You are on page 1of 337

ASES

Editor-in-chief

D.Yogi Goswami

ADVANCES IN
SOLAR ENERGY
An Annual Review of
Research and Development

volume

17

prelims

2/7/07

3:51 PM

Page i

Advances in Solar Energy


An Annual Review of Research and Development
Volume 17

prelims

2/7/07

3:51 PM

Page ii

Editor-in-Chief
D. Yogi Goswami
Clean Energy Research Center, University of South Florida, Tampa, FL

Associate Editors
Viacheslav M. Andreev
Ioff Physico-Technical Institute, St Petersburg, Russia

J. Douglas Balcomb
National Renewable Energy Laboratory, Golden, CO

Adolf Goetzberger
Fraunhofer Institut Solare Energiesysteme, Freiburg, Germany

Yoshihiro Hamakawa
Ritsumeikan University, Kusatsu Shiga, Japan

Lu Weide
China Rural Energy Industry Association, Beijing, China

Antonio Luque
Universidad Politcnica de Madrid, Madrid, Spain

Gerald R. Nix
National Renewable Energy Laboratory, Golden, CO

Morton Prince
Melrose Park, PA

Aldo Steinfeld
Swiss Federal Institute of Technology, Zurich, Switzerland

Gunnar Svedberg
Mid Sweden University, Sundsvall, Sweden

Steven Szokolay
University of Queensland, Kenmore, Australia

Lorin Vant-Hull
University of Houston, Houston, TX

Nejat Veziroglu
University of Miami, Coral Gables, FL

Carl Weinberg
Walnut Creek, CA

Roland Winston
University of Chicago, Chicago, IL

Emeritus Editor-in-Chief
Karl W. Ber
University of Delaware, Newark, DE

prelims

2/7/07

3:51 PM

Page iii

Advances in Solar Energy


An Annual Review of Research and Development
Volume 17

Edited by D. Yogi Goswami

London Sterling, VA
American Solar Energy Society, Inc., Boulder, CO

prelims

2/7/07

3:51 PM

Page iv

First published by Earthscan in the UK and USA in 2007


Copyright American Solar Energy Society, 2007
American Solar Energy Society, Inc.
2400 Central Avenue, Suite A
Boulder, CO
USA
All rights reserved
ISSN
ISBN

0731-8618
978-1-84407-314-6

Typeset by Domex e-Data, India


Printed and bound in the UK by Cromwell Press, Trowbridge
Cover design by Yvonne Booth
For a full list of publications please contact:
Earthscan
812 Camden High Street
London, NW1 0JH, UK
Tel: +44 (0)20 7387 8558
Fax: +44 (0)20 7387 8998
Email: earthinfo@earthscan.co.uk
Web: www.earthscan.co.uk
22883 Quicksilver Drive, Sterling, VA 20166-2012, USA
Earthscan is an imprint of James and James (Science Publishers) Ltd and publishes in
association with the International Institute for Environment and Development
A catalogue record for this book is available from the British Library
Library of Congress Cataloging-in-Publication Data has been applied for
The paper used for this book is FSC-certified and
elemental chlorine-free. FSC (the Forest Stewardship
Council) is an international network to promote
responsible management of the worlds forests.

prelims

2/7/07

3:51 PM

Page v

FOREWORD
Global climate change continues to be a hot topic on the world stage. However, we have
now moved away from the debate of whether global warming exists to what strategies
are needed to combat the challenge. Another ongoing debate concerns the peaking and
eventual decline of world oil production and how this will affect other conventional energy
resources including coal, natural gas and uranium. It is becoming increasingly clear that
whether we are concerned about global climate change or the peaking and eventual
decline of conventional resources, we must ramp up the use of renewable energy
resources worldwide. Indeed, Schindler and Zittel argue in their chapter in this volume of
Advances in Solar Energy that by 2050 as much as 50 per cent of our energy will have to
come from renewable energy. In our continuing series of country- and region-specific
articles we focus on the Middle East and North Africa (MENA) region Alnaser, Treib and
Knies show that by 2050 as much as 50 per cent of the electrical power for the MENA
countries may come from concentrating solar power (CSP).
At present, commercial deployment of renewable energy technologies is increasing at
rates of 2040 per cent per year. However, in order to increase the deployment of
renewable energy technologies to the levels foreseen by Schindler and Zittel, we need
continued research and development to raise the efficiency and reduce the costs of these
technologies. The chapters in this volume inform us on progress in this direction, for
example advances in quantum dot solar cells and organic solar cells. Other topics in this
volume include such applications of solar energy as solar-hydrogen production, solar
photocatalytic detoxification of water, solar drying and solar industrial process heat.
In the first chapter, Schindler and Zittel question the assumptions made by the
International Energy Agency (IEA) in forecasting the future availability of energy resources,
and present an alternative world energy outlook and a possible path towards a sustainable
future. According to the authors the world will experience the peaking of oil first, followed
by natural gas. They show that the resulting energy supply gap cannot be filled by a rising
share of nuclear energy. This gap could be filled partially by coal; however, without CO2
sequestration this would lead to emissions at unacceptable levels. Their analysis shows
that a transition in our energy resource use to renewable energy will occur. However, they
question whether there is enough time left for the transition to occur smoothly. They
recommend that the earlier we start with the transition, the better we will be able to
manage it.
Chapters 2 and 3 deal with the new developments in photovoltaic cells. Chapter 2
deals with quantum well solar cells and the progress made in this area over the last 15
years. The author, Ekins-Daukes, writes, The ability to adjust the absorption profile of a
bulk semiconductor through the inclusion of quantum wells gives rise to a number of
practical, near-term applications for quantum solar cells. Some of these applications
include multi-junction solar cells and thermophotovoltaics. The author concludes that
even though to date only the band gap has been engineered with the application of
quantum wells, it is likely that in the future it will be possible to control other properties,
such as optical transitions and phonon modes in order to produce highly efficient devices.
Chapter 3 describes the recent progress in organic photovoltaic materials and devices.
Organic and polymeric materials have the potential to reduce the cost of solar cells

prelims

2/7/07

vi

3:51 PM

Page vi

ADVANCES IN SOLAR ENERGY

making them cost competitive. At present the conversion efficiencies of organic solar
cells are relatively small (< 6 per cent), which can be attributed to a number of reasons,
such as, photon, exciton and carrier losses, poor material morphologies, and unoptimized
structures. However, there is plenty of room for improvement. The paper outlines the
approaches being followed in improving the performances.
Chapter 4 describes the progress in thermal and material characterization of
immersed heat exchangers used in solar domestic water heating systems. The authors
have made recommendations based on published literature and their own work over the
last 20 years. As manufacturers consider the use of polymers for heat exchangers, their
durability and reliability must also be considered in addition to thermal performance. The
authors also point out areas for additional research.
Chapter 5 deals with solar photocatalytic detoxification, in which there has been
tremendous research and development over the past ten years, although it is a relatively
recent environmental application of solar radiation. In this paper, the authors review the
scientific progress and applications of solar photocatalysis. Even though the process is
very attractive in remediating water pollution, the quantum efficiency is very low. The
authors describe the research and development (R&D) approaches to increase the
quantum yield, such as modifying the catalyst structure and composition and finding new
catalysts with band gaps that better overlap the solar spectrum. The authors also describe
the progress in using photo-Fenton treatment, and combined photocatalytic-biological
treatment.
At present there is considerable interest in the potential use of hydrogen for
transportation, to replace the eventually depleting supplies of oil. However, for hydrogen
to provide a viable solution, it must be produced economically and from a clean source,
such as solar energy. Solar photo-electrochemical generation of hydrogen is one such
method that has a viable potential. However, this method needs some fundamental R&D.
In Chapter 6, the authors systemically describe the solid-state chemistry-related issues
that must be considered in order to improve the viability of this method. The authors
explain the parameters such as band gap and Fermi level and the concepts of modifying
them.
Chapter 7 reviews the recent progress in using solar heat for industrial processes,
with specific examples from Europe. The authors review the technology and its potential
for use in low (< 60C), medium (60150C) and medium-high (150250C) temperature
industrial applications. They present case studies for many industries in Spain and
Portugal, arguing that as much as 7 per cent of total final energy demand in southern
European countries may be fulfilled by providing solar process heat at temperatures
below 250C. The authors state that the present costs of solar thermal systems range
from 250 to 1000/kW of thermal power, leading to average thermal energy costs in
southern Europe of 25/kWh for low temperature applications and 515/kWh for
medium temperature applications. However, according to the authors, these costs can be
reduced by 50 per cent by 2010 by mass production, reducing operation and maintenance
expenses and improving collector efficiency and collector design.
The final paper (Chapter 9) of this volume is focused on the Middle East and North
Africa (MENA) as a part of our series of country-specific articles. This article, however,

prelims

2/7/07

3:51 PM

Page vii

FOREWORD

vii

covers many countries with a common potential in using concentrating sunlight to


produce electrical power. The MENA region has the potential to provide not only for its
own electrical needs from solar energy but also for those of parts of Europe. The authors
show that with a growth rate of 2535 per cent concentrating solar thermal power (CSP)
can be used to fulfil as much as 14 per cent of the electricity demand of MENA by 2025
and 57 per cent by 2050. According to the authors, CSP may also be able to provide a
majority of the acute water needs of the region via desalination. The study also shows that
the total carbon emissions of electricity generation of all MENA countries can be reduced
from the present 770 million tons per year to 475 million tons per year by 2050, instead
of increasing to 2000 million tons.
I would like to thank all the authors for sharing their knowledge and expertise with the
rest of the world via their papers in this volume of Advances in Solar Energy. I would also
like to thank my editorial assistant Ms Barbara Graham for her tireless work in carefully
editing many articles. Finally, I would like to thank the American Solar Energy Society, its
President Ronald Larson and Executive Director Brad Collins, and the Board of Directors
for their commitment in continuing to publish Advances in Solar Energy for the benefit of
researchers and practitioners of solar energy around the world.

D. Yogi Goswami
Editor-in-Chief

prelims

2/7/07

3:51 PM

Page viii

prelims

2/7/07

3:51 PM

Page ix

Contents
Foreword
List of figures and tables
About the authors
1

Alternative World Energy Outlook 2006:


A Possible Path towards a Sustainable Future

v
xii
xxi
1

Jrg Schindler and Werner Zittel


1.1 Winds of Change: The Transition Period
1.2. Future Availability of Fossil and Nuclear Energy Sources
1.3 Alternative World Energy Scenarios
Notes and References
Annex I: Renewable Energy Potentials
Annex II: Simulation Parameters for Renewable Energy Scenarios
Annex III: Growing Share of Renewable Energy since 1990

1
7
28
38
41
42
43

45

Quantum Well Solar Cells

N. J. Ekins-Daukes
2.1 History
2.2 Quantum Well Electronic Structure
2.3 Basic Operation of the P-I-N Quantum Well Solar Cell
2.4 Near-term Applications for Quantum Well Solar Cells
2.5 Efficiency Limits
2.6 Conclusion
Acknowledgements
References

45
46
49
61
64
66
66
67

74

Recent Progress of Organic Photovoltaics

Sam-Shajing Sun
3.1 Introduction
3.2 Organic versus Inorganic Semiconductors
3.3 Organic/Polymeric Solar Cell Developments
3.4 Organic Solar Cell Fabrications
3.5 Organic Solar Cell Optimizations
3.6 Conclusions and Future Perspectives
Acknowledgements
References

Thermal and Material Characterization of Immersed


Heat Exchangers for Solar Domestic Hot Water

74
75
77
84
87
94
95
95

99

Jane H. Davidson, Susan C. Mantell and Lorraine F. Francis


4.1 Introduction
4.2 Thermal Characterization and Design

99
102

prelims

2/7/07

3:51 PM

Page x

ADVANCES IN SOLAR ENERGY

4.3 Mechanical Characterization of Polymers


4.4 Scaling of Candidate Polymers
4.5 Conclusion
Acknowledgements
References

Photocatalytic Detoxification of
Water with Solar Energy

112
118
124
125
125

130

Sixto Malato, Julin Blanco, Diego C. Alarcn, Manuel I. Maldonado,


Pilar Fernndez-Ibez and Wolfgang Gernjak
5.1
5.2
5.3
5.4
5.5
5.6
5.7
5.8

Introduction
Solar Collectors for Photochemistry
Fundamental Parameters in Solar Photocatalysis
Factors Affecting Solar Photocatalysis
Solar UV Photocatalytic Degradation of Contaminants
Evaluation of Solar UV Radiation
Installed Solar Photocatalytic Treatment Plants
Photocatalytic Detoxification of Water with
Solar Energy: Outlook for the Future
Acknowledgements
References

Solar-Hydrogen: A Solid-State
Chemistry Perspective

130
132
138
147
149
151
154
160
163
163

169

J. Nowotny, T. Bak, L. R. Sheppard and C. C. Sorrell


6.1 Introduction
6.2 Solar-Hydrogen
6.3 The Concept of Solar-Hydrogen Generation
6.4 Materials Property Requirements for Photo-Electrodes
6.5 Electronic Structure
6.6 Why Titania?
6.7 Reduced-Band-Gap Titania
6.8 Impact of Defect Chemistry on the Properties of Titania
6.9 Collective and Local Factor
6.10 Spin-off Applications of Titania
6.11 Multiphase Photo-Sensitive Systems
6.12 Solar Cell Equipped with Space-Based
Solar Energy Collector
6.13 Solar-Oxygen
6.14 Economic and Environmental Considerations
of Solar-Hydrogen
6.16 Solar-Methanol
6.17 Conclusions
Notes
Acknowledgements
References

170
174
175
178
180
190
193
194
203
204
204
205
207
207
209
210
211
211
211

prelims

2/7/07

3:51 PM

Page xi

CONTENTS

Solar Heat for Industrial Processes

xi

216

H. Schweiger, J. Farinha Mendes, Ma. J. Carvalho,


K. Hennecke, D. Krger
7.1 Introduction
7.2 Application Potential
7.3 Available Solar Collector Technology
7.4 Guidelines for Evaluation and System Design
7.5 Case Studies
7.6 Conclusions
References

Solar Energy Technology in the Middle


East and North Africa (MENA) for
Sustainable Energy, Water and Environment

216
217
220
238
247
257
258

261

W. E. Alnaser, F. Trieb and G. Knies


8.1 Introduction
8.2 Determination of the Electricity Demand in MENA
8.3 Renewable Energy Resources in EU-MENA
8.4 CSP Demand-Side Potential in MENA
8.5 The Critical Issue of Water and Energy in the GCC Countries
8.6 Water Resources and Water Demand in MENA
8.7 The Potential for Desalination by Concentrating Solar Power
8.8 Environmental Impact of Using Solar Technology in MENA
Acknowledgements
References

262
264
265
274
282
284
291
295
302
302

prelims

2/7/07

3:51 PM

Page xii

List of Figures and Tables


FIGURES
1.1

World energy demand 19712030 as given in the IEAs


World Energy Outlook 2004 and 2005
1.2
The basic dilemma Business as usual or climate policy
1.3
Oil production of countries outside OPEC and the FSU
1.4
Canadian oil production 19602030
1.5
Cumulative worldwide gas discoveries and production
1.6
Annual gas production 19202004 and extrapolation
based on a bell-shaped profile and an estimated
ultimate recovery of 12,000Tcf
1.7
Gas supply in the US
1.8
UK gas production 19702030
1.9
Natural gas supply of Europe Probable development
until 2020 according to scenario calculations by LBST
1.10 Gas production forecast for Russia
1.11 Worldwide gas production according to LBST
scenario calculations
1.12 Installed capacity of nuclear power plants and
various forecasts
1.13 World uranium reserves and cumulative uranium
consumption
1.14 Possible world coal production profiles based on 180 years
and 370 years of remaining reserves
1.15 High fossil scenarios of future production of fossil
and nuclear fuels
1.16 Low fossil scenario of future production of fossil
and nuclear fuels
1.17 Carbon dioxide emissions according to the production
profiles as outlined in the high fossil scenarios in
Figure 1.15
1.18 Principle of the logistic growth function and the meaning
of its parameters, P, T0 and b
1.19 Technical potentials of electricity production from
renewable energies
1.20 Possible market penetration of renewable energy sources
1.21 High fossil alternative world energy outlook until 2100
1.22 Low fossil alternative world energy outlook until 2100
1.23 Comparison between fossil fuel costs and renewable
installation costs
1.A.1 Cumulative installed capacity of wind energy converters,
geothermal power plants, photovoltaic modules and
solar collectors

5
7
10
11
15

16
17
19
20
21
22
23
24
25
26
27

28
29
30
32
34
35
36

43

prelims

2/7/07

3:51 PM

Page xiii

LIST OF FIGURES AND TABLES

1.A.2 Energy production from renewable sources


1.A.3 Primary energy from renewable sources
2.1
a) Layer structure showing three QWs surrounded
by barrier material; b) Band diagram for three QWs,
showing two confined states for electrons and holes;
c) Energy vs in-plane momentum dispersion relation for a QW
2.2
Density of states for a 3D bulk semiconductor g3D
and for a 2D quantum well structure g2D
2.3
a) QW p-i-n layer structure; b) Band diagram for the
QW p-i-n solar cell showing the photogeneration
and recombination processes, together with the carrier
capture and escape routes
2.4
Typical quantum efficiency for a GaAsP0.06/In0.1GaAs
75, 35 MQW structure at various biases
2.5
Quantum efficiency for an Al0.04GaAsP0.06/In0.1GaAs 75,
35 MQW showing how background impurities lead to a
collapse of the QE at forward bias
2.6
Light IV curve for a GaAsP0.06/In0.1GaAs 75, 35 MQW
structure under approximately 1 sun AM0 equivalent 3000K
tungsten halogen illumination
2.7
Plot of J0 vs absorption edge for a variety of lattice
matched, mismatched and MQW devices
2.8
Strained GaAs/InGaAs 3period MQW device:
a) Schematic diagram showing the layer configuration;
the relaxed lattice parameter is indicated laterally;
b) Electronic band-structure for the device
2.9
a) External quantum efficiency for a strained
GaAs/InGaAs MQW sample and GaAs control; b) Dark IV curves
for GaAs/InGaAs strain-balance MQW cell, a
relaxed GaAs/InGaAs MQW cell, a strained GaAs/InGaAs
MQW cell and a GaAs control
2.10 Strain-balance GaAsP/InGaAs 3period MQW device:
a) Schematic diagram showing the layer configuration;
b) Electronic band structure for the device
2.11 a) External quantum efficiency for the strain-balanced
GaAsP/InGaAs MQW sample and GaAs control; b) Dark IV curves
for GaAsP/InGaAs strain-balance MQW cell,
a relaxed GaAs/InGaAs MQW cell, a strained
GaAs/InGaAs MQW cell and a GaAs control
2.12 AM0 iso-efficiency contour plot for a monolithic
tandem structure
3.1
Lightweight and flexible thin film plastic solar cells
are very attractive for camping tents or any mobile units
3.2
Scheme of exciton Coulombic binding energy (in log scale)
versus exciton radius

xiii

43
44

47
48

50
51

52

53
54

56

57

58

59
62
75
76

prelims

2/7/07

xiv

3.3

3.4

3.5
3.6
3.7
3.8

3.9

3.10
3.11
3.12
3.13
3.14
3.15

3.16

4.1

4.2
4.3
4.4

3:51 PM

Page xiv

ADVANCES IN SOLAR ENERGY

First generation single layer type organic photovoltaic


cell or Schottky cell with: a) device structure; and
b) energy diagram
Second generation donor/acceptor bilayer type organic
photovoltaic cell or Tang cell with: a) device structure; and
b) energy diagram
Representative organic/polymeric electron donors
(p-type semiconductors)
Representative organic/polymeric electron acceptors
(n-type semiconductors)
Frontier orbital levels of representative polymeric
electron donors and acceptors
Third generation donor/acceptor blend type organic
photovoltaic cells or bulk heterojunction cells with:
a) device structure; and b) energy diagram
Fourth generation donor/acceptor bicontinuous phase
separated and nano-structured organic photovoltaic cell with:
a) device structure; and b) energy diagram
General scheme of an organic/polymer solar
cell with buffering layers
Chemical structure of PSS-PEDOT
Scheme of a DBAB- type of block copolymer
primary structure
Scheme of a potential DBAB- type of block copolymer
secondary structure
Scheme of a potential DBAB- type of block copolymer
tertiary structure
Scheme of molecular frontier orbitals and photo-induced
electron transfer as well as Dexter energy transfer
processes in a donor/acceptor binary light harvesting
system
Donor RO-PPV exciton quenching parameter, charge
recombination rate constant, and charge recombination
quenching parameter versus LUMO
offset of RO-PPV/SF-PPV-I pair
Conceptual drawing of an indirect integral solar
collector storage (ICS) system with a load-side immersed
heat exchanger for solar domestic water heating
Conceptual sketch of an immersed heat exchanger in a
vertical thermal storage tank
Three dimensional streamlines (top) and isotherms (bottom)
for a rectangular storage with a single immersed tube
Sketch of a divided indirect thermal storage vessel with
two storage compartments

78

79
79
80
80

82

83
85
85
89
89
90

92

93

101
102
106
109

prelims

2/7/07

3:51 PM

Page xv

LIST OF FIGURES AND TABLES

4.5

4.6

4.7
4.8
4.9
4.10

4.11

5.1
5.2
5.3
5.4
5.5
5.6
5.7

5.8

5.9
5.10
5.11

5.12

Predicted ratio of delivered energy of a divided storage


and an undivided storage as a function of
the dimensionless output volume and NTU
Measured ratio of delivered energy of a divided and
undivided storage as a function of ratio of the volume of
delivered hot water to the volume of storage fluid for
nominal heat exchanger NTU of 2.5 and 7
Creep compliance of PA66 in air and exposed to
hot chlorinated water
Tensile strength (top) and strain (bottom) of PA66 before
and after exposure to hot chlorinated water
SEM images of PA66: a) native unexposed surface; and
b) after 292 hours exposure at 830 40mV, 59 2C
SEM micrographs of scale formed on: a) polyamide 6,6;
b) polypropylene; c) polybutylene; and d) copper tubes
in a tube-in-shell heat exchanger
SEM micrographs of scale formed on PEX tubes after:
a) 5 hours; and b) 7.5 hours of exposure
to flowing hard water at room temperature
The various loss factors () affecting the
photon flux (I) inside a PTC photo-reactor
Structural schematic view of a DSSR reactor
showing water flowing in the transparent box
Schematic drawing of CPCs
Graphics related to the adjustment of data to an
L-H type kinetic model
Pesticide decomposition at different initial concentrations
Application of the proposed kinetic model for
mineralization of a pesticide mixture
Average UV direct and UV global (direct + diffuse)
irradiance for each month of the year at Plataforma
Solar de Almera, Spain
Overall dichloroacetic acid degradation rate (as TOC
disappearance reaction rate) comparison for a
concentrating (PTC) and a non-concentrating (CPC)
collector system
Plots of pesticide concentration as a function
of experiment time (top) and accumulated energy (bottom)
Schematic of the pilot plant batch operation
Schematic diagram of solar detoxification
demonstration plant constructed in SOLARDETOX
project at HIDROCEN, Madrid
Schematic diagram of the container washing
and photocatalytic water treatment plant concept

xv

110

111
115
116
117

121

122
133
135
138
139
140
142

143

143
153
154

157
159

prelims

2/7/07

xvi

5.13
5.14
6.1

6.2

6.3

6.4

6.5
6.6
6.7

6.8
6.9

6.10
6.11

6.12

6.13
6.14
6.15
6.16

3:51 PM

Page xvi

ADVANCES IN SOLAR ENERGY

Schematic diagram of solar detoxification demonstration


plant constructed by ALBAIDA at La Mojonera, Almera, Spain
Schematic of a solar treatment coupled with an
aerobic biological treatment
Emission of carbon dioxide over the period 17002000
according to direct atmospheric measurements and
ice probing
Emission of carbon dioxide in Australia over the period
19902000 and the emission level according to the
Kyoto target
Electrochemical cell showing the principle of
electrochemical hydrogen generation through water
photolysis using solar energy and related reactions
Electrical circuit representing the photo-electrochemical
device formed by a semiconducting photo-anode and
metallic cathode
The electrochemical chain of TiO2-x-based
photo-electrochemical device and related charge transfer
Band model of n-type semiconductor without
surface charge and involving surface charge
Effect of light in splitting of the Fermi energy level
into two quasi-levels corresponding to electrons and electron
holes for an n-type semiconductor
The schematic Jonker plot of thermoelectric
power vs log showing the critical parameters
Solar energy spectrum in terms of the number of
photons vs their energy, showing the flux density for
both commercial titania and reduced-band-gap titania
Solar energy spectrum in terms of the radiation
energy vs wavelength
Defect diagram showing the concentration of defects vs oxygen
partial pressure for undoped TiO2 at 1273K in absence of
foreign elements forming donors or/and acceptors
Effect of Fermi energy on the charge transfer
between the surface of semiconducting solid and
adsorbed species forming either acceptors or donors
Effect of cluster and grain size on electronic structure
and related band gap width
Circuit of the hybrid cell of Morisaki et al (1976) and
related electrochemical chain
Schematic illustration of the solar-hydrogen generation
system using space solar energy collector
Schematic illustration of the use of both solar-hydrogen
and solar-oxygen as input gases for the production of
electricity using a fuel cell

