cryotankers. We performed some calculations in Isprawith the help of SNAM, the Italian natural gas company,and found that the cost of transportation in large pipelinesover long distances compares with that of natural gas. Thismeans that a generating station “sees” a market as large asa continent and can grow accordingly. Certainly our nuclearreactors, optimized for the fractional electric markets,appear puny in size. Technology will strive to fit the demandbecause the rewards of size can be huge. In the 1960s,when nuclear costs were still fairly transparent, I collecteddata for stations of similar characteristics and differentsizes and found that they follow the square root rules aschemical plants often do. This means that a plant four timesas large costs double, halving the specific cost, for exam-ple, per kilowatt. Small is beautiful but big is cheap.Technology tends to trot behind the demand pull. Fordecades I was in contact with the engineers designing

xii NUCLEAR PRODUCTION OF HYDROGEN

INCSHydrogen-FM 12/8/04 8:13 AM Page xii

large electric generators, and I teased them, asking howlarge a generator can be. The answer was time invariant:double the present. And they did explain it with many con-vincing details. They were not wrong, but their time horizonwas just 6 years.

Nuclear energy seems to be here to stay, as shown ina report I prepared for the European Union Commission in1990, whose analysis is still fully valid today. It is true thatnuclear plant construction did come to a halt, but this is dueto the economic Kondratiev cycle. In fact, the cumulativepower connected to grid follows precisely the Kondratievprescription of saturating in the 1990s, as did steel and carproduction capacity, by the way. On the other hand, I repeatbecause it is central, to escape fossil fuels, one cannot justlimit nuclear to electricity production that may finally requireapproximately 50% of the primary input, because a dou-bling of energy consumption would bring the fossils back tosquare one. World energy consumption has doubled every30 years for the last 200 years, and with two-thirds of theemerging countries still running uphill, it will keep doublingfor a while.

To go nearly 100% nuclear, a chemical energy carrieris necessary, and in the 1960s, when I started meditatingabout the long-term strategies for nuclear, hydrogen didappear the inevitable choice. Distributing it to the final con-sumer is no problem, as shown by the experience with citygas, and the few thousand kilometres of high-pressurepipelines, transporting it for industrial purposes, paves theway for continental coverage. Liquefied natural gas shipscan be the blueprint for a global system by which LH2 iscarried overseas. As I have shown for electricity, spatialconsumption and transportability define the size of the gen-erators, which for hydrogen could be extremely largebecause transportability is an order of magnitude higherthan for electricity. So it is strategically appropriate to thinkbig, and my “energy island” with a hydrogen productionequivalent to the Middle East, energywise, may be the ref-erence long-term paradigm for that.

xiiiPREFACE

INCSHydrogen-FM 12/8/04 8:13 AM Page xiii

How to Make Hydrogen from Nuclear:The Strategic Choices

Once we identified the carrier in my Ispra laboratories, westarted thinking about ways to produce it from water usingnuclear heat. The ordinary solution is to use electrolysis, butwe discarded it on various grounds. First, it represents twoprocesses in series with a too-low total efficiency. Second, itappears expensive in terms of machinery. Third, and per-haps most important, it is not scalable. The energy systemis immense in size, be it 10 TW or 100 TW, and the experi-ence with the historical size evolution of chemical plantsshows that they will have production units commensuratewith the market. The drive for that is given by economies ofscale that can be as high as the square root, meaning areactor ten times as large costs three times as much, withthe specific cost reduced by a factor of approximately 3.Obviously, the technology must be mature for scaling, butone can trust engineers. In a mere 100 years, electric gen-erators did double in size every 6 years from the 10 kW ofEdison’s Jumbo Dynamo to the modern million-plus-kilowattalternators, a factor of 100 000. In reality, spatial intensity inconsumption and transportability of the medium, be it elec-tricity or hydrogen, define the optimal economic size of thegenerator and technology runs to provide it.

Based on these considerations, elementary but quin-tessential, we chose the chemical route to go from nuclearreactor heat to water splitting. We discarded the entire fam-ily of renewable energy on the basis of parallel arguments;they do not scale. It must be clear that I am speaking of thebackbone of primary energy procurement. As the successof electric batteries shows, for example, very uneconomicalways of producing electrical energy can have a commer-cially fat market share, if puny in terms of energy pro-duced. In addition, in my energy island scheme, I madeuranium “renewable” by extracting it from the seawaterused to cool the reactors, a line the Japanese successfullydeveloped in the last 30 years. Renewable means 10 000years, which is a sizable time resource to develop the nextstep, also described in the energy island paper, that would

xiv NUCLEAR PRODUCTION OF HYDROGEN

INCSHydrogen-FM 12/8/04 8:13 AM Page xiv

bring us to ten billion years of energy resources, whichshould also quiet the physiological pessimists.

The chemical processes can be seen as black boxeswe called Mark-X, where the input is water and high-temperature heat and the output is low-temperature heatplus H2 and O2 All the reactions inside run on closed cyclesso that, in principle, there is no consumption of chemicals.Inventing a good Mark is very tricky because it has to fit somany boundary conditions; the chemicals must be cheap,noncorrosive, easily separable, etc. We could see a thou-sand of them thermodynamically constructed with a com-puter program. Curiously, the dominant reason to discardthem was that nobody had ever studied most of the trivialreactions going into them. Classical chemistry is still fullyopen to exploration. The best ones, a dozen, finally emergedfrom the heads and experience of our chemists.

Thinking and experimentation reduced the number ofprocesses as the few million dollars available did not per-mit a broad-front research. The team produced an impres-sive number of data that made possible an appreciation ofthe problem technically and economically. The first ques-tion was the kind of efficiency that could be expected. Ashistory shows, all chemical processes tend to improve theirthermal efficiency during their development in applicationfollowing logistic equations. The real problem is to have adecent start. The estimates for the sulfur cycle we hadstudied most in depth with bits and pieces of a demonstra-tion plant gave a comforting 50%. After all, the Watts enginehad started the industrial revolution with an efficiency of 1%and in a couple of centuries has developed logistically toapproximately 60%.

The basic idea of a thermochemical cycle is thatbecause water cannot be cracked in a single attempt, andthe necessary temperature is too high for a nuclear reactor(although fusion researchers tried), then we must split itinto two steps or more. Intuitively, one can reason that thefirst step liberates the hydrogen and binds the oxygen sothat the free energy for the process is the difference

xvPREFACE

INCSHydrogen-FM 12/8/04 8:13 AM Page xv

between the two. The second step liberates the oxygenfrom a weaker bond than in water.

This procedure requires two reactions forward and twobackward to restore the initial state, so most processes thathave been proposed have four reactions. Because, in princi-ple, what the chemical must transfer to the water is free ener-gy, it is possible to conceive a one-reaction process in whichone molecule by heating and cooling accumulates sufficientfree energy to finally crack water. The thermal cracking ofH2SO4 almost does it, and in fact at Ispra, we did court thatreaction with various additions to close the cycle.

However, apart from some loose example, I will notenter here into the details of the water-splitting processesas the pages allotted in this review would not be enough todescribe even one. The details are well covered in thepapers and the references found on the website http://www.cesaremarchetti.org/ and elsewhere in this book.Here, I will basically delineate the strategies and the con-straints for applying these processes, because I think theyare not sufficiently treated in the current literature.

The first problem is that of a transition to very largenuclear reactors. In the 1960s, with the help of my col-leagues at General Electric and the Prof. Schulten team forhigh-temperature reactors (HTRs) in Germany, I designeda toroidal pressure vessel reactor that permits large vol-umes with limited diameters and filled it with graphite fuelballs, the German way, to produce 200 TWth. Mounted ona barge of appropriate size, together with the thermo-chemical plant, it should have been the basic generatingmodule in the energy island. It indicates that many prob-lems can be solved, but an industrial design must grow instages, as the case of electric generators clearly shows.So, one has to find a possible nursery.

Prof. Schulten had thought of chemistry to transportnuclear energy and chose methane steam reforming fol-lowed by methane synthesis to place heat at a distance.The system does not stand up to strategic analysis, but hehad developed both processes in connection with a nuclear

xvi NUCLEAR PRODUCTION OF HYDROGEN

INCSHydrogen-FM 12/8/04 8:13 AM Page xvi

reactor, and the technology could be used in a differentstrategic context.

Russia exports to Europe approximately 200 billioncubic metres of natural gas per year through a limited num-ber of very large pipelines converging in an area in front ofPoland and Slovakia. This area is interesting for variousreasons: It has water, industries, and very old and aban-doned oil fields. As I was invited to give a speech in a con-gress in Moscow dealing with energy, environment, andconservation, I tried to make everybody happy by present-ing a proposal in which some of this methane is steamreformed using nuclear heat, the hydrogen is mixed withthe flowing methane, and the CO2 is used for tertiary recov-ery of oil in the fields.

This kills several birds with one stone. The formal rea-son for the proposal was to start taking care of CO2 emis-sions, which so many people talk about and against whichnobody takes measures. In this case, CO2 disposal wouldhave a precise economic purpose, but finally it would dis-appear in the bowels of the earth by reacting with variouselements. The strategic reasons were three. The first is toprovide a nursery for the growth of nuclear reactors. Sucha large flux of methane can absorb any power, let us sayblocks of 50 GWTh, ten times larger than the largest reac-tor built. Supposedly they would be HTRs. The secondreason is that H2 would be mixed with methane and, con-sequently, spread all over Europe. The usual chicken-and-egg argument that technologies using hydrogen, such ashydrogen cars, cannot penetrate for lack of a hydrogendistribution infrastructure would lose weight. The amountsof hydrogen that can be mixed in a way compatible withexisting infrastructures can be as high as 20%, makingseparation with membranes or metallic hydrides presum-ably simple. Finally, nuclear energy would enter the pro-duction of fuels on a large scale using well-developedtechnologies.

The proposal was well accepted by the scientific com-munity and the press but had no practical consequences.

xviiPREFACE

INCSHydrogen-FM 12/8/04 8:13 AM Page xvii

However, from my studies of the penetration of importantnew ideas and technologies, I know that the process takesabout a Kondratiev cycle, i.e., approximately 50 years, sothere is no need to worry. It is important to keep the issueon the table, however, so that the system keeps absorbingit. In the case of cars or aviation, a bunch of crazy guyskept the lights on, providing important practical improve-ments to the technologies in the process, but basicallyholding them in the collective imagination.

Strategies can be seen as distant attractors’ dreams,and I tried to provide one with the concept of the energyisland. I did propose it for the first time in a paper I pub-lished in Japan in 1973, “On Hydrogen and Energy.” It hadthe basic effect of attracting the interest of the Japaneseas they saw that the miracle of energy independence,much desired at the time, was not at hand but at least tech-nically possible.

For system purposes, the energy island should have asize adequate for the market in parallel with that for chemi-cal plants. If we export LH2 in cryotankers, the island seesthe world with its present 10 TW of energy consumption.Adopting a concept often used in the electrical systems,I adopted the criterion of 10% of the perceived market, i.e.,1 TW, for this energy generator. It is a large unit, more orless on the order of those in the Middle East. In spite of thesize, I did not see insurmountable problems. I should say Ican be considered an optimist as in reaction to the screamsof my friends of the club of Rome, I wrote a paper whereI did find a luxurious installation for a trillion people, onplanet Earth—with zero exhaustion and pollution. Also, forthe island, I outlined the solutions for zero ecological impact.The location on a Pacific atoll is very helpful.

First, consider the thermal plume. There is in thePacific a strong thermal gradient. If one can pump coolingwater at, for example, 500 m depth, where the temperaturecan be 5°C, and reject it at sea surface temperature, atperhaps 25°C, there will be no visible thermal plume.Using HT graphite ball reactors, fuel manufacturing and

xviii NUCLEAR PRODUCTION OF HYDROGEN

INCSHydrogen-FM 12/8/04 8:13 AM Page xviii

reprocessing can be made automatic and possibly doneonboard, but the amount of fission products to dispose ofwill be large. So, at the Ispra center, we devised a self-sinking system in which capsules loaded with fission prod-ucts are just earthed, where the heat melts the ground andthey just sink. We made extensive experiments to checkthe equations for sinking real capsules. They can go downto, for example, 15 km in the volcanic rock of the atoll.Uranium is extracted from the cooling water, so there is nopollution from uranium mining and transportation. The bigbarges carrying all the conquibus would be armed inappropriate wharves so that the bare operations on theislands would be conducted by limited crews.

In this context, the cost of the energy produced isbasically a capital cost, and capital is basically linked toscale. Efficiency, so important when the energy source isexternal and possibly polluting, is one of the variableshere, and low efficiencies are acceptable if they bringother benefits.

Globalization of energy production also has politicalconsequences, as I stressed in my dinner speech at the firstH2 conference in Miami in 1976 (from the primeval soup toworld government). Basically, the maximum size of anempire is dictated by the speed of the transport system, andthe multinationals were fast to get the idea, but the politicalsystems are sluggish and did not. One of the possible by-products of this globalization could be the formation of areal world authority to administer the energy business andthe rest.

In the course of this analysis, I did quote the fact thatthe world forests could well support our energy needs if itwere not for an inappropriate interface, that of choppingtrees to obtain their energy content. So I wondered whethera solution could be found. It can and is described in mypaper “The Hydrogen Tree.” The inspiration came from theleguminous plants in which (nitrogen) bacteria create nod-ules in their roots, where they produce hydrogen bydecomposing the photosynthate of the plant (sugars) into

xixPREFACE

INCSHydrogen-FM 12/8/04 8:13 AM Page xix

CO2 and H2, in a sense reversing photosynthesis.Hydrogen is used to produce ammonia, sent back to theplant to pay the rent. So I reasoned, if one could constructlarge galls doing the same and collect the hydrogen via adrop irrigation system operated in reverse, then the forestcould be “milked,” producing a perfect fuel for the counter-part. Apart from developing the gall, which requires thebest in genetic engineering, the system requires no partic-ular skill in installation and operation and could be ideal forisolated villages. I did present the idea in three congressesand was much appreciated. I report it here to hold it in thestream of attention.

Cesare MarchettiLaxenburg, Austria

June 2004

xx NUCLEAR PRODUCTION OF HYDROGEN

INCSHydrogen-FM 12/8/04 8:13 AM Page xx

1. INTRODUCTION

INCSHydrogen-chap01 12/8/04 8:14 AM Page 1

1. Introduction

In his prescient plenary speech “On The Nature of NuclearPower and Its Future”1 at the Global 93 conference, WolfHäfele compared the technical, institutional, and socialopportunities for a second wave of nuclear deployments tothose which brought about the Industrial Revolution. Heargued that the first wave of nuclear deployments—forelectricity production and with an open fuel cycle—wasdestined to saturate at under 400-GWe global deploymentbecause

“Nuclear power was put into an existing technical andinstitutional infrastructure without much changing thisinfrastructure––still characterized by the use of oil inparticular but also of coal and gas.”

But to paraphrase his view of the analogy:

The Industrial Revolution exploited the factor of a mil-lion between ~1 µeV due to mass flow (of fallingwater) and ~1 eV chemical energy flow of burningcoal to achieve a revolutionary transition away fromcenturies of reliance on the water wheel and animalpower to coal-fired steam engines. The exploitation ofthe factor of a million in energy density, when enabledby reengineering the architecture of production (facto-ries, division of labor, etc.) led to the first IndustrialRevolution—which, over the ensuing two centuries, lit-erally changed the Western world (technically, institu-tionally, and socially).

He concluded:

“One must be prepared for evolution or even revolutionwhen real nuclear power (frees itself from the archi-


NUCLEAR PRODUCTION OF HYDROGEN4

tecture optimized for fossil, and) brings the factor of amillion between nuclear and chemical bond energiesto the surface––one cannot treat nuclear power likechemical power––uranium like yellow coal, so let usnot lose our perspective.”

Nuclear energy’s current configuration has left manyof its innate features unexploited: specifically, its economi-cally harvestable resource base good for a millennium ofworld energy supply by closing the fuel cycle, its capacity toopen itself to the entire primary energy market by manu-facturing hydrogen, and its capacity to break the energysecurity/nonproliferation dilemma by exploiting its incredibleenergy density to facilitate deployment of long-refueling-interval reactors supported by regional fuel cycle centers.

The objective here is to propose a reengineered worldenergy supply architecture optimized for nuclear ratherthan fossil. A central feature of the new architecture is theuse of fission heat to manufacture hydrogen from waterand the use of hydrogen as a replacement for fossil fuelchemical energy carriers. Over time, this nuclear-drivenenergy supply architecture would displace fossil and pro-vide energy to support a global energy infrastructure meet-ing all aspects of sustainable development—securelongevity, ecological compatibility, and social acceptability.2

Although it is revolutionary in concept, it is designed togradually displace the fossil architecture and the currentnuclear infrastructure by means of a many-decades evolu-tionary market penetration process. The chapters in thisbook describe incremental technological steps along thetransition pathway.

Now, let us presume that nuclear energy is finally, andlogically, committed to hydrogen production. How large anuclear energy supply would be necessary for the cominghydrogen economy era? According to a joint study by theInternational Institute for Applied Systems Analysis (IIASA)and the World Energy Council (WEC) on global energy per-spectives,3 the world final energy consumption in the year2100 will be ~20 Gtoe/yr, of which 25% is consumed by the


INTRODUCTION 5

transportation sector in the middle course (Case B).Assuming the efficiency of fuel-to-power conversion toincrease by a factor of 2 when using hydrogen fuel, therequired hydrogen amount will be 2.5 Gtoe/yr. To supplythis amount of hydrogen by nuclear energy, the number ofnuclear plants required will be, on a 1-GWe plant scale,~2500 units using the thermochemical water-splittingprocess coupled with a high-temperature reactor (50%conversion efficiency). According to the same IIASA/WECcase, nuclear electricity in the year 2100 will be 61 000TWh/yr or 5200 1-GWe plants. It is assessed to be possi-ble for nuclear energy to supply, toward the end of the 21stcentury, the total capacity (~7700 units of 1-GWe plants) ofboth electricity and hydrogen if the recycling of nuclear fuelis introduced in due time.4

Hydrogen can become an important and ecologicallybeneficial energy vector in the world economy of the future,but huge sources of carbon-emission-free primary energywill be needed. This is a job that nuclear energy can do.

The challenge is a large order. The transition to anuclear-driven hydrogen economy will require both tech-nological and institutional innovations to achieve marketpenetration. The following chapters describe incrementalmarket penetration options that will provide a basis toguarantee economic competitiveness and a suitable safe-ty level required for the nuclear production of hydrogen.Advanced nations can begin in the near term to restruc-ture their energy research and development programsbased on the vision that nuclear energy will become theworld’s major energy source—supplying hydrogen andelectricity to fuel sustainable development.


2. SYNERGISTICPRODUCTION OFHYDROGEN BY

FOSSIL FUELS ANDNUCLEAR ENERGY

The synergistic method using nuclear heat to help fos-sil fuels produce hydrogen could improve the efficiency ofconverting fossil fuels to hydrogen, thus attaining fuel savingand CO2 reduction.

For the global energy supply in the 21st century, whilethe share of renewable and nuclear energies will increase,continued use of fossil fuels to some extent will be inevitable.

This synergistic method will be a win-win arrangementfor both energy sources, especially in the intermediateterm, having environmental benefits in reducing CO2 and abenefit in using resources effectively.


2. Synergistic Production ofHydrogen by Fossil Fuels and

Nuclear Energy

1. Introduction

This chapter reviews processes and technologies to pro-duce hydrogen synergistically by the combination of fossilfuels and nuclear energy. It focuses mainly on the steam-methane (natural gas) reforming (SMR) process usingnuclear heating, as the SMR process is now dominantlyused worldwide for producing hydrogen, and technologiesto connect the SMR process with nuclear reactors arebeing developed.

The application of nuclear heat to the SMR process isestimated to be economically competitive with the conven-tional fossil-combusting technology and to be deployable inthe near future. There are other processes to producehydrogen using nuclear energy and fossil fuels, based onpetroleum, coal, and other hydrocarbons. Some of theseare also described here.

2. Steam Reforming Reactions

The reactions of methane and steam to produce hydrogenare as followsa:

aIn this chapter, the lower heating value (LHV), the value of the heatof combustion of a fuel as measured by allowing all products ofcombustion to remain in the gaseous state is used.


Reforming reaction:

CH4 + H2O = CO + 3 H2 −206 kJ/mol

Shift reaction:

CO + H2O = CO2 + H2 +41 kJ/mol

CH4 + 2 H2O = CO2 + 4 H2 −165 kJ/mol

The reforming reaction is highly endothermic, and theshift reaction is exothermic. If the exothermic heat of shiftreaction is recovered, it can be used to compensate for partof the endothermic heat of the reforming reaction and theheat balance of the reforming and shift reactions combinedis endothermic of 165 kJ/mol. In existing plants, this heat issupplied by combustion of some of the methane feed. Theconcept of nuclear-heated SMR is to supply this heat froma nuclear reactor. The beneficial effect is illustrated in Fig. 1.

As can be seen from Fig. 1, assuming an idealisticb

case in the SMR process, 3.3 mol of hydrogen is producedfrom 1 mol of methane feed in the fossil-combusting SMRprocess, while 4 mol of hydrogen is produced from 1 mol ofmethane feed in the nuclear-heated SMR process. In thisidealized case, 17% less fossil fuel is consumed to producethe same amount of hydrogen.

In the reality of an actual SMR, as efficiencyc withwhich heat is utilized is typically ~80%, so 2.7 mol of hydro-gen is produced from 1 mol of methane feed. In thenuclear-heated SMR process, as no methane is consumedfor combustion and the yield of hydrogen is nearly stoi-


bAll the heat generated by combustion of hydrocarbon is assumed tobe applied to the heat of endothermic reaction of steam reforming.Also assumed is that the endothermic reaction heat of steamreforming is fully recovered from the exothermic shift reaction.

cEfficiency is defined here as the ratio of the heat value of producedgas to that of supplied fossil fuel, i.e., (production rate of hydrogen) �(heat value of hydrogen) / (consumption rate of methane) � (heatvalue of methane) for the case of hydrogen production by steamreforming of methane. Another definition includes the electric powerconsumed by the auxiliary equipment in the denominator.


chiometric,d the nuclear-heated SMR reaction will produce4 mol of hydrogen from 1 mol of methane. Therefore, thenuclear-heated SMR process will reduce natural gas con-sumption and carbon dioxide emissions by more than 30%compared with the conventional SMR process.

More generally, the reaction formulas for steamreforming of hydrocarbons are expressed by the followingequations:

CmHn + mH2O → mCO + (m + n/2) H2 Reformingreaction

mCO + mH2O → mCO2 + mH2 Shiftreaction

CmHn + 2mH2O → mCO2 + (2m + n/2) H2

Moles of product hydrogen per C of CmHn are shownin Table 1 for the steam reforming/shift reactions of a rep-

SYNERGISTIC PRODUCTION OF HYDROGEN 11

Fig. 1. Advantageous effect of nuclear heat supply to SMR.

dThis applies to a recirculation-type membrane reformer typedescribed later in this chapter.


resentative range of fossil fuels (natural gas, petroleumproducts, and coal).

For hydrocarbons that also contain oxygen, the gener-al reforming reaction is expressed by the following equa-tion:

CnHmOk + (2n-k) H2O = nCO2 + (2n + m/2 − k) H2.

Other fossil fuels, such as petroleum products andcoal, could be reformed by steam to produce hydrogen.5

The reaction formulas, endothermic reaction heats, heats ofcombustion, and ratios of reaction heat to heat of combus-tion are tabulated for representative fossil fuels in Table 2.The ratio of heat of reaction to heat of combustion of theproduct hydrogen, which is listed in column 6, shows thepercentage of contribution of nuclear heat to the heat con-tent of the produced hydrogen. This ratio represents thepercentage of saving of fossil fuel by the nuclear-heatedsteam reforming process as compared to the idealized casefor the fossil-combusting steam reforming and shift reactionprocess.