160
161

170

172

176

176
177
182

184
188

192
192

199

201
203
205
206

207

prelims

2/7/07

3:51 PM

Page xvii

LIST OF FIGURES AND TABLES

6.17
7.1
7.2
7.3
7.4
7.5
7.6
7.7
7.8
7.9
7.10
7.11
7.12
7.13
7.14
7.15
7.16
7.17
7.18

7.19
7.20
7.21
7.22
7.23
7.24
7.25
7.26
7.27

The concept of the Australian initiative of the


solar-methanol technology
Distribution of the heat demand by temperatures
Technical and economical potential of solar industrial
process heat
Distribution of the heat demand in the industry according
to different industrial sectors
Instantaneous efficiency for different solar collector types
Evacuated tube collector
Layout of a CPC type collector with tubular absorbers
Parabolic trough collector
Possibilities for the coupling of the solar system
with the conventional heat supply
Solar system without storage
Solar system with heat storage
Yearly energy yield delivered to a process for three sites,
depending on the process temperature
Heat costs for the three systems at different
process temperatures
Solar industrial process heat plants in operation
Short-term mismatch between heat demand and
available solar heat
Global solar radiation on a horizontal surface (H) on
the Iberian Peninsula
Costs of useful heat for the different systems as a
function of the mean annual working temperature
Combination of solar thermal system and waste heat recovery
Primary energy saving with respect to a conventional
steam boiler, for 1000MWh of industrial process heat
produced either by solar thermal systems or
conventional cogeneration system
Geographical location of the proposed demonstration plants
Heat demand by processes in the malting
factory in seville
Scheme of the proposed solar system for the malting
factory in Seville
Malting factory at Poceiro, Portugal: Ground layout
Layout of the proposed solar system for the malting factory at
Poceiro, Portugal
Solar radiation in Portugal and location of the Beiral plant
Scheme of the proposed solar system for the Beiral plant
Heat, cold and domestic hot water demand of the
installations of Bodegas Mas Martinet
Total heat demand of Bodegas Mas Martinet and
solar contribution throughout the year

xvii

210
219
219
220
221
222
223
224
225
226
226
228
229
238
240
241
242
244

244
247
248
249
250
250
252
252
253
254

prelims

2/7/07

xviii

7.28
7.29
7.30
8.1
8.2
8.3
8.4
8.5
8.6
8.7
8.8

8.9

8.10

8.11

8.12

8.13
8.14
8.15
8.16
8.17
8.18
8.19
8.20

3:51 PM

Page xviii

ADVANCES IN SOLAR ENERGY

Scheme of the proposed solar system for Bodegas Mas Martinet


Autoclave process (simplified)
Solar field integration via Ruth-type (saturated steam) storage
Countries of the EU-MENA region analysed within this study
Annual direct normal irradiance for 2002
Gross hydropower potentials in EU-MENA
Temperature at 5000m depth for hot dry rock
geothermal power technology
Map of biomass productivity
Annual average wind speed at 80m above ground level
Annual global irradiation on surfaces tilted south with latitude angle
Domestic electricity demand in MENA and electricity
supplied by existing power stations, by new power plants
and by CSP for domestic consumption, export and seawater
desalination for the scenario Closing the Gap
Growth rate of CSP production in MENA in the scenario
Closing the Gap including domestic power supply,
export electricity and seawater desalination
Domestic electricity demand in North Africa and
electricity supplied by old power stations, by new power
plants and by CSP for domestic consumption, export and
seawater desalination for the scenario Closing the Gap
Electricity demand in western Asia and electricity
supplied by old power stations, by new power plants
and by CSP for domestic consumption, export and
seawater desalination for the scenario Closing the Gap
Electricity demand of the Arabian Peninsula and
electricity supplied by old power stations, by new power
plants and by CSP for domestic consumption, export
and seawater desalination for the scenario Closing the Gap
Water demand structure in MENA and its evolution
until 2050; scenario Closing the Gap
Water supply from sustainable sources and deficits in MENA
a) Water consumption growth rates in MENA;
b) Water consumption per capita in MENA
Water demand structure in North Africa and its evolution until 2050
Water demand structure in western Asia and its evolution until 2050
Water demand structure for Arabian Peninsula and its evolution
until 2050
Groundwater withdrawals as percentage of safe yield for
selected countries
Water demand, sustainable freshwater resources,
non-sustainable supply and potential future supply by
CSP via cogeneration by MED plus direct generation
via solar electricity in RO plants

255
256
256
262
268
269
270
271
272
273

274

275

278

279

280
286
287
288
289
290
291
292

293

prelims

2/7/07

3:51 PM

Page xix

LIST OF FIGURES AND TABLES

8.21

8.22
8.23

8.24
8.25
8.26
8.27
8.28

8.29
8.30

Covering the future freshwater deficits in MENA by


transitory non-sustainable sources and by CSP
plants using MED in cogeneration plus solar
electricity for RO
Capacity potential for CSP desalination plants with
MED and RO in the three main MENA regions
CO2emissions of electricity generation in million tons
per year for all countries for the scenario CG/HE and
emissions that would occur in a business as usual case
Annual per capita CO2 emissions of power generation (Scenario CG/HE)
Annual electricity demand and generation within the
analysed countries in the MED-CSP scenario
Installed power capacity and peak load within the
analysed countries in the scenario CG/HE
Share of different technologies for electricity generation in 2000
Total electricity consumption and share of different technologies
for electricity generation in the analysed countries in 2050
according to the MED-CSP scenario
Example of electricity costs and learning in the MED-CSP scenario
Projection of a future trans-Mediterranean grid
interconnecting the best sites for renewable
energy use in EU-MENA

xix

294
294

299
299
300
300
301

301
302

303

TABLES
1.1
1.2
2.1

4.1

6.1

7.1
7.2
7.3
7.4
7.5

Energy consumed for the construction of new power plants


Projected cost reductions of renewable energy technologies
AM0 Junction parameters for a GaInP top cell
together with parameters for lower junction cells
composed of GaAs p/n, strain-balanced MQW,
DBR enhanced strain-balanced MQW, strained
MQW and relaxed MQW
Scale accumulation and scaling rate for tubes
in a tube-in-shell heat exchanger as determined
by chemical analysis
Effect of Cr on the energy conversion efficiency
(ECE) of a photo-electrochemical cell involving
photo-anode made of Cr-doped TiO2
Overview of the industries studied in the POSHIP project
Potential for solar industrial process heat in the Spanish industry
Total investment costs for the solar collector field related to the
gross collector area
List of solar process heat plants
Selection of solar collector types according to
working temperatures

32
35

63

120

202
218
220
227
231
243

prelims

2/7/07

xx

7.6
7.7
7.8
7.9
7.10
7.11
7.12
8.1

8.2
8.3

8.4

8.5

8.6

8.7
8.8
8.9
8.10
8.11
8.12

3:51 PM

Page xx

ADVANCES IN SOLAR ENERGY

Evaluation criteria for solar industrial process heat


systems for the Iberian Peninsula
List of the most relevant parameters of the
demonstration projects proposals
Characteristic data of the proposed system for
solar air heating in a malting factory in Seville, Spain
Characteristic data of the proposed system for air
preheating for malt drying in Portugal
Characteristic data of the proposed
system for a textile drying process
Characteristic data of the proposed system for space
heating and cooling of a wine cellar
Annual system performance for different sites
Renewable electricity performance indicators representing
the average renewable electricity yield of a typical facility
in each country
Economic potentials of renewable energy sources in
the southern EU and MENA region
Electricity demand in MENA and generation by
conventional old and new plants and by CSP technologies
in MENA until 2050
Domestic electricity demand in North Africa and
generation of domestic power, export electricity
and power for desalination by CSP technologies until 2050
Domestic electricity demand in western Asia and
generation of domestic power, export electricity and
power for desalination by CSP technologies until 2050
Domestic electricity demand of the Arabian Peninsula
and generation of domestic power, export electricity
and power for desalination by CSP technologies until 2050
Annual per capita water consumption in 2000 and
projected demand in 2010 for the GCC countries
Solar resources in the GCC countries
Comparison of energy consumption for six types
of water desalination
The available desalination capacity in the GCC countries,
in 2000, using different desalination technologies
Global solar radiation, monthly and annual means
Areas required for renewable electricity generation in
2050 for the scenario CG/HE

245
247
249
251
253
255
257

267
267

276

278

279

281
282
283
283
284
296
297

prelims

2/7/07

10:28 PM

Page xxi

About the Authors


Diego C. Alarcn Padilla is a project researcher at the Centro de Investigaciones
Energticas, Medioambientales y Tecnolgicas (CIEMAT) at Plataforma Solar de Almera
(PSA) and has been working on different solar thermal research topics since 1994. Since
2000 he has been working within the PSA environmental applications of solar energy
area, with a special involvement in solar thermal seawater desalination projects. His first
research and development work was on computer simulation of optical devices for solar
thermal energy applications and software development for sun-tracking devices. Dr.
Alarcn is also currently a lecturer in computer science at the University of Almera,
Spain.
W. E. Alnaser holds a Ph.D. in physics and has been Professor of Applied Physics at the
University of Bahrain since 1997, where he served as the Dean of the College of Science
and also the Dean of Scientific Research. His research work is diversified but
concentrated in solar radiation modelling, measurement and application. Prof. Alnasar
devised a small mobile solar and wind energy electricity generator (2.5kW) as well as
participated in constructing a mobile solar reverse osmosis desalination unit (250 gallons
per day). He is the author of more than 100 refereed articles on material science and
superconductivity, renewable energy, astronomy and astrophysics, and environmental
physics and 30 books on physics, renewable energy and astronomy. He is currently
serving as Chairman for the Arab Section at the International Solar Energy Society
(Germany), as well as President for the International Energy Foundation for Middle East.
He is a Fellow of the Institute of Physics in London.
Professor Alnaser has received many awards at local and professional levels, including
the Shuman Prize (Jordon) for Best Youth Physicist, the Islamic Educational, Scientific and
Cultural Organization (ISESCO) Prize for Best Research, and the State Prize (Bahrain) for
Community Science Services.
He is the past editor-in-chief of the Journal of Arab Association Universities for Basic
and Applied Sciences and is now its managing director. He has been a referee for more
than 12 leading international journals and currently serves as associate editor for two
international journals.
Tadeusz Bak is a senior research fellow at the Centre for Materials Research in Energy
Conversion, University of New South Wales, Sydney, Australia. He received his Ph.D. from
the Academy of Mining and Metallurgy, Cracow, Poland in 1982. In 1996 he joined the
School of Materials Science and Engineering, University of New South Wales as a
research fellow. His expertise includes electrochemistry and materials science. His
current research interests include diffusion in ionic solids and photo-electrochemistry of
metal oxides. He has published over 100 refereed papers and 17 book chapters.
Julin Blanco Glvez earned his Ph.D. at the University of Almera in 2002 following
specialization in industrial engineering at Seville University. Dr. Blanco is a senior
researcher at the Spanish Ministry of Education and Science. Since 1995 he has served

prelims

2/7/07

xxii

3:51 PM

Page xxii

ADVANCES IN SOLAR ENERGY

as both the head of the Solar Chemistry Department of Plataforma Solar de Almera (PSA),
and the Spanish National Representative in the Task II group of the International Energy
Agency SolarPACES (Solar Power and Chemical Energy Systems). Dr. Blanco has been
involved in many national and contractual research and development projects, as well as
work for the European Union related to the development of solar technologies as applied
to waste water treatment processes and techniques. Dr. Blanco is author of 1 book and
co-author of 3 books and 8 chapters in different books, some 40 publications in indexed
international journals and more than 90 contributions to international congress and
symposiums. He has been honoured with the Jurys Grand Prix at the European Grand Prix
for Innovation Awards, 2004, in Monaco (www.european-grandprix.com/index_en.htm).
Maria Joo Carvalho has worked in solar thermal energy since the 1980s and is
currently the senior researcher in the Renewable Energy Department of the National
Institute for Engineering, Technology and Innovation (INETI), a state laboratory in
Portugal. She is the director of the Solar Collector Testing Laboratory of INETI. Dr.
Carvalho is the author of 19 papers published in international technical journals and 38
papers published in conference proceedings. She has given invited lectures on
renewable energy for universities and also seminars on the subject of solar thermal
energy with special emphasis on collector and system testing. She is a member of the
Portuguese Section of ISES.
Jane H. Davidson has been a professor in the Department of Mechanical Engineering at
the University of Minnesota since 1993. Before coming to Minnesota, she was a faculty
member at the University of Delaware and Colorado State University. Dr. Davidson is past
editor-in-chief of the Journal of Solar Energy Engineering. She was the 2004 recipient of
the ASME John I. Yellott Award for outstanding research in solar energy and is a fellow of
the American Society of Mechanical Engineering and the American Solar Energy Society
(ASES). Her publications include more than 150 papers in journals and conferences and 3
book chapters. She was the 2005 recipient of the University of Minnesotas Distinguished
Woman Scholar Award in Science and Engineering. Her research interests include solar
thermal systems and energy efficient buildings.
N. J. Ekins-Daukes is a lecturer at the University of Sydneys School of Physics, having
earned his Ph.D. from Imperial College, London in 2000 for his work on strain-balanced
quantum well solar cells. He later held a fellowship from the Japan Society for the
Promotion of Science (JSPS) at the Toyota Technological Institute in Nagoya, Japan. Dr.
Ekins-Daukes, present research involves the application of nano-structures to high
efficiency photovoltaic devices as well as contributing towards projects on the radiation
resistance of III-V solar cells and the application of III-V solar cells in concentrator systems
and the application of nano-structures to photovoltaic devices. Dr. Ekins-Daukes is a
member of the Institute of Electrical and Electronics Engineers (IEEE). He has published
papers on a range of topics covering quantum well solar cells, luminescent up-conversion,
radiation resistance and photovoltaic concentrator systems.
Joo A. Farinha Mendes is a mechanical engineer (thermodynamics), currently working
as a senior researcher and is the Head of the Solar Energy Unit in the Renewable Energy

prelims

2/7/07

3:51 PM

Page xxiii

ABOUT THE AUTHORS

xxiii

Department (DER) at the National Institute for Engineering, Technology and Innovation
(INETI), a state laboratory in Portugal. His main research activity has been devoted to solar
thermal applications, serving as national representative for several European projects,
leading to a number of papers on non-imaging optics, phase change heat transfer and
solar system design published in international technical journals and conference
proceedings. He is the Portuguese representative for the International Energy Agencys
Executive Committee on Solar Heating and Cooling Implementing Agreement and has
participated in the standardization work of solar collectors at the national, European and
international levels. He has given invited lectures on solar thermal energy for universities
and participated in numerous seminars in Portugal and other countries in Europe.
Pilar Fernndez Ibez earned her Ph.D. in applied physics from the University of
Granada in 2004, following specialization in environmental sciences and applied physics
at the University of Almera, Spain. Dr. Fernndez has ten years of work experience in
various research sectors, participating in ten European Union and two national research
and development projects related to disinfection and wastewater treatment with the help
of solar technologies development. She is co-author of three books and six chapters in
other collaborative books. She holds two patents, and is co-author of 30 international
technical journal articles and 50 contributions to different international congresses and
symposiums. She has tutored the scientific work of students on solar disinfection and
detoxification of wastewater. Dr. Fernndez has also been Invited Guest Associate Editor
for special issues of the indexed journals Catalysis Today and Solar Energy.
Lorraine F. Francis is a professor of chemical engineering and materials science at the
University of Minnesota. She joined the University in 1990 and currently serves as the
Director of Undergraduate Studies in Materials Science and Engineering. Dr. Francis
received the National Science Foundation (NSF) Young Investigator Award (19931998)
and the DuPont Young Professor Grant (19941997). Her research interests include
polymer and ceramic coatings, processing-microstructure-property relationships in
coatings and composites, interfacial studies and biomaterials. Dr. Francis is a member of
the American Ceramic Society (ACerS), the Materials Research Society (MRS) and ASM
International.
Wolfgang Gernjak earned his Ph.D. in land and water management from the University of
Natural Resources and Applied Life Sciences (BOKU) in Vienna, following specialization in
analytical and physical chemistry at the Vienna University of Technology. He has been
involved in five European Union research and development projects connected with
the application of solar photocatalysis and solar photo-Fenton for the detoxification of
industrial wastewater and disinfection of drinking water. He is author of 20 articles
in indexed international journals and 33 contributions to international congress and
symposiums.
Klaus Hennecke earned his Doctor of Engineering degree in aerospace technologies in
1978 at the University of the Federal German Forces, Munich, Germany. In 1989 he
became a manager of a series of research projects in the Solar Research Division of the
Institute of Technical Thermodynamics (ITT) within the German Aerospace Center (DLR) in

prelims

2/7/07

xxiv

3:51 PM

Page xxiv

ADVANCES IN SOLAR ENERGY

Cologne, Germany. His research is particularly involved with the field of direct steam
generation in parabolic trough collectors and solar process heat applications. As head of
the applications subdivision Dr. Hennecke is currently responsible for the transfer of
research and development results into the market. Engaged in international cooperations
under the umbrella of the International Energy Agency (IEA), implementing the Solar
Power and Chemical Energy Systems (SolarPACES) agreement, Dr. Hennecke leads the
subtask System Integration and Demonstration of the new IEA SHC Task 33/SolarPACES
Task 4 on Solar Heat for Industrial Processes (SHIP).
Gerhard Knies holds a Ph.D. in physics and has worked in the fields of high energy
physics and elementary particle research at Deutsches Elektronen Synchrotron (DESY)
Hamburg, Germany; the European Centre for Nuclear Research (CERN) in Geneva,
Switzerland; and the Stanford Linear Accelerator Center (SLAC) in both Stanford
and Berkeley, California, US. Dr. Knies developed a road map towards energy, water
and climate security by use of solar energy. In 1981 he co-founded the Hamburg Scientists
Peace Initiative, a disarmament campaign against the threatening EastWest arms race,
and made studies on the vulnerability of industrial civilization. In 1995 he founded
the Hamburg Climate Protection Foundation, which in 2003 allied with other groups
to form the Trans-Mediterranean Renewable Energy Cooperation movement for
development, climate stabilization and good neighbourhoods. Dr. Knies has been retired
since 2001.
Dirk Krueger earned his B.Sc. in energy and environmental engineering at the University
of Applied Sciences in AachenJuelich, Germany. Since 1997 he has specialized in
investigating parabolic trough collector systems for process heat applications.
Dr. Kruegers major research fields are collector testing, simulations for output forecasts
and project assistance for companies engaged in parabolic trough technology.
Sixto Malato Rodrguez is a senior researcher of the Centro de Investigaciones
Energticas, Medioambientales y Tecnolgicas (CIEMAT) at Plataforma Solar de Almera
(PSA), Spain, and has 18 years of experience in various research sectors. Since 1990 his
work at the PSA-CIEMAT has involved all the European Union research and development
projects linked to the solar detoxification of water and solar wastewater treatment
technologies. He is author of 1 book and co-author of 5 books as well as 21 chapters in
others. Dr. Malato has also co-authored more than 80 publications in indexed international
journals, 21 articles in technical journals and more than 140 contributions to 68 different
international congresses and symposiums and holds 4 patents. In 2004 he received the
Jurys Grand Prix at the European Grand Prix for Innovation Awards in Monaco
(www.european-grandprix.com/index_en.htm).
Manuel I. Maldonado Rubio earned his Ph.D. from the University of Almera
in 2001, following specialization in environmental sciences and chemistry at the University
of Granada and the Instituto de Investigaciones Ecolgicas in Mlaga, Spain.
Dr. Maldonado has worked for the the Centro de Investigaciones Energticas,
Medioambientales y Tecnolgicas (CIEMAT) at Plataforma Solar de Almera, Spain, since
2002, and has been involved in seven European Union research projects, three national

prelims

2/7/07

3:51 PM

Page xxv

ABOUT THE AUTHORS

xxv

research projects and three research and development contracts related to the
development of solar technologies applied to wastewater treatment. He has authored 29
peer-reviewed articles for international publications, co-authored 1 book and 5 chapters in
others, and given presentations at 52 international and national congresses.
Susan C. Mantell is a professor of mechanical engineering at the University of Minnesota.
Dr. Mantell has received several awards for excellence in teaching and research including
a National Science Foundation (NSF) Young Investigator Award (19941999) and a
McKnight Land-Grant Professorship (19951997). She is a leading expert in the field of
composite materials and manufacturing and most recently has been working on
development of polymer heat exchangers for solar energy applications and for use in the
transportation industry. Dr. Mantell has served as an associate editor of the Journal of
Composite Materials since 2000 and the Journal of Materials Processing and
Manufacturing Science from 1998 to 2003. She has published over 75 papers and 2 book
chapters on composite materials and polymer processing.
Janusz Nowotny received his Ph.D. in solid-state chemistry from the Institute of Physical
Chemistry, Polish Academy of Science, in 1967, and following specialized research in
diffusion in nonstoichiometric compounds earned a D.Sc. in materials science from the
Academy of Mining and Metallurgy, Cracow, Poland in 1974. Dr. Nowotny is Director of
the Centre for Materials Research in Energy Conversion, University of New South Wales,
Sydney, Australia. He has been a visiting professor at several universities in France,
Germany and Japan. Dr. Nowotny s expertise includes photo-electrochemistry, solid-state
electrochemistry, defect chemistry and the science of materials interfaces. His research
includes defect chemistry of oxide semiconductors, diffusion in ionic solids, segregation
in nonstoichiometric compounds and photo-electrochemistry. He has published over 400
refereed papers, and edited 17 books.
Jrg Schindler is a business economist and since 1984 has worked for the LudwigBlkow-Systemtechnik GmbH, becoming its managing director in 1992. His major
activities include the technical and economic analyses of market introduction of
photovoltaics and other renewable energy sources, clean drive systems for road
transport, clean fuels produced from renewable energy sources (like hydrogen), and the
future availability of fossil fuels. Mr. Schindler is currently a member of the board of
e5 The European Business Council for a Sustainable Energy Future and from 1999 to
2003 was a member of the Enquete Commission of the Bavarian House of
Representatives entitled New Energy for the New Millennium. Enquete Commissions
serve the national government by collecting scientific state-of-the-art information and then
transmitting it through publicly discussed booklets. He is also a member of the board of
the Global Challenges Network.
Hans Schweiger is director of the thermo-energetical systems and renewable energies
company energyXperts.BCN, Barcelona, Spain. Since 1991, Dr. Schweiger has been active
in the field of solar thermal energy. He has worked in collaboration with several national
and international research projects in the fields of solar thermal energy (solar collector
development, transparent insulation, building simulation and absorption cooling) and