As can be seen from the values in column 6, the per-centage of saving of feed, or percentage of reduction ofcarbon dioxide emissions, for feed petroleum products and


Table 1.Moles of Product Hydrogen per Number of C in CmHnin the Steam Reforming and Shift Reactions ofRepresentative Fossil Fuels

RepresentativeComponent’s Mole of H2 per

Fossil Molecular perFuel Substance Formula C of CmHn

Natural gas Methane C1H4 4

Petroleum LPGa C4H10 3.25Naphtha C6H14 3.17Kerosene C12H26 3.08

Coal Coal C1H0~1 2 to 2.5

aLiquefied petroleum gas.



Tabl

e 2.

Ra

tio o

fH

eat

ofR

eact

ion

to H

eat

ofC

ombu

stio

n of

Pro

duce

d H

ydro

gen

in S

team

Ref

orm

ing

Rea

ctio

nof

Rep

rese

nta

tive

Foss

il F

uels

/Hyd

roca

rbon

s

Hea

t o

fH

2H

eat

of

Hea

t o

fR

epre

sen

tati

veR

eact

ion

/C

om

bu

stio

n/

Rea

ctio

n/

Co

mp

on

ent’s

Mo

le o

fM

ole

of

H2

Hea

t o

fFo

ssil

Fu

el/

Mo

lecu

lar

Ste

am R

efo

rmin

gH

ydro

carb

on

Hyd

roca

rbo

nC

om

bu

stio

nH

ydro

carb

on

Form

ula

Rea

ctio

n F

orm

ula

(kJ/

mo

l·HC

)(k

J/m

ol·H

C)

(%)

Na

tura

l gas

CH

4C

H4

+ 2

H2O

→C

O2

+ 4

H2

165.

096

817

.0

LPG

C4H

10C

4H10

+8

H2O

→4

CO

2+

13

H2

486.

631

4615

.5

Nap

htha

C6H

14C

6H14

+ 1

2 H

2O →

6 C

O2

+ 1

9 H

273

9.3

4598

16.1

Ker

osen

eC

12H

26C

12H

26+

24

H2O

→12

CO

2+

37

H2

1433

.089

5416

.0

Coa

lC

C +

2 H

2O →

CO

2+

2 H

290

484

18.6


coal is ~15 to 19%, which is almost in the same magnitudeas that for natural gas. For hydrogen production from petro-leum products by the conventional fossil-fueled SMRprocess and in the hydrogen production from coal by theconventional steam-coal gasification process, ~70% effi-ciency is expected. If one could instead apply nuclear heatwith a membrane-type reformer, nearly all of the feedhydrocarbons would be converted to hydrogen, resulting in~30% saving of feed hydrocarbons, compared with the fos-sil-combusting processes.

3. Fossil-Combusting SMR Systems

The conventional fossil-fuel fired SMR for hydrogen pro-duction is usually composed of a steam reformer, a shiftconverter, and a hydrogen purifier based on pressure swingadsorption (PSA) as shown in Fig. 2 (Ref. 5). A mixture ofnatural gas and steam is introduced into a catalyst bed inthe steam reformer, where the steam reforming reactionproceeds over a nickel-based catalyst, typically at ~800 to900°C temperature. Then the reformed gas is supplied to ashift converter, where carbon monoxide is converted intocarbon dioxide to produce more hydrogen by the shift reac-


Fig. 2. Schematic diagram of steam reforming processes.


tion. Then the reformed gas is passed to a PSA to separatehydrogen.

An alternative fossil-combusting SMR system called amembrane reformer is composed of a steam reformerequipped with membrane modules of a palladium-basedalloy, and a nickel-based catalyst, and can perform a steamreforming reaction, a shift reaction, and a hydrogen sepa-ration process simultaneously without a shift converter andPSA, as shown in Fig. 2 (Ref. 6). By this simultaneous gen-eration and separation of hydrogen, the reformer systembecomes much more compact, provides higher efficiencythan the conventional ones, and produces 99.999% purehydrogen without any other purification equipment. Thecoincidence of hydrogen generation and separation makesthe reaction proceed at much lower temperatures, in therange from 500 to 600°C. This avoids use of expensiveheat-resistant materials and increases equipment lifetimes.A city-gas-fueled membrane reformer is now operating atone of the demonstration hydrogen stations in downtownTokyo.

4. Nuclear-Heated SMR Systems

There have been several studies7 of nuclear-heated SMRsystems, utilizing either the conventional high-temperaturereforming process or the alternative medium-temperaturereformer process.

4.1 High-Temperature Reforming

The Japan Atomic Energy Research Institute (JAERI) hasbeen developing a coupling technology to connect anuclear system with a chemical process producing hydro-gen, using the high-temperature engineering test reactor(HTTR) linked to a conventional SMR process. The HTTRis a helium-cooled reactor operating at 30-MWth outputand 950°C outlet coolant temperature.

A schematic flow diagram and drawing of the HTTRhydrogen production system is shown in Fig. 3 (Ref. 8) and



Fig. 4 (Ref. 9). This system will produce 4200 Nm3/h hydro-gen from 1800 Nm3/h natural gas using 10 MW thermalheat supply at 880°C from the HTTR. This HTTR-SMRplant is scheduled to produce hydrogen in 2008.

The merits of this reforming process are as follows:

1. no combustion of methane for the endothermicreforming reaction, and consequently no carbondioxide emission from combustion

2. built into the process, separation of the carbondioxide produced in the chemical reaction, whichwould be an advantage for future sequestrationrequirements.

4.2 Medium-Temperature Reforming

A technology to combine a membrane reformer with amedium-temperature nuclear reactor10 is shown in Fig. 5.

In addition to the noted advantages for the high-tem-perature nuclear-heated SMR process, this lower-tempera-ture membrane reformer has the following advantages:


Fig. 3. Schematic flow diagram of HTTR hydrogen productionsystem.


1. supply of nuclear heat at medium temperature (the500 to 600°C temperature range is within the nor-mal range of a sodium-cooled reactor)

2. a higher yield of hydrogen from the natural gasbecause the recirculation of residual hydrogenwith the closed-loop configuration minimizeshydrogen waste


Fig. 5. Schematic diagram of nuclear-heated recirculation-typemembrane reformer.

Fig. 4. Drawing of JAERI’s HTTR hydrogen production system.


3. more compact reformer with a smaller membranefor diffusion (recirculation of residual hydrogenincreases the average driving force for hydrogendiffusion through the membrane).

A conceptual design of a nuclear-heated recirculation-type membrane reformer is shown in Fig. 6. This utilizes asmall-scale sodium-cooled fast reactor of 240 MW thermaloutput to supply heat at 565°C to the SMR process, whichproduces 200 000 Nm3/h hydrogen from 50 000 Nm3/hnatural gas.

5. The Significance of Synergistic Production ofHydrogen for Resource and Environment

Within the context of an expanding hydrogen economy, onemust consider the feasibility of producing hydrogen com-mercially using nuclear heat directly with zero or reducedemission of carbon dioxide. One alternative is to applynuclear-generated electricity to conventional low-tempera-ture electrolysis of water to produce hydrogen at the site ofdemand. Another alternative is processes for the thermo-chemical decomposition of water. At present, these require


Fig. 6. Conceptual design of a fast reactor membrane reformerplant.


high temperatures, so it is necessary to develop both anappropriate reactor type and thermochemical processes,and to solve issues such as temperature transients andpressure limitations on the chemical process side andmaterials suitably resistant to the severe conditions of hightemperature and chemical corrosiveness.

Compared to these thermochemical processes, hydro-gen production by nuclear-heated steam reforming of natu-ral gas is considered to be much closer to commercializa-tion and is viewed as an intermediate step to nuclear-drivenhydrogen production from water.11 Cost estimates of themain nuclear hydrogen processes show that only thenuclear-heated SMR process could compete economicallywith the fossil-combusting SMR process for hydrogen pro-duction.12 Estimates of the total cost, including the energy,distribution, and sequestration for a localized demand of amodel city also favor the nuclear-heated SMR process.13

The foremost feature of nuclear-heated SMR is reduction inconsumption of fossil fuels, and consequently in emission ofcarbon dioxide, per unit of hydrogen produced. In separat-ing carbon dioxide, it also simplifies any requirements forsequestration.

Membrane reforming with recirculation of reactionproducts in a closed-loop configuration is a particularlypromising nuclear application. Because the nuclear heat isneeded at below 600°C, it employs a compact membraneand reformer and provides efficient conversion of thehydrocarbon feed and high-purity hydrogen without addi-tional processing.

The net ratio of endothermic reaction heat supplied bynuclear energy to the heat value of the hydrogen producedis between ~15 and 19%. That means that most of theenergy in the hydrogen that is produced comes from thechemical energy of carbon. Table 3 shows approximateconversion factors (in percent) of reactor heat to hydrogenheat value (ratio of heat value of the hydrogen product tothe heat supplied from a nuclear reactor) and conversionfactors (in percent) for natural gas heat content to hydrogenheat value (ratio of heat value of product hydrogen to heat




Tabl

e 3.

App

roxi

ma

tion

ofH

eat

Con

vers

ion

Fac

tor

in N

ucle

ar a

nd F

ossi

l Pro

duct

ion

ofH

ydro

gen

Nuc

lear

Nuc

lear

Ene

rgy

Ele

ctric

ityE

lect

ricity

Nuc

lear

Hea

tN

ucle

ar H

eat

Na

tura

l Gas

Sou

rce

and

Hea

tH

eat

(Ele

ct./H

eat

(Hig

h Te

mp)

(Med

ium

Tem

p)a

= 3

2 to

50%

)(H

igh

Tem

p)(C

ombu

stio

n)

Raw

ma

teria

lW

ate

rW

ate

rW

ate

rN

atu

ral g

as

Na

tura

l gas

wa

ter

wa

ter

Pro

duct

ion

Ele

ctro

lysi

sH

ot

The

rmoc

hem

ical

Ste

am

Ste

am

proc

ess

elec

trol

ysis

refo

rmin

gre

form

ing

H2

hea

t/25

to

40%

45%

50%

330%

a—

Hea

tnu

clea

r he

at

conv

ersi

onfa

ctor

(%

)H

2he

at/

——

—11

5%a

80%

met

hane

hea

ta F

or a

rec

ircul

atio

n-ty

pe m

embr

ane

refo

rmer

.E

ffic

ienc

y of

reac

tor

hea

t ut

iliza

tion

is 6

0%;

hydr

ogen

yie

ld f

rom

met

hane

is 9

5%.

Hyd

roge

npo

duct

ion

proc

ess


content of feed natural gas) for the important hydrogenproduction processes.

As Table 3 shows, the heat conversion factors ofnuclear-heated SMR process are 330% for H2 heat/nuclearheat and 115% for H2 heat/methane heat. The larger arethe conversion factors, the smaller is the energy consump-tion for producing a fixed quantity of hydrogen. As thenuclear-heated SMR process uses both fossil and nuclearenergy inputs, the conversion factors become larger thanother processes that use only one energy source. Thatmeans this process consumes a smaller quantity of bothenergy sources to produce a fixed quantity of hydrogenthan required by other processes. The heat conversion fac-tor in terms of the hydrogen heat to total heat input, H2heat/(nuclear heat + methane heat), is 85% in the conditionshown in Table 3.

Most discussions of future global energy supply antic-ipate continued use of fossil fuels to some extent for a cen-tury or so while reducing carbon dioxide emission.14 Thesynergistic production of hydrogen using both energysources could enable nuclear energy to improve the effi-ciency with which valuable fossil fuel resources are con-sumed. While the long-term worldwide supply/demandanalyses of nuclear fissile materials shows that transitionto plutonium recycling will have ultimately to occur ifnuclear energy is to have a major role in supplying energy,efficient use of nuclear fuel by such an effective process asthe nuclear-heated SMR remains desirable.15 During thenuclear power expansion phase in the 21st century, thequantity of fissile materials available from natural uraniumresources and fuel recycle could emerge as a constraint.Avoiding the uncertainty of relying on speculativeresources of uranium, careful consideration on a worldwidebasis for efficient utilization of fissile materials would beneeded over the process of hydrogen/electricity productionand the blend of reactor types and fuel cycles.

With all these benefits, the synergistic blending of fos-sil fuels and nuclear energy to produce hydrogen can be an



effective solution for the world until nuclear water-splittingprocesses are available to supply hydrogen in large vol-umes. For both the fossil fuels industry and the nuclearindustry, this approach offers a viable symbiotic strategy forthe coming era of the hydrogen economy with a minimalimpact on resources, the environment, and the economy.



3. HYDROGEN PRODUCTIONTHROUGH

HIGH-TEMPERATURE ELECTROLYSIS

High-temperature electrolytic water splitting yields bet-ter performance than the conventional water electrolysisprocesses due to decreased electrode overpotentials andincreased oxygen ion diffusivity.

The high-temperature electrolysis of steam (HTES)process, supported by nuclear heat and electricity froma high-temperature reactor, could achieve a thermal-to-hydrogen conversion efficiency in the range of thoseattained from thermochemical processes.

Since the HTES process utilizes the solid-oxide tech-nology developed for fuel cells, it appears to be a viablemeans for producing hydrogen in the intermediate term.

INCSHydrogenchap03 12/8/04 9:54 AM Page 23

3. Hydrogen ProductionThrough High-Temperature

Electrolysis

1. Introduction

A leading candidate technology for large-scale hydrogenproduction is high-temperature electrolysis of steam16

(HTES). HTES uses a combination of thermal energy andelectricity to split water in a device very similar to a solid-oxide fuel cell. Solid-oxide fuel cells (SOFCs) have beenstudied extensively for their application to power generationsystems.17,18 SOFCs operate at high temperature (600 to1000ºC), which allows for internal reforming, promotesrapid kinetics with nonprecious materials, and produceshigh-quality process heat as a by-product.19

Higher temperature operation yields better perform-ance due to decreased electrode overpotentials andincreased oxygen ion diffusivity. In addition, from a thermo-dynamic standpoint, the electrical energy demand for elec-trolysis, represented by the Gibbs free energy change decreases with temperature. However, for large-scalehydrogen production, operating temperatures will probablybe limited to the range of 800 to 850ºC, even with couplingto a high-temperature advanced reactor.

In comparison to SOFCs, solid-oxide electrolysis cells(SOECs) have received relatively little attention in the liter-ature. However, the feasibility of operating solid-oxide cellsat high temperature in the electrolysis mode has beendemonstrated for both tubular20 and planar systems.21

¢gf


Results from several studies are abstracted in a compre-hensive report published by Japan Atomic EnergyResearch Institute22 (JAERI). In addition, the JAERI reportcovers the thermodynamics of HTES and results fromrecent laboratory tests of both tubular and planar solid-oxide cells.

High-temperature electrolytic water splitting supportedby nuclear process heat and electricity has the potential toproduce H2 with an overall system efficiency near thoseof the hydrocarbon and the thermochemical processes,but without the corrosive conditions of thermochemicalprocesses and without the fossil fuel consumption andgreenhouse gas emissions associated with hydrocarbonprocesses.16 Specifically, a high-temperature advancednuclear reactor coupled with a high-efficiency high-temper-ature electrolyzer could achieve a competitive thermal-to-hydrogen conversion efficiency of 45 to 55%.

A conceptual schematic of an advanced nuclear reactorcoupled to a hydrogen production plant is shown in Fig. 1.The reactor [in this case, a high-temperature gas-cooled


Fig. 1. High-temperature electrolysis using a gas-cooled reactor.


reactor (HTGR)] supplies thermal energy to drive the powercycle and to heat steam for the electrolysis process. Thehigh-temperature heat exchanger supplies superheatedsteam to the cells at a temperature of 750 to 950ºC, and apressure of 1 to 5 MPa. The input gas contains both steamand hydrogen to maintain reducing conditions at the elec-trolytic cathode. While the high-temperature electrolysis(HTE) plant would be devoted primarily to the production ofhydrogen, it is important to note that output of the electricalgenerator could also be sent to the grid, if demanded.Conversely, the steam flow rate could be increased, and theelectrolyzer could be operated at a higher current densityto profitably accept power from the grid for increased hydro-gen production in times of low electrical demand. Thus,HTE can play an important additional role in matching theconstant output of future reactors to varying electricaldemands.

The electrolytic cell consists of a solid-oxide elec-trolyte, such as yttria-stabilized zirconia (YSZ), with porouselectrically conducting electrodes deposited on either sideof the electrolyte. As shown in Fig. 2, a mixture of steam

HYDROGEN PRODUCTION THROUGH ELECTROLYSIS 27

Fig. 2. Components of an HTE cell.


and hydrogen at 750 to 950ºC is supplied to the cathodeside of the electrolyte. The oxygen ions are drawn throughthe electrolyte by the electrochemical potential, liberatingtheir electrons and recombining to form molecular O2 onthe anode side. The entering steam-hydrogen mixture maybe as much as 90 vol% steam. Similarly, the exiting mixturemay be as much as 90 vol% H2. The water and hydrogengas mixture is passed through a condensing or membraneseparator to purify the hydrogen. While present experi-ments and fuel cells operate at pressures near atmospher-ic, future cells may operate at pressures up to 5 MPa.

A basic thermodynamic analysis can be applied to thewater-splitting process to determine the thermodynamicefficiency limit as a function of temperature. A schematic ofa generic thermal water-splitting process is shown in Fig. 3.Water enters the control volume from the left. Since theultimate feedstock for any large-scale water-splitting oper-ation will be liquid water, it is reasonable to consider thecase in which water enters the control volume in the liquidphase at temperature T and pressure P. Pure hydrogen andoxygen streams exit the control volume on the right, also atT and P. Two heat reservoirs are available, one at temper-ature TH and one at temperature TL. Heat transfer betweenthese reservoirs and the control volume is indicated in thefigure as QH and QL. There is no work crossing the control-volume boundary.


Fig. 3. Schematic of a thermal water-splitting process operatingbetween temperatures TH and TL.


From a chemical reaction standpoint, the water-split-ting process corresponds to the dissociation of water:

The first and second laws of thermodynamics can beapplied to this process as follows:

first law: (1)

second law: (2)

where ∆HR is the enthalpy of reaction and ∆SR is theentropy change of the reaction. The thermal-to-hydrogenefficiency of thermal water-splitting processes can bedefined in terms of the net enthalpy change of the workingfluid (can also be thought of as the energy content of theproduced hydrogen), divided by high-temperature heatadded to the system:

(3)

Combining the first- and second-law equations for thereversible case and substituting into the efficiency definitionyields

(4)

Note that the water-splitting process defined in Fig. 1 issimply the reverse of the combustion of hydrogen with oxy-gen. Therefore, the enthalpy of reaction for the water-split-ting process is the opposite of the enthalpy of combustion,which by definition is equal to the “heating value” of thehydrogen. Since for our process, we have assumed that thewater enters the control volume in the liquid phase,

(5)

where HHV is the high heating value of hydrogen. If we fur-ther assume that T and P represent standard conditionsand that TL = To ,

¢HR � HHV,

hT,max �1 � TL �TH

1 � TL¢SR �¢HR.

hT �¢HR

QH.

¢SR � QH

TH�

QL

TL,

QH � QL � ¢HR;

H2O S H2 � 12 O2.



(6)

such that the efficiency expression can be rewritten as

(7)

The HHV of the hydrogen and the standard-state Gibbsenergy of formation for water are fixed quantities such thatthe second factor on the right side is a constant.

Note that this result applies to both thermochemicalcycles and to HTES. For HTE, the entire power cycle wouldbe internal to the control volume and the high-temperatureheat would be shared between the power cycle and theelectrolysis process.

A plot of this result is presented in Fig. 4 for TL = 20°C.The top curve represents the maximum possible water-splitting efficiency result given by Eq. (7). The bottom curveis simply 65% of this thermodynamic limit. The 65% valueis based on a typical percentage of Carnot efficiency thatcan be achieved with a well-engineered modern power

hT � a1 �TL

THb a HHV

�¢G0f,H2O

b � a1 �TL

THb a 1

0.83b.

¢HR � TL¢SR � �¢G 0f,H2O


Fig. 4. Theoretical thermal water-splitting efficiencies.


cycle. If we assume that 65% of the maximum possibleefficiency might also be achievable with a well-engineeredwater-splitting process, then efficiencies of the magnitudegiven in the lower curve of Fig. 4 should be expected.

A comparison of expected efficiencies of severalthermochemical and HTE water-splitting processes wasprepared by Yildiz and Kazimi.23 Figure 5 was taken fromthis study. In this figure, the energy efficiencies are plottedfor the leading candidate thermochemical cycle, SI, and forthe HTES process coupled to two advanced-reactornuclear power cycles. GT-MHR is the gas-turbine modularhigh-temperature reactor and AGR-S-CO2 is an advancedgas-cooled reactor supercritical CO2 power cycle. TheHTES efficiency values presented in this figure are similarto the 65% values presented in Fig. 4 over the entire tem-perature range. For temperatures above 800°C, the pre-dicted efficiencies of the SI process are similar to those ofthe HTES processes. The HTES technology coupled tothe direct-cycle supercritical-CO2-cooled AGR promiseshigher efficiency values at lower operating temperaturesthan the SI and HTES coupled to the GT-MHR technology.


Fig. 5. Comparison of thermal-to-hydrogen efficiencies of HTEand SI thermochemical processes.23


The possibility of achieving high efficiency at lower tem-peratures is a significant advantage in favor of HTES.Furthermore, for HTES, the power cycle could be usedboth for supplying power to the grid and for hydrogen pro-duction, providing increased flexibility.

2. Thermodynamics of HTE

From a Gibbs-free-energy viewpoint, high temperaturesreduce the amount of electrical energy required to breakthe chemical bonds in the water molecules.16,21 Therefore,the electrical energy demand ∆G for electrolysis decreaseswith increased temperature, as shown in Fig. 6. The ther-mal energy requirement T∆S, however, increases withincreasing temperature. The total energy demand ∆Hincreases very slightly with temperature. The ratio of ∆G to∆H is ~93% at 100ºC and ~70% at 1000ºC.

Higher reactor operating temperatures result in higherthermal efficiencies for electrical power generation, and anoverall analysis of the efficiency of hydrogen productionwith nuclear energy should consider the total thermal ener-


Fig. 6. Overall energy requirements for HTE.


gy required per kilogram of hydrogen. If we assume thatthe power generation cycle efficiency is 65% of the Carnotmaximum, then the total thermal energy requirementdecreases significantly with temperature.

Operation of the electrolyzer at high temperature isalso desirable from the standpoint of kinetics and elec-trolyte conductivity, both of which improve at higher operat-ing temperatures. Therefore, HTES is favored over solidpolymer electrolysis from both thermodynamic and kineticstandpoints.

2.1 Electrolysis Efficiency

An electrolysis efficiency ηe can be defined for electrolysiscells, analogous to the fuel cell efficiency definition pre-sented in textbooks on fuel cells (see Refs. 17, 18, and 19).The thermal efficiency quantifies the heating value of thehydrogen produced by electrolysis per unit of electricalenergy consumed in the stack. Based on this definition,

(8)

and since the stack electrical current is directly related tothe molar production rate of hydrogen by

(9)

where F is the Faraday number (F = 96 487 J/V·mol), theelectrolysis efficiency can be expressed in terms of celloperating potential as

(10)

The efficiency for the fuel-cell mode of operation is theinverse of Eq. (10).

Note that the value of the efficiency defined in thismanner for electrolysis can exceed 1.0. As an example, forthe reversible stoichiometric case (which occurs only atzero current density), the cell potential approaches theopen-cell value yieldingEo � �¢go

f >2F,

he ��¢hf>2F

V.