prelims

2/7/07

xxvi

3:51 PM

Page xxvi

ADVANCES IN SOLAR ENERGY

numerical methods for heat transfer and fluid dynamics (software development for CFD
and radiation heat transfer). He served as project coordinator of the European Union
research project on solar industrial process heat, POSHIP (NNE5-1999-0308).
Dr. Schweiger also teaches masters and post-graduate courses on solar thermal energy
at the Catedra UNESCO (Polytechnical University of Catalonia). He was co-founder of the
engineering cooperative Aiguasol Engineering (Barcelona, Spain), where he was active
from 1999 to 2005. He earned his degrees from the Politechnical University of Catalonia
(UPC, Terrassa, Barcelona, Spain), specializing in the optimization of solar thermal
absorber elements with transparent insulation.
Leigh Sheppard is a postdoctoral fellow at the Centre for Materials Research in Energy
Conversion, School of Materials Science and Engineering, University of New South Wales,
Sydney, Australia. In 2006 he received the Deutsche Akademische Austausch Dients
award of the Hahn-Meitner Institute to work with Professor H. Tributsch.
Dr. Sheppards research interests include photo-electrochemistry of TiO2. He has
published 16 refereed papers and 2 book chapters.
Charles C. Sorrell is Professor of Ceramic Engineering, School of Materials Science
Engineering, University of New South Wales, having earned his Ph.D. in ceramic
engineering at the University of New South Wales in 1987. His main research interest areas
include ceramic processing, phase equilibria, microstructure and performance. Professor
Sorrells other research activities are focused on titania-based materials for
environmentally friendly applications, bioceramics and microwave heating, glass-ceramics,
refractories, high temperature superconductors, graphite, silicon nitride, zirconia, and
gemstone heat treatment. Professor Sorrell is a fellow of the Australian Institute of Energy
(AIE) and the Institution of Engineers/Australia (IEAUST). He has been editor of the Journal
of the Australasian Ceramic Society, Advances in Applied Ceramics, Metals and Materials
International, and Ceramics International. Professor Sorrell has published over 350 papers
and supervised 75 B.Sc. projects, as well as 35 postgraduate students.
Sam-Shajing Sun obtained his Ph.D. degree in polymer/materials chemistry from the
University of Southern California in 1996. After a postdoctoral appointment at the Loker
Hydrocarbon Research Institute, Dr. Sun joined the Norfolk State University in early 1998.
He is currently leading an organic and polymeric materials research programme at the
Center for Materials Research (CMR) at Norfolk State University. Dr. Sun is also directing
a Center for Research and Education in Advanced Materials sponsored by NASA. Dr.
Suns research interests include the moulding, design, synthesis, processing and
characterization of novel organic and polymeric electronic and photonic materials.
Current research projects of his lab focus on the development of novel organic/polymeric
photovoltaic materials for potential plastic solar cells and photodetector applications.
Dr. Sun has published more then 50 refereed papers and presented more then 40 papers
and lectures relevant to organic/polymeric optoelectronic materials at international
scientific communities and academic/governmental organizations. He has recently
edited a new CRC Press book titled Organic Photovoltaics: Mechanisms, Materials and
Devices, a comprehensive book covering the state of the art of the field. Dr. Sun is a
member and technical referee of a number of major scientific organizations.

prelims

2/7/07

3:51 PM

Page xxvii

ABOUT THE AUTHORS

xxvii

Franz Trieb has worked in the field of renewable energies since 1983. After the
implementation of hydrogen storage for an autonomous renewable energy system at the
University of Oldenburg, Germany, he obtained specialized training in renewable energy
at the National University of Tacna, Peru. Since 1994, Dr. Trieb has worked as project
manager at the Institute of Technical Thermodynamics of the German Aerospace Center
(DLR), working on solar energy resource assessment by satellite remote sensing, market
strategies for concentrating solar power and renewable energy scenarios.
Two of Dr. Triebs recent studies look at the future demand and the sustainable supply
of electricity using the potential of renewable energy sources in the countries of the
Mediterranean Region and Europe (www.dlr.de/tt/med-csp and www.dlr.de/tt/trans-csp).
Werner Zittel obtained his Ph.D. in physics in 1987 from the Max-Planck Institute for
Quantum Optics, Garching, and at the Technical University of Darmstadt, both in Germany.
He has worked as a scientific collaborator for the German Aerospace Center (DLR) and at
the Institute for Technical Physics of the German Research Center for Aircraft and Space
Technology (DFVLR, today DLR) in Stuttgart and at the Fraunhofer Institute for Solid State
Technology, in Munich, Germany. Since 1989, he has worked for Ludwig-BlkowSystemtechnik GmbH, Ottobrunn, Germany.
Dr. Zittel actively researches climate change effects on the trace gas methane
(especially in Russia); the external and social costs of energy use in the public transport
sector; the energetic amortization of renewable energy generation technologies;
advanced hydrogen storage technologies; the use of hydrogen in existing natural gas
grids; hydrogen production from biomass; analysis of the existing hydrogen markets;
introduction scenarios for hydrogen into the energy and transport economy; fundamental
questions of the power industry; analysis of fossil resource availability; and advanced
hydrogen storage systems for automotive use.
He is a member of the Association for the Study of Peak Oil, Germanwatch e.V., and
Global Challenges Network e.V.

prelims

2/7/07

3:51 PM

Page xxviii

ch-01

2/6/07

7:36 PM

Page 1

Alternative World Energy Outlook


2006: A Possible Path towards
a Sustainable Future
Jrg Schindler and Werner Zittel

Abstract
This paper sketches a possible path towards the future energy supply in the light of
foreseeable challenges and opportunities. First, the shortcomings of the World Energy
Outlook, which is biennially published by the International Energy Agency, are revealed.
The deficiencies of these reports are the reason why a more progressive approach and
alternative scenarios are needed. An alternative approach must address both the
challenges as well as promising pathways towards a sustainable energy future. Second,
we briefly address the challenges and discuss those non-sustainable approaches which
represent todays conventional wisdom in the energy community. Finally, and most
important, we want to describe the real potential of renewable energies in a scenario
called Alternative World Energy Outlook 2006.
It is not the intention of this paper to draw a complete and exact picture of what our
energy future will look like, as every attempt to do so is certain to fail. However, the aim
is to describe the limits of what is possible and what is impossible simply due to
geological and physical restrictions. Most relevant in this respect is to take into account
the peaking of oil production in the near future and the peaking of natural gas production
in the mid-term future.

Keywords World Energy Outlook; renewable energy; climate change; economic principles; logistic
growth

1.1 WINDS OF CHANGE: THE TRANSITION PERIOD


The main thesis of this paper is that we are at the beginning of a structural change of our
economic system. This change will be triggered by declining fossil fuel supplies and will
influence almost all aspects of our daily life.
Climate change will also force mankind to change the energy consumption pattern
away from fossil fuel combustion. This is a very serious problem. However, the focus of
this contribution is on resource depletion aspects, as these are much less transparent to

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

7:36 PM

Page 2

J. SCHINDLER AND W. ZITTEL

GLOSSARY OF TERMS USED IN THIS PAPER


API
ASPO
AWEO
bbl
boe
BGR
BP
BTU
CO2
DoE
DTI
EIA
EUR
EWEA
FSU
Gb
GoM
Gtoe
GW
IAEA
IEA
IHS
IPCC
LBST
LNG
M
Mb
Mtoe
NEB
NPD
OECD
OPEC

R/P
SO2
SOT
t
toe
Tcf
UK
USA
WEO
yr

American Petroleum Institute


Association for the Study of Peak Oil
Alternative World Energy Outlook
barrel (1 barrel = 159 litres)
barrel oil equivalent
Bundesanstalt fr Geowissenschaften und Rohstoffe (German federal agency
for resources)
company name (originally standing for British Petroleum)
British thermal unit
carbon dioxide
Department of Energy
Department of Trade and Industry, UK
Energy Information Administration
estimated ultimate recovery
European Wind Energy Association
former Soviet Union
gigabarrel (=109 barrels)
Gulf of Mexico
gigatons of oil equivalent (= 109 tons)
gigawatt (= 109 watts)
International Atomic Energy Agency
International Energy Agency
IHS-Energy, company name
Intergovernmental Panel on Climate Change
Ludwig-Blkow-Systemtechnik GmbH, company name
liquefied natural gas
million
million barrels
megatons of oil equivalent (= 106 tons)
National Energy Board, Canada
National Petroleum Directorate, Norway
Organisation for Economic Co-operation and Development
Organization of Petrol Exporting Countries (OPEC member states are:
Algeria, Indonesia, Iran, Iraq, Kuwait, Libya, Nigeria, Qatar, Saudi Arabia,
the United Arab Emirates and Venezuela)
reserve to production (ratio)
sulphur dioxide
solar thermal power plants
ton
ton oil equivalent (1 toe = 7.1 boe)
teracubic feet (= 1012 cubic feet)
United Kingdom
United States of America
World Energy Outlook
year

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

7:36 PM

Page 3

Alternative World Energy Outlook 2006

the public. The following scenarios try to figure out if the present resource depletion
problem eventually could be mastered just by increasing the extraction rate of other
conventional energy sources.
This transition period probably has its own rules which are valid only during this phase.
Things might happen which we have never experienced before and which we may never
experience again once this transition period is over. Specifically our way of dealing with
energy topics may change completely, which will have consequences for our economic
systems.
The International Energy Agency (IEA) denies that such a fundamental change of our
energy supply is likely to happen in the near or medium-term future and therefore does
not give a warning that our economic system is in danger. Therefore the views of the IEA
are briefly analysed. It will be shown that the business as usual approach of the biennially
published World Energy Outlook does not describe our energy future appropriately.
After that the main drivers determining the transition to a changing energy future are
discussed. It is important to understand that we are entering such a transition period,
leading to fundamental changes. The acknowledgement of this development and the
resulting mindset are necessary preconditions for coping with this situation in an
appropriate way. This will lead to completely different behaviour on the part of individuals,
companies and governments. The imminent transition is not a voluntary act in which
people might or might not engage because changing boundary conditions will force us to
adapt our energy system and also our way of life.

1.1.1 CRITIQUE OF THE WORLD ENERGY OUTLOOK 2004/2005


The IEA biennially publishes a World Energy Outlook (WEO) describing the most probable
development of energy supply and demand over the next 20 to 30 years.1 These outlooks
differ only in minor details from year to year and follow this general pattern:
Demand is projected by extrapolating past experience. According to these demand

projections, energy consumption will rise tremendously until 2030. There will be no
sign of rising supply problems or an imminent peak of oil production.
Fossil fuel use will dominate the supply. Renewable energy sources (apart from
hydropower or traditional biomass use) will contribute only a minor share and will
only rise marginally over the next 25 years. Even by 2030 the contribution will still be
only about 2 per cent.
A more detailed analysis reveals that the secured supply of this rising demand over the
next 30 years relies on a few basic assumptions which are not questioned by the IEA.2, 3
The main assumptions are:
On the one hand it is claimed that oil resources are sufficient to supply the projected

rise of consumption. At the same time, the IEA recognizes that proved reserves are
not sufficient and new reserves must be discovered. In addition, it is stated that the
uncertainty and poor quality of data is of growing concern to all involved in the oil
industry. Even with this background, the IEA concludes that growth of oil

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

7:36 PM

Page 4

J. SCHINDLER AND W. ZITTEL

consumption is possible and likely disregards the uncertainties. More detailed


analyses by critical observers (for example by the Association for the Study of Peak
Oil) show that it is much more probable that the existing production basis is
shrinking and the peak of oil production is very close. The evidence provided by
independent oil geologists and analysts is not discussed seriously by the IEA and is
more or less ignored.
The IEA states that many more financial resources than in the past will have to be
invested in the upstream sector in order to compensate for the rising decline of the
existing production base. Thus the future availability of sufficient resources is
reduced to a purely financial problem rather than lack of resource base. At the same
time the IEA observes that it is very uncertain whether oil companies and oilproducing states will increase their upstream investments sufficiently.
As a consequence of the assumed rise of fossil fuel supply, the future growth of
renewable energies is projected based on unchanged conditions and thus assuming
no need to raise their share. Even 25 years from now the IEA foresees only a
marginal contribution from renewable energy sources. However, it must be noted
that all previous WEO reports have considerably underestimated the actual growth of
wind energy. The now projected future growth rates still fall far behind the
experienced rise of all renewables in the past 1020 years.
The IEA approach shows the fundamental difference between observers influenced by
economic reasoning and those influenced by geological reasoning: the economic
observers completely ignore the influence of geological restrictions, instead seeing
possible future supply problems simply as a consequence of insufficient investments in
the upstream sector.
In the latest WEO report, the IEA argues that oil prices will decline as soon as oil
companies invest more money in the upstream sector. Independent of whether oil
resources are sufficient or not, it is hard to imagine that oil companies will invest more
money to reduce prices.
Regarding future oil prices, many observers have changed their views over the last
year. Hardly anybody still believes that future oil prices will return to low levels of below
US$4050/bbl.
In the following paragraphs some results of the latest two WEOs are discussed.
Figure 1.1 shows the projections of the WEO as published by the IEA in 2004 and 2005.
It shows the development of the primary energy supply since 1965 and the forecast for the
next 25 years (until 2030). The fossil energy sources (oil, coal and natural gas) together
provide almost 80 per cent of present world energy needs. Oil is the most important fuel with
a share of almost 40 per cent. Biomass, mostly used in an unsustainable manner for cooking
and heating in developing countries, covers between 10 and 15 per cent of the world energy
demand. Nuclear and hydropower produce roughly the same amount of electricity at world
level, each with a share of about 18 per cent of world electricity production. In the graph the
share of nuclear energy is exaggerated due to the primary energy supply calculation method,
which assumes an efficiency of 33 per cent for primary nuclear fuel to electric power
provision and almost 100 per cent for hydropower to electricity. Finally, the smallest

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

7:36 PM

Page 5

Alternative World Energy Outlook 2006

Mtoe
6000
oil
5000

2005

WEO 2004
WEO 2005
WEO 2005-altern.
WEO 2005-low invest
gas

4000

coal
3000
2000
biomass
1000
Nuclear (x3)
0
1965

Other-WEO2005-alt
Other-WEO2005
Other-WEO2004

hydro

1985

2005

2025

Year
Source: Historical data BP Statistical Review of World Energy Outlook International Energy Agency 2004/2005

FIGURE 1.1 World energy demand 19712030 as given in the IEAs World Energy Outlook 2004 and 2005

contribution, with about 1 per cent is derived from all other renewables such as wind, direct
solar energy, geothermal energy and wave power. The contribution from these sources might
rise to a share of 2 per cent in 2030, according to the WEO projection.
The graph tells us that oil will remain the most important fuel. Its consumption and
production levels will increase over the next 25 years by the same amount as they did
during the last 40 years. The message to consumers is that in the next two decades we
will see no major changes. Not shown in the graph, the price of oil is assumed to rise to
US$27/bbl by 2030 in WEO 2004 and to US$39/bbl in WEO 2005.
The latest WEO 2005 report also discusses a low investment scenario and an
alternative policy scenario. The projected oil consumption for these two alternative
scenarios is also shown in Figure 1.1. According to the report, the low investment
scenario would result in an oil price rise of up to US$52/bbl by 2030. In contrast to these
scenarios, however, oil was traded at US$6070/bbl for several months in 2005 and early
in 2006 and has risen above US$75/bbl.
Another IEA message is that natural gas and coal supplies will compete for the second
most important position behind oil, with natural gas supplies probably developing more
rapidly than coal: the supply of natural gas is expected to more than double within the
next 25 years.
At present, North America and Europe are the biggest natural gas markets, consuming
more than half of the worlds natural gas production. In both regions natural gas
production has already peaked. North America is experiencing a decline of its domestic
supply. Only very small amounts of natural gas are imported from foreign sources due to
limited import capacities. There is no doubt that large quantities of natural gas are still in
the ground. However, it is doubtful that future global extraction rates will exceed past

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

7:36 PM

Page 6

J. SCHINDLER AND W. ZITTEL

levels when production in North America and Europe is already in decline. Any new natural
gas supplies must first close this gap before there can be a net increase. This may turn
out to be a big challenge!
A few comments can also be made on the other energy sources. The growth of biomass
use is supposed to continue as in the past. On nuclear energy the report spends just 1 out
of more than 400 pages, claiming without any further justification that total production will
remain almost at present levels. New renewable energies like wind or solar will almost triple
their contribution within the next 25 years. This is based on an expected average growth rate
of about 4.5 per cent per year. However, this rate is far below the actual experience: during
the last 15 years the average growth rate of wind energy was about 3035 per cent per year,
for solar thermal energy about 1015 per cent and for solar electricity production 1520 per
cent and strongly accelerating in the last few years.
It is obvious that the IEA systematically overestimated the future role of fossil energy
sources and underestimated that of renewable energy sources.

1.1.2 DRIVING FORCES OF FUTURE ENERGY DEMAND/SUPPLY


What are the driving forces that influence our present energy supply and may even grow
in importance in the years to come? Three main drivers can be identified.
Most prominently, in the last 20 years mankind has learned a lot about the relationship
between fossil fuel combustion and its role in heating the Earths atmosphere. Basic
physics teaches us that the more we pollute the atmosphere with carbon dioxide, the
more we change the radiation balance. This results in heating the Earths surface
temperature with the consequence of increased evaporation and rainfall. An increase of
water vapour in the air is almost identical to more energy in the air and consequently must
result in stronger rainfalls and storms. These effects today are beyond any doubt. The
remaining uncertainties are: To what amount can competing effects dampen the rise in
temperature and how large is the range of the natural variability of our climate? How much
does the temperature still need to rise until we are convinced that the observed rise is
related to anthropogenic influence?
Figure 1.2 sketches the dilemma: economic principles imply that gross domestic
product needs to grow each year in order to maintain the living standard of the population.
All efforts are directed towards this growth. The main advice to developing countries is to
accelerate their economic growth. But economic growth is almost completely linked to
the growth of material products based on rising consumption of raw materials and fossil
energy. All this happens in spite of the indisputable but ignored fact that unrestricted
growth of limited resources is simply not sustainable.
The best way to illustrate this basic dilemma is to consider climate change. Figure 1.2
shows the history of world energy consumption, which is directly related to the level of
greenhouse gas emissions. The projected alternative trends exhibit schizophrenia. On the
one hand it is believed that a growing consumption of fossil fuels is needed to keep the
economies running and to increase welfare; on the other it is known that only a strong
reduction of fossil fuel use will save our planet.
Another potent driver influencing the energy future is the natural endowment of the
Earth with fossil energy resources. Oil is the first resource whose availability is beginning

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

7:36 PM

Page 7

Alternative World Energy Outlook 2006

Gtoe/yr (Gigatons oil equivalent/year)


20
18
16

Business as usual: economic growth

14
12
other

10
8
6

coal

gas

oil

1920

1940

1960

1980

Emission reduction:
ecological welfare
2000

2020

2040

Year

Source: BP Statistical Review of World Energy

FIGURE 1.2 The basic dilemma Business as usual or climate policy; growing fossil
energy consumption is tantamount to growing air pollution and reinforced greenhouse effect

to decline. Within a few years it will be followed by the decline of natural gas production
and after that by the geological restrictions of uranium and coal supply. The question is
not whether we like this or not depletion will occur even if one tries to ignore it. The
limited resource base will be detailed in the following section.
The third major driver is more positive. During recent years we have seen lots of
improvements and technical innovations in promoting the use of renewable energy
sources. This positive driver will make it easier to manage the transition from the
unsustainable present to a more sustainable future. However, this is not an automatism.
How fast the major economies of the world especially will move in such a direction
depends on many players these are the political leaders, the industry leaders, and of
course the consumers and voters. The extent and speed of action will be strongly
influenced by the assessment of the present situation.

1.2 FUTURE AVAILABILITY OF FOSSIL AND NUCLEAR


ENERGY SOURCES
1.2.1 OIL
The conditions for the worlds oil supply have entered into a new phase, with increasing
demand pressures, worries about the security of supply in important oil-producing
countries, speculative factors, and clear indications that limitations on the supply side
have caused unexpected and high price increases. In view of the fact that increasing oil
production is obviously becoming more and more difficult, it is now almost irrelevant
whether the peak of oil production has already been reached or whether growth of
production just cant keep pace with rising demand anymore.
With accumulating evidence we will soon be able to decide between the conflicting
views held by the optimists and pessimists. According to the doctrine of the optimists

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

7:36 PM

Page 8

J. SCHINDLER AND W. ZITTEL

(mostly economists), rising prices will induce a fast increase of oil exploration and
production which in turn will lead to a relaxation in the oil market in the near future. In
contrast, the pessimists (mostly influenced by geological considerations) expect that it
will become increasingly difficult to balance the increase in demand by a sufficient rise in
supply. As a consequence production will not be able to keep up with demand and, after
a short phase of stagnation, will inevitably decline.
Many events during the last five years vindicate the theses of the pessimists. But it
would be much more appropriate to leave these misleading categories behind and to
speak only about realistic and unrealistic views. Facts and not beliefs will decide the
issue.
In the following sections the present state of the worlds oil supply will be outlined.
(For a more detailed analysis see Notes 4, 5 and 6.)

1.2.1.1 General pattern of oil production


The different phases of oil production can be described schematically as follows. In the early
phase of the search for oil, the easily accessible oil fields are found and developed. With
increasing experience the locations of new oil fields are detected in a more systematic way.
This leads to a boom in which more and more new fields are developed, initially in the
primary regions, later all over the world. Those regions which are more difficult to access
are explored and developed only when sufficient new oil can no longer be found in the easily
accessible regions. As nobody will look for oil without also wanting to produce it, in general
shortly after the discovery of new fields their development will begin.
With increasing production the pressure of an oil field diminishes and the water levels
rise, and after some time the production rate begins to decline. This trend can be
controlled to a certain extent so that the decline of the production rate is delayed or
reduced: by injecting gas or water into the reservoir in order to increase the pressure, by
heating the oil or by injecting chemicals in order to reduce the viscosity of the oil.
In every oil province the big fields will be developed first and the smaller ones later.
As soon as the first big fields of a region have passed their production peak, an increasing
number of new and generally smaller fields have to be developed in order to compensate
for the decline of the production base. From there on, it gets increasingly difficult to
sustain the rate of production growth. A race begins which can be described as follows.
More and more large oil fields show declining production rates. The resulting gap has to
be filled by bringing into production a larger number of smaller fields. However, these
smaller fields reach their peak much faster and then contribute to the overall production
decline. As a consequence, the regions production profile, which results from the
aggregation of the production profiles of the individual fields, becomes more and more
skewed, the aggregate decline of the producing fields becomes steeper and steeper.
This decline has to be compensated for by the ever faster connection of more and more
ever smaller fields.
This pattern can be observed very well in many oil provinces. However, sometimes this
general pattern has not been followed, either because the timely development of a
favourable region was not possible for political reasons or because of the existence of
huge surplus capacities so production was held back for a longer period of time. However,

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

7:36 PM

Page 9

Alternative World Energy Outlook 2006

the more existing spare capacities are reduced, the closer the production profile will follow
the described pattern.
In the history of oil production, which is now extending over more than 150 years, we
can identify some fundamental trends:

the worlds largest oil fields were all discovered more than 50 years ago;
since the 1960s annual oil discoveries have decreased tangentially;
since 1980 annual consumption has exceeded annual new discoveries;
till now more than 42,000 oil fields have been found, but the 400 largest oil fields
(1 per cent) contain more than 75 per cent of all oil ever discovered; and
the historical maximum of oil discoveries must be followed by a maximum of oil
production (the peak).
How close to the peak have we already got? How steep will the decline be after the peak?
These are the crucial questions remaining.

1.2.1.2 Countries outside OPEC and the former Soviet Union


At the global level, the development of different oil regions took place at different times
and at varying speeds. Therefore today we are able to identify production regions being
in different development stages and with this empirical evidence we can validate with
many examples the simple considerations described above.
Looking at the countries outside the former Soviet Union (FSU) and OPEC, it can be
noticed that their total production increased until about 2000, but since then total
production has been declining. A detailed analysis of the individual countries within this
group shows that most of them have already reached their production peaks and that only
a very limited number of countries will still be able to expand production, particularly Brazil
and Angola.
Responsible for the stagnation of the oil production in this group of countries was the
peaking of the oil production in the North Sea which occurred around 2000 (1999 in Great
Britain, 2001 in Norway). Worldwide onshore oil production had reached a plateau much
earlier and has been declining since the mid-1990s. This decline could be balanced by the
fast development of offshore fields, which now account for almost 50 per cent of the
production of all countries in this group. The North Sea alone has a share of almost 40 per
cent of the total offshore production within this group. The peaking of the North Sea was
decisive because the production decline could not be compensated any more by a timely
connection of new fields in the remaining regions it was only possible to hold the
plateau for a few years.
Crucial for the further development was the production peak of Cantarell in Mexico, the
worlds biggest offshore field. This field, discovered in 1978, even today contributes one
half of Mexican oil production. It reached a plateau for some years and started to decline
in 2005. Furthermore, the quality of the oil produced in Mexico has degraded steadily; the
share of light oil has halved since 1997.
This steady degradation of the quality of the oil produced can be observed in almost
all regions which have passed their peak and poses an additional challenge for the

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

7:36 PM

10

Page 10

J. SCHINDLER AND W. ZITTEL

Mb/day
history

40
35
30

Ecuador 99
Colombia 99
Argentina 98

25
20
15
10
5

Mexico 04

Norway 01
Oman 01
Australia 2000
United Kingdom 99

Malaysia 97
Gabon 97
Syria 95
India 95
Egypt 93
Indonesia 77
NGL
Romania 76
Alaska
Canada (conv.) 74
Rest-USA 7 1
Germany 67
Austria 55
Texas 71

1900 10

20

30

40

50

60

70

80

Yemen
Neutral Zone
Brazil
Angola
China

89
GOM

90 2000 10

Source: IHS 2003, BP Stat Rev 2005: Denmark, UK, Canada, Norway, Alaska, USA, Mexico, Brazil, Argentina:
2005 estimate based on Jan-Nov data from government statistics, Analyses and Forecast LBST

FIGURE 1.3 Oil production of countries outside OPEC and the FSU

existing downstream infrastructures: refineries have to operate with oil of decreasing


quality. The share of lower oil qualities is steadily increasing this will additionally drive
upwards the prices for the remaining good oil grades.
Particularly interesting is the example of Indonesia, the only OPEC member state which
is included in this group of countries as it will probably soon leave OPEC, because in March
2004 for the first time more oil was imported in this country than exported.
Oil production in regions having passed their peak can be forecasted with some
certainty for the next 10 years. Even if it is assumed that the remaining regions with growth
potential (Angola, Brazil and the Gulf of Mexico) will considerably expand their production
by 2015 (in accordance with the forecasts of the companies operating in these regions),
total oil production for this group of countries will decline by 1015Mb/day by 2015.
As the production of conventional oil is declining, this group of countries will be able to
supply additional amounts only from non-conventional sources. Non-conventional oil sands
in Canada and Venezuela will contribute 34Mb/day in 2015, provided that the already
announced expansion plans will be realized without any further delay.

1.2.1.3 The former Soviet Union


Oil production in the FSU peaked reaching a production rate of more than 12Mb/day at
the end of the 1980s. Production then collapsed by almost 50 per cent within five years.
The production peak at the end of the 1980s had been forecasted by Western geologists
based on the depletion patterns of the largest oil fields. However, the following production

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

7:36 PM

Page 11

Alternative World Energy Outlook 2006

11

Mb/day
5

history

4
Bitumen
(4Gb)

3
Synthetic
crude oil
(6.7Gb)

Sasketchwan
Alberta (5.5Gb)

1
0

Total conv
New Foundland
oil production (18Gb) (1.4Gb)

1960

1970

1980

1990

2000

2010

2020

2030

Tarsands
(<10API)

Heavy
oil (1017API)
NGL-Alberta
Convcrude
oil production

Source: 19752004 data National Energy Board, Canada; 19601974 data USDoEEnergy Information Administration 2005:
Estimate by NEB, January 2006; 20062015 Forecast, tar sands based on NEB-study, May 2004, conventional and heavy oil
based on LBST estimate

FIGURE 1.4 Canadian oil production 19602030. The rising contribution of non-conventional oil production
(heavy oil below 17API and tar sands below 10API) substitutes declining conventional supply

collapse during the economic breakdown turned out to be much steeper than expected.
After the liberalization of the oil market, Russian companies were able to stop this decline
and to increase production levels again at double-digit rates in some years during the
last five years with the help of international cooperation and investments. However, this
fast recovery has now come to an end as the easily accessible fields have been developed
and the financial and technological backlog has caught up.
The double-digit growth rates in Russia contributed to compensating for the inescapable
production decline in other regions of the world. But despite this strong revival of Russian
production, oil prices have remained under pressure and have been rising slowly but
continuously and have even exceeded US$75 per barrel, a sixfold increase since 1999.
The two other important oil regions of the FSU are Azerbaijan and Kazakhstan.
Azerbaijan is the oldest industrial oil region of the world. Its highest production rates were
reached 40 years ago. Today, we can expect an expansion of production only in the offshore
areas. There, especially, the field complex Azeri-Chirag-Guneshli, has to be mentioned. Once
fully developed, this field probably will reach its maximum in 2008 or 2009, with a production
rate of 1Mb/day. Soon thereafter the production rate will decline very fast to almost negligible
amounts within 1015 years. The total production of this region, however, will increase by a
smaller amount as 150,000bbl/day are already produced from Azeri-Chirag-Guneshli today
and as the production from other fields will drop noticeably in coming years.
For some years Kazakhstan was considered to be a potential counterbalance for Saudi
Arabia. Today we know that these hopes were exaggerated. They were nurtured by

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

12

7:36 PM

Page 12

J. SCHINDLER AND W. ZITTEL

speculations of the US Energy Information Administration (EIA), which estimated the oil
and gas reserves in the Caspian Sea region to be up to 300Gb of oil equivalent. Realistically,
only about 45Gb of oil are likely to be recoverable, about half of this amount located in
already developed fields.
High expectations regarding their future production potential concentrate on three
fields: Tengiz, Kamchagarak and Kashagan. Tengiz and Kamchagarak have been
producing oil for some years. All three fields contain oil with a high sulphur content, the
development of which jeopardizes the environment and is very expensive. In Tengiz
alone, more than 4500 tons of sulphur are separated from the produced oil each day and
stored in the surrounding area, polluting the environment. Plans for an extension of
production are delayed due to high development costs and difficult geological
conditions. In 2000 the third big oil field, Kashagan, was found. It is assumed that
production can be increased considerably from 2006 on; however, there are big doubts
whether this will be possible. The high sulphur content, a high deposit pressure of more
than 1000 bars and an unfavourable geographical location far away from any
infrastructure make it difficult and expensive to develop. It is certainly no coincidence
that two of the big companies involved in the discovery of the field (BP and Statoil) have
withdrawn from the consortium developing the field. After an analysis of the first
exploration drilling, it was communicated that the companies internal criteria for
development were not fulfilled.
Azerbaijan and Kazakhstan will probably be able to double their production rate by
2010 from 1.3Mb/day to 2.52.6Mb/day but to expect more seems unrealistic.
According to this assessment, the whole region may be able to increase its production
in the coming years, but the very big expansion expected by many people will not occur. A
production increase of 23Mb/day is probably already on the high side.

1.2.1.4 OPEC member countries


The conclusion of the previous analyses is that the expected production decline in the
group of countries described initially is partly offset by a possible expansion in Russia and
the Caspian Sea. But there still remains a gap of 510Mb/day which has to be closed to
keep world oil production constant until 2015. There are only the OPEC member countries
left to fill this gap. If world oil consumption is to grow further, the additional amounts
would have to come from OPEC as well.
Conventional wisdom has it that this will be easily possible for OPEC. However, a
production growth of 510Mb/day within ten years does constitute a problem, particularly
as it is widely accepted that (apart from Iraq which cannot be considered to be a reliable
oil producer for the time being) only Saudi Arabia is supposed to be able to increase its
oil production significantly. This would require an expansion of at least 50 per cent of
Saudi Arabian oil production within very few years. This is a very ambitious goal, even for
a country with an abundance of oil.
Moreover, in recent years the suspicion has grown that conditions for oil production
in Saudi Arabia are no longer as favourable as is commonly assumed and are getting more
and more difficult. In assessing the future production potential of Saudi Arabia, Ghawar,
the worlds biggest oil field, plays a key role. This field was discovered in 1948 and has

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

7:36 PM

Page 13

Alternative World Energy Outlook 2006

13

now been producing oil for more than 50 years. It is a fact that more water is pumped into
the field than oil is extracted, and it seems quite possible that the production rate will
decline in the near future. Anyway, it is certain that Ghawar cannot contribute to an
expansion of the Saudi Arabian production.
There is an ongoing debate on whether Saudi Arabia will be able to increase its
production significantly. This debate was initiated in early 2004 by Matthew R. Simmons, an
American investment banker from Houston.7 Simmons very much doubts the possibility of a
significant growth of production. His assessment is based on a comprehensive in-depth
analysis of technical papers in the public domain addressing the problems of oil production
in Saudi Arabia and on a great number of interviews with engineers working on site and also
a visit to the oil fields in Saudi Arabia.8
Simmons has provoked comments by Abdul-Baqi and Nansen Saleri, senior
executives of the state-owned company Saudi Aramco.9 But their comments have rather
fuelled existing fears instead of assuring the world. First, it was admitted that the big old
oil fields are in decline, and that already the Abqaiq field has been depleted by 73 per cent
and Ghawar by 48 per cent. Moreover it was indirectly confirmed that the proved reserves
do not amount to 262Gb, as is widely assumed. The proved reserves amount to only
130Gb while another 130Gb have been counted as reserves already because it is regarded
probable that they can be developed eventually. If one were to apply the same criteria
which are common practice with Western companies, then Saudi Aramcos statement of
proved reserves should be devalued by 50 per cent. This was confirmed indirectly by
another Saudi Aramco executive.10
Furthermore, Saudi Aramco executives tried to counter Simmons fears by stating
that a production of 10Mb/day could be upheld until 2042. In doing this they had to
assume that the above mentioned reserves of 260Gb are proved reserves (which they
definitely are not). Saudi Aramco went on to state that in case of a more aggressive
development of the remaining reserves, production could be increased to 12Mb/day by
2016 and then could be maintained constant until 2033.9 But even this scenario put
forward by the Saudis is hardly reassuring in view of the projections of the IEA, which
assume that in the longer term an additional 20Mb/day are supposed to come from
those regions.
The analyses of Simmons and others (for example Bakhtiari11) make the point that
Saudi Arabias potential to increase production will soon reach its limits. The world is
nearing the moment of truth. The next few years will reveal whether those who believe
that OPEC has no more spare capacity left are right, or whether the peak of world oil
production can be delayed for a few more years.
Should world oil production be increased, it will be taken by many as evidence that oil
production can be increased for many more years to come. However, in reality if the oil
production increases, 1) the remaining oil will be consumed that much faster and 2) an
increase of production in the short term will in effect further increase the consumption of oil,
resulting in a steeper decline than otherwise necessary.
Recent developments are in obvious contrast to the assertions of the optimists which
do not foresee any problems in the availability of oil within the next 2030 years. But at
least they now acknowledge that price increases might be possible.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

14

7:36 PM

Page 14

J. SCHINDLER AND W. ZITTEL

Two oil supply scenarios are defined to be used later, these characterized by different
production profiles:
1 the high fossil scenario based on the Association for the Study of Peak Oil (ASPO)
production profile, with a peak before 2010 and a moderate decline rate of 23 per
cent per year;12 and
2 the low fossil scenario based on the fears that future global decline rates will be
higher than in the old mature oil regions (for example that decline rates in Alaska and
UK are 510 per cent versus 3 per cent in the lower 48 states of the USA).
Accordingly, after a peak in 2008 and plateau until 2010 a decline rate of 5 per cent
per year is assumed for 20102020, 3 per cent for 20202040, 2 per cent until 2050
and 1 per cent thereafter.

1.2.2 NATURAL GAS


Presently the world consumes much less gas than oil. But the substitution of declining oil
supplies by natural gas would result in a drastic increase of gas consumption.
Natural gas reserves and production rates can be analysed similarly to oil. This leads
to an estimate of possible future gas production rates and the probable timing of the peak
gas supply.
The main conclusion of this analysis is that the worlds natural gas supply might peak
about 20 years from now. If the production is increased at a faster rate or future
discoveries and stated proved reserves are smaller than assumed in this analysis the
peak might come earlier. If production is increased at a slower rate, only a few years
would be saved until the inevitable start of decline.
This global topdown analysis is somewhat theoretical because, unlike oil, natural gas is
supplied and consumed in regional markets. Only about 7 per cent of the total natural gas
supply is traded globally. Production in these regional markets is determined by differing and
specific supply conditions, which are not correlated. Individual regional markets are better
described with bottomup analyses of producing fields, ranking fields by size and their status
of depletion. A full analysis also has to include the evaluation and interpretation of regional
creaming curves of past discoveries (time series of the success ratio of exploration drillings).
Such an analysis has been performed elsewhere but is beyond the scope of this paper.13,14
At present, North America is the largest regional market, with an annual volume of
about 780 billion m3, amounting to almost one third of the world market. This market
already experienced its production maximum in the early 1970s, followed by a second
smaller peak in 2001. But also in Europe, the second largest gas market, further expansion
of consumption is getting more and more difficult as aggregate domestic supplies already
have passed their peak. For example UK gas production started to decline in 2001.

1.2.2.1 The topdown scenario of the global availability of natural gas


Figure 1.5 shows the history of cumulative discoveries and production of natural gas for
the world.15 Annual discoveries peaked around 1970 and have been slightly declining
since that time. In contrast, annual consumption is still rising and has already exceeded
annual discoveries for several years.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

7:36 PM

Page 15

Alternative World Energy Outlook 2006

15

Tcf [10,000 Tcf = 270,000 billion m]


12,000
10,000
8000
6000

Production forecast
(growth rate in 2010: 1.9 %)
Cumulative
discoveries

4000
2000

Depletion mid-point ~2025


Cumulative
production

1920

1940

1960

1980

2000

2020

2040 2060
Year

2080

Source: History: IHS-Energy; Projection: LBST based on logistic grown function

FIGURE 1.5 Cumulative worldwide gas discoveries and production. The extrapolation of the data supports
the forecast that in total about 12,000Tcf will be discovered. The peak of production can be expected from this
rough topdown analysis when about half of the total volume is produced, that is around 2025

Extrapolating the declining rate of new discoveries results in the assumption that in
total about 12,000Tcf (~325,000 billion m3) might be discovered until 2100. Proved gas
reserves are comparable in size to the proved oil reserves (~160Gtoe), but the already
produced share is smaller (one third for gas against half for oil). Therefore the global
production peak can be expected to happen later than for oil. A rough topdown approach
suggests that gas production will peak around 2025.
A bell-shaped production profile which fits the historical production pattern until 2004
and assumes the estimated total of 12,000Tcf, results in a smoothly diminishing annual
production growth rate (which is currently at 2.5 per cent and is expected to be at 1.9 per
cent in 2010). If future growth rates are larger, the expected production peak would
happen sooner and vice versa.
This extrapolation assumes that about 75 per cent of world gas reserves have already
been discovered (compared to 90 per cent for oil). Future reserve reassessments and
upward revisions of (older) producing gas fields will only marginally influence this pattern
as they usually do not result in changes of the production profile of these fields.
This topdown analysis does not take into account the quality of the gas fields. Especially
it does not make allowances for the so-called stranded gas situated far away from existing
transport infrastructures or for low quality gas with a high content of CO2 or SO2. Today,
stranded gas is not used for economic reasons. However, this could change with increasing
gas prices. But this would also imply much higher costs for the development of these fields

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

16

7:36 PM

Page 16

J. SCHINDLER AND W. ZITTEL

Tcf/year [100 Tcf = 2700 million m3]


180
160
140
120
100
80
60
40
20
1920

1940

1960

1980

2000

2020

2040

2060

2080

Year
FIGURE 1.6 Annual gas production 19202004 and extrapolation based on a bell-shaped profile
and an estimated ultimate recovery of 12,000Tcf. Currently 3000Tcf have been consumed and proved
reserves are at 6300Tcf16, 17

and for the conditioning of the produced gas, especially for liquefaction or upgrading into
other liquid hydrocarbon fuels. The conditioning, transport and upgrading of this gas would
consume between 2050 per cent of its energy content, depending on whether the gas is
liquefied or transformed into synthetic crude oil, ammonia, methanol or hydrogen. Of course,
this conditioning would reduce the available gas reserves correspondingly.
These additional problems and the lead times for the construction of the necessary
infrastructure make it probable that the calculated production rate in this scenario
provides an upper (optimistic) limit for the gas extraction curve, shown in Figure 1.6.
Probably the peak will be sooner, followed by a plateau lasting several years.
Natural gas liquefies at temperatures of below -160C and ambient pressure. At ambient
conditions it is gaseous and can be transported best in pipelines. Therefore pipelines are the
backbone of the established gas markets, which have developed over time and connect the
major consumers with their supply regions. This regionalization is the main difference to
crude oil. Oil can be transported very easily, which helps to equalize regional imbalances.
The importance of the pipeline infrastructure leading to separate regional markets
makes it questionable whether the above sketched topdown approach provides the
basis for a meaningful interpretation. Probably this scenario will never be realized.
Regional supply scenarios have a much higher importance as they reflect regional supply
problems. Though liquefaction is possible and will be expanded in the future, a much higher
effort is necessary regarding energy, materials, investments and lead times. Today, about 7
per cent of the traded natural gas is liquefied; about 90 per cent is transported via pipelines.
In the following the most important regional markets are sketched North America and
Europe. Both regions together produce about 45 per cent of the worlds gas and consume

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

7:36 PM

Page 17

Alternative World Energy Outlook 2006

17

about 55 per cent. In both markets the supply situation has dramatically deteriorated during
the last few years, coming as a complete surprise for economics-oriented gas market
analysts. The analysis of these regional markets provides a better understanding of
possible future supply restrictions than the optimistic topdown approach sketched above.

1.2.2.2 The gas market in North America


In North America natural gas has been used almost as long as oil has been extracted. The
natural gas market grew in parallel to the large rise of oil production. Production and
consumer regions are distributed all over North America. During the last 100 years a large
distribution and transport grid grew with almost two million km of transport pipelines.
Figure 1.7 illustrates that natural gas production peaked in the US in 1972 and declined
rapidly in the years after. But more and more wells were drilled in mature and new areas,
almost doubling the number of active wells from about 200,000 in 1985 to more than
300,000 ten years later. This helped to reverse the decline from 1985 on. But since the
mid-1990s this second production increase came to a halt, and only imports from Canada
still grew. Finally, in 2001 the production of North America as a whole passed its peak and
entered the decline phase, though in the year 2004 more than 400,000 active wells were
in operation.18
The gas consumption in North America exhibits strong seasonal fluctuations, with
a pronounced maximum during the winter season. Residential use and industry use
is predominantly for heating. The gas consumption for electricity production
reaches a seasonal maximum during the summer days, when a high demand for air

Tcf/year
25
Imports

20
15

US production
Louisiana

10

Oklahoma

Texas

0
1935 40 45 50 55 60 65 70 75 80 85 90 95 2000 05
Year
FIGURE 1.7 Gas supply in the US. Since 2001 production has been in decline; even imports from
Canada cannot substitute declining domestic supplies as production in Canada is also past its peak19

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

18

7:36 PM

Page 18

J. SCHINDLER AND W. ZITTEL

conditioning exists. On average, in summer about 40 per cent less gas is consumed than
in winter.
The gas production, however, is almost independent of the season. Therefore huge
amounts of gas are stored during the summer. The weekly storage additions are used by
commodity market analysts as a measure of a tight or relaxed supply situation.
Over the last three decades gas has been traded on the commodity markets at
between US$1 and US$2 per million BTU (this corresponds to 3.77.4 cents/m3). As late
as the summer of 2002 the EIA forecast that gas demand and supply will rise by 50 per
cent until 2020.20 There was no hint at possible future supply problems by the agency.
According to this forecast, the gas price was expected to rise to about US$2 per million
BTU until 2020. But critical observers could foresee mounting problems already in the
winter of 2000/2001. Due to first supply restrictions well-head gas prices then climbed by
a factor of 3.5 up to a monthly average of US$6.7 per million BTU in January 2001.18
This sudden price rise had grave effects for some consumers: small and medium
enterprises with a high gas consumption (for example gardeners with large greenhouses)
went into bankruptcy; companies with long-term contracts (for example ammonia and
methanol producers21) stopped the production of their products and instead resold the
gas with a larger profit margin. The unexpected combination of high oil and gas prices
triggered the economic recession in 2001 which spread over to Europe and many other
countries. Since that time gas prices have stabilized at a level of US$68 per million BTU
during summer with spikes far above US$10 per million BTU during winter. At the
beginning of 2006 the price stood around US$10 per million BTU (~30 cents/m3).18
The declining domestic oil and gas production, in combination with limited import
capacities (new liquefied natural gas, LNG, terminals are planned but have long lead times
to construct), will probably increase the supply problems in the years to come. Over the
next years the domestic production will continue to decline. Canada will need more gas
for the extraction and upgrading of tar sands, which presently consumes about 5 per cent
of the domestic gas production.19
Hidden from the general public, many market observers and politicians, an energy
crisis looms over the horizon with grave consequences for the North American energy
markets, the whole economy and probably even for other regions.