N #H2

� I>2F,

he ��¢hf N

#H2

VI,



(11)

which for steam electrolysis at 850ºC is equal to 1.34. Forcases with variable gas concentrations, the open-cellpotential is given by the Nernst Eq. (1), and the correspon-ding efficiency limit varies accordingly. It is not desirable tooperate an electrolysis stack near this efficiency limit, how-ever, because the only way to approach this limit is to oper-ate with very low current densities. There is a trade-offbetween efficiency and hydrogen production rate in select-ing an electrolysis stack operating voltage. Similarly, in thefuel-cell mode, there is a trade-off between efficiency andmaximum power production. Maximum power productionfor SOFCs occurs for operation at ~0.5 V, whereas maxi-mum efficiency occurs at the open-cell potential, ~1.0 V.Depending on optimization parameters, a good operatingpoint usually occurs at ~0.7 V. Similarly, for electrolysisoperation, the trade-off between maximum hydrogen pro-duction rate and maximum efficiency tends to predict opti-mal operating potentials near 1.3 V.

The Faraday efficiency quantifies the maximum elec-trical energy value of the hydrogen produced by electroly-sis per unit of electrical energy consumed. It can also bedefined in terms of cell operating potential and hydrogenproduction rate:

(12)

The maximum theoretical Faraday efficiency is 1.0, corre-sponding to reversible operation. Again, the Faraday effi-ciency for the fuel-cell mode of operation is the inverse ofEq. (12).

2.2 Thermal-Neutral Voltage

Operation of a planar solid-oxide stack in the electrolysismode is fundamentally different from operation in the fuel-cell mode for several reasons, aside from the obviouschange in direction of the electrochemical reaction. From

hF �¢gf>2F

E�

¢N #H2

¢gf

EI .

he,max �¢hf

¢gof,



the standpoint of heat transfer, operation in the fuel-cellmode typically necessitates the use of significant excessairflow to prevent overheating of the stack. The potential foroverheating arises from two sources: The hydrogen oxida-tion reaction is exothermic, and the ohmic heating is asso-ciated with the electrolyte ionic resistance. The heated airproduced from a fuel-cell stack can be used beneficially asa source of high-temperature process heat.

Conversely, in the electrolysis mode, the steam reduc-tion reaction is endothermic. Therefore, depending on thecurrent density, the net heat generation in the stack may benegative, zero, or positive. This phenomenon is illustratedin Fig. 7. The figure shows the respective heat fluxes in aplanar solid-oxide stack associated with the electrochemi-cal reaction and the ohmic heating. The net heat flux is alsoshown. A stack-average area-specific resistance of 1.25,an operating temperature of 1200 K, and hydrogen mole


Fig. 7. Thermal contributions in electrolysis and fuel-cell modesof operation.


fractions of 0.1 and 0.95 at the inlet and outlet, respective-ly, were assumed for these calculations. The net heat fluxis always positive and increases rapidly with current densi-ty in the fuel-cell mode. In the electrolysis mode, the netheat flux is negative at low current densities, increasing tozero at the thermal-neutral voltage, and positive at highercurrent densities. The thermal-neutral voltage can be pre-dicted from direct application of the first law to the overallsystem:

(13)

Letting Q = 0 (no external heat transfer), and invokingFaraday’s law,

(14)

Note that substitution of this value for Vtn into Eq. (10)yields an electrolysis efficiency of 100% for operation atVtn.

Since the enthalpy of reaction ∆HR is strictly a functionof temperature, the thermal-neutral voltage is also strictly afunction of temperature, independent of cell area-specificresistance (ASR) and gas compositions. The particular val-ues of net cell heat flux at other operating voltages do,however, depend on cell ASR and gas compositions. Thethermal-neutral voltage increases only slightly in magnitudeover the typical operating temperature range for solid-oxidecells, from 1.287 V at 800°C to 1.292 V at 1000°C. Notethat these voltage values are very close to the optimaloperating potential of 1.3 V discussed previously. At typicalsolid-oxide electrolysis stack temperatures and ASR val-ues, operation at the thermal-neutral voltage yields currentdensities in the 0.2 to 0.4 A/cm2 range, which is very closeto the current density range that has yielded successfullong-term operation in SOFC stacks. Operation at or nearthe thermal-neutral voltage simplifies thermal managementof the stack because no significant excess gas flow isrequired. In fact, in the electrolysis mode, since oxygen isbeing produced, there is also no theoretical need for airflowto support the reaction at all. In a large-scale electrolysis

Vtn � �¢HR>2F.

Q � W � ¢HR.



plant, the pure oxygen produced by the process can besold as a valuable commodity. Careful consideration mustbe given, however, to the choice of materials for containingpure oxygen at elevated temperatures.

3. Conceptual HTE Plant Design

An HTGR conceptual design that could be used for simulta-neous electrical power and hydrogen production was pre-sented in Fig. 1. In this design, the primary helium coolantis heated in the reactor to outlet temperatures in the rangefrom 850 to 1000ºC. A portion of the hot helium outletstream serves as the working fluid in a gas-turbine powercycle, and a separate helium stream flows through a high-temperature heat exchanger, providing process heat to asteam-hydrogen gas mixture. This gas mixture is then fed toa high-temperature steam electrolysis device, such as aregenerative solid-oxide cell, which produces one outletstream of pure oxygen and another steam-hydrogen outletstream that is significantly enriched in hydrogen. Both thesteam-hydrogen stream and the oxygen stream are main-tained at pressures of ~5 MPa . The steam-hydrogen outletmixture is passed through a heat exchanger and a separa-tor (e.g., steam condenser) to yield a high-purity hydrogengas product.

The focus of the research discussed here is thehydrogen-production process, including the development ofreliable high-temperature SOECs, identification of appro-priate hydrogen-permeation-resistant materials for thehigh-temperature heat exchanger, and the development ofan energy-efficient and cost-effective means for separationof hydrogen gas from the steam-hydrogen mixture.

As indicated in Table 1, the inlet gas flow is a 90:10(molar or volume) mixture of steam and hydrogen at850°C. The inlet mixture composition is dictated by twoconsiderations. First, the gas must provide sufficient steamfor electrolysis across the entire width of the cell; second,some hydrogen must be present at the inlet to maintainreducing conditions and to prevent electrode oxidation. The



target outlet mixture composition on the steam-H2 side willbe about ~10% steam by volume (10 vol%) and 90 vol%H2 Other steam-hydrogen ratios can be achieved with vari-ation of flow rate, current density, and cell size, but the out-let steam fraction should be kept high enough to ensurethat local steam starvation will not occur anywhere in thestack. The gases, including steam, are well approximatedas ideal gases at these conditions. The volumetric flow onthe steam-H2 side of the cell remains constant, while theoxygen flow on the cathode side of the cell increases dur-ing its passage across the face of the cell.

The current density adopted in this conceptual designis 0.2 A/cm2 (2000 A/m2), consistent with current densitiesthat have yielded 40 000-h operation in SOFCs. Based onthis current density, a total active cell area of 120 000 m2 isnecessary.

Table 2 indicates some of the cell configuration detailsthat have been adopted for this conceptual design. The


Table 1.Full-Scale Electrolyzer Flow Rates and Operating Conditions

Reactor power 600 MW thermal

H2 productionOverall process efficiency 50%Rate based on LHV 2.5 kg/s 238 tonnes/day

1667 � 106 sccm 84.8 � 106 std ft3/day

Water consumption 22.3 kg/s 354 gpm

Electrolysis cell operating conditions

Temperature 850°C 1560°FPressure 5 MPa 702.5 psiInlet mole fractions, H2O/H2 0.9/0.1Outlet mole fraction, H2O/H2 0.1/0.9Volumetric H2 + H2O flow rate 2.90 m3/s 2084 � 106 sccmVolumetric O2 flow rate

(generated) 1.16 m3/s 833.6 � 106 sccmCurrent density 0.2 A/cm2

Cell operating voltage 1.1 V


individual cells are assumed to have a 100- � 100-mmactive area. This relatively small size is determined by thethermal expansion compatibility between the electrolyteand the electrodes. Advances in electrode materials to per-mit a more conforming bond between the electrolyte andelectrodes could allow larger cell sizes. The use of an elec-trode-supported configuration, in which the electrolyte is onthe order of 10 to 30 µm thick, may also allow for larger cellsizes.

The cell layer thicknesses indicated in Table 2 repre-sent an anode-supported design. The electrolyte is 10 µmthick, the anode is 1500 µm, and the cathode is 50 µm. Thebipolar plate, which provides flow passages for the steam-hydrogen mixture and separate passages for the producedoxygen, is 2.5 mm thick.


Table 2.Cell Configuration

Cell Area

Individual cell width 10 cmIndividual cell active area 100 cm2

Total number of cells 12 � 106

Total active cell area 120 000 m2

Cell Layer Thickness

Electrolyte 10 µmAnode 1500 µmCathode 50 µmBipolar plate 2.5 mmTotal cell thickness 4.06 mm

Stack Dimensions

Cells/stack 1000Stack height 4.06 mStack volume 0.041 m3

Stack volume with manifolding 0.162 m3

Number of stacks 12 000Total volume of all stacks 486 m3

Hot volume 1944 m3

Stacks per row 75Number of rows 160Hot volume height 5 mHot volume width 15 mHot volume length 25.9 m


In estimating the total hot volume for the plant, wehave assumed that the volume necessary for manifolding isthree times the cell volume. Thus, the total cell volume is486 m3, and the overall hot volume is 1944 m3.

The channels in the bipolar plate are assumed to besemicircular in cross section, with a 1.0-mm diameter anda spacing of 2.0 mm. Table 3 shows the preliminary gas-flow parameters, which indicate that the average flowvelocity would be ~1.23 cm/s in the steam-hydrogen chan-nels and ~0.493 cm/s in the oxygen channels. Much moredetailed modeling of the gas flow in an individual cell is nowbeing performed at Argonne National Laboratory.

4. Modular Electrolysis Plant Concept

As part of this conceptual design, it is interesting to envi-sion in a practical sense how a hydrogen-production facili-ty may be integrated with a nuclear reactor. For assembly,maintenance, and repair purposes, the hydrogen plantshould be modular and movable. Properly sized modularunits could then be moved as needed to other locations.Each modular hydrogen production unit would be com-posed of a matrix of SOEC stacks, a heat exchanger forpreheating the makeup water, and a condenser-separator-dehydrator for the hydrogen production stream. Units couldbe added or removed from the facility, depending on therequired production rate of hydrogen.


Table 3.Gas Distribution Parameters

Gas-Flow Distribution in Bipolar Plate Channels

Channel diameter (semicircular) 1.0 mmChannel spacing 2.0 mmChannel cross section 0.393 mm2

Channels per plate 50Total number of channels 599 � 106

H2 + H2O flow rate per channel 14.5 cm3/minH2 + H2O channel flow velocity 1.23 cm/sO2 outlet flow rate per channel 5.81 cm3/minO2 outlet channel flow velocity 0.493 cm/s


Figure 8 depicts side and top views of the way thatthese components could be mounted on a rail-trans-portable module. A typical rail-transportable module size is18.5 m long by 2.9 m wide by 4 m high. Cell stacks couldbe 3 m in height, permitting 1 m above for manifolding. Foran overall cell thickness of 4.05 mm, 740 cells would beincluded in each stack. If one assumes that improvementsin SOEC manufacturing technology permit production ofcells with active areas of 500 � 500 mm, one could envi-sion 48 stacks oriented in three rows. The gas manifoldingfor the stacks would be arranged in parallel to minimize thevolumetric flow rate and pressure drop through each stack.Each production unit would require a 0.22 m3/s inlet flowrate of 90% steam/10% hydrogen mixture at 850°C and 5-MPa pressure. This flow would require an ~6-in.-innerdiameter (15-cm diameter) heavy-walled pipe at the maininlet to each module.

Stacks would be electrically connected in series tominimize the production unit amperage requirements. Eachstack would require 814 V and 500 A, or 39 kV and 500 Aper module (19.5 MW). For the 600-MW reactor discussedearlier, 14 modules would be required to produce 2.5 kg/sH2. To operate the electrolyzers at reasonable efficiencies,complete steam depletion must be avoided. Therefore, theoutlet gas mixture from the electrolysis units will invariablyinclude a significant steam mole fraction, perhaps 10%.The hot electrolyzer exit stream will initially flow through theheat exchanger to preheat the makeup water. Heat removalfrom the hydrogen-steam mixture will result in condensa-tion, yielding a heat-exchanger outlet dewpoint tempera-ture close to the temperature of the incoming makeupwater. The condensate can be recycled into the makeupwater. For a dewpoint temperature of 25°C at 5-MPa pres-sure, the corresponding steam mole fraction is 0.065%.This steam concentration will generally not be low enoughfor storage or transport because the storage or transportpressure is generally higher than 5 MPa. So an additionalstage of dehydration will be required. Several industrial gasdehydration processes are available (e.g., glycol dehydra-




Fig

.8.

Mod

ular

hyd

roge

n pl

ant

usin

g H

TE

.


tion, dessicant adsorption) to reduce the steam content tovery low levels. A hydrogen-steam separator unit is shownin Fig. 8 with this in mind. The specific design for the sepa-rator would be dependent on the final intended use for thehydrogen.

5. Conclusions

High-temperature electrolysis using solid-oxide technologyappears to be a viable means for producing hydrogen usingnuclear energy. Laboratory-scale experiments in the pastyear have shown that this technology can produce hydro-gen at close to the theoretical parameters.

The conceptual design of an electrolytic plant to beattached to a 600-MWth reactor has been developed, sug-gesting that the plant would be of moderate size and thatthe parameters of cells would be reasonable. A rail-trans-portable, modular unit has been described such that themodules could be manufactured in a factory and installedat the reactor site. The modules would each produce ~0.18kg (2000 normal liters) of hydrogen per second and requirean electrical input of ~20 MWe.



4. THERMOCHEMICAL PRODUCTION OF

HYDROGEN

Bulk chemical processes benefit from economy ofscale and may turn out to be the best for very-large-scalenuclear production of hydrogen for a mature global hydro-gen economy.

Thermochemical cycles produce hydrogen through aseries of chemical reactions in which the net result is theproduction of hydrogen and oxygen from water feedstock—with all chemicals in the process being fully recycled.

Thermochemical cycles, in which the iodine-sulfurprocess is considered the most promising and the best suit-ed to coupling with a high-temperature reactor, have thepotential of high conversion efficiency from nuclear heat tohydrogen and are receiving research and developmentattention worldwide.


4. ThermochemicalProduction of Hydrogen

1. Introduction

Thermochemical cycles produce hydrogen through a seriesof chemical reactions, where the net result is the produc-tion of hydrogen and oxygen from water at a much lowertemperature than in direct thermal decomposition (T higherthan 2600°C). Energy is supplied as heat for pure thermo-chemical cycles in the temperature range necessary todrive endothermic reactions, generally in the 750 to 1000°Crange or higher, or even as a combination of heat and elec-trolysis for so-called hybrid processes. All chemicals usedin the process are fully recycled. This solution for massiveproduction of hydrogen seems very attractive, because itfulfills two fundamental criteria of “sustainable develop-ment”: Production is supplied from an “inexhaustible” anduniformly distributed resource, water, and the production isgreenhouse gas emission free, as long as heat is suppliedaccording to this criterion, which is the case of nuclearenergy supply. Nevertheless, the thermochemical optionwill lead to a viable process only if it respects the two crite-ria of economic competitiveness: demonstrated technicalviability and efficiency. From the point of view of efficiency,thermochemical cycles may have theoretical advantagesover the alternative of high-temperature electrolysis,because their efficiency is not burdened with the detraction



of the efficiency of electricity. Thermochemical processesclearly have attractive scaling characteristics for large-scale applications. As will be described further, thermo-chemical cycles for massive hydrogen production from anuclear high-temperature source were studied and evalu-ated extensively in the 1970s in the context of the 1973 oilcrisis, but none led to an industrial application. The mainreasons were (a) the difficulty of acquiring reliable thermo-dynamic and kinetic data in the fields of interest for properevaluation of the viability of an industrial process and (b)the lack of tools, especially simulation tools for flow-sheetgeneration necessary to screen and survey solutions forindustrial feasibility. Another reason was the disappearancein the 1980s of the compelling need that justified researchand development (R&D) effort. The important conclusion ofthis early work was selection of emerging families of ther-mochemical cycles and, from these, the most promisingoptions.

The challenge in the coming years, driven by the per-spective of the transition to the hydrogen economy, is toturn the ideas and the promises of past decades intoindustrial reality. This means that thermochemical cycleshave now to prove viability and feasibility leading to thenear-term steps of demonstration included in the roadmapof generation IV reactor development. The first priorityshould be intensive efforts to acquire reliable data in thefields of thermodynamics, kinetics, and material behaviorto support simulation and flow-sheet optimization. Effortsmust also be directed at evaluation and optimization oftargets, which will in turn drive innovation, technologicalimprovement and breakthroughs, and technoeconomicevaluations. The objectives are extremely important andambitious, and the lead times are very short, which meansthat they need to be shared within the framework of inter-national collaboration. The objective of this chapter is topresent the fundamental characteristics of thermochemi-cal cycles, to discuss emerging options, to address diffi-culties and questions that research may answer, and todefine the scope for improvement.


THERMOCHEMICAL PRODUCTION OF HYDROGEN 49

Fig. 1. Schematic view of the overall process of water-splitting.

2. Principal Considerations for ThermochemicalCycles

2.1 Cycle Efficiency

In the calculation of the thermochemical production cost,the overall efficiency of a process and the investments forthe plant are the most important parameters. The primaryraw material is energy. The major target will be to attain anefficiency in the 40 to 50% range, i.e., beyond the efficien-cy of industrial current alkaline electrolysis, which current-ly (including the conversion of heat to electricity) is in the20 to 23% range or advanced high-temperature electrolysis(HTE) (28 to 32%) (Ref. 24).

The overall process (Fig. 1) can be considered as ablack box, where the input/output are as follows:

1. water at temperature Ta (usually ambient tempera-ture) and at atmospheric pressure

2. heat Qc provided at temperature Tc

3. heat Qc rejected at temperature Ta lower than Tc.

4. work W (for pumps, compressors), driven by elec-tricity produced by a turbine with an electrical effi-ciency ηel.

Necessary work can be converted into an equivalent heatsource Qw using the relation



(1)

where ηw is the efficiency of the work to be provided.

There are many definitions of efficiency; we define anenergetic efficiency as

(2)

where ∆G0(Ta = 25°C) = −237 kJ/mol is the free enthalpy ofwater formation.

Rather than ηe, the thermal efficiency ηth is usuallyemployed:

(3)

where ∆H0(Ta = 25°C) = −286 kJ/mol is the enthalpy ofwater.

According to the second law of thermodynamics, ηe islimited by the Carnot efficiency therefore,

(4)

This inequality (4) shows that it is possible to reach overallefficiency (up to 89% for Tc = 1000°C; see Fig. 1), but onlyif heat is provided at high temperature and work is mini-mized to avoid the double conversion of heat ⇒ electricity,electricity ⇒ work within the process. Endothermic reac-tions should be handled at the highest temperature com-patible with the heat source.

This inequality (4) is only an upper-bound efficiency.Effective efficiencies of thermochemical cycles and HTEare much below this efficiency limit for the following rea-sons:

1. To avoid having reactants mixed with the products,a large excess of energy is often necessary (preliminarycalculations published to date have not emphasized thispoint). This is particularly critical in cycles with multiple

hth �¢H

01Ta2Qc � Qw

6 a1 �Ta

Tcb ¢H

01Ta2¢G01Ta2 .

11 � Ta>Tc 2;

hth �¢H

01Ta2Qc � Qw

,

he �¢G01Ta2Qc � Qw

,

W � hwhelQw,



Fig. 2. Carnot efficiency limit (Ta = 50°C).

reactions. Indeed, thermochemical cycles with too manyreactions will probably never be competitive.

2. Substantial heat is usually needed to upper-heatand eventually vaporize the reactants. While this heat isassumed to be recovered from cooling of the products, inmost cases, especially if hydrogen is produced at high tem-perature, most of the available heat is at too low tempera-ture to be useful.

3. Chemical reactions are often quite incompleteeven with catalysts, especially at low temperatures.

4. The thermochemical cycles have been designed tobe coupled to a nuclear reactor, and more particularly to ahigh-temperature reactor (HTR). From a thermodynamicpoint of view, it is important to fit the Q(T ) curve of the heatavailable from the reactor to the required Q(T ) curve of theprocess to maximize efficiency. Reactions occurring at afixed temperature do not match well with HTR heat. Forexample, for the UT3 process (Fig. 3), which has two reac-tions at fixed temperature (near 725 and 575°C), HTR cou-pling will imply a maximum temperature for the helium loopof more than 875°C and a minimum temperature exceed-ing 600°C.



Fig. 3. UT3 deficient coupling with HTR heat.

2.2 Main Features for Evaluation of ThermochemicalCycles

2.2.1 Constraints on Economic Viability

The main problem with the economic viability of thermo-chemical cycles lies in the large recirculation of materials.Typically, for 1 t of hydrogen produced, the minimum amountof material recirculation is between 500 and 10 000 t (Fig. 4).Therefore, cycles involving only liquids or gases and liquidsare preferred because transport of such large amounts ofsolid materials would consume too much energy.

Processes with multistage chemical operations withpotentially slow kinetic steps may lead to very importantholdup of materials, which may adversely affect the plantinvestment cost. The purity of hydrogen and oxygen pro-duced is an important issue with a direct influence on cost,as the molar value of the intermediate material is oftenmuch higher than the hydrogen produced (for example, bya factor of 1000 for iodine). The hydrogen purity to be sup-plied to fuel cells, for example, must also be specified.

Side reactions may occur; they have to be observedand quantified because they make the optimization of the



Fig. 4. Typical amount of chemicals per ton of produced H2 bydifferent processes: electrolysis, HTE, and some thermo-chemical cycles (see Sec. 4 for description).

thermochemical process more complex. Also, the toxicity ofthe compounds can be considered as go- or no-go criteria.The quite corrosive property of strong acids at high tem-perature implies tests and developments of advancedmaterials that will directly affect the feasibility of theprocess (corrosion products may not be poisons) and theinvestment cost.

2.2.2 Emerging Issues

There are two possible valuable by-products of thermo-chemical cycles: oxygen, which is also produced by con-ventional electrolysis, and heat (Qa), if the temperature Tais higher than 100°C. This heat provides an opportunity toproduce massive amounts of potable water, which could bea high priority in a large part of the world before the end ofthe century. As almost all losses of efficiency in thermo-



chemical processes are heat losses, it means that, for aprocess with 40% efficiency for production of hydrogen,one can recover more than 50% from the HTR heat to pro-duce drinkable water.

Storage of hydrogen is a major weakness of thehydrogen economy. Even if high-pressure storage and liq-uefaction are envisaged, the most advantageous method ofstorage could be a hydrogenated compound (methanol,ammonia, hydride). This hydrogenated compound could bethe final product of the thermochemical cycle, which wouldincrease the efficiency of the whole transformation anddecrease the global cost. For example, it is possible to pro-duce methanol using water and carbon dioxide accordingto the following transformation25:

CO2 + 3 CH4 + 2 H2O ⇔ 4 CO + 8 H2

4 CO + 8 H2 ⇔ 4 CH3OH

3 CH3OH ⇔ 3 CH4 + 3/2 O2 (electrolysis, low voltage),

which yields

CO2 + 2 H2O ⇔ CH3OH + 3/2 O2.

The methanol can be stored and reformed to yield 3 H2 and1 CO2, the CO2 being retained but occupying a far smallerstorage volume than the associated H2.

3. History of Thermochemical Cycles Studies

The concept of thermochemical production of hydrogenfrom water was first studied thermodynamically by Funkand Reinstrom26 in 1964. The first major program was atthe European Community Joint Research Center (Ispra),beginning in the late 1960s and continuing till 1983 (Ref.27). The goal of this work was to identify thermochemicalcycles that could be potentially coupled to a high-tempera-ture gas-cooled reactor. A three-phase program investi-gated 24 cycles. In phase I, thermochemical cycles weredeveloped on a base of mercury, manganese, and vanadi-um. Mercury was abandoned because of its toxicity. In



phase II, nine cycles based on iron and chlorine wereinvestigated, because the chemistry of iron and chlorinewas already well known and the reactants were cheap andnot particularly noxious. The iron-chlorine cycles wereabandoned because of the difficulty encountered with ther-mal decomposition of ferric chloride and because detailedefficiency calculations were much lower than the firstapproximations. Phase III focused on several sulfur-basedcycles with a laboratory demonstration of the sulfur-bromine process. Associated with these laboratory effortswere parallel activities involving corrosion testing, design oflarge-scale equipment, and development of industrial flowsheets. The 1.5-yr duration of the hybrid sulfur-brominecycle laboratory tests remains the most extensive demon-stration of any thermochemical cycle.

In the United States, the Gas Research Institute fund-ed a long-term program that systematically analyzed ther-mochemical cycles. During the 9-yr program, 200 distinctthermochemical cycles were examined, approximately 125cycles were considered feasible based on thermodynamicconsiderations, and the 80 most promising were tested inlaboratory. Of these, 15 were found operable using batchtechniques with reagent-grade chemicals, and 8 were suc-cessfully operated with recycled materials to achieve proof-of-principle. In particular, the sulfur-iodine (SI) cycle wasextensively studied by the General Atomics Company, i.e.,individual reactions, corrosion, bench scales of the threeparts of the cycle, and full-scale flow-sheet development.27

Russia too carried out some major research on thermo-chemical cycles and finally built a small demonstration loopof the hybrid S process.

Japan has focused its efforts on two cycles: the UT3cycle (iron, calcium, and bromine) and the SI. In bothcases, the reactions were studied separately, and demon-stration loops were successfully operated at laboratoryscale. Japan is the only country that has maintained unin-terrupted activity in this field.28 With the decrease in oilprices in the 1980s, all these efforts were abandonedexcept for a small effort in Japan (Fig. 5).



Fig. 5. Evolution of the number of scientific papers on thermo-chemical production of hydrogen published between 1973and 2004.

It appears that a large panel of cycles has then beentested, and the feasibility of some promising thermochem-ical cycles has been demonstrated. A recent review29 eval-uates these cycles based on the following criteria: numberof reactions, number of chemical steps, number of ele-ments, least abundant element in the process, corrosionissues, degree to which the process is continuous and theflow of solids is minimized, maximum temperature of theprocess, degree to which the chemistry of the cycle hasbeen demonstrated, and the extent to which cost informa-tion is available. This review must be treated cautiouslybecause it neglects some industrial-scale issues like heatexchangers and size of equipment and because it tends tooverpenalize the cycles lacking relevant thermodynamicdata. Four cycles clearly emerge from this review: first, theWestinghouse cycle and then, three close challengers, theIspra hybrid cycle, the UT3 cycle, and the SI cycle. We willdescribe them with some remarks concerning their indus-trial scaleup.



4. Main Known Thermochemical Cycles

4.1 The UT3 cycle

The UT3 cycle (Fig. 6) was developed at the University ofTokyo in the 1980-1990s (Ref. 30). A small pilot plant wassuccessfully operated for a few days.

Stages 1 and 3 produce reactants for stages 2 and 4.The process has four reactors connected in series inside aloop; after one cycle, the reactors are switched off and thedirection of the flow is reversed. This cycling may causecoupling problems with a continuous HTR source becauseof variations in the temperature and heat demand of thefour reactors.

The physicochemical approach forecasts some diffi-culties because of the sintering of the solid reactants. Withregard to the second and fourth reactions, thermodynamicsis favored by operation at very low pressure and high tem-perature, whereas volatility of the bromide moleculesnecessitates the opposite conditions. One way could be tooperate with overstoichiometric conditions, which implieslarge volumes of gases and will probably lead to exces-sively large reactors because of the low pressure levelrequired by thermodynamics.

Fig. 6. The UT3 cycle.



Many other problems remain unresolved, such as thehigh cost of fabrication of reactants and bromine toxicity.31

4.2 The Hybrid S and Ispra Cycles

The three other cycles belong to the sulfur cycles family.Decomposition of H2SO4 meets the constraints explainedpreviously and is well suited for the oxygen generation stepof thermochemical water-splitting:

1. The main data are known for this thermochemicalreaction as sulfuric acid decomposition is now an industrialprocess.

2. Sulfur is a common product, and recirculation ratein the flow sheet of decomposition of H2SO4 is low(between 1 and 1.5 molar).

3. Heat coupling with HTR heat is excellent, shown inFig. 7, enabling up to 70% thermal efficiency for thisprocess.

The variants within the family of sulfur-cycle processes allconcern the hydrogen production section of the process.

Fig. 7. Good matching of heat needed for H2SO4 decomposition(solid line) and heat delivered by the HTR (dashed line).Energy required for H2SO4 distillation is obtained byrecovering heat from the cooling of reactants and recom-bination of the remaining SO2.



Fig. 8. Hybrid S cycle.

The simplest (but perhaps not the most efficient) wayis to electrolyze SO2 to H2SO4 and H2; this is the hybrid Scycle (Fig. 8) (also called Westinghouse cycle in the UnitedStates):

SO2 (g) + 2 H2O (l) = H2 (g) + H2SO4 (aq)

(electrolysis, 30 to 90°C)

H2SO4 (g) = H2O(l) + SO2 (g) + 1/2 O2 (g)

(thermochemical, 800 to 950°C).

The main features of this cycle are that it consists ofonly two reactions and it allows the hydrogen production tobe separated from the nuclear plant. The main drawback isthe use of an electrolytic step. The most important issuesfor increasing the competitiveness of this cycle are reduc-tion of the cell voltage (currently ~0.6 V compared with atheoretical 0.17 V) and the cost of the electrolyzers.Special caution has to be taken to avoid sulfur or H2S for-mation in the electrolytic step.32

In general, processes involving an electrolytic step arenoncompetitive because of the overpotential and theinvestment in electrolyzers. However, this cycle may be theonly exception because of the very low theoretical energydemand and because it allows use of only one chemicalstep, thereby avoiding most of the problems of purityencountered in other thermochemical cycles. The calculat-ed industrial efficiency24 are in the 37 to 40% range, which,despite its electrolytic part, places it as a very good chal-lenger to the electrolysis of water.



Another variation is the hybrid sulfur-bromine Ispra cycle:

SO2 (g) + 2 H2O (l) + Br (g)

= 2 HBr(g) + H2SO4 (aq) (30 to 120°C)

H2SO4 (g) = H2O(l) + SO2 (g) + 1/2 O2 (g)

(thermochemical, 800 to 950°C)

2 HBr(g) ⇒ H2 (g)+ Br2 (g) (electrolysis, 70° to 320°C)

E = 1.066 V

Relative to the hybrid S cycle, this cycle has two advan-tages: It tends to minimize H2S and S formation, and muchless water is produced in the H2SO4. The main problem isthe voltage needed with a theoretical amount of 1.066 V atambient temperature, which is too high for good efficiencyof this cycle. Nevertheless, this voltage can be consider-ably lowered by operating at a higher temperature (theIspra demonstration operated at 0.8 V). A serious supple-mentary problem is the toxicity of bromine.

4.3 The SI Cycle

The SI cycle is a pure thermochemical cycle initially pro-posed in the mid-1970s. It consists mainly of the three fol-lowing reactions:

1. I2(s) + SO2(g) + 2H2O(1) → 2HI + H2SO4

2. 2HI → H2 + I2

3. H2SO4 → SO2 + H2O + 1/2 O2.

The first reaction, known as the Bunsen reaction, proceedsexothermically as an SO2 gas absorption reaction with amixture of solid iodine and liquid water. The second reac-tion, in which hydroiodic acid is decomposed into iodineand hydrogen, is slightly endothermic. In the third reaction,sulfuric acid is decomposed into a mixture of sulfur dioxide,water, and oxygen, as for the hybrid cycles previouslydescribed. Except for hydrogen and oxygen, the productsof the two last reactions are recycled as reactants in theBunsen reaction (see Fig. 9).



Fig. 9. Current version of the SI cycle.

In the 1980s, General Atomics27 described a process-ing scheme for the SI cycle in which the reactants and prod-ucts were fluids only. Iodine’s melting point being ~114°C,the Bunsen reaction was then carried out at ~120°C. Thisreaction lacks favor at such a temperature (∆G400 K =82 kJ·mol-1), but conversion can be improved by divergingfrom the stoichiometric Bunsen conditions, e.g., by increas-ing initial amounts of I2 and H2O. Adding I2 helps to shiftequilibrium 1 toward the formation of hydroiodic and sulfu-ric acids. Furthermore, it was also stated in the 1970s thatthe products of Bunsen reaction can spontaneously sepa-rate into an aqueous solution of sulfuric acid (light phase)and an aqueous solution of polyhydroiodic acids (heavyphase called HIx) when an excess of iodine is present in themedium. This phase separation, arising from the formationof polyhydroiodic acids in which iodide anions are solvatedby molecular diiodine, is crucial for the succeeding steps ofthe process, and this property constituted the first keybreakthrough in the SI cycle. Moreover, when water is intro-duced in excess, apart from the fact that it tends to shiftequilibrium 1 toward the right, it causes a substantialchange in the reaction enthalpy due to acid dilution, accord-ing to the following reactions:



H2SO4(1) + 4H2O(1) → (H2SO4 + 4H2O)(aq)

∆G400 K ≈ −66 kJ⋅mol-1

2HI(g)+ 8I2(1) + 10H2O(1) → (2HI + 10H2O + 8I2)(aq)

∆G400 K ≈ −104 kJ⋅mol−1

Under these conditions, the global modified Bunsen reac-tion (as described previously) then becomes thermody-namically favored:

4. 9I2 + SO2 + 16H2O → (2HI + 10H2O + 8I2)

+ (H2SO4 + 4H2O)

∆G400 K ≈ −88 kJ·mol−1.

However, this excess of water makes the reaction quiteexothermic (∆H400 K ≈ −90 kJ·mol−1), which dramaticallylowers the maximum efficiency of the cycle.33 Indeed, itwould be beneficial to limit the energy loss linked to thenegative value of ∆G400 K by devising a variation of freeenergy during Bunsen reaction that was closer to zero.Moreover, an excess of I2 and H2O is quite unfavorable forthe following HIx section (as is described later). Hence,research and development efforts are being devoted totesting new designs of the Bunsen reaction with lesseramounts of I2 and H2O to find the best compromisebetween thermodynamic improvement, phase separation,and energy loss.34 One can also try to improve the Bunsenreaction at 120°C by increasing SO2 pressure to limit theexcess of iodine and water and therefore limit the negativeconsequences on the HIx section (Fig. 10).

In the SI cycle, HI decomposition according to reaction2 must be achieved from the HIx mixture produced in themodified Bunsen reaction described in reaction 4. The HIxsection is the most crucial step for the efficiency of thecycle as it presents the lower rate. Three main difficultiesmust be overcome: the presence of an azeotrope (Fig. 11)in the mixture, which prevents efficient distillation of HI, theincomplete and very slow decomposition of HI in H2 and I2,and the very large energy exchange, more than one order



Fig. 10. Bunsen section improvement on SI efficiency.

Fig. 11. Azeotropic lines of HIx mixture (♦: experimental data;full lines: model).

of magnitude greater than for the sulfuric acid decomposi-tion, due to the large heat capacity of the HIx mixture.

The direct decomposition of HI in gaseous phase isnormally an incomplete process, making it necessary torecirculate too large an amount of product and to vaporizetoo much water. The original scheme of General Atomics27

(GA) was to use phosphoric acid, but this is costly and has



poor efficiency; besides, it introduces a third element in thecycle.

In the 1980s, RWth Aachen35 proposed a concept ofreactive distillation of HIx, allowing distillation and decom-position of HI in the same reactor at 350°C. A liquid-gasequilibrium is obtained in the middle of the column,gaseous H2 is extracted at the top of the column, and I2 issolubilized in the lower liquid phase. This process is cur-rently under investigation, the objectives being to try to limitthe excess of H2O and I2 reactants initially introduced in theBunsen section (to reduce energy consumed) and to raiseHI molarity through water evaporation and iodine separa-tion from the HIx mixture.

The use of permselective membranes is also consid-ered for the HIx section of the cycle. One way to proceedmay consist in using membranes to separate HI from theHIx mixture before performing HI decomposition. Anothermethod would be to heat up the HIx mixture to ~600°C inorder to have a homogeneous gaseous phase and to usepermselective membranes to separate the slightest amountof H2 produced through catalytic activation in the reactor inorder to shift the equilibrium described in reaction 2 towardthe decomposition of HI.

Preliminary schedule improvements in global efficien-cy of the whole process are reported Fig. 12. Reactive dis-tillation of the HIx mixture and the use of permselectivemembranes are now under study.36 To correctly estimatethe potential of these processes, it is essential to measurethe partial pressure of the different compounds present inthe liquid-gas equilibrium of HIx mixtures over a largerange of temperatures. This is a real challenge because ofthe opacity of iodine and the corrosive properties of thechemical medium that is encountered.

To improve the global SI cycle, it appears essential toconcentrate effort on the HIx section, necessitating modifi-cations to the preceding Bunsen section. Experimentaldata are required to optimize both HIx and Bunsen sec-tions. Research and development programs will involve, on



Fig. 12. Expected global SI cycle efficiency for different tech-niques of HI decomposition.

the one hand, design and construction of appropriate reac-tors36 and, on the other hand, design of new analyticalmethods to determine the composition of the solution,mainly using spectroscopic techniques (ultraviolet-visiblespectrophotometry, inductively coupled plasma–atomicemission spectroscopy, etc.), and of the gas phases, usingoriginal optical diagnostics (IR, Raman, etc.).

The cost of iodine also has a great influence on theeconomic viability of this cycle: Care must be taken not tolose more than 0.02% iodine in the products and to mini-mize the holdup of iodine in this cycle.

To conclude, as Fig. 13 shows, challenging improve-ments are still required, either from a thermodynamic pointof view (optimization requiring data acquisition) and/orfrom a theoretical point of view (process improvements,technological breakthroughs).

5. Conclusion

Large-scale hydrogen production through thermochemicalcycles powered by nuclear heat is becoming popular againbecause of two major context changes: the prospects of a



transition to a hydrogen-based economy and the prospectsfor development of fourth-generation, high-temperaturegas reactors. Among the thermochemical cycle options, theSI option, which is fully thermochemical and best suited tocoupling with a nuclear heat source, is the most promisingone in terms of efficiency and cost. Nevertheless, in spiteof all the efforts made in the past and small-scale labora-tory demonstrations, this technology is not yet mature.Demonstration of viability remains to be established on thebasis of reliable experimental data (thermodynamic andkinetic data). Process optimization will need to achieve anefficiency on the order of 50% for economic competitive-ness (even with residual heat utilization for applicationssuch as seawater desalination and economic value real-ized from the oxygen coproduct). This is the basis of all thecurrent R&D efforts, a key point being the development ofanalysis and diagnostic methods allowing access to meas-urements. Once this viability is secured, it will be necessaryto focus the effort on the feasibility phase, taking intoaccount process optimization, including a few technologicalbreakthroughs on key issues, materials selection, a solu-

Fig. 13. Evaluation for SI progress margins: ____ Carnot limit,– – – R&D max Commissariat à l’Energie Atomique33

(CEA), • • • GA 2003 (Ref. 37), CEA 2003.


tion for coupling to the reactor and optimization of heat dis-tribution within the cycle, technoeconomical analysis of thewhole process, and demonstration phases at scales repre-sentative of industrial reality. The challenge—within theshort time frame of approximately the next decade—reallypertains to basic science and technology. An importantadvantage for thermochemical cycle development is the cli-mate of international collaborations and the sharing ofresearch results, which adds synergy to individual efforts.



5. SAFETY ISSUES OFNUCLEAR HYDROGEN

PRODUCTION

Handling of hydrogen has a long commercial experi-ence, and safety technology for hydrogen in industrialapplication is now a mature field.

What is new for the field of nuclear hydrogen produc-tion are safety issues arising for co-sited hydrogen andnuclear plants where explosive hazards of hydrogen affectnuclear safety.

In the design, safety assessment, and operation ofnuclear hydrogen production plants, the experiences ofhydrogen safety for light water reactors, as well as theaccumulated experiences of industrial hydrogen practice inchemical plants, will play an important role.


5. Safety Issues of NuclearHydrogen Production

1. Introduction

In 1766, the very famous and reclusive scientist HenryCavendish discovered a gas that he called “flammable air.”In those early times of chemistry as a science, Cavendishwas an outstanding researcher with a very limited capabil-ity to communicate his research results. In 1781 he wasable to demonstrate that flammable air, in combination withoxygen, produced water. Two years after that (1783),Lavoisier proposed the name hydrogen for Cavendish’sflammable air, placing it in the emerging systematic contextof chemistry. Some decades later, hydrogen becameaccepted as the lightest of all the chemical elements. Itsconstitution was definitely established early in the 20th cen-tury as an electron orbiting a proton, but its simplicity hadalready been identified in the Mendeleyev periodic table ageneration before. By then, it was evident that all the hydro-gen in our planet occurs in compound molecules, mainlywater. To set it free, it must be chemically reduced fromthose molecules. Once isolated, hydrogen clearly deservesCavendish’s name, flammable air, because its combustionis easily initiated, is extremely energetic, and releases ahuge amount of heat. This creates very high temperaturesand can produce explosive overpressures. Warnings aboutit started in the pioneering work by Cavendish andLavoisier, but hydrogen found its early way to industrialapplications in the first decade of the 19th century in theform of manufactured gas.



A typical way to produce gas for lighting and heatingwas based on the anaerobic reaction between water vaporand very hot coal. After condensing excess steam, the non-condensable gas was mainly made of H2 and CO. Eventoday, this is the basis of coal gasification, a clean technol-ogy to burn coal, where a gas-steam combined cycle canbe used to increase overall efficiency. [The most powerfulplant of this technology is Elcogas in Puertollano, Spain,with 315 MWe (Ref. 38).]

For many years and in many places, different types ofmanufactured gases containing H2 were widely distributedand used for industrial, commercial, and residential appli-cations, but most of these have now been displaced bynatural gas. The safety record of the manufactured gasindustry was very good, as it has generally been in the nat-ural gas industry. However, in some cases the switch frommanufactured gas to natural gas was not without hazard. Inthe 1970s, for instance, the deployment of natural gas inBarcelona, Spain, was plagued with a short but intenseseries of accidents involving several casualties. Sub-sequent deployments of natural gas in other Spanish citieslearned from this experience and were uneventful. Sincethen, safety records of the natural gas industry and com-mercialization have been very good. The unfortunate pio-neering experience in Barcelona could be related to thelack of well-trained personnel, because safety standardswere well developed at that time. However, many workersinvolved in changing piping, valves, and, above all, gasburners, had no previous experience in this field. Thisexample is of general interest because a future deploymentof the hydrogen economy will have to appreciate very seri-ously that a highly flammable gas will be used by millionsof people without previous experience of it. At present,hydrogen is used only by specialists. Most of it is theso-called captive hydrogen, “captive” in the sense that it isproduced in the same facility where it is used, as in petro-leum refineries. In some specific areas (Ruhr Valley inGermany and La Porte industrial complex in Texas) hydro-gen pipelines are used to connect production installations


SAFETY IN NUCLEAR HYDROGEN PRODUCTION 73

to industrial consumers. However, the hydrogen pipelines’total length (in the world) is still below 1000 km, which istrivial compared to the total length of natural gas networks.

Approximately 400 billion m3 of hydrogen is producedevery year in the world.39 Petroleum refineries use thebiggest share of it for desulfuration and for hydrogenationand cracking of heavy molecules. It is also used in makingfertilizers and in food industries (also to increase the hydro-gen content of molecules that originally have much lesscommercial value). In all these applications, the safetyrecord is high, and the existence of technical legislation [forinstance, 29 CFR 1910 (Ref. 40)], standards, and safetyguides is a guarantee for future developments. However, itis worth noting that more than 40% of accidents arecaused by human errors.39 This reinforces the warning thatdeployment of the hydrogen economy needs to be sup-ported with a strong commitment to training and qualifyingpersonnel.

There are two hydrogen applications in which safety isa primary issue: cooling of large alternators and, evenmore, propulsion of large rockets, including shuttles of theNational Aeronautics and Space Administration (NASA).

A 500 MV·A alternator typically has a hydrogen gascharge of 70 Nm3 (or even more) acting as main coolant forrotor and stator coils. Both to fill and remove it, an inertatmosphere of CO2 must be used. Obviously, such a hydro-gen content is a hazard, and safety measures are takenboth in design of the machine and during operation to avoidrisks of fire or explosion.

Large rockets for space missions use very highamounts of liquid hydrogen (and liquid oxygen) for propul-sion, and very secure systems must be designed, con-structed, and maintained to provide a reliable propulsionmechanism. Besides that, hydrogen is used inside spaceshuttles to produce electricity (and water) in fuel cells.Again, safety is a major concern, though NASA’s record inthis regard is encouragingly good.



The use of hydrogen in rockets is explained in termsof weight and payload. One kilogram of H2 has a heat con-tent equal to 2.8 kg of gasoline or 2.4 kg of methane.However, in other applications, volume can be more impor-tant than weight, and H2 is less attractive. For instance, inenergy content, 1 � of liquid H2 is equivalent to 0.27 � ofgasoline, but liquid hydrogen has the additional drawback ofneeding ultralow temperatures. Similarly, 1 � of gaseous H2at 350 bar (and room temperature) is equivalent to 0.1 � ofgasoline or 0.3 � of methane (also at 350 bar).

2. Hydrogen Properties Related to Safety

Tables 1 and 2 gather some relevant information aboutgaseous and liquid hydrogen. Among the most importantdata are the limits of flammability in air. In molar (volume)percentage, the lean limit is 4.1% H2, and the rich limit75%. This is the widest range of flammability for all com-bustible gases. The very wide detonability range, between18.3% and 59%, is also worth pointing out, because it isrelated to the combustion speed and the overpressure inthe wavefront, which can reach 2 MPa (20 bar).

Liquid hydrogen requires very low temperatures, andit is the second coldest cryogen after liquid helium. In thiscontext, only helium can provide an inert atmosphere overliquid H2, because any other inert fluid (as CO2 or N2)would condense.

Liquid H2 has a very small specific heat of vaporiza-tion (<0.5 MJ/kg), which means that evaporation will occursuddenly after any liquid hydrogen spill or leak. In a fire ofliquid H2, the heat from the flames will strongly boost theboiling process, and the main inhibitor against accelerationof the fire will be the lack of oxygen, because it will taketime for the oxygen to diffuse (enhanced by convection)and approach the fire front from the surrounding air.

One of the means of characterizing the potential con-sequences of a hydrogen explosion is by the Fireballmodel. Of course, those models cannot predict the behav-



ior of hydrogen explosions in all geometries and ventila-tion/confinement conditions. It is obvious that in theabsence of enough oxygen, no hydrogen explosion willoccur. In the classical combustion triangle, three things are

Table 1.Gaseous H2 Properties

Property Value

Molecular weight 2.02Boiling point (K) 20.3Critical temperature (K) 33Density of gas (kg/m3) 0.0838Viscosity of gas at NTPa (g/cm⋅s) 8.9 � 10�5

Stoichiometric composition (vol%) 29.53Self-ignition temperature (K) 8.58Minimum energy for ignition (mJ) 0.02Limits of flammability in air (vol%) 4.1 to 75

Limits of detonability in air (vol%) 18.3 to 59Flame temperature (K) 2318Detonation velocity (km/s) 1.48 to 2.15Detonation overpressure (kPa) 1470

Lower heating value (MJ/kg) 120Higher heating value (MJ/kg) 142Burning velocity at NTP (cm/s) 265 to 325Percent thermal energy radiated (%) ~21

Heat release rate (kJ/cm2⋅s) 1.53 � 10�2

Energy of explosion (kg TNT/m3) 2Buoyant velocity at NTP (cm/s) 1.2 to 9Diffusion coefficient at NTP (cm2 /s) 0.61

aNormal temperature and pressure.