1.2.2.3 The gas market in the UK


The economic strength of the UK is based to a large extent on the production and export
of hydrocarbons. It was therefore a shock when in September 2003 the UK became a net
importer of oil for the first time in almost three decades.22 In two or three years time the
UK may permanently become an oil importing country as the domestic oil production is
rapidly declining, at an annual rate of about 10 per cent.
But developments in the gas market are also following a similar pattern. Figure 1.8
shows that each year the newly developed fields reach their peak faster and then add to
the growing decline rate of the production base. The new fields are getting smaller and
smaller. Because these newly developed fields can no longer compensate for the decline
of the base production, UK gas production, which has been in decline since 2001, has
already decreased by 20 per cent.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

7:36 PM

Page 19

Alternative World Energy Outlook 2006

19

Source: DTI, April 2005; Forecast: LBST

FIGURE 1.8 UK gas production 19702030. Each area represents the contribution from all new
fields which are developed within one year.23 The dashed line provides a possible future production
profile if expected future discoveries and remaining untapped reserves are connected in time

Since new field developments are rare, the future production profile can be estimated
with a high level of confidence. This is outlined in Figure 1.8. By 2010 the total production
will be 50 per cent below the level of 2000 and by 2015 another 3050 per cent below the
level of 2010, depending on the success of finding new fields.
Very soon the gas supply of the UK will depend on imported natural gas. A major
possible source is the Norwegian gas field Ormen-Lange, which will be directly connected
with the UK via pipeline. However, this field will supply gas at the earliest in late 2007. In
view of several downward revisions of the size of this field by the developing companies
and the retreat of BP from the consortium it is doubtful whether the initially envisaged
production can be achieved.24
The gas flow in the five-year-old pipeline connecting the UK with the continent has
been reversed. Originally the pipeline was built to export gas but is now used for
importing gas from the continent. Yet its capacity is rather limited. A large expansion of
import capacities will have to be provided by newly built LNG terminals.
The oil and gas supply situation of the UK should be seen as a warning which demonstrates how soon the days of surplus production of oil and gas, with their corresponding
high export revenues, can be followed by the necessity of steadily rising imports.

1.2.2.4 The gas market in Europe


Figure 1.9 illustrates the historical and future gas supply of Europe. The historical data are
taken from official statistics.25 Future production is based on a detailed field-by-field

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

20

7:36 PM

Page 20

J. SCHINDLER AND W. ZITTEL

Billion m3/year
history

forecast

500

LNG: +5% per yr

400

140 billion m3/year


Import capacity
Must be constructed

Imports:
+5% per yr
LNG Imports

300

Imports from
Russia, North Africa
Germany
(900 billion m3)

200
100

Imports are constant

Italy
(1000 billion m3)

Norway
(4500 billion m3)

Netherlands (3550 billion m3)


UK (EUR: 3600 billion m3)

0
1960

70

80

90
Year

2000

10

20

Source: OECD 2004, BP 2004; Forecast: LBST 2004

FIGURE 1.9 Natural gas supply of Europe probable development until 2020 according to scenario
calculations by LBST26

analysis of the UK, Norway and The Netherlands. Since 2004 the still growing supply from
Norway can no longer compensate for the declining supply from the UK and other
countries Europe has passed its gas production peak. Even a 50 per cent increase of
Norwegian gas production cannot stop the overall decline. Europe is in need of rapidly
rising imports from Russia, North Africa or other parts of the world.
According to this analysis, gas imports must rise at an annual rate of 5 per cent until
2020 just to keep the supply base flat. Future growth of demand would require even
higher rates. Even this zero growth scenario requires the construction and operation of
about four or five new pipelines with a net import capacity of about 30 billion m3/year until
2020.
In Europe, only Norway can still substantially expand its production capacity. A few
years ago its reserve to production ratio (R/P-ratio) amounted to more than 60 years.
However, the strongly rising production and diminishing new discoveries have now
reduced the R/P-ratio to about 30 years. Norway therefore will presumably reach its gas
production peak within the next 10 to 20 years. If Europe wants to increase its gas
demand according to the forecast, gas imports must at least double by 2020. More
precisely, new import capacities of about 150200 billion m3 must be built and put into
operation within the next 15 years. To put this required effort into perspective: the just
proposed and planned pipeline through the Baltic Sea connecting Russia with Germany
will have a capacity of about 28 billion m3 in the first phase (one line) and of 55 billion m3
in the second and final phase with two lines. It is expected that the first line will come into
operation at the earliest in 2010, but more likely in 2011 or 2012. However, by 2020 import
capacities about five times greater will be needed in Europe.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

7:36 PM

Page 21

Alternative World Energy Outlook 2006

history

Tcf/year
40
35
30
25
20
15
10
5

21

forecast

Zappolyarnoye (60Tcf)

+3% p.a.

Kharampur (12Tcf)
Yamburg (170Tcf)
Small fields
(29Tcf)
Astrahan (10Tcf)
Urengoy Severnyy(28Tcf)
Konsomolskoye (28Tcf)
Vyngapur (12Tcf)
Bolshoy Gubkin (16Tcf)
Orenburg(48Tcf)
Medvezhye (75Tcf)
Vuktyl(12Tcf)
Yubilneynoye (12Tcf)
Urengoy
Samotlar (9Tcf)
(250Tcf)
Old fields
(30Tcf)

0
1960

1970

1980

1990

2000

+1% p.a.
ca. 10 Fields (10-15Tcf)
Karasovey (26Tcf)
Leningradskoye (50Tcf)
Shtokmanovskoye (55Tcf)
Semokovskoye (15Tcf)
Rusanovskoye (25Tcf)
Bovanenko (70Tcf)

2010

2020

Year

Source: Laherrere, LBST estimate

FIGURE 1.10 Gas production forecast for Russia25

1.2.2.5 Russia
The required additional amounts of gas cannot come from Russia as the three largest
fields (Urengoy, Yamburg and Medvezhye) containing about one third of the
discovered gas are already in decline.27 It is very questionable whether Russian gas
production still can be expanded for a longer time period. Figure 1.10 shows the fieldby-field analysis of Russian gas production since 1960 and the forecast until 2020.16 This
forecast is based on already known but not yet developed fields and supposes a
hypothetical time schedule for their development. The analysis leads to the conclusion
that probably an annual production increase of 1 per cent can still be realized in the next
10 to 15 years. However, this requires a timely development of new fields. These fields
are situated further north (in the Barent Sea and Kara Sea) or further east; therefore
development will be much more time consuming and much more expensive than the
development of the already producing fields. Also the question remains how much of
this gas will be available for export to Europe as Russian domestic consumption is
expected to rise and East Asian countries (China, Korea and Japan) will compete for
imports.

1.2.2.6 Global analysis


Performing such an analysis for all gas-producing countries leads to the conclusion that
probably worldwide gas production will peak when Russian gas production peaks.
Though some world regions will still expand their production beyond 2020 (for example
Quatar and possibly Iran), the decline in North America and Russia probably cannot be
compensated for. A probable scenario for the global gas production until 2030 is shown
in Figure 1.11. The graph shows the production volume of each major region. The different
regions of the world are defined according to the IEAs WEO.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

22

7:36 PM

Page 22

J. SCHINDLER AND W. ZITTEL

Billion m 3
5000
4500
Latin America

4000
3500
3000
2500
2000

Africa

South Asia
East Asia

Middle East

China
OECD Pacific
OECD Europe

1500

Transition Economies

1000
500

OECD North America

0
1960 65 70 75 80 85 90 95 2000 05 10 15 20 25 30 Year
Source: IHS-Energy, BP Statistical Review of world Energy Projection: LBST 2005

FIGURE 1.11 Worldwide gas production according to LBST scenario calculations

For the formulation of global energy scenarios later in this paper two alternative global
gas production scenarios are used:
1 future gas production according to the Association for the Study of Peak Oil12 it
exhibits a production plateau for 20152040 at 3000Mtoe (=3400 billion m3 or 35 per
cent above the 2004 level); and
2 a second low fossil scenario assuming that this peak production plateau can only be sustained for ten years until 2025 and then will be followed by a decline of 3 per cent per year.

1.2.3 NUCLEAR ENERGY


Some people are convinced that nuclear energy can solve our energy problems for a long
time to come. However, there are not many justifications for such a view. Even if one were
to neglect environmental, legal or financial aspects or lacking public acceptance, some
simple considerations make it difficult to believe in that technology as a long-term
sustainable solution.
In 1975 and later years the International Atomic Energy Agency forecast that
worldwide nuclear power capacity would rise to more than 1000GW by 2000. This proved
to be far too optimistic. Even the current forecast with a growth of 1030 per cent until
2030 probably will be much too optimistic because existing reactors are ageing. Most of
todays reactors were built between 1965 and 1985. This was the golden age of nuclear
energy. In recent years the number of new reactors fell below five annually. Most reactors
have taken several years for construction (about ten years on average) before being

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

7:36 PM

Page 23

Alternative World Energy Outlook 2006

23

Data source: IAEA June 2005, scenario: LBST 2005

FIGURE 1.12 Installed capacity of nuclear power plants and various forecasts

connected to the grid. Figure 1.12 shows the development of the cumulative nuclear
power capacity worldwide. Of the 550 reactors built more than 100 have already been
decommissioned; on average, old reactors have been decommissioned after less than 25
years of operation.
Assuming that the still operating reactors will be decommissioned after 40 years on
average, world nuclear power capacity will decline rapidly in the next 25 years, from 370GW
now to less than 100GW in 2030.25 In the upper right part of the figure a recent forecast of
the International Atomic Energy Agency is sketched.28 The next dotted line below is the
projection of the IEA in its most recent WEO.1 The shaded area below is the LBST scenario
describing what might be achieved if all efforts are undertaken to bring about a renaissance
of nuclear power: at best one probably could uphold the present level. The new capacity
line in the figure indicates how many new reactors must be built to fit this scenario. As the
number and timing of the required new reactors in this scenario seems to be very ambitious,
it is much more realistic to expect a decline of the installed nuclear capacity.
Figure 1.13 shows the worlds uranium resources based on public statistics.29, 30 The
upper part sketches the uranium used for nuclear weapons by assuming that this amount
will remain constant for the next 25 years. The area below shows the amount which is
already consumed by nuclear power plants in operation. If the present capacity remains
constant for the next 25 years a simple extrapolation shows that the proved uranium
reserves will be consumed by 2030, including the amounts stored from nuclear
disarmament and from reprocessing plants. Even if reasonable additional assured
uranium resources (believed to be producible at costs below US$130/kg) become
available, about 80 per cent of these reserves and resources would be exhausted by 2030.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

24

7:36 PM

Page 24

J. SCHINDLER AND W. ZITTEL

Data source: BGR 2003

FIGURE 1.13 World uranium reserves and cumulative uranium consumption

Additional estimated possible resources shown in the graph have a probability of


realization of between 5 and 50 per cent. Most of these resources are highly speculative
and probably will never be extracted. Assuming that about 25 per cent of these speculative
resources will be transferred into proved reserves within the next 25 years leads to the
result that by 2030 about two-thirds of the available uranium will have been extracted if
world nuclear capacity remains at the present level. This makes an increase of nuclear
power in the future almost impossible.
Only a very fast introduction of nuclear breeder reactors could be a way to overcome
the resource restrictions. Yet at present no move in such a direction can be observed;
instead, almost all major nations have either cut back or even stopped these activities. A
switch to thorium fuelled reactors also would not solve the problem, as the thorium
reserve situation is similar to uranium, and a fast conversion of reactors to thorium within
the next few years is impossible. The scenario calculations later on include the phase-out
of nuclear energy within the next 30 years, as from a world energy perspective it is almost
irrelevant whether nuclear will keep its present level, rise for a short time or decline.

1.2.4 COAL
At the present production level, so-called proved coal reserves would last for another 180
years.17, 29 A bell-shaped production pattern (which is fitted to this reserve number and to
the historical growth pattern) allows an estimate that coal production will peak around
20502060. At peak, the production would be about 6070 per cent higher than the
present production rate. This profile is sketched in Figure 1.14.
This figure also shows another much more optimistic production pattern based on an
unrealistic doubling of present reserves. In this case the production peak could be delayed
until 2080 while reaching higher production levels. This hypothetical case is intended to

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

7:36 PM

Page 25

Alternative World Energy Outlook 2006

25

Data Source: Historical Data: BP Statistical Review of World Energy, BGR Scenario: LBST 2005

FIGURE 1.14 Possible world coal production profiles based on 180 years and 370 years
of remaining reserves

demonstrate that even the assumption of such an optimistic production profile cannot
avoid the peaking of the supply of all fossil fuels within the next 50 years, as will be shown
below. Much more realistic seems to be an R/P-ratio of 180 years, which is maybe still too
optimistic in view of the bad experiences with the poor quality of oil and gas reserve data.
Furthermore, these production profiles do not take into account that coal is now
mainly used for electricity production. Additionally, in the future coal will also have to
substitute declining oil supplies. But this implies further transformation losses in the order
of 4050 per cent as each attempt to transform coal into a transport fuel (either into
synthetic crude oil via the Fischer-Tropsch process or into hydrogen) results in such losses
of the original energy content. Coal is a very dirty energy source emitting large amounts
of carbon dioxide and other pollutants. With huge efforts these emissions eventually
might be avoided the necessary technologies are available in principle, if not yet in
reality. But doing so provided that the reservoirs for the storage of the carbon are
available would reduce the usable net energy by another 2030 per cent.
The conclusion is that the coal production profiles shown in the figure provide an
upper limit of the future availability of coal and one which is not identical to the marketable
net energy.

1.2.5 FOSSIL SCENARIOS, NET ENERGY BALANCES


AND CARBON DIOXIDE EMISSIONS
The aggregation of the production histories and scenario projections for the different
fossil and nuclear energy sources is shown in Figure 1.15. This is probably a view of how

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

26

7:36 PM

Page 26

J. SCHINDLER AND W. ZITTEL

Data Source: Oil, Gas: Colin Campbell/ASPO 2005; Coal, Nuclear Scenario: LBST 2005

FIGURE1.15 High fossil scenarios of future production of fossil and nuclear fuels, based on peak oil
before 2010 and a decline after peak of about 23 per cent per year and a gas peak around 2040

the supply situation could develop in the next few decades which is rather biased to the
high side. Oil production is assumed according to ASPO projections; gas production is
taken from ASPO with a production plateau for 20152040 at a 35 per cent higher
production level than today; coal projections are taken from the bell-shaped profiles
sketched above, based on the more probable R/P-ratio of 180 years and on an upper limit
of 370 years.
Even within this upper limit scenario a further growth of world energy supply comes
to a halt as soon as gas production cannot rise any more. A further decline is unavoidable
in the following decades.
The production of nuclear energy cannot reverse this trend. Projections for nuclear
energy show the energy production from existing reactors and their phasing out after 40
years on average. For this figure, the conversion from nuclear electricity to primary energy
is based on 33 per cent efficiency. However, more appropriately, nuclear should be treated
as primary electricity, thus directly comparable to renewably produced electricity later on.
The low share of only a few per cent and the above sketched limited resource base makes
nuclear energy irrelevant for the global energy supply situation whether one tries to keep
its share or not. Yet from a financial view it is very relevant whether budgets are directed
to the nuclear industry or to other energy sources.
Due to the mature status of oil exploration and production, the energy needed to
supply high quality crude oil to the markets is increasing (through energy losses).
Therefore the difference between produced energy and supplied net energy increases

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

7:36 PM

Page 27

Alternative World Energy Outlook 2006

27

Data Source: Oil, Gas: LBST 2005; Coal, Nuclear Scenario: LBST 2005

FIGURE 1.16 Low fossil scenario of future production of fossil and nuclear fuels, assuming oil
peaking in 2010 and decline rates of 5 per cent per year

over the years, thus reducing the energy that can be supplied to the markets below the
values shown in these figures. The accurate assessment of the future development of the
energy losses, which today average about 10 per cent for oil, is very problematic.
In order to get a better understanding of the upper and lower boundaries of possible
future development, a low fossil scenario is also formulated. Such a scenario is
presented in Figure 1.16.
Oil starts to decline at 5 per cent annually between 2010 and 2020, followed by a
softer decline of 3 per cent until 2040, then declines of 2 per cent until 2050 and 1 per
cent thereafter. The plateau of world gas production is expected to end in 2025 (and not
2040 as in the high fossil scenario) with an annual decline rate of 5 per cent in the
following ten years and 3 per cent thereafter. The growth of coal production is also
seen to be smaller than in the high fossil scenario above with a plateau at
3600Mtoe (not 4600Mtoe) lasting from 2020 to 2050 and followed by an annual decline of
1 per cent.
In this low fossil scenario the paramount importance of peak oil is demonstrated. Once
oil peaks and then declines at 5 per cent per year, no other energy source will be able to
stop the aggregate decline of supplies, even for a few years. This scenario is not unlikely.
The high fossil scenario (Figure 1.15) is regarded as being positive from an
economists point of view but is certainly regarded as negative from an ecologists point
of view. The carbon dioxide emissions related to this scenario are shown in Figure 1.17.
Even with fossil fuel consumption peaking around 2010, these high emission levels
might remain constant for 30 years before they begin to decline. And the worst scenario
with respect to greenhouse gas emissions (with an assumed R/P-ratio for coal of 370
years) maintains this emission level until the end of this century.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

28

7:36 PM

Page 28

J. SCHINDLER AND W. ZITTEL

Data Source: LBST 2005

FIGURE 1.17 Carbon dioxide emissions according to the production profiles as outlined in the
high fossil scenarios in Figure 1.15

But both scenarios are not as bad as the business as usual scenarios calculated by the
Intergovernmental Panel on Climate Change (IPCC). These scenarios are not based on a
corresponding available resource base. In contrast, the above described low fossil
scenario (Figure 1.16) would even be a help for a climate protection policy as total carbon
dioxide emission could drop by about 40 per cent by 2050 and by about 5560 per cent
by 2100.

1.3 ALTERNATIVE WORLD ENERGY SCENARIOS


1.3.1 RENEWABLE ENERGY SCENARIOS
Renewable scenarios can be built in various ways by extrapolating past trends and
incorporating expected future changes. The IEA builds its scenarios primarily using
economic considerations in which no drastic price changes for fossil and renewable fuels
are assumed. Therefore it is no surprise that no incentive for a rising contribution from
renewables is foreseen.
In contrast, the following scenarios are calculated by extrapolating observed past
growth trends and then limiting the further growth by taking into account the estimated
total supply potential for the different technologies. These are in effect market penetration
scenarios, but they do not represent a forecast, rather they describe what would be
possible under favourable market conditions (while still starting from the empirical data
describing the past development). No assumptions are made regarding actual future
market conditions. The resulting scenarios therefore show the technical limitations of
possible future developments, disregarding economic limitations.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

7:36 PM

Page 29

Alternative World Energy Outlook 2006

29

The main purpose of these scenarios is not so much to show how global energy
demand might be supplied in 2100, but to sketch the major characteristics of the
transition period between todays fossil-fuel-based energy economy and a possible future
based on renewable energies.
The following scenarios are modelled with so-called logistic functions which grow
exponentially in the beginning and then approach the ultimate potential attributed to
these sources with continually decreasing growth rates. The general pattern of the
function is fitted to the historical data by adapting two relevant parameters, namely,
T0, the time at which the growth rate is at maximum and the parameter b, which
provides a measure for the growth rate. This procedure is sketched in principle in
Figure 1.18.
For each of the ten world regions as defined in Figure 1.11 and each renewable
energy technology a growth scenario was modelled and adapted to the historical
development in that region. The parameters used for this modelling are listed in Annex II.
In Annex III the historical development of renewable energy generation over the last ten
years is provided. Wind energy has shown by far the largest growth rates, of about
3040 per cent, for the last two decades. The average growth rate of solar electricity
generation is between 20 and 30 per cent, increasing over the last few years to almost 40
per cent.
In the short term the growth is influenced by the experienced growth rate of
recent years. However, long-term installations are highly sensitive to the assumed potential.
Therefore the potential was estimated for each region and each technology individually. Due
to the broad variation of the estimates, an upper and lower figure was assumed for most

FIGURE 1.18 Principle of the logistic growth function and the meaning of its parameters, P, T0 and b

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

30

7:36 PM

Page 30

J. SCHINDLER AND W. ZITTEL

FIGURE 1.19 Technical potentials of electricity production from renewable energies; keep in
mind that the thermal biomass potential is three times larger than its shown potential for
electricity generation.3143 (The potential for solar thermal power in China and Asia is also very
large; however, it was not considered in this study.)

technologies. A graphical summary of the potentials of different renewable energy sources


is shown in Figure 1.19. These data are also provided in the annexes.
The assumptions for the calculation of these potentials are based on literature sources
and our own calculations.3143 For the scenario calculations a figure close to the minimum
value of the respective potential was always used.
The calculation of this so-called technical potential was based on the following
methodological approach13, 31 and the resource studies indicated:
Geothermal energy: according to common practice the potential for electricity

production was calculated by multiplying the number of active volcanoes within the
specified region with a scaling factor.32, 33
Hydropower: the results are based on a literature survey.32, 34
Photovoltaics: the potential was calculated by multiplying the forecast number of
inhabitants and the gross domestic product with a specific factor which takes care
of available roof areas and solar radiation. The link between population and gross
domestic product was used to estimate the available roof area in 2100. This
methodology is described in detail in Notes 13 and 35. Therefore the basic
assumption is that PV modules are only roof and faade mounted. Obviously, this
underestimates the full PV potential.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

7:36 PM

Page 31

Alternative World Energy Outlook 2006

31

Biomass: biomass is the most discussed renewable resource as the potential varies

by a factor of ten, depending on the assumptions regarding arable land area and
fertilizer use. The results are based on a literature survey and on our own
calculations.36, 37
Wind energy: the results are based on literature survey and our own calculations.3840
Solar thermal power plants: for North Africa and the Middle East the results of the
Mediterranean study (DLR) were used. These results were adapted to other
regions.4143
In Annex I the detailed data are provided, including the potential for thermal energy
conversion.
Nobody today can accurately forecast how much of this potential will be used in the
future. This depends on many aspects which are not of a technical nature. Most important
will be the acceptance by the public in the face of increasing environmental and supply
problems. Society has to and will decide the value of environmental aspects on the one
hand and the value of producing and using energy on the other hand. To partly account
for such decisions, the scenario calculations were based on the minimum values of the
estimates of the technical potential, which by definition include environmental restrictions
as they are obvious today but do not include additional restrictions from lacking public
acceptance.
The aggregate renewable energy scenario is shown in Figure 1.20. It represents the
aggregation of the possible market share potentials for each technology in each world
region. In general it is found that in the long run plenty of renewable energy will become
available. The capacity increase of renewable energies might eventually reach its
maximum around 2060. But most important: primary renewable energies will supply most
energy in the form of electricity and not as fuel or heat.
The focus of the analysis is on the next 20 to 30 years, which are identified as
the critical period. The accurate size of the renewable potential doesnt influence the
growth rates during the next 20 to 30 years, so is of minor importance in the present
context.
Another factor influencing the net energy balance is the energy needed for the
construction of the new generation capacity. For a rough estimate relevant data of today
are used. These are summarized in Table 1.1 and are derived from detailed life cycle
analyses based on todays energy mix in Europe. Since the dominant factor in this balance
is the consumption for extraction and production of the raw materials, most of the energy
needed is in the form of electricity. For these calculations, the electricity consumption is
traced back to the primary fuel consumption based on the present European electricity
mix. Since these data change in the same way as the energy mix and technological
progress, they might be regarded to be a worst case estimate of the energy consumption
during the period of construction.
The detailed calculation including re-powering of the decommissioned power
plants reveals that until 2020 less than 200Mtoe are consumed annually for the
construction of new generation capacity. This figure increases to about 500Mtoe around
2040 and reaches its maximum of about 1000Mtoe around 2070. About two-thirds of the

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

32

7:36 PM

Page 32

J. SCHINDLER AND W. ZITTEL

Data Source: LBST 2005

FIGURE 1.20 Possible market penetration of renewable energy sources

TABLE 1.1 Energy consumed for the construction of new power plants31

Hydropower
Wind energy
Geothermal electricity
Geothermal heat
Biomass (electricity)
Biomass (heat)
Photovoltaic (EFG-Si)
Solar thermal electricity

COAL
Mtoe/GW
0.92
0.12
0.13
0.022
0.13
0.06
0.2
0.18

OIL
Mtoe/GW
0.2
0.046
0.026
0.004
0.027
0.013
0.074
0.045

GAS
Mtoe/GW
0.026
0.019
0
0
0
0
0.042
0.073

NUCLEAR
Mtoe/GW
0.088
0.009
0
0
0
0
0.26
0.002

TOTAL
Mtoe/GW
1.24
0.19
0.14
0.026
0.157
0.073
0.58
0.37

energy is consumed as coal and one tenth as oil. As already mentioned, in the long run
the energy efficiency of the production processes as well as the energy mix may improve.
Therefore the total energy consumption for capacity building does not seem to be a
bottleneck restricting the growth rates of renewables during phases of declining oil or gas
supplies.
The full potential of the renewable energy sources was not utilized for the calculations
shown in Figure 1.20. If this were done it would mainly influence the shares of primary
energy supply towards the end of this century (or even beyond) but does not much
influence the development in the coming decades.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

7:36 PM

Page 33

Alternative World Energy Outlook 2006

33

These are the main conclusions concerning the renewable energy scenario:
For technological reasons it is obvious that most of the renewable energy potential

will be produced as electricity.