Table 2.Liquid Hydrogen Properties

Property Value

Temperature of liquid at NBPa (K) 20.3Heat of vaporization (MJ/kg) 0.46Density of liquid at NBP (kg/m3) 71Density of vapor at NBP (kg/m3) 1.34Viscosity of NBP liquid (g/cm⋅s) 13.56Vaporization rate of liquid pools without burning (cm/min) 2.5 to 5.0Energy of explosion (g TNT/cm3 NBP liquid fuel) 1.71

aNormal boiling point, i.e., at 1 atm.



needed: a fuel (hydrogen), an oxidizer (oxygen), and anigniting point, or hot point. Flammable mixtures of H2/airneed very small amounts of heat to start ignition.

To have a high-speed deflagration (250 to 320 m/s) ora detonation (~2000 m/s), an H2/air mixture is needed. Theworst case can happen in closed volumes capable of con-fining an overpressure as the mixture is formed. Then a hotspot can start ignition and a detonation wave can belaunched with overpressures that can destroy the sur-rounding walls, pipes, and structures.

There is a well-known set of safety principles andrules to minimize the hydrogen risk to the levels of anyother industry. There are three main areas of work in thiscontext: inherently safe design, personnel training, andinstrumentation and control. Safe design of hydrogen facil-ities must include the fail-safe rule for any part of the equip-ment (valves, pumping system, vacuum devices, etc.) andmust also include such other principles as adequate venti-lation, elimination of potential ignition points, and someother items that will specifically be addressed in the follow-ing sections of this chapter. Some of the safety analysistechniques and principles are general (for instance, the useof fault-tree analysis and consequences analysis), butsome issues are specific to the hydrogen industry, such asthe embrittlement produced by hydrogen absorption inmetallic components.

Operator training is a fundamental prerequisite in allactivities with some level of risk, as is the case consideredhere. Last, but not least, monitoring and control of hydro-gen facilities is very likely the main point in operationalsafety. Like radioactivity, hydrogen can be detected easily,and early warnings can be triggered by H2 concentrationswell below the lean limit of flammability or by identifying asmall leak. In both cases, monitoring of the evolution of theaccident is essential for a suitable emergency response. Inmany accidents, natural or enhanced ventilation is the pre-ferred option. For enclosed spaces, catalytic combustioncan be a better option. Early fire detection and quenchingare other fundamental objectives.



3. A Brief Overview of Hydrogen Risks inLight Water Reactors

Hydrogen risks in light water reactors (LWRs) were first for-mally recognized in 1964 by the U.S. Advisory Committeeon Reactor Safeguards in its report on the proposed con-struction of the 1473-MWth Connecticut Yankee pressur-ized water reactor. The report alluded to the need for astronger containment system “if a large fraction of theZircaloy cladding would undergo metal-water reaction in acore meltdown, releasing heat and hydrogen, which couldburn adding more heat.” Since that statement, the study ofhydrogen risks has been considered worldwide by regula-tory authorities and the industry, and a great deal ofprogress has been made to understand and control suchrisks. Now it can be said that the knowledge acquiredthrough research and from the analysis of the operatingexperience is sufficient to guarantee the safety of suchpower plants from that type of risk.

The early theoretical considerations were soon provencorrect. The accident at the Three Mile Island unit 2 (TMI-2)nuclear power plant (March 28, 1979) involved the oxidationof 45% of the zirconium fuel cladding, generating some 460kg of hydrogen and reaching ~7.9% volume concentrationin the containment atmosphere that also contained 35%steam. A hydrogen deflagration took place, the pressurepeak reached 3 bar, but no major damage was found in thecontainment.41 Likewise, in the 1986 Chernobyl-4 nuclearaccident, substantial amounts of hydrogen were also pro-duced by metal-water reactions; hydrogen detonations anddeflagrations were largely responsible for severe damage tothe system.

All these experiences clearly demonstrate that in caseof a severe accident, the integrity of the containment build-ing could be threatened by the formation and ignition offlammable hydrogen-air-steam mixtures.42 Violent combus-tion regimes could lead to the loss of leaktightness of thisultimate barrier, while stabilized diffusion flames could dam-age components or systems essential to safe operation.Prediction of the formation and behavior of such flammable



mixtures and their possible combustion constitutes funda-mental information for the design and implementation of dif-ferent mitigation devices and procedures.43 As a result oflarge investments in experiments, theory, and computing,substantial background knowledge, an experimental data-base, and powerful analytical tools for simulations havebeen developed in this area.

On their side, regulatory organizations have devel-oped appropriate regulations to counter the hydrogen prob-lem through specific design requirements (10 CFR 50Appendix A) and to mitigate its consequences in the unlike-ly event that such metal-water reactions are produced.Nevertheless, no generally agreed-on set of solutions hasuniversally been adopted.

4. Risk Analysis of Hydrogen Facilities

At present, 65% of the hydrogen is consumed in oil refiner-ies. Fertilizer production facilities, food industries, and otherchemical specialties are also substantial hydrogen con-sumers. In most cases, hydrogen is produced in the sameplace that it is consumed—so-called captive hydrogen.Hydrogen for dispersed applications such as fuel cells isstill a very minor fraction of total world production.

Although each oil refinery defines and implements itsown safety procedures, most follow the standardsapproved by several organizations [American NationalStandards Institute, American Society of MechanicalEngineers, Compressed Gas Association, Group onAnalysis and Management of Accidents of the Committeeon the Safety of Nuclear Installations Working Group,International Organization for Standardization, NationalFire Protection Association (NFPA), National Institute forOccupational Safety and Health, National HydrogenAssociation, Occupational Safety and Health Administra-tion in the United States, TÜV from Germany, KHK fromJapan, and others]. Some of those standards (for instance,NFPA) are addressed to fire protection, and others apply topressurized tanks of 35 MPa and 70 MPa [ISO-11439 andNGV-2 (United States)]. Tanks tested at the Institute of



High Pressure for Gas Applications (KHK, Japan) achieved160 MPa before rupture (i.e., 2.35 times as high as the 70-MPa reference value).

More than 4000 institutions, universities, companies,and professional associations have worked with hydrogenand have contributed to the substantial body of knowledgeon hydrogen safety. This is a very valuable resource.However, most of those institutions and companies operatewith very high professional standards. Laymen have notparticipated so far in the hydrogen sector. If the hydrogeneconomy initiative seeks to extend hydrogen applications tothe lay level, new safety standards will be required. Fillingan alternator with hydrogen, for instance, is a routine taskfor specialized professionals, but it would be an irresponsi-bly dangerous mission for untrained personnel. A key pointin hydrogen safety for the future is how to cope with thisproblem of extending hydrogen application to everyone.

5. Integrated Safety in Nuclear Production of Hydrogen

5.1 Electrolytic Production of H2

In this case, an energy carrier, electricity, is used to connect(and to geographically decouple) the nuclear power plant(NPP) and the hydrogen facility. Although many scenarioscan be devised about this idea, a very sound one wouldrely on a dedicated NPP. This means that those NPPswould not sell their kilowatt-hours to the general grid but toa specific site (or pool of sites) of hydrogen facilities. In thisway, the nuclear-chemical system could be better optimizedas a unified entity.

However, the opposite scenario would also beacceptable, with the hydrogen facilities just working dur-ing hours of very low electricity prices. Note that H2 isstorable and should be stored for commercialization. So,discontinuous operation of the electrolytic facilities is apossibility to be considered, although continuous opera-tion seems to be more adequate from the electrochemicalpoint of view.



Regardless of how nuclear power is converted intohydrogen, no special safety features arise, provided thereis enough physical distance between the NPP and thechemical plants. As already noted in relation to the gener-ation IV initiative, scenarios of nuclear H2 production wouldneed very safe and economically competitive reactors. Inturn, this requires a suitable nuclear fuel cycle, as studiedin Advanced Fuel Cycle Initiative of the U.S. Department ofEnergy. To become relevant in the hydrogen economy,nuclear fission will require the following:

1. proliferation resistance for the whole fuel cycle

2. enhanced efficiency in exploiting nuclear ores (nat-ural uranium, thorium)

3. minimization of radiotoxicity contained in the long-term nuclear wastes

4. very high safety standards, ruling out any nuclearaccidents that are not essentially confined withincontainment.

Ways and means to cope with those challenges arebeing envisaged in new conceptual designs and new repro-cessing techniques, such as UREX+. It could be said thatthe aforementioned principles have inspired the develop-ment of nuclear fission since its beginning (as was seen,for instance, in the International Nuclear Fuel CycleEvaluation, 1978–1980); however, they must be followedmore tightly in the quest for a new, much broader deploy-ment of nuclear energy in a context of sustainable devel-opment, including the hydrogen economy.

5.2 Thermochemical Production of H2

Ways to produce molecular hydrogen using endothermicreactions are described in Chap. 4. In general, they involvemultistage chemical reactors requiring heat at very hightemperatures (in the range from 700 to 900°C). This impliesa physical, thermal connection between the reactor and thehydrogen plant, although an intermediate heat exchanger(IHX) is included in almost all concepts.



Thermally insulated pipes would obviously be used fortransferring heat from the reactor (or the IHX) to the hydro-gen production facility (the chemical reactor). The distancecan be a few hundred metres, which is not sufficient to con-sider them as independent facilities. On the contrary, theymust be considered as belonging to the same installation (orsite) where accidents in one can affect the other. Forinstance, fires have been a safety issue since the very begin-ning of nuclear energy (the Windscale accident in the UnitedKingdom in 1957 started with a fire, the Vandellos accidentin Spain in 1989 was propagated by a fire, and many otherinstances could be cited). In such a nuclear-chemical com-pound, fires will have enhanced importance in its safetyanalysis. Moreover, the possibility of hydrogen fireballs ordetonation waves will have to be taken into account. Thecomplexity of this subject places it beyond the scope of thisreport, though that should not be taken to imply that it will beunmanageable. The existing expertise in nuclear safetyaddressing the hazards from hydrogen produced in nuclearaccidents is more than sufficient to deal with this problem.

Very likely, reactors dedicated to H2 production will bevery different from current water-cooled designs. However,all the experience gained with LWRs will be very relevantfor the future, particularly regarding safety, just as currentexperience with hydrogen is relevant to understanding andmanagement of hydrogen-related risks.

Last, but not least, some specific safety issues willappear in relation to the reactor coolant needed to reachtemperatures as high as required for hydrogen production.Candidate coolants are gases (particularly, inert gases),molten salts, and molten metals (particularly, lead).

In the case of gases, very high pressures needed foreffective cooling can present a threat to the mechanicalintegrity of some containment barriers. Moreover, residualheat removal is difficult to achieve in some types of acci-dents after a depressurization.

Molten metals and salts can work at atmospheric pres-sure, but they can present severe problems of corrosion.



The neutronic and thermal-hydraulic advantages of moltenlead, for instance, are well known, which is why it has beenproposed for some futuristic reactors (including accelerator-driven reactors, such as the energy amplifier by CarloRubbia and co-workers at CERN). However, corrosion ratesincrease enormously with temperature, and cladding andstructural materials (including IHX) are likely to require unfa-miliar materials.

Another advantage of molten lead (and similarcoolants with a very low Prandtl number) is the possibilityof using natural convection for cooling, at least to removeresidual heat. Passive safety features play a very importantrole in the development of all new reactor designs.

6. Conclusions

In the systematic review that will be needed for the safetyassessment of new reactor types, the accumulated experi-ence with LWRs will play a very important role across thefull range of potential reactor types that are being consid-ered for the energy requirements of the 21st century.Ensuring nuclear safety is the first priority in this evolution-ary revolution. Notwithstanding the TMI-2 and Chernobyl-4accidents, the safety record of the nuclear industry hasbeen very good. Even so, there is considerable scope forenhancements by using new materials (new fuels, materi-als, and coolants) and new designs (based on very stablereactors with strongly negative reactivity feedback, passivecooling, unconditional subcriticality in the case of acci-dents, etc.).

Safety experience in the production and use of hydro-gen in refineries and other chemical facilities is a veryimportant source of knowledge for addressing the safetyissues of the hydrogen economy. Several studies havecompiled most of the relevant data on hydrogen accidents.For instance, most of the facts on hydrogen accidents havebeen compiled44 from insurance records. From anotherviewpoint, Los Alamos National Laboratory carried out asafety analysis on liquid hydrogen shipment by highway


truck trailers.39 In a more complex context, NASA hasmade several safety evaluations of the use of hydrogen indifferent applications (direct combustion, fuel cells, andothers). All these studies will be extended in the near futureunder the umbrella of the International Partnership on theHydrogen Economy. It is believed that new studies will beneeded to specifically address some of the cases pro-posed by the generation IV initiative.

It is impossible to claim that all possible safety issueshave been anticipated in the analysis of nuclear productionof hydrogen by all possible methods. It can be stated, how-ever, that the principles, criteria, rules, and a methodologyto make safety assessments of any nuclear installationdevised to produce hydrogen, either by electrolysis or bythermochemical reactions, do all exist. Nuclear safetymethodologies and practices can be extended to hydrogenoperations to reach safety levels as high as reasonablyachievable.

BIBLIOGRAPHY

1. G. Brundtland, Our Common Future (TheBrundtland Report), Oxford University Press (1987).

2. A. L. Camp, J. C. Cummings, M. P. Sherman, Ch.F. Kupiec, R. J. Healy, J. S. Caplan, J. R. Sandhop, and J. H. Saunders, “Light Water Reactor Hydrogen Manual,”NUREG/CR-2726, SAND82-1137, Sandia National Lab-oratories (Aug. 1983).

3. “Standard for Hydrogen Piping Systems atConsumer Locations,” CGA G-5.4, Compressed Gas As-sociation, Arlington, Virginia (1992).

4. Proc. GENES/ANP Conf., Kyoto, Japan, Septem-ber 2003, H. Chung et al., “Nuclear Hydrogen Productionand Its Safe Handling,” Paper 1043; see also, K.Verfondern and W. Von Lensa, “European Research andDevelopment on HTGR Process Heat Applications,” Paper1126; see also, L. M. Crosbie and D. Chapin, “HydrogenProduction by Nuclear Heat,” Paper 1143.



5. Proc. 2nd Information Exchange Mtg. NuclearProduction of Hydrogen, Argonne, Illinois, October 2–3,2003, NEA 2003.

6. D. Wade et al., “STAR-H2: A 400 MWth Lead-Cooled, Long Refuelling Interval Reactor for HydrogenProduction,” paper ICONE11-376576, Proc. Int. Conf.Nuclear Energy, ICONE-11, Tokyo, Japan, April 20–23,2003, American Society of Mechanical Engineers.

7. L. Wolf, A. Rastogi, D. Wennerberg, T. Cron, andE. Hansjosten, “Detailed Assessment of the HDR Hydro-gen Deflagration Experiments E12,” Nucl. Technol., 125,136 (1999).

8. www.energiasostenible.net (English file).



6. HYDROGEN ECONOMICSFOR AUTOMOTIVE USE

Widespread conversion of transportation fuel tohydrogen could have a huge beneficial impact on the envi-ronment; however, it will be most effective only if the pri-mary energy used to produce hydrogen itself has lowemissions. Previous chapters discuss future options fornuclear energy to meet that need.

Even so, there is no need to wait for starting a transi-tion because low-temperature electrolysis is a proventechnology and existing types of nuclear reactors (lightwater and heavy water reactors) are commercially com-petitive and have attained an important market share ofelectrical production.

Electrolysis at distributed stations will be the produc-tion process of choice for the near-term supply of hydrogento the transport market. This will arise principally becauselocal electrolysis can minimize any distribution costs byrelying on off-peak electricity delivered over existing grids.


6. Hydrogen Economics forAutomotive Use

1. Introduction

Most hydrogen production today is produced by steam-methane reformers (SMRs), typically in world-scale plantsproducing 200 to 300 tonnes/day.a This scale of plant pro-duces enough hydrogen to fuel approximately half a millionlight vehicles. An advanced reactor of 1 GW using a directhydrogen-generation process of 50% efficiency would havecomparable capacity. In 2040, when the conversion of free-ranging vehicles to hydrogen could be nearing completion,approximately four production plants of this size could sup-ply the entire transport needs of a city of two million peo-ple through a network of distribution pipelines. But for theearly stages of conversion to hydrogen fueling, this supplypattern is excessively large and would be unconditionallyexpensive. For these early stages, distributed generation ofhydrogen by water electrolysis can provide a low-cost, flex-ible approach.

2. Overview

Water electrolysis is not a newcomer to hydrogen produc-tion. Although SMRs are almost invariably used to producehydrogen in bulk from methane (natural gas), electrolysis isthe current norm for production of small quantities. This

aTo help conversion from energy to mass, 1 tonne hydrogen = 142GJ [higher heating value (HHV)]; 1 tonne/day = 1.64 MW.



occurs because hydrogen is expensive to distribute untilthe quantities are large enough to justify a pipeline, andpipeline costs are almost scale-independent below ~0.5GW of capacity.45 As is already apparent from the first scat-tering of hydrogen filling stations around the world, elec-trolysis will be the preferred method for supply to the earlystages of transport conversion. Indeed, as this study willshow, water electrolysis is likely to be the prevalent methodof hydrogen production until the transition to hydrogen-fueled vehicles penetrates 5 to 10% of the vehicle market.To be the lead technology for full-scale adoption of hydro-gen as a vehicle fuel, it is essential that electrolytic hydro-gen be delivered at the point of vehicle supply at a pricethat is reasonably competitive with oil-based fuels. Toachieve this, low costs are important both for the capitalcost of the cells and for the electricity.

In this study, the relative costs of centralized and dis-persed electrolytic hydrogen production are compared for ahypothetical, but typical, city. Around the world, the densityof car ownership in conurbations is less variable than onemight expect (Table 1). It seems reasonable to approximatea typical city as one that is 40 km in diameter and has adensity of 1200 cars/km2 (1.5 million vehicles in all) and toassess how to supply a growing proportion of hydrogen-fueled vehicles in that context. Figure 1 gives a simplifiedlayout of this typical city. If one were to provide hydrogen-fueling service to this city through ten service stations, itwould require roughly the pattern of Fig. 1—a ring of nine

Table 1.Car Densities and City Scales for Selected Cities

ApproximateDiameter Population Cars/1000 Density

City (km) (millions) People46 (cars/km2)

Toronto 40 2.2 430 753Atlanta 35 3.5 475 1729Paris 48 11 425 2585Stockholm 27 1.9 390 1295Delhi 44 9.4 200 1237Rio de Janeiro 50 9 180 825


HYDROGEN ECONOMICS FOR AUTOMOTIVE USE 89

stations each serving a 7-km radius plus one central sta-tion.b Since 3 kg of hydrogen will meet the current expec-tation of ~400 km of driving between fuelings, these areneeded on average once a week, based on the averageannual driving distance of Canadian cars of 21 000 km.The analysis assumes that the city has reached the stageof having a total of 24 000 hydrogen-fueled cars—a level ofsignificant penetration with almost one vehicle in 60 oper-ating on hydrogen. Total demand for hydrogen is 10tonnes/day, the output of a small, but practicable SMR.Expressed in energy terms, hydrogen is being suppliedwith an energy content of 17 MW. (Note that the peak elec-trical demand of such a city would be >2 GW, so electroly-sis would add not much more than 1% to total demand onthe city’s electrical grid.)

Fig. 1. Distribution pattern for a hypothetical urban area.

b A perfect city would use a ring of six stations, but reality will beimperfect.



The scale chosen for this evaluation is arbitrary, but asmaller scale of conversion to hydrogen fuel would merelyreinforce the case for electrolytic hydrogen production.

Hydrogen can be distributed within the city either bypipeline or by tube-trailer truck.

For the pipeline option, the diagram shows a plausiblenetwork of supply pipelines from an SMR on the city’s edgeto a central station and a ring-line linking the other stations.The total length is ~107 km. Alternatively, a trucking optionmust cover ~450 km (assuming straight-line delivery andtwo-way journeys). Two trucks operating 16 h/day shouldsuffice.

3. Hydrogen Production Economics

All economic calculations are in U.S. dollars and use15%/yr return on investment and 10-yr amortization—equivalent to an annual capital charge of 20%.

3.1 ElectrolysisRecognizing the importance of low capital cost for elec-trolysis, Stuart Energy Systems (the predominant worldsupplier of electrolytic capacity) is developing a completeelectrolysis unit (including compression) that, in massproduction, will cost no more than 300 $/kW. The actualcell uses 1.8 V, but total energy usage for the system(including compression) is equivalent to 2.1 V. Beyond theminimum sales threshold needed for mass production,electrolysis is intrinsically a modular process, so its eco-nomics are almost scale-invariant. Electrolysis would bedeployed where the hydrogen is to be dispensed. The onlyadditional cost is for local storage to meet the daily cycleof demand and, preferably, to avoid production duringtimes of peak electricity demand and price.

For continuous operation, low-cost cells contribute nomore than 20% of the cost of electrolytic hydrogen. Low-costelectricity is much more important, which is why the use ofoff-peak electricity to produce hydrogen has been widelyendorsed. Note, though, that the charge for cell capital willgrow in importance if low-cost electricity has a really low



availability factor. Thus, coupling wind generation (availa-bility ~30%) with electrolysis would triple the cost contributionof cells and would also add to storage costs, especially ifstorage had to accommodate windless periods.

Where electricity prices have been freed from regula-tory control, variation in supply prices appears to be ac-centuated. In 2001, the Alberta electricity market wasderegulated, and after an initial period of extreme pricefluctuation, prices had settled down in 2002. The averagebuying price by the Alberta Power Pool (APP) in 2002 was29.3 $/MWh (converted from the Canadian price at thethen-appropriate rate of 1.5 Can$ = 1 U.S.$). (Coinciden-tally, but importantly, this average is very close to AtomicEnergy of Canada Limited’s 30 $/kWh target for itsAdvanced Candu Reactor (ACR), a Generation III+ nuclearreactor design.) This APP information was used to deter-mine the optimum strategy for electrolytic hydrogenproduction.47

Examination of the APP price details (shown in Fig. 2)revealed that electricity sold to the grid for an average of only22.4 $/MWh for all but 5% of the year. Of course, selling tothe APP and buying by a customer requires a price markup.The markup is difficult to estimate but should be constrainedboth by electrolysis being disconnected at all times of high

Fig. 2. Hourly selling prices for electricity to Alberta Power Pool.



demand—requiring transmission only off-peak—and by theexpectation that an increasing proportion of generatingcapacity will in future be suited to base-loading.Consequently, a modest distribution markup of 10 $/MWhwas used in evaluating the cost of electrolytic hydrogen.

Actual hourly data were used to cost a guaranteedsupply of hydrogen. The study varied the price level forelectricity above which electrolysis would be interrupted. Asthe threshold price is lowered, capacity for both electrolysisand storage must rise. Table 2 summarizes the outcome,showing a minimum of 2447 $/tonne H2 with a cut-off at55 $/MWh. Electrolysis capacity is 20% above that requiredfor continuous operation, and 17 h of storage capacityc isrequired—comfortably above the minimum needed toaccommodate the daily cycle of demand.

By using two price thresholds—a lower one that is nor-mally applied and a higher one to be used when stored

Table 2.Storage Requirement and Hydrogen Cost with Varying Electrolysis

Electrolysis Hours of Storage Required forRequired Electricity Threshold per Megawatt-hour of

(% ofcontinuous) 48 $ 50 $ 55 $ 60 $ 65 $

110 119 92 65 51 45115 68 52 30 24 19120 37 29 17 16 16125 32 26 16 14 14130 28 24 16 13 13135 25 23 15 12 12140 24 21 14 12 12

Total H2 Cost ($/tonne H2)

110 3278 3046 2826 2714 2671115 2848 2715 2539 2499 2464120 2597 2535 2447 2450 2460125 2576 2531 2460 2455 2463130 2561 2533 2481 2466 2475135 2554 2545 2492 2477 2486140 2575 2547 2503 2498 2506

cAssessed at 400 000 $/tonne H2.



hydrogen reserves were very low—it was shown that a min-imum production cost of ~2400 $/tonne hydrogen (17 $/GJ)could be achieved with this configuration of electrolysis andstorage and storage equivalent to an average of 17 h ofdemand (comfortably above the minimum needed to meetdaily cycles).

With electrolytic production, the layout of the city isunimportant.

3.2 Steam-Methane Reforming

Locating small SMRs at hydrogen filling stations is judged tobe uncompetitive mainly because of the high cost of a col-lection system to take the CO2 to sequestration.47 So theSMR option will rely on a centralized SMR and a distributionnetwork either by pipeline or by trucks using tube-trailers.While electrolysis has costs that are effectively independentof scale and location, with hydrogen produced by SMR,one must use an assumed scale of production and patternof distribution. Since a tube-trailer truck can deliver 1 tonneof hydrogen, the study chose an SMR producing 10tonnes/day (112 000 Nm3/day) supplying ten stations in thelayout of Fig. 1 and each dispensing 1 tonne/day.

Hydrogen-fueled light vehicles are assumed to usefuel cells and have ranges similar to today’s automobiles; atypical 40-� gasoline or diesel fillup (~28 kg) will bereplaced by 3 kg of hydrogen (since the fuel cell has threetimes the efficiency and hydrogen has three times the ener-gy density per unit of mass). So this scale leads to a fillingstation that dispenses 330 fuelings per day—quite a highnumber.

3.2.1 Production Costs

The SMRs typically convert 1 mol of natural gas into ~2.8mol of hydrogen and 1 mol of carbon dioxide (CO2).Typical capital cost for a world-scale SMR producing 2.6million Nm3/day of hydrogen (33 500 GJ/day) is approxi-mately $65 million. The SMR capital cost varies with sizeraised to approximately the 0.66 power. So an SMR pro-ducing 10 tonnes/day (1420 GJ/day) would have a capital



cost of approximately $8 million. Applying 20% as theannual capital charge, SMR capital contributes 3.1 $/GJ tothe hydrogen cost.

As with electrolysis, the cost of the energy input—inthis case natural gas—dominates the cost of hydrogen.Unlike oil, natural gas is expensive to move between conti-nents, so the price varies somewhat, being fairly typically1 $/GJ cheaper in Europe than in North America. In theNorth American context, a hub price of 5 $/GJ appears tobe becoming the accepted norm for a long-term average,held down by the cost of importing liquefied natural gasand supported by limited remaining reserves in NorthAmerica. Natural gas at 5 $/GJ was the basis for this studywith a distribution charge added. Enbridge, a majorCanadian supplier of natural gas, publishes its distributionfees for large industrial users.48 Using the 2002 conversionrate to U.S. dollars (to maintain comparability with electrol-ysis), the incremental, average distribution charge is 1.67$/GJ. The conversion efficiency of an SMR depends onhow much use can be made of by-product, low-grade heat.Assuming that this heat is of no appreciable value, 79%conversion49 is expected. So energy cost for the hydrogenis 8.4 $/GJ.

3.2.2 Cost of CO2 Sequestration

The CO2 from the SMR will have to be sequestered (or thebenefit of converting to hydrogen-fueled vehicles would belargely foregone). The cost of this, as yet unproven, tech-nology is difficult to assess and will vary with accessibilityto a suitable underground storage site. A reasonable recentestimate of the likely cost of CO2 sequestration is 135$/tonne C (30 $/tonne CO2). This is based on the U.K. gov-ernment’s Department of Trade and Industry positionpaper,50 which uses £84/tonne C for the typical cost for CO2sequestration in a deep geological aquifer 300 km from thesource. This could be an underestimate since an SMR pro-duces CO2 in two very different streams: ~70% as a fairlypure stream separated from the hydrogen and ~30% as aflue-gas stream in which CO2 is <20% of the total volume.No cost has been estimated for processing the latter to



extract the CO2. One tonne of hydrogen by SMR produces7.75 tonnes of CO2 or 232 $/tonne H2. This contributes 1.6$/GJ to the cost of hydrogen.

3.2.3 Distribution Costs: A Case Study for aHypothetical City

Hydrogen can be distributed within the city either bypipeline or by tube-trailer truck.

For the trucking option, trucks must cover ~450 kmeach day (assuming straight-line delivery and two-way jour-neys). Two trucks operating 12 h/day should suffice.Alternatively, the pipeline option of Fig. 1 requires a totallength of ~107 km.

Padró and Putsche46 have reviewed the costs ofdistribution by pipeline or truck. They quote data from Oneyet al.,51 and these are summarized in Table 3. For consis-tency with the rest of this study, Oney’s figures have beenconverted to the HHV. (These cost estimates may under-estimate pipeline costs since Amos52 gives pipeline coststhat are ~50% higher.)

Based on Oney et al.’s figure of 1.71 $/GJ for trans-mission of 0.154 GW of hydrogen energy over 161 km,Padró and Putsche concluded that pipelines give lowercosts. But 0.154 GW (4.1 million GJ/yr) is a quite prodi-gious scale, sufficient to fuel 180 000 cars or almost tenmillion refuelings annually. Further, Padró and Putsche’sfigures show that the cost of pipelines is already becomingnearly independent of size below 0.5 GW.d For the smallestscale of distribution, the relationship of cost and distance is

dSince the cost of pipeline distribution systems will be dominated bythe cost of the small-scale laterals, the filling-station pattern ofvehicle fueling may endure at least until hydrogen widely penetratesthe energy market. For filling stations, economies of scale obvious-ly improve the economics of pipeline distribution, but there are toomany unknowns to predict whether that will ever become the norm.Yet finer distribution would seem to require extension of hydrogenas a fuel beyond transportation. This is certainly possible; hydrogenwas the major constituent of the town gas that preceded wide-spread use of natural gas.



linear with an intercept of 0.27 $/GJ, so the total cost for a107-km pipeline network of 0.154-GW size can be calcu-lated as

Since the required delivery is only 1270 GJ/day(0.0147 GW), the unit cost for pipeline delivery is 12.9 $/GJ.

The truck transport alternative would likely operatewith compressed gas, although hydride storage has alsobeen proposed. Amos gives the cost of compressed gastransport by tube-trailer in the 4.2 to 9.3 $/GJ (HHV) rangefor distances of 16 and 160 km, respectively. BecauseAmos used a tube-trailer cost of $140 000 rather than theactual current cost of $400 000, it seems reasonable toraise the cost by 0.2 $/GJ. So, depending on the distributiondistance, the delivery charge ranges from 4.4 to 9.5 $/GJfor 16 and 160 km, respectively. By simple proportion, forthe average trucking distance (one way) of 22 km from a

$>GJ � 5 � 106 GJ>yr � 6.13 M$>yr.

a0.27 �107161

� 11.71 � 0.272b

Table 3.Pipeline Costs from Ref. 51 and Adjusted to HHV*

Transmission Rate Distance Cost

(GW) (GJ/day) (tonne/day) (km) ($/GJ) ($/km⋅tonne)

0.154 13 300 79 161 1.71 1.44805 7.49 6.33

1609 14.71 12.43

0.497 42 900 255 161 0.70 0.59805 2.43 2.05

1609 4.60 3.89

0.994 85 900 509 161 0.70 0.59805 2.43 2.05

1609 4.60 3.89

1.49 128 800 764 161 0.70 0.59805 2.43 2.05

1609 4.60 3.89

*All figures revised to the HHV for hydrogen.



single SMR, the cost should be 4.6 $/GJ. Trucking appearsto be the better choice on this scale than a pipeline.

However, note that the pipeline option could carrymore than an order of magnitude greater hydrogen flowsfor no additional cost and so an inverse proportionate fall inunit cost. As demand for hydrogen grows, pipelines willbecome increasingly competitive.

SMRs do not load-follow over large ranges, so storagewill have to be provided to handle the daily cycle ofdemand. For the pipeline option, 12 h of capacity has beenassumed at $200 000, adding 0.07 $/GJ. The truck optionneeds rather more than double that capacity since deliveryis once per day and a full truckload has to be available atthe SMR.

4. Cost Comparisons

4.1 Cost Comparisons with the Principal Scenarios

The costs of the filling station options can now be summa-rized in Table 4.

Of course, water electrolysis only avoids CO2 emission(and the cost of sequestration) where the new use of elec-tricity for hydrogen production is supplied with essentiallyzero CO2 emission. Since high capacity is also needed, theobvious source is nuclear power.

For production of hydrogen on this scale, electrolysisusing time-of-day pricing appears to give the lowest costs.Electrolysis operated continuously is more expensive, andthe margin is almost certainly greater than shown since theargument for a low markup charge for distribution of inter-ruptible electricity does not apply. For the remote SMR,hydrogen distribution by truck is decisively cheaper thandistribution by pipeline. It is also, obviously, more easilyadapted to supplying new stations.

As already noted, the pipeline distributing hydrogen isgreatly oversized, and the case summarized in Table 4 pro-vides fueling for only 1 in 60 light vehicles. As demand for



Tabl

e 4.

Com

paris

on o

fth

e O

ptio

ns f

or H

ydro

gen

Sup

ply

Rem

ote

Rem

ote

Lo

cal

Lo

cal

SM

R w

ith

SM

R w

ith

Ele

ctro

lysi

sE

lect

roly

sis

Pip

elin

eTr

uck

Usi

ng

Off

-O

per

atin

gD

istr

ibu

tio

nD

istr

ibu

tio

nP

eak

Po

wer

Co

nti

nu

ou

sly

Uni

t pr

oduc

tion

size

(to

nne

H2/

day)

1010

11

SM

R o

r el

ectr

olys

is c

apita

l cos

t (M

$)O

ne a

t 8

One

at

8Te

n a

t 0.

844

Ten

at

0.70

3S

tora

ge c

apita

l (M

$)Te

n a

t 0.

2E

leve

n a

t 0.

4Te

n a

t 0.

28Te

n a

t 0.

2P

rodu

ctio

n an

d st

orag

e ca

pita

l (M

$)10

12.4

11.2

9.0

Cap

ital c

harg

e fo

r pr

oduc

tion

+ s

tora

ge (

M$/

yr)

2.0

2.5

2.2

1.8

Cap

ital c

harg

e ($

/GJ)

3.9

4.6

4.2

3.5

Ene

rgy

cost

($/

GJ)

8.4

8.4

12.8

>15

.6D

istr

ibut

ion

cost

($/

GJ)

12.9

4.6

00

Car

bon

char

ge (

$/G

J)1.

61.

60

0

Tota

l ($/

GJ)

26.8

19.4

17.0

>19

.1

Tota

l ($/

tonn

e H

2)38

0027

5324

14>

2712



hydrogen expands, the economics of pipeline distributionimprove relative to the other options. Table 5 explores thiseffect.

The main effect on costs in Table 5 stems from largerflows through the existing pipeline network. However, thenetwork would also have to be expanded somewhat to sup-ply the additional stations, and the costing assumes anadditional 3.5 km for each station added—a reasonable pro-vision in the context of Fig. 1. The costs apply to initialinstallation of a larger SMR rather than adding small SMRsas the system expands. This is debatable but a fairly smalleffect, responsible for saving at most 1.3 $/GJ. Savings onthe SMR capital are the only possible benefit to the truckingoption, and it would only achieve minor benefit. Local elec-trolysis remains completely uninfluenced by scale.

The obvious inference from Table 5 is that centralizedproduction and pipeline distribution within a city becomescompetitive at approximately five times the scale used forthe basic comparison in Table 4. In other words, when con-version reaches the equivalent of approximately 1 lightvehicle in 12, centralized hydrogen production and pipelinedistribution emerge as competitive.

Table 5.Effect of Larger Scale on SMR Supply Options

Number of filling stations 10 20 50 25 SMR production size (tonne/day) 10 20 50 50Pipeline length (km) 107 142 247 177SMR or electrolysis capital cost One at One at One at One at

(M$) 8 12.6 23.1 23.1Storage configuration and capital 10 at 20 at 50 at 50 at

(M$) 0.2 0.2 0.2 0.2Production and storage capital (M$) 10 16.6 33.1 33.1Capital charge for production + 2.0 3.3 6.6 6.6

storage (M$/yr)

Capital charge ($/GJ) 3.9 3.2 2.6 2.6Energy cost ($/GJ) 8.4 8.4 8.4 8.4Distribution cost ($/GJ) 12.9 7.6 5.3 3.8Carbon charge ($/GJ) 1.6 1.6 1.6 1.6

Total ($/GJ) 26.8 20.8 17.8 16.3

Total ($/tonne H2) 3800 2955 2534 2321



4.2 Comparison with Existing Vehicle Technology

But, one must ask, what prospect has hydrogen of oustingoil-based fuels? While one can simply argue that hydrogenis the only choice—barring an unforeseen breakthrough inelectricity storage—because CO2 emissions from vehiclesare far too large to be acceptable, the economics for ahydrogen economy developed in this chapter also lookfavorable.

Today’s typical 21 000 km/yr automobile uses ~2400 �of gasoline costing close to $1000/yr plus $700/yr for col-lateral CO2 emissions (5.2 tonnes CO2/yr). However, thecompetition by 2020 should probably be highly efficientdiesel-electric hybrids, and their fuel cost could be lowerby a factor of 3 for a fuel cost of $330/yr plus $240/yr forcollateral carbon emission (1.75 tonne CO2/yr). If poweredby a fuel cell, the same automobile is estimated to need161 kg/yr of hydrogen. On that basis, the hybrid and thehydrogen-fuel cell break even if the base cost of hydrogenis just over 3500 $/tonne or 24.65 $/GJ. And provided theenergy supply to the production process does not emitCO2, the hydrogen route is pollution-free. The estimatedcost of producing hydrogen in Table 4 strongly suggeststhat hydrogen will not require sustained government sup-port to displace oil-based fuels. It should be possible todevise a fiscally neutral tax regime that favors but does notnecessarily exempt hydrogen.

4.3 Cost SensitivitiesEnergy costs are the most sensitive inputs for both SMRand electrolytic production. An increase in natural gas priceof 1 $/GJ increases the hydrogen production cost by 1.3$/GJ. An increase in electricity cost of only 2.3 $/MWh hasthe same effect.

If, as suggested earlier may be the case, the cost ofpipelines is underestimated by one-third, cost attributed topipeline distribution for the base case would rise by 6.5 $/GJ.

One opportunity for lowering the cost of the electrolyt-ic approach is to realize value for the oxygen coproduct. Themarket in the United States for oxygen exceeds53 23 million



tonnes/yr at ~27.5 $/tonne. For each tonne of hydrogen, 8tonnes of oxygen is produced. For the base case, 80 tonnesof oxygen is produced each day. So, assuming 100 suchcities (an order-of-magnitude approximation for the UnitedStates) would produce 3 million tonnes/yr suggests that theoxygen market could likely absorb by-product oxygen com-mercially, giving a potential credit of 220 $/tonne H2 or 1.55$/GJ less costs for compression.

4.4 Larger-Scale Production after 2020

Considering the radical nature of the shift in technology,even the envisaged ~1.7% conversion of light vehicles tohydrogen fuel is a substantial undertaking that may rea-sonably be projected to be achieved around 2020.However, this would only be a stepping stone to much larg-er scales of hydrogen use beyond 2020. By then, theadvanced thermochemical processes for hydrogen produc-tion now entering development could be contenders.Costing for these is particularly speculative, but even thefuture capital costs of the conventional processes must beuncertain so far into the future, perhaps excluding SMRtechnology, which could be considered mature technology.The cost of conventional water electrolysis is projected tofall to close to 150 $/kW with economies of scale and expe-rience. Given the uncertainty over capital costs, we haveattempted a limited assessment of the technologies basedon noncapital costs.

Either a world-scale SMR or a nuclear reactor of ~0.85GWe (converting electricity to hydrogen with 70% efficiency)would produce ~350 tonnes/day of hydrogen and so wouldmeet the hydrogen needs of the hypothetical city when con-version of its transport (light and heavy) reaches 30% of themarket. It is this scenario that is in Sec. 4.4.1.

eTechnically, whether the reactor produces hydrogen thermochemi-cally or produces electricity for use in electrolysis or uses somecombination of heat and electricity is less important than one mayexpect. At the high temperatures likely to be required for indirectwater dissociation, conversion to electricity becomes almost as effi-cient as the efficiency expected of thermochemical processes.


4.4.1 Comparison of Operating Costs

On the 350 tonnes/day scale, pipeline delivery would be farsuperior to truck delivery, and the cost of distribution wouldbe far less important than for the earlier case. Filling sta-tions would now be only ~3 km apart, and the pipelines fora main network, originally built to minimum cost but withgreat overcapacity, would no longer be oversized. Only thefinest part of the network, the final connection laterals,would be small enough for their cost to have become inde-pendent of the flows handled, but these would be short. Ifone envisages repeated clusters within the city structure ofFig. 1, a reasonable pattern of distribution could use theoriginal layout and add a final, finer (built at the 0.13-GWminimum size) network superimposed. This finer networkwould typically carry gas over only an average of ~3 km tothe individual stations. Final laterals would not need anybooster compression.

From Table 3, based on 0.42 GW (handling half theflow), it seems reasonable to use 0.70 $/GJ for the cost ofthe main grid plus 0.30 $/GJ for the final lateral (in theabsence of any need for further compression, the final 3 kmis simply prorated to 3 km from the costs in Table 3). Thetotal of 1.0 $/GJ is equivalent to 142 $/tonne H2.

While advanced processes may now be contenders,we consider here only the likely position of a centralizedSMR and centralized water electrolysis. The SMR clearlyenjoys economies of scale, but electrolysis can now usethe cost of electricity without markup for electrical distribu-tion, and savings could be achieved both in capital cost andpower consumption by having the reactor produce dedi-cated direct-current (dc) power. As well as using 150 $/kWfor the cell cost, it would be reasonable to lower the cost ofgeneration to 29 $/MWh (for dc power) and the voltageequivalence to 1.9 V to recognize the effects of higher-pressure electrolysis and elimination of alternating-currentrectification.

All of the processes under consideration use inputs ofchemical, electrical, or high-temperature thermal energy or



combinations of these. As before, since the most versatileform of these three forms is electricity, they can be com-pared by converting the energy consumption of all of theprocesses to their electrical equivalents using appropriateconversion efficiencies. Natural gas’s chemical energy isassumed to convert to electricity at 60% efficiency. Heat at850°C is assumed to convert to electricity with 55% effi-ciency. The value used for electricity is 3 ¢/kWh and isbased on the expected generating cost from new reactors,such as the ACR.

High-temperature electrolysis has been attributed anelectrical usage of 1.4 V and the SI and UT-3 processesassumed to use heat with 50% efficiency. Heat and chem-ical energy for the nuclear-assisted SMR is distributed inthe usual proportion for SMRs. The analysis (Table 6) con-siders energy, distribution, and sequestration costs for this0.70-GW scale.

In an electricity system with a much larger componentof nuclear, the effect of time-of-day pricing would beexpected to become even more pronounced than the 24%cost reduction for power for 95% of the time shown inAlberta in 2002. Only the conventional distributed electrol-ysis can take advantage of that.

Assessed for energy use, distribution, and sequestra-tion, Table 6 shows all processes within ~20% of 1400$/tonne. On this scale and provided CO2 can be effectivelyand economically sequestered, the SMR routes look prom-ising. Electrolysis can probably only compete if it can exploitits interruptibility and obtain low-cost off-peak power.Nuclear heat appears best deployed as a supplementarysource of energy to SMRs, though we note that it uses a sig-nificantly smaller nuclear source—perhaps not a real disad-vantage since large centralized units will have to be locatedon a pipeline grid for security of supply. Conventional elec-trolysis using electricity at off-peak prices is comparablypriced. The competitive positions of SI and UT-3 thermo-chemical processes and of high-temperature electrolysissuffer from far less efficient use of energy than SMRs.




Tabl

e 6.

Com

paris

on o

fP

roce

sses

bey

ond

2020

Ele

ctro

lysi

sw

ith

Act

ual

H2

Tim

e-o

f-D

ayS

MR

+

SI

or

Hig

h-t

emp

Pro

cess

En

erg

yS

MR

Ele

ctro

lysi

sP

rici

ng

Nu

clea

rU

T-3

Ele

ctro

lysi

s

Ele

ctric

al e

nerg

y (G

J/to

nne

H2)

203a

203

135

Che

mic

al e

nerg

y (G

J/to

nne

H2)

142

156

111

Hea

t en

ergy

(G

J/to

nne

H2)

4528

485

Ele

ctric

al e

quiv

alen

t (G

J/to

nne

H2)

9420

320

391

156

182

Ele

ctric

al e

quiv

alen

t (M

Wh/

t H

2)26

5656

2543

51C

O2

emis

sion

(to

nne

C/to

nne

H2)

2.11

1.50

Ele

ctric

ity (

$/to

nne

H2)

780

1692

1284

761

1302

1515

Dis

trib

utio

n ($

/tonn

e H

2)14

20b

0b14

214

214

2S

eque

stra

tion

($/to

nne

H2)

285

00

203

00

Tota

ls (

$/to

nne

H2)

1307

1692

1284

1106

1444

1657

a Usi

ng t

otal

ele

ctric

ity in

put

equi

vale

nt t

o 2.

1 V.

b Ele

ctric

ity is

like

ly t

o ha

ve a

dis

trib

utio

n co

st a

s w

ell,

but

it is

har

d to

est

ima

te.

Ifit

is u

tiliz

ed o

nly

whe

n th

ere

is f

avor

abl

e tim

e-of

-day

pric

ing,

itco

uld

be lo

w.

Cos

t of

inpu

t en

ergy

’s d

istr

ibut

ion

also

has

to

be b

orne

by

proc

esse

s us

ing

natu

ral g

as a

s fu

el.


Table 6 is very speculative. Though use of electricity asan energy currency eliminates the distortion introduced byrelative fluctuations in the price of different energy sources,it does require judgment on an appropriate cost for electric-ity. While it is arguable that this should be increased toinclude distribution costs, so long as electrolysis does notuse power at periods of peak load, the incremental cost ofdistribution could be quite small. The competitive position ofconventional electrolysis is highly dependent on low-costelectricity, but in an increasingly nuclear system, hydrogenfrom electrolysis has useful load-leveling attributes thatcould fully offset the cost of distribution. It also seems prob-able that massive experience with electrolysis in the earlystages of hydrogen deployment—when we have shown thatit is likely to be at a cost advantage—will have lowered theoperating voltage.

Storage costs have not been considered here. Locallyproduced hydrogen will obviously need storage but so willthe larger systems based on centralized production, andsince the distribution system has largely been sized to itsaverage continuous capacity, local storage will be neededor the main pipeline system enlarged to accommodate thediurnal variation in demand.

5. Scale of Hydrogen Supply

In Canada by 2030, total energy demand for all forms oftransportation is projected to have increased by 30% fromtoday’s figure to reach ~2900 PJ/yr. Allowing the superiorefficiency of fuel cells, this would require ~1000 PJ/yrexpressed as hydrogen energy or 48 GWe. So if the hydro-gen were manufactured centrally, it would require ~23reactors or world-scale SMRs to supply their fuel. In thiscontext, distribution distances of 100 km are realistic.

6. Conclusions

Distributed, low-temperature electrolysis will be the produc-tion process of choice for the initial supply of hydrogen to thetransport market provided that an intermediate generation



of nuclear reactors comes on-line between 2010 and 2020and delivers electricity close to the target cost of 3 US¢/kWh.This will arise principally because local electrolysis can min-imize any distribution costs by relying on off-peak electricitydelivered over existing grids.