The production of heat will be more restricted to the local heat demand.
The production of fuels might become the most challenging task in a renewable

scenario, as it must be produced primarily from electricity.


Biomass is the most discussed renewable source as some people assume a ten

times larger potential than anticipated in this analysis. Other observers see biomass
as a very problematic energy source as it competes with food production and as the
potential might diminish with progressing climate change. How much of the biomass
will enter the market in the form of heat, fuel or electricity will strongly depend on
the technological progress, regional implementation of technologies, cost
competitiveness and demand. This question is left open in these scenario
calculations. Biomass is seen as heat production (different from Figure 1.19). The
conversion of biomass into fuels or electricity would reduce these figures by
between 10 and 70 per cent.

1.3.2 ALTERNATIVE WORLD ENERGY OUTLOOK


Figure 1.21 provides an overview of the high fossil alternative world energy
scenario, including the declining supply of fossil resources at the bottom, which is
steadily substituted and extended by the growth of renewable energy additions. In
this graph, nuclear energy is treated with the same conversion factor from
secondary to primary energy as the renewables to give evidence of its limited
importance.
From this analysis the conclusion can be drawn that a transition from todays fossilfuel-based energy system to a sustainable long-term strategy based on renewable
energies will be possible. It is also obvious that no other long-term alternatives are
foreseeable. It does not matter much whether the energy supply at the end of this
century is somewhat above or below the stated figures because this will not alter the
general picture. A possible worldwide demand rise must necessarily be restricted by
the available supply, which requires that our consumption habits and technologies
must adapt.
The basic trends in the foreseeable transition show that the bottleneck of energy
availability will emerge during the first quarter of this century. While the availability of the
primarily used fossil fuels begins to decline, our consumption patterns would not have
adapted to the diminishing supplies and the use of renewable energies would still be in
its infancy. Therefore the transition period will be crucial and we have to evaluate possible
bottlenecks in more detail.
To show the possibility of an even more demanding transition period in the next 2030
years, a low fossil alternative world energy outlook is presented in Figure 1.22. This is
based on a 5 per cent annual decline of oil between 2010 and 2020 and a less progressive
rise of natural gas with a peak plateau lasting from 2025 until 2035 (and not 2045 as in the
ASPO scenario used in the high fossil outlook). Furthermore, the growth of coal supply is

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

34

7:36 PM

Page 34

J. SCHINDLER AND W. ZITTEL

Data Source: LBST 2005

FIGURE 1.21 High fossil alternative world energy outlook until 2100 the transition from
fossil fuels to renewable energy sources, based on the high fossil scenario (Figure 1.15)

restricted to 55 per cent (4200Mtoe/yr) and not to 75 per cent or 4700Mtoe/yr as before.
The probability of such a scenario is relatively high as world oil supply is now declining at
far more than 5 per cent per year in almost all offshore regions which have passed peak
production. Modern technology has helped to extract oil at a faster rate but this implies
that the eventual decline is even steeper.
In this scenario the next 2030 years will probably see a declining primary world energy
supply. The potential growth rates of renewable energies are large, but the absolute share
is still too small to substitute an oil decline in the order of 5 per cent annually.
Rising energy costs are to be expected because of rising fossil fuel prices and the
more expensive renewable energy technologies. However, renewable energy will become
cheaper as its market share grows. This effect can be estimated by a scaling function
which models declining prices. Typical industrial learning curves have led to a cost
reduction of some 1020 per cent for each doubling of the total cumulative production
volume since the beginning of industrial production.
Typical installation costs of renewable energy technologies and their possible cost
degression potential are given in Table 1.2. The fundamental difference in the utilization of
renewable and fossil energy sources is that the major part of the costs for renewable
energies occur during the installation as capital costs (this is to some extent also true for
nuclear energy), while the costs of fossil sources are dominated by the fuel costs.
Therefore it is legitimate to compare fossil fuel costs with installation costs for a rough
assessment of the financial impacts. The only exception is biomass, which is also

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

7:36 PM

Page 35

Alternative World Energy Outlook 2006

35

Data Source: LBST 2005

FIGURE 1.22 Low fossil alternative world energy outlook until 2100 the transition from
fossil fuels to renewable energy sources, based on the low fossil scenario (Figure 1.16)

TABLE 1.2 Projected cost reductions of renewable energy technologies


ELECTRICITY
HYDRO WIND PV
SOT
Invest (/kW) 2004
3500
1400 6900 3800
Cumulative produced volume 2005 (GW) 860
48
4
0.3
Cost reduction at doubling (%)
0
15
20
4
Life time (years)
80
25
30
30

GEO
4000
5
30

HEAT
GEO
SOLAR
800
1200 (/m2)
9
100 million m2
10
20
30
30

dominated by fuel cost. Figure 1.23 provides a cost comparison of total energy supply
costs on this basis.
Past prices are adjusted for inflation to 2005. The projection is made on the basis that
oil prices will rise to US$100/bbl, gas prices to 0.28/m3, coal prices to 140/t and
biomass prices to 5 Euro-cents/kWh. The conclusion is that energy costs will rise, but this
rise will not be all that dramatic. The most pronounced increase is due to rising fossil fuel
costs. This phase will be followed by declining renewable installation costs, which in turn
will lead to growing market shares.
The detailed cost figures shown in Figure 1.23 should not be regarded as pretending
to be precise quantitative forecasts. The purpose of this rough scenario is only to
demonstrate qualitatively what could be expected. Generally, in the case of possible

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

36

7:36 PM

Page 36

J. SCHINDLER AND W. ZITTEL

We are here!

Data Source: LBST 2005

FIGURE 1.23 Comparison between fossil fuel costs and renewable installation costs; the
increase of fossil fuel prices sets the order of magnitude; with growing market penetration
renewable energies get cheaper, spurring further growth of market share

energy shortages caused by insufficient fossil fuel supplies, the prices of fossil fuels will
probably rise more than those of renewable energies. In such a situation higher overall
energy price levels must be expected.

1.3.3 TRANSITION TOWARDS SUSTAINABILITY


The world is entering a crucial transition period: from nearly total reliance on fossil
energies to a sustainable energy future based on renewable energy sources. The world
will experience the peaking of oil first, followed by natural gas. The resulting energy
supply gap cannot be filled by a rising share of nuclear energy and can only partly be filled
by an increased use of coal. But without CO2 sequestration, this would increase carbon
dioxide emissions to unacceptable levels.
The limits of growth will then become apparent and this will probably happen rather
soon. Renewable energies will have to fill the gap they have the necessary potential.
This is clearly shown by the alternative world energy outlooks until 2100. But even when
renewables are fully used, the transition remains a difficult task.
The arrival of peak oil will lead to rising and eventually permanently higher energy
prices. However, this is by no means a catastrophe. High energy prices will initially hurt
the economy, which is not adapted to such an environment. Yet higher energy prices must
be seen as a necessary precondition for the transition to a sustainable energy future.
In the longer run high energy prices are not the problem but the first step towards a
solution.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

7:36 PM

Page 37

Alternative World Energy Outlook 2006

37

One important aspect is the efficient use of energy. In our present system, with low
energy prices, efficiency is not really attractive and the economic rewards are weak.
This will certainly change with higher prices and also the social valuation of efficient
energy use will increase. This is the case of private virtue versus public responsibility.
Declining oil and natural gas supplies for stationary applications (like electric power
generation and the provision of heat) can relatively easily be substituted by energy
efficiency measures and by strongly rising contributions from renewable energy sources.
Probably there will be plenty of renewable electricity available but not enough fuels for
transport. A serious deficit of fuels will develop over time which can only partly be
compensated by improved efficiency along the whole transport chain.
There exist various options for producing alternative fuels: hydrogen produced with
electricity, hydrogen or synthetic liquid fuels produced with biomass, and hydrogen or
synthetic liquid fuels produced with the remaining coal instead of using it for power
generation. But whatever option is chosen eventually, fundamental changes in lifestyles
and in consequence in the whole economy will be caused by limited supplies and higher
energy prices.
The transport intensity will have to be reduced without endangering mobility, as will
the energy intensity of the production and use of goods. Less available energy and higher
fuel costs will increase transport resistance. There will be a stronger preference for
activities in the neighbourhood and there will be a penalty on longer distances. This will
eventually change land use patterns suburbia in many cases will not work anymore
and perhaps there will be a revival of the cities in the USA. The dominance of the private
car will recede and the perception of what properties an attractive car should have will
also change. Other modes of transport will gain importance and social acceptance.
The industrialized economies rely on growth. Once growth rates are shrinking this is
seen as a symptom of crisis no growth is a disaster. So nobody is prepared for a future
with continually declining fossil energy supplies. This scenario is not on the radar of public
awareness. Rising energy prices will change the perception and only then will society look
for sustainable pathways into the future. It will be a major task to adapt our way of life,
our economies and our political and social systems to this situation. There are no ready
solutions. A comprehensive public debate has not started yet. The outcome is open.
Another question is: What level of energy consumption is sustainable? What level of
human energy use can nature bear without destroying the basis of (human and other) life
on Earth? One line of thought is the idea of a 1.5kW society every human being may only
use this amount of energy all the time. If this level of energy use were to be distributed
justly among all human beings it would mean a substantial reduction of energy use in
industrialized countries but also a substantial rise of energy use in poor countries.
To achieve such a goal, Germany would have to cut average energy use by a factor 4
(down from 5.5kW) and the USA by a factor 8! Half of this reduction probably is achievable
through more efficient technologies, but the other half will have to be brought about by a
change in lifestyles. Both approaches will be needed.
The transition certainly will not be easy and there will be many conflicts and many losers.
One reason is that the build-up of new energy supply structures needs very long lead times:
it is very doubtful whether there is still enough time for a transition without turmoil.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

38

7:36 PM

Page 38

J. SCHINDLER AND W. ZITTEL

The challenge will be to master this transition period in a compatible and controllable
manner. The viable path between economic disruptions on the one hand and ecological
disruptions on the other is very narrow. The later we start to realize the real challenges and
the necessity for drastic changes, the more difficult it will be to cope with the challenges
lying ahead.

AUTHOR CONTACT DETAILS


Jrg Schindler, Ludwig-Blkow-Systemtechnik GmbH, Ottobrunn, Germany
e-mail: schindler@lbst.de
Werner Zittel, Ludwig-Blkow-Systemtechnik GmbH, Ottobrunn, Germany
e-mail: zittel@lbst.de

NOTES AND REFERENCES


1 International Energy Agency (2005) World Energy Outlook 2005, Paris.
2 J. Schindler and W. Zittel (2004) The countdown of world oil production And what are the views of the most
important international energy agencies, www.energybulletin.org, published 2 October 2004.
3 J. Schindler and W. Zittel (2005) comment on the IEAs World Energy Outlook 2005, www.energiekrise.de, November.
4 C. Campbell, F. Liesenborghs, J. Schindler and W. Zittel (2003) wechsel!, Deutscher Taschenbuchverlag, 2nd edition,
Munich, Germany.
5 W. Zittel and J. Schindler (2005) Oil depletion in V. Lauber (ed) Switching to Renewable Power A Framework for the
21st Century, Earthscan, London and, Sterling, VA, pp2161.
6 J. Schindler and W. Zittel (2004) Energieversorgung am Wendepunkt, Schriftenreihe Club Niedersterreich, no 89,
Wien, December 2004.
7 M. Simmons (2004) The Saudi Arabian Oil Miracle, presentation at the Center for Strategic and International Studies
(CSIS), Washington, 24 February.
8 M. Simmons (2005) Twilight in the Desert: The Coming Saudi Oil Shock and the World Economy, John Wiley & Sons,
Hoboken, New Jersey, US.
9 Mahmoud Abdul-Baqi, Aramco vice president, exploration, and Nansen Saleri, Aramco manager, reservoir management
(2004), Fifty year crude oil supply scenarios: Saudi Aramcos perspective, presentation at the Center for Strategic and
International Studies (CSIS), Washington, 24 February, on the occasion of a discussion with M. Simmons.
10 Sadad al-Husseini, Saudi Aramco: At the current depletion rate of 3 billion bbl/y, which represents 2.3 per cent of the
remaining 130 billion bbl of proven developed reserves, quoted in K. Aleklett (2004) From Paris to Berlin Steps
towards the final countdown to peak oil & gas, presentation at the 3rd International Workshop on Oil and Gas Depletion,
Berlin, May 2526.
11 S. Bakhtiari (2004) World oil production capacity model suggest output peak by 200607, Oil & Gas Journal, 26 April.
12 C. Campbell, Association for the Study of Peak Oil, see www.peakoil.net or www.peakoil.ie.
13 P. Schmidt, W. Weindorf, R. Wurster, M. Zerta and W. Zittel (2005) Global energy market analysis, task 1 in Large scale
hydrogen production in Patagonia, unpublished report, LBST, Munich, Germany.
14 J. Schindler and W. Zittel (forthcoming) Das Frdermaximum von Erdl und Erdgas.
15 Petroleum Exploration and Production Statistics PEPS, HIS-Energy, Geneva and London, 2004.
16 Own analysis by extrapolating the historical production data from HIS-Energy and adapting the total production volume to
the field size as provided in The worlds gas potential 1995, Petroconsultants, Geneva, 1995.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

7:36 PM

Page 39

Alternative World Energy Outlook 2006

39

17 BP Statistical Review of World Energy at www.bp.com, 2005.


18 US Department of Energy, Energy Information Administration, see statistics at www.eia.doe.gov.
19 National Energy Board (2004) Canadas oil sands: Opportunities and challenges to 2015, May, Publications Office,
National Energy Board, Calgary, Alberta.
20 See Figure 6 in Annual Energy Outlook 2002 with Projections to 2020 by the Energy Information Administration, US
Department of Energy, 2002 at www.eia.doe.gov/oiaf/archive.html#aeo.
21 Peck Hwee Sim (2001) Grim outlook for US methanol producers, Chemical Week, 28 March.
22 G. Serjeant (2003) UK dips toe in nightmare future of disappearing oil, The Times, 12 September.
23 Figures based on UK government statistics of oil and gas fields production, see internet information of the UK
Department of Trade and Industry at www.og.dti.gov.uk/pprs/pprsindex.htm.
24 See detailed analyses of the downward revisions of the proved reserves of Ormen-Lange at www.energiekrise.de/news
from 15 April 2004.
25 Power reactor information system (PRIS), IAEA, online statistics at www.iaea.org/DataCenter/statistics.html, July 2005.
26 Statistics are taken from the homepage of DTI (UK), NPD (Norway), TNO (The Netherlands) and BP Statistical Review of
World Energy.
27 Analysis of field production profiles by J. Laherrere. The data are taken from Petroleum Exploration and Production
Statistics (PEPS) from IHS-Energy.
28 International Atomic Energy Agency Energy (2005) Electricity and nuclear power estimates for the period up to 2030,
IAEA-RDS-1/25, Vienna, July.
29 P. Gerling et al (2003) Reserven, Ressourcen und Verfgbarkeit von Energierohstoffen 2002, Bundesanstalt fr
Geowissenschaften und Rohstoffe, Hanover, Germany.
30 International Atomic Energy Agency, internet database PRIS, see www.iaea.org.
31 LBST calculations with the E3 Database software, which allows a full lifecycle analysis concerning materials
consumption; energy consumption; pollutant emissions; including carbon dioxide; and cost. The calculations were
performed by W. Weindorf.
32 J. Bjornsson, I. Fridleifsson, Th. Helgason, H. Jonatansson, J. M. Mariusson, G. Palmason, V. Stefansson and L.
Thorsteinsson, (1998) The potential role of geothermal energy and hydropower in the world energy scenario year 2020,
Paper presented at the 17th World Energy Congress, Houston, Texas, September 1998.
33 V. Stefansson (2005) World Geothermal Assessment, proceedings of the World Geothermal Congress, Antalya, Turkey,
2429 April.
34 World Energy Council (WEC) (2001) Survey of energy resources: Hydropower, www.worldenergy.org/wecgeis/publications/reports/ser/hydro/hydro.asp.
35 V. Quaschning (2000) Systemtechnik einer Klimavernderlichen Elektrizittsversorgung in Deutschland fr das 21.
Jahrhundert, Fortschritt-Berichte VDI, Reihe 6: Energietechnik, VDI-Verlag GmbH, Dsseldorf, Germany.
36 M. Hoogwijk (2004) On the Global and Regional Potential of Renewable Energy Sources, Proefschrift Universiteit Utrecht,
Faculteit Scheikunde, The Netherlands.
37 M. Kaltschmitt and H. Hartmann (eds) (2001) Energie aus Biomasse Grundlagen, Techniken und Verfahren, Springer,
Berlin, Heidelberg and New York.
38 EWEA and Greenpeace (2004) Windforce 12 A blueprint to achieve 12 per cent of the worlds electricity from
windpower by 2020, May.
39 D. L. Elliot, L. L. Wendell and G. L. Gower (1991) An assessment of available windy land area and wind energy potential
in the contiguous United States, Pacific Northwest Laboratory, Richland, Washington, PNL-7789.
40 M. Kruska, D. Ichiro, M. Ohbayashe, K. Takase, L. Tetsunari, G. Evans, S. Herbergs, H. Lehmann, K. Mallon, S. Peter, A.
Sekine, K. Suzuki and D. Assmann (2003) Energy Rich Japan, Institute for Sustainable Solutions and Innovations (ISUSE),
www.energyrichjapan.info.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

40

7:36 PM

Page 40

J. SCHINDLER AND W. ZITTEL

41 H. Klaiss and F. Staiss (eds) (1992) Solarthermische Kraftwerke im Mittelmeerraum, Deutsche Forschungsanstalt fr Luftund Raumfahrt/Zentrum fr Sonnenenergie und Wasserstoffforschung, Springer, Berlin, London and New York.
42 F. Trieb (ed) (2005) Concentrating solar power for the Mediterranean region, final report by the German Aerospace
Center (DLR), Institute of Technical Thermodynamics, study commissioned by the Federal Ministry for the Environment,
Nature Conservation and Nuclear Safety, Germany, 16 April.
43 A. Leitner (2002) Fuel from the sky: Solar powers potential for Western energy supply, BDI consulting / NREL/SR-55032160, National Renewable Energy Laboratory, Golden, Colorado, July.
44 W. Weiss, I.Bergmann and G. Faninger (2006) Solar Heat Worldwide, Solar Heating & Cooling Program, International
Energy Agency, Paris, France.
45 European Solar Thermal Industry (ESTIF), see www.estif.org.
46 European Photovoltaics Industry Association (EPIA), see www.epia.org.
47 Global Wind Energy Council, see www.gwec.net.
48 IEA (various years) Energy statistics and balances of OECD countries, Paris.
49 IEA (2005) Energy statistics and balances of non-OECD countries, Paris.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

7:36 PM

Page 41

Alternative World Energy Outlook 2006

41

Annex I: Renewable Energy Potentials13, 3143


Hydro
OECD North America
Min
1.5
Scenario in 2100
1.3
Max
1.5
OECD Pacific
Min
0.2
Scenario in 2100
0.15
Max
0.2
OECD Europe
Min
0.8
Scenario in 2100
0.7
Max
1
Transition economies
Min
2.2
Scenario in 2100
2.2
Max
2.2
China
Min
1.9
Scenario in 2100
1.7
Max
1.9
East Asia
Min
0.9
Scenario in 2100
1
Max
0.9
South Asia
Min
0.7
Scenario in 2100
0.7
Max
0.7
Latin America
Min
2.9
Scenario in 2100
1.9
Max
2.9
Middle East
Min
0.2
Scenario in 2100
0.2
Max
0.2
Africa
Min
1.8
Scenario in 2100
1.3
Max
1.8

ELECTRICITY [1000Twhheat/yr]
Wind
PV
SOT

Geo

HEAT [Twhheat/yr]
Biomass
Geo

Solar

14
5.5
14

3.2
4.9
5.1

3.4
1.2
3.4

0.1
0.4
0.8

5.3
5.6
23.9

>1
0.9
>8

~0.3
0.3
>0.3

3.6
2.7
3.6

1.9
2.1
3

6.8
2.5
6.8

0.1
0.24
1

0.6
1.1
15.3

>1
1.2
>10

~0.1
0.1
>0.1

3.8
4.0
4.2

1.2
1.3
1.9

2.6
1.5
2.8

0.1
0.4
0.3

2.3
2.2
5.2

>1
1
>3

~0.2
0.2
>0.2

10.6
6.0
10.6

3.5
3.5
5.6

<0.1
0
<0.1

0.2
0.7
1.7

2.7
3.5
38.9

>2
2
>17

~0.1
0.3
>0.1

4.6
4.2
4.6

2.2
2.4
3.6

<0.1
0
<0.1

<0.1
0.03
0.1

2
3.2
30

4.6
2.5
4.6

1.1
1.1
1.8

0
0
0

0.4
0.79
1.7

0.9
1.1
4.9

4.6
2.8
4.6

1.5
1.8
2.5

0.3
0.25
0.3

<0.1
0.01
<0.1

0.8
1.7
8.6

5.4
5.2
5.4

3.6
4
5.8

0.2
0.1
0.2

0.4
0.7
2.7

6.1
7
31.4

<0.1
0.1
0.9

1.2
1.7
1.9

8.3
9
224

<0.1
0.05
0.3

0.1
0.1
3.8

1
0.1
>3

~0.1
0.1
>0.1

10.6
8.4
10.6

6.3
6.5
10

0.2
0.35
1.3

5.7
5.8
39

>2
0.4
>13

~0.5
1.3
>0.5

40
44
414

1
0.6
1
>4
4
>17
1
0.4
1
>4
6
>27

~0.5
1
>0.5
~0.8
0.5
>0.8
~0.6
0.3
>0.6
~0.3
1
>0.3

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

42

7:36 PM

Page 42

J. SCHINDLER AND W. ZITTEL

Annex II: Simulation Parameters for Renewable


Energy Scenarios
ELECTRICITY
Hydro Wind
PV
OECD North America
P (GW)
400
T0 (year)
2010
b (years)
30
OECD Pacific
P (GW)
90
T0 (year)
2020
b (years)
30
OECD Europe
P (GW)
300
2020
T0 (year)
b (years)
65
Transition economies
P (GW)
700
T0 (year)
2040
b (years)
20
China
P (GW)
500
T0 (year)
2030
b (years)
17
East Asia
P (GW)
360
T0 (year)
2039
b (years)
16
South Asia
P (GW)
250
T0 (year)
2031
b (years)
15
Latin America
P (GW)
700
T0 (year)
2040
b (years)
24
Middle East
P (GW)
80
T0 (year)
2052
b (years)
20
Africa
P (GW)
500
T0 (year)
2063
b (years)
20

SOT

Geo

Biomass1

HEAT
Geo2

Solar3

2500
2040
6.5

2900
2055
5.7

500
2050
6.2

50
2030
11

5000
2030
20

750
2050
10

300
2045
20

900
2040
5.5

1500
2036
4.5

1000
2060
10

30
2035
10

100
2045
12

600
2046
6.5

110
2026
12

1500
2023
5

1100
2035
4.4

600
2050
5.5

50
10
2040

200
2025
22

500
2040
10

250
2027
10

2000
2060
5.7

3500
2070
5

100
2050
6.5

300
2037
11

1000
2060
8.4

300
2050
3.5

1500
2040
4.8

1600
2045
4

3
2039
7

200
1950
40

300
2040
8

1000
2020
5.5

1000
2060
5.4

770
2050
4

100
2047
12

100
1970
11

2000
2060
5

500
2040
6

1000
2050
8

900
2040
3.8

100
2035
3.5

1
2030
8

150
1990
7,5

2
2035
15

400
2030
6

1500
2045
4.5

2000
2060
4.5

50
2040
4

100
2070
12

600
2020
10

3000
2065
7

1000
2030
6

30
2050
6

680
2050
3.8

3000
2060
5.5

6
2040
10

10
2020
13

30
2050
10

100
2035
7

3000
2060
6

4000
15,000
2090
2070
6.8
5.5

60
2060
9

500
2005
16

200
2060
8

1000
2045
7

1 Potential in Mtoe
2 Potential in GW thermal energy
3 Potential in million m2

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

7:36 PM

Page 43

Alternative World Energy Outlook 2006

43

Annex III: Growing Share of Renewable Energy Since


1990

Year
For reasons of comparison the installed collector area is converted into power by the factor 0.7kW per m2. The data have been
collected over the years by the LBST from various different statistics, including data from Notes 4349. Though the absolute share
of photovoltaics is very small, its growth rate is accelerating, exceeding 40 per cent over recent years. Its absolute contribution is
lagging about ten years behind wind energy.

FIGURE 1.A.1 Cumulative installed capacity of wind energy converters, geothermal power plants,
photovoltaic modules and solar collectors

~910% p.yr.

Year
Photovoltaics, geothermal, wind and biomass are restricted to electricity generation. For reasons of comparison the heat
generation from solar collectors is also included. Not included are contributions from hydropower or heat from geothermal
sources and biomass. Fuels production from biomass is also not included. The data have been collected over the years by the
LBST from various statistics including data from Notes 4349.

FIGURE 1.A.2 Energy production from renewable sources

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-01

2/6/07

44

7:36 PM

Page 44

J. SCHINDLER AND W. ZITTEL

Wind power
Biomass (incl. traditonal use)

Year
Year
Biomass includes so-called traditional use in developing countries. Electricity from hydropower and other sources is directly
converted into Mtoe. The share of other renewables is relatively small, but exhibits by far the largest growth rates. The data
have been collected over the years by the LBST from various statistics including data from Notes 4348. The conversion from
capacity to energy production was performed by the LBST.

FIGURE 1.A.3 Primary energy from renewable sources

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 144

ch-02

2/6/07

7:36 PM

Page 45

Quantum Well Solar Cells


N. J. Ekins-Daukes

Abstract
The quantum well solar cell represents the highest efficiency nano-structured solar cell
demonstrated to date. The concept and practical constraints for this cell are discussed
together with its potential application in high efficiency multi-junction solar cells and
thermophotovoltaic devices. Finally the ultimate efficiency limit of the quantum well
solar cell is discussed.

Keywords quantum well; uncertainty principle; density of states; recombination; voltage


enhancement; thermophotovoltaics; electroluminescence; entropy

2.1 HISTORY
Since its inception, the quantum well (QW) has been applied widely in semiconductor
devices. Its roots stem from the resonant tunnel diode and it has found applications in a
large number of electronic devices, such as photodetectors, light-emitting diodes, lasers,
optical modulators and high mobility transistors. The almost universal utility of QWs is due
to the ability to engineer the optical and electronic properties of a QW structure to suit a
particular application. It is then reasonable to suppose that some advantages may exist
when using QWs in photovoltaic structures.
Early work (Chaffin et al, 1984) suggested that a superlattice could provide a means of
achieving highly efficient multi-junction solar cells, allowing the constituent band gaps to
be matched to the solar spectrum, without recourse to lattice mismatched techniques or
complex quaternary materials. However, the most decisive step in bringing QWs to
photovoltaics was made in the early 1990s with the proposal of the multiple quantum well
(MQW) solar cell (Barnham and Duggan, 1990). Here it was suggested that the QWs
should be located in the intrinsic region of a pin diode and the proposal was quickly
followed by a successful experimental demonstration in the AlGaAs/GaAs material system
(Barnham et al, 1991) showing a considerable increase in short-circuit current (Jsc).
The ability to adjust the absorption profile of a bulk semiconductor through the
inclusion of QWs gives rise to a number of practical, near-term applications for QW solar
cells, such as a component of a multi-junction solar cell, or in thermophotovoltaics where
low band-gap materials are required.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

ch-02

2/6/07

46

7:36 PM

Page 46

N. J. EKINS-DAUKES

QWs and other low dimensional structures also offer some possibilities for testing
unconventional concepts in photovoltaics. For example, the proposal by Barnham and
Duggan also predicted that some advantage in the open-circuit voltage (Voc) could be
attained over bulk semiconductor material. This aspect has stimulated an interesting
debate over what electronic properties are necessary to achieve a photovoltaic voltage
enhancement in low dimensional semiconductors.
Many excellent review papers have already been written on lattice matched QW solar
cells (Barnham et al, 2002; Nelson, 2001 and 1995) and are accompanied by a large volume
of journal papers and conference proceedings. To distinguish this review paper from earlier
ones, the role of strain-balance is emphasized and is presented as a means of broadening
the range of materials that can be usefully applied in QW photovoltaic devices.

2.2 QUANTUM WELL ELECTRONIC STRUCTURE


Over the past 20 years an extremely large volume of literature has been published on the
properties of QWs. Most pertain to devices other than solar cells, but the general
properties of the QW remain the same, irrespective of the application. A number of books
have now been published on the physics of QW devices (Weisbuch and Vinter, 1991;
Bastard, 1988) and one book has a particular emphasis on QW solar cells (Barnham and
Vvedensky, 2001). What follows is a brief description of the properties of QWs and the
reader is referred to the detailed texts above for a full discussion.
In essence, a QW is a thin layer of material possessing a lower band gap than the
surrounding barrier or host material. The QW must be sufficiently thin to bring about
quantum effects through the confinement of electrons and holes, so its dimensions are
of the order of the electron and hole de-Broglie wavelength, typically a few nanometres
wide. The structural properties of the QW should be similar to the barrier, meaning that
there is no discontinuity in crystal symmetry and ideally the interfaces should be abrupt.
Under these conditions, the QW potential can be treated as a perturbation of the crystal
potential, allowing delocalized states to extend across the two materials. The properties
of these states closely resemble those of a quantum 1D potential well, and they can be
described in an analogous fashion using the envelope function approximation (Barnham
and Vvedensky, 2001).
Figure 2.1a shows how the QW and barrier layers are physically stacked. The resulting
electronic structure when the QWs are made sufficiently thin is shown in Figure 2.1b. The
envelope functions for each confined energy level are shown for two bound electron and
hole levels in each QW. The lowest bound level in the QW does not lie at the bottom of the
QW, but exists at an elevated energy. This is a direct result of confining the electron and
hole and can be simply understood through Heisenbergs uncertainty principle x p h/2.
Decreasing the well width, x increases the uncertainty in the momentum p and hence
increases the energy of the lowest bound level. The absorption edge of the QW can
therefore be controlled simply by varying the well width. The parity of the confined levels
is well defined and therefore restricts the allowed optical transitions, so for example the
e1h1 will be a strong transition, while e1h2 is weak. As the electron and hole are only
confined in one dimension, the carriers are free to move in a 2D plane, giving continuous
in-plane momentum dispersion curves for each bound level. This is illustrated in Figure 2.1c

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

ch-02

2/6/07

7:36 PM

Page 47

Quantum Well Solar Cells

47

k||

FIGURE 2.1 a) Layer structure showing three QWs surrounded by barrier material; b) Band diagram for
three QWs, showing two confined states for electrons and holes; c) Energy vs in-plane momentum dispersion
relation for a QW, showing two confined electron levels (e1 and e2) and the first heavy hole (hh1) and the first
light hole (lh1) levels

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

ch-02

2/6/07

48

7:36 PM

Page 48

N. J. EKINS-DAUKES

for two bound electron levels (e1 and e2) and the first heavy and light hole levels. The
degeneracy in the hole levels at k=0 is lifted when the QW is strained (Barnham and
Vvedensky, 2001).
The quantum confinement of carriers also serves to restrict the density of states.
Figure 2.2 compares the step-like behaviour of the QW 2D density of states to the
continuous 3D density of states; the expressions for the 2D and 3D densities of states are
expressed below:

m
h2

(1)

1 2m 3
(
)2 E
2 2 h2

(2)

g2D (E) =

g3D (E) =

where m* is the carrier effective mass. It is clear that in the 2D case, the density of states
is independent of energy for each bound level and depends only on the effective mass. It
is this simple dependence of the QW density of states that gives rise to the many
applications for QWs in electronic devices. For example, the strength of optical transitions
is governed by the density of states, so materials can be chosen to optimize the oscillator
strength of a particular transition, and the well width can be adjusted to achieve the
desired transition energies.
Finally, the number of bound states in a QW is of particular importance for the
QW solar cell, as each bound level adds further states from which absorption can

FIGURE 2.2 Density of states for a 3D bulk semiconductor g3D and for a 2D quantum well structure g2D

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

ch-02

2/6/07

7:36 PM

Page 49

Quantum Well Solar Cells

49

take place. The number of bound states in a QW is given by the following expression
(Bastard, 1988):

n(L) = 1 + Int[

2m Vb L2
]
2 h2

(3)

where Vb is the potential of the confining barrier, L is the well width and Int returns the
lowest integer value of the argument. The role of the effective mass is again important
and can lead to situations, such as in an Al0.3GaAs/GaAs QW, where the band offset gives
a deeper well for the electrons than for holes, yet the larger hole mass more than
compensates for the band offset, giving more heavy hole bound states than bound
electron states.
When a series of QWs is grown, if the barriers are thick, such that only a negligible
fraction of each QW wave function penetrates into the neighbours, then each QW can
be considered isolated and the structure is termed a multiple QW (MQW). The properties
of the MQW can be considered simply as a linear multiple of single QWs. However, when
the barriers separating each well are sufficiently thin to allow sufficient overlap of the
wave functions from adjacent wells, then the QWs become coupled and a superlattice
is formed (Dingle et al, 1975). Under these circumstances, the energy levels
characteristic of an isolated QW broaden to form mini-bands that extend throughout the
superlattice. There is no clearly defined point at which an MQW becomes a superlattice,
as narrowing of the barriers leads to an ever increasing broadening of the confined
levels. Strictly it is only by defining an arbitrary bandwidth that the transition from MQW
to superlattice can be determined. However, for most structures discussed in this review,
the thermal broadening of the confined levels at room temperature is greater than any
superlattice effects, so the structures can be safely called MQW devices. Some
consideration is given to superlattices in section 2.3.4.
Before concluding this section, it is important to note that the QW band alignment
considered above is one where both the electron and hole are confined and is called a
type-I alignment. On occasions, the band alignment of the QW and barrier material are
such that only the electron or hole is confined, but not both, and the QW is then said to
be type-II. To date only type-I QW arrangements have been pursued for photovoltaic
applications as the optical transitions are generally stronger. Nevertheless, type-II band
alignments have some desirable properties, such as the suppression of Auger
recombination (Tang et al, 1995), and have been exploited for optoelectronic applications
(Gevaux et al, 2001).

2.3 BASIC OPERATION OF THE P-I-N QUANTUM WELL SOLAR CELL


The QW solar cell proposed by Barnham and Duggan (1990) is shown schematically in
Figure 2.3a. The device consists of a p-i-n structure with the QWs grown in the i-region.
Epitaxial techniques such as molecular beam epitaxy (MBE) and metal organic vapour
phase epitaxy (MOVPE) can be used to fabricate QW structures with monolayer accuracy
(Barnham and Vvedensky, 2001; Stradling and Klipstein, 1989). Figure 2.3b shows the
band diagram for the QW solar cell and shows the main carrier photogeneration and
recombination paths in the barrier and QW layers.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

ch-02

2/6/07

50

7:36 PM

Page 50

N. J. EKINS-DAUKES

FIGURE 2.3 a) QW p-i-n layer structure; b) Band diagram for the QW p-i-n solar cell showing the
photogeneration and recombination processes, together with the carrier capture and escape routes

2.3.1 PHOTOGENERATION AND CARRIER ESCAPE


For the QWs to contribute to the photocurrent, the carriers must escape from the QW
through a combination of thermal and tunnelling processes (Nelson et al, 1993); this is
shown schematically in Figure 2.3b together with the corresponding carrier capture
process. The carrier escape is efficient providing that there is sufficient thermal energy
(Barnes et al, 1996; McFarlane et al, 1999) and the transverse electric field is sufficiently
strong (Nelson et al, 1993; Serdiukova et al, 1999). In most cases studied to date, there is
sufficient thermal energy at room temperature for the carriers to escape the QW and in
the presence of a strong electric field the probability for escape is unity under short-circuit
conditions.
An example of efficient photogenerated carrier collection is shown in the quantum
efficiency (QE) profile in Figure 2.4 for the GaAsP/InGaAs MQW p-i-n device whose
structure is depicted in Figure 2.3. The GaAsP barriers are responsible for the majority of
the QE from 400 to 860nm and the InGaAs QW extends the solar cell response from 860
to 920nm. In this example only one electron level is bound and the first transition in the
step-like density of states is obscured by the presence of exciton levels giving rise to
sharp peaks in the absorption. These exciton levels persist at room temperature, due to
the increased exciton binding energy in QW layers (Barnham and Vvedensky, 2001). It is
clear that, apart from a modest Stark-shift in the QW QE (Monier et al, 1999), the collection
efficiency remains high from short-circuit through to forward bias.
Nevertheless, if the background doping is too high, then the field across the QWs
becomes insufficient to remove carriers efficiently and the carrier collection collapses,
both in the bulk and QW. A spectacular example of this is demonstrated in Figure 2.5,
showing bias-dependent QE for a strain-balance structure composed of Al0.04GaAsP0.06/

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

ch-02

2/6/07

7:36 PM

Page 51

Quantum Well Solar Cells

51

FIGURE 2.4 Typical quantum efficiency for a GaAsP0.06/In0.1GaAs 75, 35 MQW structure at various biases

InGaAs. The inclusion of just 4 per cent Al into the GaAsP barriers introduces sufficient
background impurities1 to give an almost negligible spectral response in forward bias.
Similar results are observed if the QWs are placed in the doped regions of the device, at
which point the diffusion length is reduced to the order of the barrier width (Ekins-Daukes,
2000; Barnham et al, 2002).
There are also instances where there is insufficient thermal energy to allow efficient
carrier escape from the QWs, for example in InP/InGaAs QWs (Zachariou et al, 1998a) and
in particularly deep strain-balanced InGaAs/InGaAs QWs (Rohr et al, 2002a). However, if the
electric field is maintained across the QWs and the temperature is sufficiently high, then
efficient carrier escape is maintained beyond the operating point for most MQW solar cell
devices. This is illustrated in the case of the GaAsP/InGaAs device by the light IV curve
shown in Figure 2.6. The temperature coefficient for MQW solar cells has been found to
be marginally better than for equivalent p-i-n solar cells made from the well and barrier

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

ch-02

2/6/07

52

7:36 PM

Page 52

N. J. EKINS-DAUKES

FIGURE 2.5 Quantum efficiency for an Al0.04GaAsP0.06/In0.1GaAs 75, 35 MQW showing how background
impurities lead to a collapse of the QE at forward bias

material (Ballard et al, 1998 and 2001). Historically the AlGaAs/GaAs material system was
the first material system in which the p-i-n QW solar cell was demonstrated (Barnham et
al, 1991) and was followed by other lattice matched material systems, such as InGaP/GaAs
(Osbourne, 1994; Zachariou et al, 1998b), InP/InGaAs (Freundlich et al, 1994; Zachariou et
al, 1996; Zachariou, 1996) and InGaAsP/InP (Rohr et al, 1998 and 1999; Raisky et al, 1998;
Rohr et al, 2000a; Raisky et al, 2001). A number of strained material systems have also
been investigated, for example GaAs/InGaAs (Barnes et al, 1994; Ragay et al, 1994a;
Barnes et al, 1996; Barnes, 1994), InP/InAsP (Freundlich et al, 1994), InP/InGaAs
(Serdiukova et al, 1997; Freundlich, 2000) and also some strain-balance structures based
on GaAsP/InGaAs (Ekins-Daukes et al, 1999), GaAs/InGaAs (Ekins-Daukes, 2000; EkinsDaukes et al, 1998), InGaAs/InGaAs (Rohr, 2000; Rohr et al, 2000b and 2002b). The strained
and strain-balance materials are discussed further in section 2.3.3.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

2/6/07

7:36 PM

Page 53

Quantum Well Solar Cells

Current / mAcm2

ch-02

SC=32.1

53

mAcm2

Voc = 09.55V
FF = 82.9%
Eff=18.7%

Bias/V
Source: Bushnell (2002)

FIGURE 2.6 Light IV curve for a GaAsP0.06/In0.1GaAs 75, 35 MQW structure under approximately 1 sun
AM0 equivalent 3000K tungsten halogen illumination

Despite the complexity of the absorption profiles of the MQW solar cell, it has been
possible to model the spectral response accurately using the envelope function
approximation outlined earlier (Paxman et al, 1993; Renaud et al, 1994; Nelson, 1995).
Computer modelling has enabled advanced cell structures to be designed, featuring
compositionally graded layers and back surface mirroring, to improve minority carrier
transport and optical absorption (Connolly, 1997; Connolly et al, 1998).

2.3.2 CARRIER RECOMBINATION AND DARK IV


Recombination takes place across both the barrier and QW material and can be
understood using standard recombination models for radiative and non-radiative SRH
recombination (Nelson, 2003). Experimentally, it is generally found that the Voc of a lattice
matched QW solar cell is lower than that of a control cell made of the barrier material, but
higher than that of a control cell made of the well material. If a comparison is made
between the Voc of barrier and well p-i-n control cells grown in the same MOVPE or MBE
machine, one observes that the QW cell has a marginally higher Voc than for a bulk cell of
equivalent absorption edge (Barnham et al, 1996 and 2002), estimated from a linear
interpolation between well and barrier p-i-n control cells. This effect is commonly called
voltage enhancement, but the term sometimes leads to confusion; it is the voltage that is
enhanced with respect to a bulk p-i-n cell composed of the well material, not the barrier
material.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

ch-02

2/6/07

54

7:36 PM

Page 54

N. J. EKINS-DAUKES

To give a broad indication of the recombination behaviour that can be expected from
a QW solar cell, Figure 2.7 shows values of J0 plotted against the cell absorption edge for
a large number of solar cells, with band gaps spanning the solar spectrum. A theoretical
estimate of the radiative limit is also plotted, showing the lower limit for J0, assuming unity
absorptivity (Shockley and Queisser, 1961).

Ln(J0)

FIGURE 2.7 Plot of J0 vs absorption edge for a variety of lattice matched, mismatched and MQW devices

The values of J0 have been extracted by a simple process, taking experimental values
for Jsc and Voc and assuming that the solar cell behaves as an ideal diode. Under these
conditions J0 is estimated as follows:

Jsc = J0 e

qVoc
kT

(4)

where q is the electronic charge, k is Boltzmanns constant and T is the device


temperature. Extracting J0 in this way only gives an approximate means of comparing PV
devices. In particular, it assumes that the fill-factor is recombination limited and that
radiative recombination (unity diode ideality) dominates. Under 1 sun illumination, most
QW and p-i-n solar cells have a diode ideality of two, whereas good quality p/n cells
usually have an ideality of one. Such differences in ideality have been neglected here, as
the comparison in J0 is made across many orders of magnitude. Light IV data has been
taken from a large number of sources: AlGaAs/GaAs MQW (Paxman, 1992),
GaAsP/InGaAs MQW (Bushnell, 2002), InGaAs/InGaAs MQW (Connolly and Rohr, 2003),

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

ch-02

2/6/07

7:36 PM

Page 55

Quantum Well Solar Cells

55

InGaAsP/InGaAsP MQW (Connolly and Rohr, 2003), bulk GaInP (Yang et al, 1997), bulk
GaAs (Tobin et al, 1990), lattice mismatched InGaAs (Takamoto et al, 2000; Hoffman et al,
1998), bulk GaInNAs (Friedman et al, 1998; Li et al, 2000), bulk In0.53GaAs (Takamoto et al,
1997), bulk InGaAsP (Takamoto et al, 1997) and bulk Ge (Nagashima et al, 2003).
It is clear from Figure 2.7 that similar performance is observed from most of the MQW
structures as bulk cells. Of particular interest are the low band-gap MQWs (Connolly and
Rohr, 2003) that show exceptionally low recombination, in comparison with the
neighbouring bulk p/n cells, making the cells attractive for thermophotovoltaics, as
discussed in more detail in section 2.4.2. It is also noteworthy that the strain-balanced
GaAsP/InGaAs cells (Ekins-Daukes et al, 1999) have a similar J0 to the lattice mismatched
InGaAs cells, leading to possible applications in multi-junction solar cells, discussed in
section 2.4.1.
When attempting to model the Voc or dark IV of an MQW cell, it is usually assumed
that the MQW system is sufficiently close to quasi-thermal equilibrium to allow electron
and hole quasi-Fermi levels to be used to describe the carrier populations in the QW. The
extent to which quasi-thermal equilibrium is established in an MQW is discussed in
section 2.5, but nevertheless, the dark currents in a number of QW systems have been
modelled accurately (Nelson et al, 1995; Nelson, 1995) including the effects of
asymmetric QW location in the i-region (Nelson et al, 1999).
The success of computer modelling of MQW structures has led to more general
MQW device optimization studies (Connolly et al, 2000 and 2001). While actual device
performance clearly depends on a large number of material parameters, these models
now have a certain degree of predictive power that make them useful for assessing
suitable MQW device designs.