Consequently, a network of hydrogen fueling for vehi-cles ought to be well launched by the time that newprocesses based on high-temperature nuclear reactors areavailable around 2020. Distributed production is a nearlyessential precursor to distribution networks that could uti-lize larger-scale production processes.

Whether competitive processes based on high-tem-perature nuclear heat will subsequently take over willdepend on their capital costs and on what has happened tocapital and operating costs for conventional electrolysisduring its window of early opportunity. For large-scale pro-duction, suitable for a maturing hydrogen market, SMRsalso will be strong competitors if both the cost of securesequestration is in line with today’s projected estimates andsecure supplies of natural gas are available.



7. A SUSTAINABLENUCLEAR FISSION–

BASED HYDROGEN ENERGYSUPPLY ARCHITECTURE

Nuclear energy, if its supply architecture is properlyconfigured, could drive a global hydrogen economy.

An ordered sequence of energy carriers is proposed:nuclear fuel shipped from regional fuel cycle centers to thenuclear power plants, hydrogen and water piped from thenuclear plants to the district load centers, and electricitywired from distributed production centers to end use wouldbe organized sequentially in the order of their energy den-sity and their associated power-carrying capacity throughpractical-sized conducts (e.g., ships/trains, pipelines/trucks,wires, respectively).

Over time such an architecture could achieve marketpenetration and evolve in response to economy-of-scaledrivers for central production of hydrogen at the regionalcenters vis-à-vis the energy security drivers for distributedhydrogen production.


7. A Sustainable NuclearFission–Based Hydrogen

Energy Supply Architecture

1. Introduction

Nuclear energy’s current configuration has left many of itsinnate features unexploited: specifically, its economicallyharvestable resource base good for a millennium of worldenergy supply by closing the fuel cycle, its capacity to openitself to the entire primary energy market by manufacturinghydrogen, and its capacity to break the energy security/non-proliferation dilemma by exploiting its incredible energydensity to facilitate deployment of long-refueling-intervalreactors supported by regional fuel cycle centers.

The objective here is to propose a reengineered worldenergy supply architecture optimized for nuclear ratherthan fossil. It is intended for global energy supply in themid–21st century market conditions, where 80% of theworld’s population of ~10 billion people reside in cities,where electricity and hydrogen serve as complementaryenergy carriers, where a closed fuel cycle fully exploits theearth’s uranium resource, and where ecologically neutralclosure of the energy supply chain is attained by eliminat-ing carbon and by sending only fission products to waste.

Although it is revolutionary in concept, it is aimed togradually displace the fossil architecture and the currentnuclear infrastructure by means of a many-decades evolu-tionary market penetration process.



2. An Ordered Sequence of Energy Carriers Basedon Their Power-Carrying Capacity

The proposed nuclear-based energy supply architectureresponds to foreseen midcentury needs and market con-ditions within the constraints imposed by the principles ofsustainable development. It is targeted for support ofurban centers in both developing and developed countries.It uses nuclear fuel and hydrogen as long-distance energycarriers and distributed electricity generation at the citydistrict level to mesh with existing-style urban energy dis-tribution infrastructures based on grid delivery of electrici-ty, hydrogen, potable water, and communications througha common grid of easements. These attributes for a worldenergy supply architecture fit naturally into a hierarchicalhub/spoke architecture.

Regional fuel cycle centers operating under the con-trol of the client nations—but with international nonprolifer-ation oversight—supply front- and back-end services tothousands of long-refueling-interval battery heat sourcereactors deployed throughout the region. The long-refuel-ing-interval battery reactor plantsa are sited near cities toprovide process heat for the manufacture of hydrogen andoxygen in a thermochemical water cracking plant and man-ufacture of potable water from a desalination bottomingcycle.

The hydrogen and water will be piped or trucked to citydistricts through a grid of distribution conduits. At district-level distribution hubs, the hydrogen will be partitioned tomeet energy service needs:

1. A third will be dispensed for hydrogen-fueled trans-portation services.

2. A third will be distributed by pipe throughout the dis-trict for heating homes, apartments, offices, and factories.

aThe reactors are referred to as “batteries” because they store 20 yrworth of heat and they load follow by passive means—deliveringheat when it is requested by the balance of plant and passivelyshutting off when the request stops.


WORLD ENERGY SUPPLY ARCHITECTURE 111

3. A third will be converted in fuel cells and/or micro-turbines to electricity for distribution throughout the district.

The electricity will in turn be distributed by wire to enduses varying from electric motors for motive force, toamperage for light generation, or to microamperages forinformation transmittal and storage.

The energy carriers—nuclear fuel shipped from theregional centers to the battery nuclear power plants,hydrogen and water piped from the battery nuclear plantsto the district load centers, and electricity wired from dis-tributed production center to end use—are organizedsequentially in the order of their energy density and theirassociated power-carrying capacity through practical-sizedconducts (e.g., ships/trains, pipelines/trucks, and wires,respectively).

Figure 1 illustrates this hierarchical energy deliveryinfrastructure at an abstract level. The hubs represent

Fig. 1. Hierarchical hub/spoke energy architecture.



where one energy carrier (nuclear fuel, hydrogen, electrici-ty) is converted into the successive energy carrier along thesupply chain—a carrier better suited to the required func-tion. The spokes represent the transmission channels ofthe energy carrier from its source point to its point of use.The widths of the spokes in Fig. 1 suggest the power-carrying capacity of practical conduits for each energy car-rier; the fractal-type expansion of the architecture as itprogresses from the uranium ore energy resource to thepoint of energy service end use reflects the diminishingenergy-carrying capacity and corresponding multiplicity ofcarrier conduits in the sequence of energy carriers.

For example, a 2-week voyage to deliver a singlewhole core refueling cassette good for 20 years (at a capac-ity factor of 0.9) in a 400-MWth battery reactor representsa 188-GWth power transmission conduit. A single ship car-rying ten cassettes on an itinerant 1-month delivery voyagecould supply nearly 1000 GWth (1 TW·yr/yr) to its serviceregion. A fleet of ten ships could provide 10 TW·yr/yr(for perspective, the world’s entire primary energy use cur-rently is ~16 TW·yr/yr).

Marchetti54 has observed that the economical scaleof equipment sited at the hubs will expand to match theenergy demand in the geographical area circumscribed bythe spokes. The practical lengths of the spokes depend onthe energy density of the energy carrier and the cost andloss-rate per mile of transport. Because of the enormousenergy density of nuclear fuel contained in the refuelingcassettes, the reach of the nuclear supply spokes throughpractical-sized transport conduits can be thousands ofmiles, and as a result, the fuel cycle facilities at the region-al centers can be sized for economy of scale to service thevery large demand arising from a significant global region.Assume a midcentury world of ten billion souls consumingenergy at an average of 4 toe/capita/yr (typical of primaryenergy use in Europe):

� 53.5 TW # ˛yr>yr,

10 # ˛109 people � 4

toeperson yr

� 12.04 kWthyr

9 toe



and suppose the nuclear battery deployments could pro-vide one-fifth of this. Then the nuclear architecture wouldbe required to produce ~10 TWth·yr/yr.

A spent cassette will have delivered 400 MWth � 365days/yr � 20 yr � 0.95 = 2 774 000 MWth days of energy atdischarge. At 100 MWth days/kg average discharge burnup,it will contain 27.74 tonnes of heavy metal to be recycled.

A regional fuel cycle center designed for 4000 tonnesheavy metal/yr throughput capacity (approximately twiceLa Hague) could recycle 4000/27.74 = 144 cassettes/yr. Ata cassette refueling interval of 20 yr, the single fuel cyclecenter could service 20 � 144 = 2880 regional battery reac-tors or 2880 � 400 MWth = 1.15 TWth·yr/yr. By 2050 itwould be necessary to have built ten regional fuel cyclecenters of 4000 tonnes/yr throughput to service 20% ofestimated total world primary energy.

Recycling facilities benefit from economy of scale andhave historically been built in 400 to 2000 tonnes/yrthroughput. Pyrometallurgical recycle technology is innate-ly modular and could be deployed in even 50 tonnes/yrscale. Therefore, even if providing the world’s entiredemand (~50 TWth·yr/yr) by midcentury, no more than adozen such fuel cycle centers using economy-of-scalefacilities could meet the world’s needs. In that sense, theycould be viewed as the 21st century analogue to the oilfields of the 20th century.

The astounding energy density of nuclear fuel makesit a suitable energy carrier for an eminently scaleablenuclear-based energy supply architecture—facilitatingaffordable and manageable multidecade growth of deploy-ment starting from incremental market penetration.

The reach of the next link in the supply chain—thehydrogen pipeline spokes—would reflect their several-gigawatts carrying capacity55 and would service regions ofseveral 100-mile dimension through pipeline grids such asare currently used to distribute natural gas to load centers.Pipeline grids would carry hydrogen from the battery reac-tor plants to district centers scattered throughout the city



and its surrounding population region. At a primary energyuse rate of 4 toe/capita/yr (i.e., ~5.5 kWth day/person-day)a 5.5-GW hydrogen pipeline could service a city and itsenvirons with a population of a million people.

The reach of the electricity distribution wires startingat district microturbine or fuel cell converters of hydrogen toelectricity and taking the electricity to final use in lighting,motors, and information management would be of thescale of city districts and skyscrapers—as is the currentusage. This last stage of distribution would use the existingelectrical and water distribution network (where it alreadyexists) and would thereby make the conversion to the newenergy architecture nearly transparent to the end user ofenergy services.

By midcentury, district-level conversion of hydrogen toelectricity—as opposed to conversion of heat to electricity atthe STAR reactor sited at the city perimeter—is envisionedfor several reasons. The first—and the one which is alreadydriving a transition—is supply reliability. Microturbines and(imminently) fuel cells can provide secure electricity at a dis-trict level, even if the broader grid suffers a shutdown,because they run on a storable supply—currently naturalgas, but eventually hydrogen. Some planners believe thatdistributed generators will, in fact, eventually drive the grid.The second driver is that the hot water produced as thewaste from conversion of hydrogen to electricity at districthubs can be used in support of the city’s hot water needs.This will increase billable product for the owner of the con-version equipment, but more importantly, it will reduce thewater vapor and thermal plume ecological footprint of theconversion step. This sets the scale of electricity productionat a district level because of the limited reach of hot waterdistribution spokes.

Carrier conduit crossconnections of the user hubs tomultiple supplier hubs and energy storage buffers providedby the storable nuclear fuel and hydrogen energy carrierswould provide for robustness of energy security and forprotection against monopolistic pricing.



Hydrogen would serve all primary energy markets(including transportation) that are currently serviced by fos-sil fuels. Over time, the hydrogen would gradually displaceoil, gas, and coal, and the new carbon-free, nuclear-basedarchitecture would gradually replace the current fossil-based world energy supply infrastructure. The transforma-tion has already started at the district level; microturbineemplacement (driven by natural gas) is occurring as back-up electrical supplies for factories and skyscrapers.Stationary fuel cells for offgrid locations (driven by bottledgas) are offered by utilities to selected clients. And hun-dreds of millions of dollars of private investment is goinginto fuel cell development for the transportation sector. Theproposed new architecture will mesh seamlessly with thesetransitions already starting at the user level.

3. Battery Heat Source Reactor Plants for Hydrogenand Water Production Sited Near Cities

Modular-sized long-refueling-interval battery heat sourcereactors sited near cities to manufacture hydrogen (anddesalinated water) are proposed. Such fast neutron spec-trum reactors are currently being designed at a concep-tual level (see Refs. 56, 57, and 58). They operate atpower densities typical of light water reactors (LWRs)—with liquid metal natural-circulation cooling and with pas-sive load following and passive safety response character-istics. A 400-MWth sizing retains natural-circulation capa-bility in a rail-shippable reactor vessel size and facilitatespassive decay heat removal. The fast neutron spectrumdesign with internal conversion ratio of unity yields zeroburnup reactivity loss over a full 20-yr burnup interval sothat minimal reactivity is vested in control rods, which isthe key to enable passive load following/passive safety.Passive safety response is an essential prerequisite to sit-ing tens of thousands of battery reactors near cities world-wide. Since passive safety/passive load follow enables useof a balance of plant having no safety function, this allowsfor indigenous construction and operation of the balanceof plant.



The reactor heat source supplies process heat at800°C core outlet temperature to a balance of plantthrough an intermediate low-pressure fused salt loop anddrives a thermochemical water cracking plant59 for hydro-gen and oxygen production. Process heat for other usescan be supplied using heat rejected from the thermochem-ical cycle at relatively high temperature (600°C). Alter-nately, it can drive a supercritical CO2 Brayton cycle60 tosupply on-site electricity requirements.

A desalinization bottoming cycle is used for productionof potable water for feedstock to hydrogen production andfor potable water sales.61 Alternate bottoming cycles canbe used at landlocked sites.62 Such bottoming cycles areintended both to provide energy services and to minimizethe ecological (thermal plume) footprint of the power plant.

Figure 2 illustrates a battery heat source plant con-cept. Converting heat to hydrogen at 45% efficiency andusing the reject heat to produce potable water, one 400-MWth plant can meet all primary energy and all nonagri-cultural water needs of a city of 25 000 people.

4. Ecologically Neutral Closure of the EnergySupply Chain

The application of nuclear heat to produce hydrogen as areplacement for fossil fuels achieves an essentially green-house-gas-free energy supply chain extending fromresource to end use, and it allows nuclear fission energy tomove beyond electricity to service all sectors of primaryenergy usage.

Processes convert one energy carrier to another atthe hubs (nuclear fuel to hydrogen and hydrogen to elec-tricity) or to energy services (nuclear heat to potable water,electricity to motive force, etc.). These conversion process-es generate wastes. The proposed architecture providesfor an ecologically neutral closure of the entire energy sup-ply enterprise through recycle of these wastes. Closure isobtained on electricity production and use by electronreturn through ground. Closure is obtained on thermo-



Fig

.2.

Ba

ttery

hea

t so

urce

pow

er p

lant

.



chemical water cracking hydrogen production and its use infuel cells or microturbines by nature’s oxygen and watercycles. Potable water closure could be obtained by pipelinereturn of sewage from city districts to the battery powerplants where the O2 by-product of hydrogen productionwould be put to productive use for sewage treatment.Closure of the nuclear fuel cycle would be obtained bydesigning the battery reactors for fissile self-sufficient oper-ation and transuranic recovery and recycle at the regionalfuel cycle centers. This is represented by the return linesalong the spokes in Fig. 1.

In an ideal sense, the net ecological effect of theentire energy supply chain would be the consumption ofuranium ore and the creation of waste fission products,whereas everything else in all links of the energy supplychain would be recycled. The known-plus-speculative eco-nomically recoverable ore of ~15 million tonnes of uranium,when fully fissioned, could supply the world’s energy needsfor a millennium. The envisioned hierarchical hub/spokenuclear-based global energy architecture would be ecolog-ically neutral and fully sustainable in both its resource avail-ability and waste management aspects.

5. Managing the Transition and Fueling a Millenium of Sustainable Development via Symbiotic Fuel Cycles

The emplacement of a dozen or less regional fuel cyclecenters will provide the mechanism to manage the world’sinventories of fissile material and the nuclear waste pro-duced throughout a transition involving coexistence of asymbiotic mix of reactor types serving different market seg-ments. Clearly, open-cycle LWRs will maintain a growingand significant global nuclear market share for decades.Working inventories for new battery deployments will initial-ly come from reprocessing LWR spent fuel. This symbioticfuel cycle allows for incremental market penetration of bat-teries by beneficially managing the waste from the current



once-through LWR cycle (such that only fission products goto a repository) while simultaneously beneficially providingfissile transuranic feedstock for initial working inventories ofbattery deployments. Since the distributed battery heatsource reactors are fissile self-sufficient, once started up,their refueling cassette refurbishment requires only 238Ufeedstock, and each battery and its replacements wouldmaintain a steady energy supply for many centuries whilefed only by 238U. However, for fueling a growing deploymentof batteries, the LWR source of feedstock will not providefor more than several decades of growth until the econom-ically recoverable uranium ore reserves driving the LWRopen cycle are depleted. Sustained growth will ultimatelyrequire the presence of fast breeder reactors. In the archi-tecture proposed here, fast breeder reactors will ultimatelybe sited at the regional fuel cycle centers to manufactureexcess fissile material to fuel new deployments in a grow-ing economy after the source of fissile from LWR spent fuelis exhausted. The heat from their operation will be convert-ed to hydrogen for shipment to regional consumers.

At the end of the transition period, a sustainable,growing, fissile self-generating nuclear energy architecturewill be driven by the world’s 238U resource base. The ener-gy shares of battery plants versus breeders in the enter-prise will satisfy a simple fissile balance equation whereinthe fissile generation rate integrated over all reactors in thecomplex must be sufficient to refuel the existing fleet andalso to produce an excess sufficient to provide the workinginventories for the next round of new deployments.

This fissile balance requirement will mean that in agrowing economy, the energy production fraction in distrib-uted batteries will be larger vis-a-vis the breeders sited atregional centers:

1. the slower is the energy demand growth rate

2. the shorter is the breeder doubling time

3. the smaller is the battery fissile working inventorycompared to the breeder working inventory.



6. Centralized Hydrogen Production at Fuel CycleCenters––Marchetti’s Canton Islands

Unless the world economic growth rate stagnates, a signif-icant fraction of the fission heat released in the growing fis-sile self-generating nuclear enterprise will be released atthe regional fuel cycle centers by the breeders. This heatwill be used to generate the next best (carbon-free) long-distance energy carrier after nuclear fuel, i.e., hydrogen.

The hydrogen will be shipped from the regional fuelcycle centers to users—much like oil is currently shippedfrom oil fields to users worldwide. As virgin uranium oreavailability diminishes in the decades ahead, the regionalfuel cycle centers will partially evolve into the CantonIslands nuclear architecture63 proposed by Marchetti in the1970s, wherein breeders manufacture hydrogen at fuelcycle center locations for shipment to regional consumers.This may be a less secure energy supply route than is theone based on distributed battery power plants because ofthe vastly different energy densities of nuclear fuel com-pared with hydrogen as an energy carrier:

1. 1 kg H2 after burning has released 1.39·10-3MWthdays

2. 1 kg of nuclear fuel has released 100 MWth days.

The nuclear fuel is nearly 105 times higher in energy-carry-ing capacity than hydrogen; therefore, 105 fewer shipmentsare required to deliver the same energy to regional cus-tomers. Said another way, 105 longer intervals betweenshipments can provide the same energy security.

7. Market Force Mediation of the Global Fissile Inventory

The partitioning of the hydrogen supply between centralbreeders and distributed batteries will be determined by thefissile material production requirement for a growing econ-omy.

In an ideal situation, the ratio of battery to breeders willbe such that the world’s fissile inventory would be totally


tied up in working inventories of power-producing reactors(and/or in the recycle pipeline having a several-year lagtime). No fissile material in excess of need would be pro-duced and none would accumulate in waste or would buildup in interim storage. Ideally, the function of the breeders atany regional fuel cycle center will be to balance the supplyand demand for fissile material for the service region.

Realistically, in a commercial enterprise, the fractionsof batteries and breeders will be strongly influenced byprofit motives and the relative prices that can be chargedfor fissile material compared with hydrogen. If new batteryreactors are being built and fissile becomes scarce so thatits price rises, then market forces will tend to increase thebreeder fraction in the enterprise such that fissile supplycan meet demand. If at the same time the cost to the con-sumer of hydrogen production from battery plants is lessthan that from breeders, there will be pressure to increasebattery fraction in the enterprise over that of breeders.These two market responses will mediate against eachother to tend to keep fissile supply and demand in balancein the service region.

But if hydrogen produced in central breeders ischeaper than that produced in distributed batteries, thenmarket pressure would tend to increase the breeder frac-tion at the expense of batteries, and this would be an auto-catalytic trend.b This will produce fissile for which there isno market. However, the breeding ratio for any given breed-er is adjustable without changing its hydrogen productionrate, so its core layout could be changed to lower breedingratio and avoid production of fissile for which no marketdemand exists. The financial incentive to do so is to avoidproduction of excess fissile, which will impose inventorycharges on the owners of the fuel cycle facility.

Because of the nearly global reach of the refuelingcassette spokes in the hub/spoke architecture, it is expected


bClient nations will value the energy security attained with locallysited batteries vis-à-vis centrally sited breeders via market forcesthrough the price of fissile versus the price of hydrogen.


that a global price for nuclear fuel cassettes will exist as itcurrently exists for oil. This should equalize the price ofnuclear fuel worldwide and work to mediate fissile produc-tion among the several regional fuel cycle centers and thusfor the world as a whole.

It appears that both technical means and free-marketeconomic incentives can be put into place to maintain fis-sile supply and demand in balance. World fissile inventoriescan be mediated by market forces without governmentalintervention. Thereby, the current situation experienced forthe once-through cycle, leading inexorably to an economi-cally motivated ever-growing fissile inventory in spent-fuelstorage, will be avoided.

8. Conclusion

The discovery of nuclear fission has provided mankind witha further factor of a million (over fossil fuels) in usable ener-gy density for service of mankind. Just as when coal andsteam engines—replacing animal and water power—washarnessed to fuel the Industrial Revolution, this further fac-tor of a million in energy density holds potential to fuel asocietal change of similar scope—a global sustainabilityrevolution.

The Industrial Revolution transition from renewable toa chemical energy resource required that the architectureof production be reengineered to concentrate capital andlabor to exploit the million times higher energy density ofchemical fuel; so too the architecture of energy supplymust now be reengineered to exploit the energy density offission. First, the use of heat of fission to manufacturehydrogen from water will permit nuclear to support all sec-tors of primary energy use. Second, the closure of thenuclear fuel cycle will extend the recoverable energy con-tent of the earth’s endowment of economically recoverableuranium ore to a millennium of global supply. And third, theproposed hierarchical hub/spoke energy delivery architec-ture will help break the energy security/nonproliferation



dilemma and will facilitate a pathway for incremental mar-ket penetration of nuclear deployments worldwide.

Last, as was the case for the Industrial Revolution,reengineering technology is by itself not sufficient; the tran-sition to nuclear-fueled sustainable development in the 21stcentury will require new business strategies and new insti-tutional arrangements as well. Among them are institution-al measures for normalizing worldwide standards ofnuclear facilities safety and operations and (2) institutionalmeasures enabling establishment of regional fuel cycleand waste management centers operating under interna-tional nonproliferation oversight.

Eliminating carbon emissions from every link of theenergy supply chain, recycling the water from hydrogen com-bustion, and consigning only fission products to the wastestream will enable nuclear fission to fuel an ecologically sus-tainable world energy supply for many tens of centuries.



8. CONCLUDING REMARKS


8. Concluding Remarks

1. Nuclear Role in Hydrogen Production

At present, hydrogen is mostly used for industrial processes,and its production corresponds to ~2% of the world primaryenergy demand. The demand for hydrogen will become anorder of magnitude larger by the middle of the century asthe use of hydrogen expands—not only for industrialprocesses but also to transportation.

Eventually, electricity and hydrogen are expected tobecome the two major energy carriers. Approximatelyone-third of the world’s primary energy is converted toelectricity at present. While the ratio to be used for elec-tricity production is forecasted to increase to about half oftotal primary energy in the future, hydrogen is consideredto be the most promising energy carrier produced from theremaining half of primary energy.

Hydrogen, as well as electricity, could be produced bymany kinds of primary energy, namely, fossil fuels, renew-able energy sources, and nuclear energy. Nuclear energyhas the merits of capability for sustainable bulk supply,advantageous environmental effects of minimal carbondioxide emissions, and high energy density leading to en-ergy security for many countries. Nuclear energy holds thepotential to play a major role in fueling sustainable devel-opment by producing hydrogen as well as electricity.


2. Supply Capability of Nuclear Energy and Effect of Nuclear Hydrogen Production

According to the estimates of the World Energy Council64

(WEC), the world primary energy demand in 2100 would beabout four times that of 1990 in the case of the middlecourse (B case) with nuclear energy expected to supply24% of total primary energy for electricity production. Thisamount of nuclear supply corresponds to a capacity of~5200 units of 1000-MWe plants. The supply of fissile fuelto these plants is feasible, assuming the ultimate resourcesof natural uranium 16.3 Mton according to the NuclearEnergy Agency/International Atomic Energy Agency RedBook65 and the recycling use of plutonium (or all transuran-ics) by fast breeder reactors (FBRs) with a breeding ratio of1.2 to 1.3 introduced from 2030 to 2050 (Ref. 66).

To reduce global greenhouse gas emissions andbegin displacing fossil fuels, optimizing the recycling use ofplutonium in FBRs could increase energy supply by nuclearenergy to one and one-half times in 2050 and two times in2100 of the WEC-B case estimates. By effectively utilizingnuclear energy for hydrogen production, this excess supplycapacity of nuclear energy over the WEC-B case couldreplace fossil fuels share as shown in the Table 1 (Ref. 66).

In such a scheme, the global supply capacity of fossilin 2100 would become smaller than it was in 1990, thusattaining stabilization of atmospheric carbon dioxide con-


Table 1.Primary Energy Supply for 1990–2100WEC-B Case → Proactive Nuclear Deployment CaseEnergy in Gtoe(109 ton Oil Equivalent)

1990 2050 2100

Fossil 6.9 12.7 → 11.4 15.0 → 5.0 Nuclear 0.45 2.7 → 4.0 8.3 → 18.3Hydro + renewables 1.6 4.4 11.4

Total 9.0 19.8 34.7


centration even in the face of global growth of energy useby a factor of 4.

3. Hydrogen Production Processes Using Nuclear Energy

For the production of hydrogen using nuclear energy, manyprocesses have been proposed. The leading processesnow under research and development are nuclear-heatedsteam reforming of natural gas or other hydrocarbons, ther-mochemical splitting of water by nuclear heat, high-temper-ature electrolysis of steam by nuclear electricity and heat,and electrolysis of water by nuclear electricity (Fig. 1).

As for the coupling of nuclear reactor and hydrogenproduction processes, a variety of concepts have beenproposed so far as shown in Table 2.

Still new processes of nuclear hydrogen production arejoining the race. It is considered that targeted exploratoryresearch would be worthwhile for broadening the possibilityof reaching the goal of efficient and economic hydrogenproduction by nuclear energy.

CONCLUDING REMARKS 129

Fig.1. Methods of hydrogen production by nuclear energy.



Tabl

e 2.

Var

ious

Con

cept

s of

Nuc

lear

Hyd

roge

n P

rodu

ctio

n

Pro

du

ctio

n M

eth

od

Raw

Mat

eria

lsE

ner

gy

Rea

cto

r Ty

pe

(Typ

ical

)

Ele

ctro

lysi

s of

wa

ter

Wa

ter

Ele

ctric

ityP

ower

rea

ctor

s(L

WR

,H

WR

,et

c.)

Hig

h te

mpe

ratu

re e

lect

roly

sis

ofst

eam

Wa

ter

Ele

ctric

ity +

hea

tH

igh-

tem

pera

ture

(hig

h te

mpe

ratu

re)

gas-

cool

ed r

eact

ors

The

rmoc

hem

ical

spl

ittin

g of

wa

ter

Wa

ter

Hea

t (h

igh

tem

pera

ture

)H

igh-

tem

pera

ture

gas-

cool

ed r

eact

ors

The

rmoc

hem

ical

spl

ittin

g of

wa

ter

Wa

ter

Hea

t (h

igh

tem

pera

ture

)H

igh-

tem

pera

ture

liqui

d-m

etal

- an

d m

olte

n-sa

lt-co

oled

rea

ctor

sT

herm

oche

mic

al s

plitt

ing

ofw

ate

rW

ate

rH

eat

(med

ium

tem

pera

ture

)S

odiu

m-c

oole

d fa

st r

eact

ors

The

rmoc

hem

ical

spl

ittin

gof

wa

ter

(hyb

rid)

Wa

ter

Hea

t (m

ediu

m t

empe

ratu

re)

Sod

ium

-coo

led

fast

rea

ctor

s+

ele

ctric

ityT

herm

al d

ecom

posi

tion

ofm

etha

neN

atu

ral g

asH

eat

(hig

h te

mpe

ratu

re)

Hig

h-te

mpe

ratu

reliq

uid-

met

al-c

oole

d re

acto

rsS

team

ref

orm

ing

ofm

etha

neN

atu

ral g

as +

wa

ter

Hea

t (h

igh

tem

pera

ture

)H

igh-

tem

pera

ture

gas-

cool

ed r

eact

ors

Ste

am r

efor

min

g of

met

hane

Na

tura

l gas

+ w

ate

rH

eat

(med

ium

tem

pera

ture

)S

odiu

m-c

oole

d fa

st r

eact

ors

(mem

bran

e)S

team

ref

orm

ing

ofD

ME

Dim

ethy

l eth

er +

wa

ter

Hea

t (lo

w t

empe

ratu

re)

LWR

Rad

ioca

taly

sis

ofw

ate

rW

ate

rG

amm

a ra

yR

adio

activ

e w

aste


4. Prospects of Commercialization and Future Energy Systems

While it is unclear what course the commercialization ofnuclear production of hydrogen will take, a typical prospectbased on the current state of knowledge could be as follows:

1. In the near term, coupling of electricity generationby light water reactor/heavy water reactor (LWR/HWR) andwater electrolysis can be commercialized, in some casesby using off-peak power, as the relevant technologies arealready proven.

2. In the intermediate term, nuclear-heated steamreforming of natural gas using medium-temperature reac-tors could be utilized because of its advantage in economiccompetitiveness, in spite of some carbon dioxide emissions.Also, high-temperature steam electrolysis coupled withhigh-temperature reactors could join the line with higherconversion efficiency and lower barrier in materials.

3. In the long term, coupling of high-temperaturereactors and thermochemical water splitting would ulti-mately be deployed, as the bulk chemical processes bene-fit from economy of scale and may turn out to be the bestfor very-large-scale nuclear production of hydrogen for amature global hydrogen economy.

Here, we have reviewed production technologies, safe-ty of nuclear hydrogen production, economic competitive-ness, and hydrogen energy supply systems. However, forcommercialization of the nuclear production of hydrogen,numerous additional issues in addition to the issuesreviewed in the previous chapters must be solved. Thoseissues include an appropriate size of production plant tomeet a market demand, public acceptance of nuclear-derived hydrogen, and partner industries to develop nuclearhydrogen production with the nuclear industry. While theyare outside the scope of this booklet, these are importantissues to be solved in the course of development.

In the very long term (in the inevitable event of inade-quate fossil and renewable options for world energy supply),

CONCLUDING REMARKS 131


the world may require a very large-scale global deploymentof nuclear fission–generated electricity and hydrogen. Twoalternative energy architectures can be contemplated—bothexploiting the world’s uranium ore resources by closing thenuclear fuel cycle at a dozen or fewer regional fuel cyclecenters. Breeder reactors located at the regional centerswould convert the fertile 238U resource into fissionable mate-rial to fuel a growing fission-based energy economy. Thesebreeders would utilize the heat released in fission to crackwater for hydrogen manufacture. In one alternative—a high-ly centralized architecture that emphasizes economy ofscale in both nuclear fuel cycle and hydrogen production(discussed in the Preface)—liquid hydrogen would beshipped from the regional centers to energy consumersworldwide. In an alternate distributed hub/spoke architecture(discussed in Chap. 7), fissionable fuel manufactured bybreeders at the regional centers would be shipped tonuclear reactors sited near energy load centers worldwide,where it would be used for the local manufacture of electric-ity and hydrogen. In this case—in an architecture emphasiz-ing supply diversification and incremental market penetra-tion—the economy of scale in nuclear fuel cycle operationsis maintained while hydrogen manufacture is sharedbetween the centralized breeder reactors and the distributedlocal reactors. Both architectures would achieve an ecologi-cally neutral world energy supply by eliminating carbonemissions at every link of the energy supply chain and byself-consuming all long-lived radioactive species. In thatcase (only) fission products are consigned to waste—andthey decay to a radiotoxicity level slightly below that of theoriginal uranium ore after 200 to 300 yr of sequestration.

Market forces for affordable energy and political pres-sures for energy security and nonproliferation assuranceswill shape the evolutionary pathway toward one or the otherof these architectures. However, in either case, the nuclearproduction of hydrogen from water as a synthetic chemicalenergy carrier complementary to electricity offers a path-way to a global energy supply that meets the longevity andecological tenets of sustainable development.



REFERENCES

1. W. Häfele, “On the Nature of Nuclear Power andIts Future,” Proc. Global ’93, Future Nuclear Systems:Emerging Fuel Cycles and Waste Disposal Options,Seattle, Washington, September 12–17, 1993, p. 3.

2. G. Brundtland, Our Common Future, OxfordUniversity Press, Oxford, United Kingdom (1987).

3. N. Nakicenovic, A. Grubler, and A. McDonald,“Global Energy Perspectives—A Joint IIASA–WEC Study,”Cambridge University Press (1998) �http://www.iiasa.ac.at/�

4. M. Hori, “Role of Nuclear Energy in the Long-Term Global Energy Perspective,” Proc. 1st InformationExchange Mtg. Nuclear Production of Hydrogen, Paris,France, October 2–3, 2000, p. 25 (2001).

5. M. Hori et al., “Synergistic Production of Hydro-gen Using Fossil Fuels and Nuclear Energy—Application ofNuclear-Heated Membrane Reformer,” presented atOECD/NEA 2nd Information Exchange Mtg. NuclearProduction of Hydrogen, Argonne, Illinois, October 2–3,2003.

6. I. Yasuda et al., “Development of MembraneReformer for Hydrogen Production from Natural Gas,” pre-sented at 14th World Hydrogen Energy Conf., Montreal,Canada, June 9–13, 2002; see also Y. Shirasaki et al.,“New Concept Hydrogen Production System Based onMembrane Reformer,” presented at Fuel Cell Seminar,Palm Springs, California, November 18–21, 2002.

7. S. Shiozawa et al., “Present Status of JAERI’s R&Don Hydrogen Production Systems in HTGR,” Proc.OECD/NEA 1st Information Exchange Mtg. Nuclear

INCSHydrogen-references 12/8/04 8:21 AM Page 133


Production of Hydrogen, Paris, France, October 2–3, 2000,p. 57, Organization for Economic Cooperation and Develop-ment; see also H. Endo et al., “Potential of Multi-purposeLiquid Metallic Fuelled Fast Reactor (MPFR) as a HydrogenProduction System,” Proc. OECD/NEA 1st InformationExchange Mtg. Nuclear Production of Hydrogen, Paris,France, October 2–3, 2000, Organization for EconomicCooperation and Develoment; see also M. Tashimo et al.,“Concept of FR-MR,” presented at 11th Int. Conf. NuclearEngineering, Tokyo, Japan, April 20–23, 2003, ICONE-11-36321.

8. Y. Inagaki et al., “Research and DevelopmentProgram on HTTR Hydrogen Production System,” present-ed at Int. Conf. Advanced Nuclear Power Plants and GlobalEnvironment (GENES4/ANP2003), Kyoto, Japan, Septem-ber 15–19, 2003, paper 1062.

9. Y. Inagaki et al., “Present Status and Future Planof HTTR Project,” presented at OECD/NEA 2ndInformation Exchange Mtg. Nuclear Production ofHydrogen, Argonne, Illinois, October 2–3, 2003.

10. M. Tashimo et al., “Advanced Design of FastReactor-Membrane Reformer (FR-MR),” presented atOECD/NEA 2nd Information Exchange Mtg. NuclearProduction of Hydrogen, Argonne, Illinois, October 2–3,2003.

11. “Nuclear Energy for Hydrogen Generation,”Position Statement 60, American Nuclear Society (June2003).

12. S. Shiozawa et al., “Present Status of JAERI’sR&D on Hydrogen Production Systems in HTGR,” Proc.OECD/NEA 1st Information Exchange Mtg. NuclearProduction of Hydrogen, Paris, France, October 2–3, 2000,p. 57, Organization for Economic Cooperation andDevelopment.

13. A. I. Miller, “Electrochemical Production ofHydrogen by Nuclear Energy,” Nuclear Production ofHydrogen––Technologies and Perspectives for Global


REFERENCES 135

Deployment, Chap. 4, Table 6, American Nuclear Society(2004).

14. N. Nakicenovic et al., “Global Energy Perspec-tives—A Joint IIASA–WEC Study,” Cambridge UniversityPress, �http://www.iiasa.ac.at/cgi-bin/ecs/book_dyn/bookcnt.py�

15. M. Hori, “Role of Nuclear Energy in the Long-TermGlobal Energy Perspective,” Proc. OECD/NEA 1st Infor-mation Exchange Mtg. Nuclear Production of Hydrogen,Paris, France, October 2–3, 2000, p. 25, Organization forEconomic Cooperation and Development; see also M.Tashimo et al., “Role of Nuclear Produced Hydrogen forGlobal Environment and Energy,” presented at OECD/NEA2nd Information Exchange Mtg. Nuclear Production ofHydrogen, Argonne, Illinois, October 2–3, 2003.

16. “Hydrogen as an Energy Carrier and ItsProduction by Nuclear Power,” IAEA-TECDOC-1085,International Atomic Energy Agency (May 1999).

17. Fuel Cell Technology Handbook, G. Hoogers, Ed.,CRC Press, New York (2003).

18. J. Larminie and A. Dicks, Fuel Cell SystemsExplained, Wiley, New York (2000).

19. Fuel Cell Handbook, 6th ed., DOE/NETL 2002/1179, EG&G Technical Services, Science ApplicationsInternational Corp. (2002).

20. N. J. Maskalick, “High Temperature ElectrolysisCell Performance Characterization,” Int. J. HydrogenEnergy, p. 563 (1986).

21. R. Hino and Y. Miyamoto, “Hydrogen Productionby High-Temperature Electrolysis of Steam,” Proc.Technical Committee Mtg., Oarai, Japan, October 19–20,1992, IAEA-TECDOC-761, p. 120, International AtomicEnergy Agency (Aug. 1994).

22. R. Hino, H. Aita, K. Sekita, K. Haga, and T. Iwata,“Study on Hydrogen Production by High TemperatureElectrolysis of Steam,” JAERI-Research 97-064, JapanAtomic Energy Research Institute (Sep. 1997).



23. B. Yildiz and M. S. Kazimi, “Nuclear EnergyOptions for Hydrogen and Hydrogen-Based Liquid FuelsProduction,” MIT-NES-TR-001, Massachusetts Institute ofTechnology (Sep. 2003).

24. G. Beghi, “A Decade of Research on Thermo-chemical Hydrogen at the Joint Research Centre, Ispra,”Int. J. Hydrogen Energy, 11, 12, 761.

25. W. Weirich, K. F. Knoche, F. Bert, and H. Barnert,“Thermochemical Process for Water-Splitting—Status andOutlook,” Nucl. Eng. Des., 78, 285 (1984).

26. J. E. Funk and R. Reinstrom, “Energy Require-ments in the Production of Water,” Ind. Eng. Chem.,Process Design Develop., 5, 3, 336 (1966).

27. G. E. Besenbruch et al., GA-A18257, GeneralAtomics (1982).

28. Y. Miyamato et al., “R&D Program on HydrogenProduction System with High Temperature CooledReactor,” Proc. Int. Hydrogen Energy Forum 2000, Munich,Germany, September 2000, Vol. 2, p. 271.

29. J. E. Funk, “Thermochemical Hydrogen Produc-tion: Past and Present,” Int. J. Hydrogen Energy, 26, 185(2001).

30. K. Yoshida, H. Kameyama, and T. Aochi, “ASimulation Study of the UT3 Thermochemical CycleHydrogen Production Process,” Int. J. Hydrogen Energy,15, 3, 171 (1990).

31. F. Lemort, C. Lafon, and F. Girold, “Thermo-dynamic and Technological Assessment of the IS Cycle,”EHEC 2003.

32. B. D. Struck, R. Junginger, D. Boldersdorf, and J.Gehrmann, “The Anodic Oxidation of Sulfur Dioxide in theSulfuric Acid Hybrid Cycle,” Int. J. Hydrogen Energy, 5,487.

33. S. Goldstein, J. M. Borgard, and X. Vitart,“Upperbound and Best Estimate of the Efficiency of the ISCycle,” Int. J. Hydrogen Energy (accepted paper).


REFERENCES 137

34. X. Vitart, J. M. Borgard, S. Goldstein, and S.Colette, “Investigation of the IS Cycle for MassiveHydrogen Production,” Argonne National Laboratory(2004).

35. M. Roth and K. F. Knoche, “ThermochemicalWater Splitting Through Direct HI Decomposition fromH2O/HI/I2 Solutions,” Int. J. Hydrogen Energy, 14, 8, 545(1989).

36. G. Besenbruch, S. Goldstein, L. Brown, D.Powers, E. Funk, P. Pierrard, J. M. Borgard, and P. Pickard,“An International Laboratory-Scale Demonstration of theSulfur-Iodine Thermochemical Water-Splitting Cycle,”AIChE J. 2003.

37. K. Schultz, “Thermochemical Production ofHydrogen from Solar and Nuclear Energy,” San Diego,California, April 14, 2003.

38. M. Treviño, “Technología de gasificacíon integra-da en ciclo combinado,” Elcogas, Club Español de laEnergia, Madrid (2003).

39. T. K. Bose, J. Ohi, and R. Hay, Sourcebook forHydrogen Applications, TISEC, Canada (1998).

40. “Occupational Health and Safety Standards,” Title29, Parts 1900 to 1910, Code of Federal Regulations(1993).

41. H. O. Henrie and A. K. Postma, “Lessons Learnedfrom Hydrogen Generation and Burning During the TMI-2Event,” GEND-061, U.S. Department of Energy (May1987).

42. “Hydrogen in Water-Cooled Nuclear Power Reac-tors (State-of-Art-Report),” EUR 14037 EN, Commission ofEuropean Communities (Oct. 1991).

43. W. Breitung and P. Royl, “Procedure and Tools forDeterministic Analysis and Control of Hydrogen Behaviourin Severe Accidents,” Nucl. Eng. Des., 202, 249 (2000).

44. R. G. Zalosh and T. P. Short, “Compilation andAnalysis of Hydrogen Accident Reports,” COO-4442-4,



Factory Mutual Research Corp., Norwood, Massachusetts(1978).

45. C. E. G. Padró and V. Putsche, “Survey of theEconomics of Hydrogen Technologies,” NREL/TP-570-27079, U.S. Department of Energy (Sep. 1999).

46. World Bank data, available at http://people.hofstra.edu/geotrans/eng/ch6en/conc6en/transitvscaruse.html

47. A. I. Miller and R. B. Duffey, “Meeting the Near-Term Demand for Hydrogen Using Nuclear Energy inCompetitive Power Markets,” Proc. 2nd InformationExchange Mtg. Nuclear Production of Hydrogen, Argonne,Illinois, October 2–3, 2003, Nuclear Energy Agency.

48. Enbridge, December 2003, available at http://www.egd.enbridge.com/LVRC/html/rate6.html

49. P. L. Spath and M. K. Mann, “Life CycleAssessment of Hydrogen Production via Natural Gas SteamReforming,” NREL/TP-570-27637, National RenewableEnergy Laboratory, Golden, Colorado (Feb. 2001).

50. “DTI Initial Contribution to the PIU Energy PolicyReview,” United Kingdom Department of Trade andIndustry, Working Paper, July 2001, http://www.cabinetoffice.gov.uk/innovation/2001/energy/EPQw.pdf

51. F. Oney, T. N. Veziroglu, and Z. Dulger, “Evaluationof Pipeline Transportation of Hydrogen and Natural GasMixtures,” Int. J. Hydrogen Energy, 19, 10, 813.

52. W. Amos, “Costs of Storing and TransportingHydrogen,” NREL/TP-570-25106, National RenewableEnergy Laboratory, Golden, Colorado (May 1998).

53. Chem. Eng. News (June 24, 1996) available athttp://pubs.acs.org/hotartcl/cenear/960624/prod.html

54. C. Marchetti, “Energy Islands in the FinalConfiguration of the Hydrogen Economy,” Plenary Lectureat the World Hydrogen Conference, June 21–26, 1998.


REFERENCES 139

55. J. Nitsch and C. Voigt, “Hydrogen in a FutureEnergy Supply System,” Hydrogen as an Energy Carrier,C. Winter and J. Nitsch, Eds., Springer Verlag (1988).

56. D. Wade et al., “STAR-H2: A 400 MWth Lead-Cooled, Long Refueling Interval Reactor for HydrogenProduction,” paper ICONE11-376576, Proc. Int. Conf.Nuclear Engineering, ICONE-11, Tokyo, Japan, April20–23, 2003, American Society of Mechanical Engineers.

57. J. J. Sienicki, A. V. Moisseytsev, D. C. Wade, M. T.Farmer, C. P. Tzanos, J. A. Stillman, J. W. Holland, P. V.Petkov, I. U. Therios, R. F. Kulak, and Q. Wu, “The STAR-LM Lead-Cooled Closed Fuel Cycle Fast Reactor Coupledto a Supercritical Carbon Dioxide Brayton Cycle AdvancedPower Converter,” Global 2003, New Orleans, Louisiana,November 16–20, 2003, American Nuclear Society (2003).

58. N. W. Brown, C. F. Smith, E. Greenspan, A.Minato, D. C. Wade, J.-S. Choi, W. G. Halsey, and D. Vogt,“Liquid Metal Cooled Reactors and Fuel Cycles forInternational Security,” paper ICONE 11-36339, Proc. 11thInt. Conf. Nuclear Engineering, ICONE-11, Tokyo, Japan,April 20–23, 2003, American Society of MechanicalEngineers (2003).

59. R. Doctor, D. Matonis, and D. Wade, “HydrogenGeneration Using a Calcium-Bromine ThermochemicalWater-Splitting Cycle,” Proc. 2nd Information ExchangeMtg. Nuclear Production of Hydrogen, Argonne NationalLaboratory, October 2–3, 2003 (to be published by OECD-NEA).

60. A. Moisseytsev, J. Sienicki, and D. Wade, “CycleAnalysis of Supercritical Carbon Dioxide Gas TurbineBrayton Cycle Power Conversion System for Liquid Metal-Cooled Fast Reactors,” paper ICONE 11-36023, Proc. Int.Conf. Nuclear Engineering, ICONE-11, Tokyo, Japan, April20–23, 2003, American Society of Mechanical Engineers(2003).

61. R. Faibish, “Seawater Desalination as aBottoming Cycle for a Hydrogen-Producing Nuclear Plant



Complex,” accepted for publication in Proc. Int. YouthNuclear Conf., Toronto, Canada, May 9–13, 2004.

62. Y. Kato and T. Nitawoki, “Zero Waste HeatRelease Nuclear Cogeneration System,” Paper 3313, Proc.ICAPP, Cordoba, Spain, May 4–7, 2003.

63. C. Marchetti, “On Hydrogen and EnergySystems,” Int. J. Hydrogen Energy, 1, 1 (1976).

64. N. Nakicenovic, A. Grubler, and A. McDonald,Global Energy Perspectives––A Joint IIASA–WEC Study,Cambridge University Press.

65. Joint Report on Uranium: Resources, Productionand Demand, Organization for Economic Cooperation andDevelopment/Nuclear Energy Agency and InternationalAtomic Energy Agency (1997).

66. M. Hori, “Role of Nuclear Energy in the Long-Term Global Energy Perspective,” Proc. 1st InformationExchange Mtg. Nuclear Production of Hydrogen, Paris,France, October 2–3, 2000, p. 25, Organization forEconomic Cooperation and Development/Nuclear EnergyAgency (2001).


nuclear production of hydrogen · in addition, hydrogen had a long and successful career as a final...

Documents