2.3.3 STRAINED AND STRAIN-BALANCED QW SOLAR CELLS


Changing the QW width allows the absorption threshold to be adjusted to higher
energies, but there are situations where it is desirable to lower the absorption threshold.
When using GaAs or InP barriers, these lower band-gap materials are usually strained, so
for example InGaAs is a particularly good choice for devices based on GaAs (Barnes et al,
1994; Ragay et al, 1994a; Barnes et al, 1996; Barnes, 1994) and InP (Serdiukova et al,
1997; Freundlich, 2000; Freundlich et al, 1994; Freundlich, 1998).
As an example, a device composed of compressively strained InGaAs QWs and GaAs
barriers grown on a GaAs substrate is considered; the associated layer structure and band
diagram is shown in Figure 2.8. Note that the relaxed lattice parameter is indicated
laterally, illustrating the compressive strain required to grow the InGaAs QWs on the GaAs
lattice. In Figure 2.9, the quantum efficiency and dark IV characteristics for two strained
QW GaAs/InGaAs solar cells and a GaAs p-i-n control cell are shown. The increased
photocurrent due to the QWs is evident from 900nm to 1000nm; however, this is
accompanied by a very large rise in dark current due to defects introduced through strain
relaxation taking place at the top and bottom of the MQW stack (Griffin et al, 1996; Mazzer
et al, 1996). In the case of the 23 QW cell, the strain relaxation is so great as to degrade
the short wavelength QE. Unless measures are taken to passivate the defects associated
with the strain-induced misfit dislocations (Okada et al, 2002; Suzuki et al, 1999), the

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

ch-02

2/6/07

56

7:36 PM

Page 56

N. J. EKINS-DAUKES

FIGURE 2.8 Strained GaAs/InGaAs 3period MQW device: a) Schematic diagram showing the layer
configuration; the relaxed lattice parameter is indicated laterally; b) Electronic band-structure for the device

effects of strain relaxation are so severe that strained GaAs/InGaAs structures can never
offer a substantial increase in current over a GaAs cell without suffering a catastrophic
loss in voltage (Ekins-Daukes et al, 2001).
In the strained GaAs/InGaAs structure, strain builds up with each InGaAs layer.
However, it is possible to minimize the build-up of strain by incorporating alternate tensile
and compressively strained layers. Through an appropriate choice of alloy composition
and layer thickness, and accounting for differences in elastic constant, it is possible to
achieve structures which are locally strained but exert no net force on the substrate or
neighbouring repeat units. The layer thickness required to achieve a strain-balanced
structure can be estimated using the following expression (Ekins-Daukes et al, 2002):

a0 =

A1 t1 a1 a22 + A2 t2 a2 a21
A1 t1 a22 + A2 t2 a21

(5)

where a0 is the substrate lattice parameter, a1 and a2 are the respective QW and barrier
lattice parameters, t1 and t2 the respective QW and barrier layer thicknesses, and A1 and
A2 constants for the QW and barrier and defined through the elastic stiffness coefficients
Cxx as follows:

A = C11 + C12

2
2C12
C11

(6)

In principle, the critical thickness of a strain-balanced structure is infinite. In practice,


however, there will always be some shortcomings, both in the growth and design of a

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

ch-02

2/6/07

7:36 PM

Page 57

Quantum Well Solar Cells

57

FIGURE 2.9 a) External quantum efficiency for a strained GaAs/InGaAs MQW sample and GaAs control; b)
dark IV curves for GaAs/InGaAs strain-balance MQW cell, a relaxed GaAs/InGaAs MQW cell, a strained
GaAs/InGaAs MQW cell and a GaAs control

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

ch-02

2/6/07

58

7:36 PM

Page 58

N. J. EKINS-DAUKES

strain-balance stack. The layer thicknesses determined from Equation 5 are based on
linear elastic theory and Vegards law (Vegard, 1921) is usually used to relate the material
composition to lattice parameter. These are good approximations, but will become
increasingly inaccurate at high strain. Finally, the material parameters used to determine
the strain-balance condition are usually those measured at room temperature, yet the
device is grown at an elevated temperature (up to 650C), and differences in thermal
expansion between the component layers ensure that no structure can be exactly strainbalanced at both the growth temperature and room temperature.
A typical strain-balanced structure is shown in Figure 2.10. Here GaAsP is used as the
tensile material to strain compensate the compressive InGaAs layer. Such strain-balanced
structures have demonstrated unprecedented photovoltaic performance for QW solar
cells (Ekins-Daukes et al, 1999). The quantum efficiency and dark IV are shown in
Figure 2.11. The quantum efficiency plot indicates the additional absorption due to the
QWs over a GaAs p-i-n control sample. The dark current profiles show excessively high
rates of recombination for the strained and relaxed GaAs/InGaAs cells, but the
strain-balanced cell has a dark current profile comparable to the GaAs p-i-n control cell.
Therefore, by using the strain-balanced QW approach, strained, lower band-gap
semiconductors can be incorporated into a solar cell with no additional misfit dislocations
and can therefore maintain a high Voc. A remaining challenge is to increase the absorption
of the MQW stack, as the 20 MQW sample only has a QE of approximately 20 per cent
due to the overall low optical thickness of QW absorbing material. This could be solved
by simply growing more QWs into the structure, and the current state-of-the-art device is
a 50 MQW GaAsP/InGaAs device (Bushnell et al, 2003b). However, one cannot increase

FIGURE 2.10 Strain-balance GaAsP/InGaAs 3period MQW device: a) Schematic diagram showing the layer
configuration; b) Electronic band structure for the device

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

ch-02

2/6/07

7:36 PM

Page 59

Quantum Well Solar Cells

59

FIGURE 2.11 a) External quantum efficiency for the strain-balanced GaAsP/InGaAs MQW sample and
GaAs control; b) Dark IV curves for GaAsP/InGaAs strain-balance MQW cell, a relaxed GaAs/InGaAs MQW cell,
a strained GaAs/InGaAs MQW cell and a GaAs control

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

ch-02

2/6/07

60

7:36 PM

Page 60

N. J. EKINS-DAUKES

the number of QWs indefinitely. First, increasing the number of QWs places an ever
tighter tolerance on the growth process (Ekins-Daukes et al, 2002), so a growth technique
capable of producing a 20 QW structure may not be suitable for a 100 QW sample.
Second, the necessity to maintain a field across the QW for efficient carrier escape places
a stringent demand on the background doping level in the i-region. The devices discussed
above typically have i-regions approximately 1m wide, which through a simple
electrostatic calculation (Ekins-Daukes, 2000) requires a background doping level below
1 1015cm-3. If the background doping level rises to a point that the MQW is no longer
fully depleted, then the catastrophic loss in photocurrent observed in Figure 2.5 is
experienced. Third, the strained QW and barriers must not exceed their individual critical
thickness (Matthews and Blakeslee, 1974), thereby placing some limits on the well width
and number of bound levels in the QW. Indeed, even when the well and barrier remain
below their critical thickness, strain-induced composition modulation can degrade the
MQW under certain growth conditions (Bushnell et al, 2003b; Bushnell, 2002).
Such problems lead one to conclude that MQW devices could benefit from
light-trapping schemes (Yablonovitch and Cody, 1982; Steinke, 1996; Hepp, 1998). One
effective technique has been to incorporate a distributed Bragg reflector (DBR) into the
structure (Bushnell et al, 2003a), thereby doubling the optical path through the QW layers.
To approach unity absorption, however, the optical path length needs to be extended
further. Textured coatings and grating structures have been proposed to increase the
optical path through the QWs and some practical means for their formation investigated
(Bushnell et al, 2002; Bushnell, 2002).

2.3.4 SUPERLATTICE STRUCTURES


In comparison to the MQW, relatively little experimental work has been performed on
applying superlattices for photovoltaics, but nevertheless there are some interesting
opportunities in this area. While the superlattice maintains the advantages offered by
MQWs for engineering the band structure, it also allows transverse carrier transport
within the mini-bands, alleviating the need for carriers to escape the QWs. The early work
by Chaffin (Chaffin et al, 1984; Chaffin, 1987) proposed using superlattices in multijunction solar cells to obtain an optimal match to the solar spectrum.
In this proposal the superlattice was doped to form a p/n junction and this concept is
currently being pursued in an attempt to create an all silicon Si/SiO2 multi-junction solar
cell (Green, 2000). Superlattices have also been proposed to improve transport when
working with difficult materials (Varonides and Berger, 1997; Kojima et al, 2000) and in
particular to provide a 1eV junction for a multi-junction solar cell using InGaAsN (Bedair et
al, 2000).

2.3.5 RADIATION RESISTANCE OF QW SOLAR CELLS


The radiation resistance of InP/InAsP p-i-n QW solar cells has been investigated,
comparing the MQW against an InP p-i-n control cell (Walters et al, 2000a). The Jsc of the
QW is more robust against 3MeV protons than the InP pin control cell. However, this
advantage appears to be offset by a greater loss in Voc in the QW cell over the control,
leading to approximately the same power degradation in both cells. A detailed study

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

ch-02

2/6/07

7:36 PM

Page 61

Quantum Well Solar Cells

61

showed that the proton irradiation introduced a mid-gap trap that is likely to be associated
with the QW or QW Barrier interfaces (Walters et al, 2000b and 2001). Further work on the
radiation resistance of MQW solar cells is required, especially on GaAs-based QW
devices, which could usefully enhance the GaInP/GaAs/Ge multi-junction space solar cell,
discussed in section 2.4.1.

2.4 NEAR-TERM APPLICATIONS FOR QW SOLAR CELLS


Due to the high cost of launching a spacecraft, typically US$70,000/kg, the space
photovoltaic industry is able to use expensive and specialized photovoltaic technology,
with the primary aim of achieving a high specific power (power per unit mass). In general,
the rigid panel structures on which the PV cells are mounted dominate the specific power,
so the panel specific power scales with the efficiency of the solar cell. This makes high
efficiency cells such as the GaInP/GaAs/Ge triple junction particularly attractive for space
use (Fatemi et al, 2000). QW solar cells could therefore find practical applications as a
component of a multi-junction space solar cell (Freundlich and Serdiukova, 1998;
Freundlich, 2000 and 2002; Ekins-Daukes et al, 2000) and this application is discussed in
more detail below.
The terrestrial photovoltaic industry is continually looking for lower cost materials in
an attempt to bring down the cost of photovoltaic power generation. Nevertheless, there
are opportunities for highly specialized devices such as QW solar cells in high
concentration photovoltaics, where at sufficiently high concentration the cell cost
becomes a negligible component of the overall concentrator system cost (Yamaguchi and
Luque, 1999). QWs have been considered for concentrator applications (Barnham, 1992)
and, as for the multi-junction space solar cell, QWs hold some advantages for a
concentrator multi-junction cell (Tibbits et al, 2003).
In general, any photovoltaic application where precise tuning of the absorption edge
is required will lend itself to the use of QW structures. Examples include
thermophotovoltaics (Connolly and Rohr, 2003) (discussed below) or wavelength splitting
schemes such as the luminescent concentrator (Barnham et al, 2000).

2.4.1 MQW MULTI-JUNCTION SOLAR CELLS


There have been plenty of proposals for using MQWs in multi-junction solar cells, but it is
only recently that experiment has shown results that could be promising for multi-junction
applications. A possible near-term application is in the InGaP/GaAs solar cell, where it is
well known that the cell could be improved if the lower cell junction had a band gap
somewhat lower than that of GaAs. The problem can be overcome to some extent if the
InGaP layer is thinned, in order to match the photocurrents (Kurtz et al, 1990), but the cell
still operates with a compromised voltage. There is then an opportunity to engineer a
material using the strain-balance approach that could provide an ideal low band gap
bottom junction for a tandem solar cell. Figure 2.12 shows how the efficiency of a two
terminal tandem solar cell varies as a function of the top and bottom cell band gaps. The
extent to which the existing InGaP/GaAs is not optimally matched is clear, as is the
efficiency enhancement if the lower cell possessed an absorption edge around 1.3eV. This
approach can be considered an alternative to the popular bulk lattice mismatched layers

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

2/6/07

62

Upper Cell Band Gap/eV

ch-02

7:36 PM

Page 62

N. J. EKINS-DAUKES

Monolithic tandem lso-Efficiency Contours


One Sun; AM 0 Illumination
Temperature = 300K

Lower Cell Band Gap/eV


FIGURE 2.12 AM0 iso-efficiency contour plot for a monolithic tandem structure; the calculation is
described in Ekins-Daukes et al (2000)

(Dimroth et al, 2000), the advantage being that strain relaxation and the associated
degradation in material quality due to dislocations are avoided altogether.
To give an indication of the performance of these strain-balanced structures in a
tandem solar cell, Table 2.1 shows the projected Voc for both the lower cell and the
tandem structure, together with the Jsc and the tandem efficiency values. The calculation
is based on published data taken from a GaInP top cell (Yang et al, 1997) and a variety of
MQW candidates for the lower junction (Bushnell et al, 2003a; Ekins-Daukes, 2000).
Values for J0 are estimated from Jsc and Voc as described earlier, using equation 4. The
GaInP junction is modelled with J0 =5 1023mAcm2 and can deliver 20mAcm-2, making
the 2J cell current limited by the lower junction in all cases.
The GaInP/GaAs efficiency provides the baseline for comparison at 26 per cent and is
in reasonable agreement with experimental values (Takamoto et al, 2001). An improved
performance is achieved with both the GaInP/SB MQW and the GaInP/DBR SB-MQW, the

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

ch-02

2/6/07

7:36 PM

Page 63

Quantum Well Solar Cells

63

TABLE 2.1 AM0 Junction parameters for a GaInP top cell together with parameters for lower junction
cells composed of GaAs p/n, strain-balanced MQW, DBR enhanced strain-balanced MQW, strained MQW and
relaxed MQW; the thick GaInP cell is modelled with J0 = 5 1023 mAcm2 and can deliver 20mAcm2,
making the 2J cell current limited by the lower junction in all cases.

GaAs p/n
SB-MQW
DBR SB-MQW
Strn. MQW
Rlxd. MQW

LOWER CELL ONLY


J0mAcm2
1.7 1016
6.0 1015
6.0 1015
1.9 1013
9.6 1010

V0C/V
1.006
0.920
0.923
0.831
0.611

2J GaInP/ LOWER CELL


JSC/mAcm2
VOC/V
15.8
2.411
17.1
2.326
18.3
2.328
17.1
2.236
17.1
2.016

EFF.
27.1%
26.1%
28.8%
25.9%
23.2%

latter giving an efficiency in excess of 28 per cent AM0. These figures illustrate the relative
importance of an increase in Jsc over a loss in lower junction Voc. The GaInP/strained and
GaInP/relaxed MQW cell experiences an efficiency loss, due to the excessive dark current
evident from Figure 2.11b.
The figures projected in Table 2.1 look encouraging for the MQW cell as a component
of a multi-junction solar cell. However, the efficiency estimates are somewhat generous
to the MQW in terms of the fill factor. While J0 has been chosen to fit the Jsc and Voc, it
has been assumed that the diode ideality is unity, which is reasonable for p/n devices and
yields a fill factor around 90 per cent for the structures described in Table 2.1. However,
in practice the diode ideality is closer to two for MQW and p-i-n devices under 1 sun
illumination. This serves to reduce the fill factor to around 82 per cent and reduces the
efficiency advantage for the DBR cell to a negligible margin.
Before the introduction of the strain-balance technique, the MQW Voc was the limiting
factor in a multi-junction device. Now the strain-balance MQW devices give good Voc, so
work should proceed to raise the fill factor and identify the processes responsible for the
high MQW diode ideality. In addition, the absorption in the strain-balanced structure is still
quite weak, even with the DBR structure, and the QW current could be raised to match
the GaInP cell by employing the more sophisticated light-trapping techniques discussed
in section 2.3.3.

2.4.2 THERMOPHOTOVOLTAICS
In a thermophotovoltaic system the radiative emission from a heat source is collected
using a photovoltaic device (Coutts and Fitzgerald, 1998). The heat source is usually
followed by a selective emitter, which radiates strongly only over a narrow energy band.
It is therefore desirable to match the solar cell to the selective emitter, and for this
purpose the band-gap engineering offered by QW structures can be useful (Connolly and
Rohr, 2003). The first practical demonstration of a QW thermophotovoltaic cell was based
on lattice matched InGaAs QWs on InP (Griffin et al, 1997 and 1998) and showed an
absorption edge at 1.55m. The InGaAsP material system is extremely versatile as it offers
a wide range of band gaps that are lattice matched to InP. A particular virtue of this
material is that both the QWs and barriers can be grown from InGaAsP. This facilitates the

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

ch-02

2/6/07

64

7:36 PM

Page 64

N. J. EKINS-DAUKES

MOVPE growth process as neither the group III nor group V sources are shut off
completely, as is necessary when growing InP/InGaAs structures. InGaAsP MQW devices
can be designed to possess an absorption edge between 0.9 and 1.6m (Rohr et al, 1998,
1999 and 2000a). Such devices are particularly suitable for erbia selective emitters
operating at temperatures around 2000K.
There are some environmental advantages to lowering the selective emitter
temperature to 1500K, and for this purpose an emitter operating at longer wavelengths is
required (Abbott et al, 2002). Suitable emitters for this temperature include thulia and
holmia, requiring photovoltaic cells that can absorb emission at 1.75m and at 1.95m
respectively. Strained InGaAs QWs on InP are suitable for these long wavelengths
(Serdiukova et al, 1997; Freundlich, 2000) and the same strain-balance technique can be
used to ensure that the device is free of dislocations (Rohr, 2000; Rohr et al, 2000b and
2002b). A strain-balanced InGaAs/InGaAs MQW device grown on an InP substrate has
recently been demonstrated, with an absorption edge at 2m (Rohr et al, 2002a).

2.5 EFFICIENCY LIMITS


It has already been demonstrated that the QW solar cell has practical advantages in terms
of band-gap engineering. However, the issue of whether a QW solar cell has some
inherent efficiency advantage over traditional materials, such as GaAs, has been the
subject of some debate. In the original proposal, Barnham and Duggan (1990) predicted
that the Jsc of a QW solar cell would be enhanced over a solar cell made from the barrier
material. As discussed above, this result is well established and marks one of the highly
useful aspects offered by QW solar cells. However, they also predicted that the upper limit
for the Voc of the QW solar cell would be dominated by the high band-gap barrier material.
This prediction was controversial as it implied that the thermodynamically rigorous limit
for photovoltaic power conversion established by Shockley and Queisser (1961) could be
broken. A vigorous debate ensued which is still ongoing in certain areas. What follows is
a brief summary of the various theories that have been proposed and the objections that
have been raised. The terminology used here follows a recent review of the arguments for
efficiency enhancements in QW solar cells (Anderson, 2002). Briefly, a global efficiency
enhancement is defined as a theoretical limit for ideal solar cells, where the QW cell
outperforms a conventional bulk cell, when both are operating under idealized conditions.
A relative efficiency enhancement is a practical demonstration of a superior efficiency for
a real QW solar cell when compared with a conventional bulk solar cell. The applications
for the QW solar cell described in section 2.4 fall into a category defined as ancillary
advantages and are not of concern to us here.
In support of the original paper by Barnham and Duggan, a number of device
level predictions have been made indicating a global efficiency advantage (Corkish and
Green, 1993; Anderson, 1995; Rimada and Hernandez, 2001a and b) from QW solar cells.
In general they all assume that the relationship between absorption and recombination
can be broken. However, these device models are found to be thermodynamically
inconsistent when carriers are in quasithermal equilibrium in only the conduction and
valance band (Arajo et al, 1994; Ragay et al, 1994b; Arajo and Marti, 1995; Luque and
Marti, 1997a).

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

ch-02

2/6/07

7:36 PM

Page 65

Quantum Well Solar Cells

65

However, an important feature of the early thermodynamic consistency arguments


was that they assumed that the carrier mobility is infinite throughout the device and that
the carrier population in the QW is described by the same quasi-Fermi level as the barrier.
However, if carriers in the QW are subjected to an additional driving force, that is not in
equilibrium with the lattice, then the QWs can possess their own quasi-Fermi levels and
the possibility of a global efficiency enhancement emerges once more (Kettemann and
Guillemoles, 1995). This concept has been shown to be thermodynamically consistent in
a general form, where a third intermediate band is postulated from which carriers are
excited optically via solar photons into the conduction and valance band (Luque and Marti,
1997b).
The operation of an intermediate band QW device (Bremner et al, 1999) is similar to
that of a QW infrared photodetector (Levine, 1993). To date a number of mechanisms have
been proposed through which a three-band, QW device could show an efficiency
enhancement (Green, 2000; Honsberg et al, 2002); this has been extended to cover
multiple bands (Brown and Green, 2003; Peng et al, 2003). The subject of thermodynamic
consistency (Luque et al, 2001) has again been discussed, as has the detrimental effect
of transitions between intermediate bands in > 3-band models (Wrfel, 2003), but there
is broad agreement that, at least in theory, a global efficiency advantage is possible if the
MQWs can establish a third band. Alternatively, if a temperature difference can be
maintained between the QW and barrier material, a global efficiency enhancement is also
possible (Kettemann and Guillemoles, 2002).
The question then arises as to whether in practice a third band, or something
approximating it, exists in an MQW solar cell. The experimental evidence that shows an
enhanced Voc (Barnham et al, 1996) is a strong indication of a relative efficiency
enhancement that the MQW p-i-n can outperform a bulk p-i-n cell but it is important
to note that it does not confirm the existence of a global efficiency enhancement.
Nevertheless, a more detailed picture emerges when modelling the dark current (Nelson
et al, 1995) and electroluminescence (Nelson et al, 1997) from a QW solar cell. In both
cases the quasi-Fermi level in the QW has to be suppressed by approximately 20meV with
respect to the barrier material to obtain a good fit to experimental results. It is important
to note that the suppressed quasi-Fermi level is observed in the dark, where electrically
injected carriers are falling into the QW. So the quasi-Fermi level suppression may be due
to some transport-limiting process leading to a quasi-Fermi level gradient. In this case, if
light is absorbed in the QW and generates a net photocurrent, then the gradient in the
quasi-Fermi levels must be reversed to describe the flow of current and should lead to an
inflated quasi-Fermi level in the QW and therefore higher recombination (Luque et al,
2001). This is then an example of how a gradient in a quasi-Fermi level will always
generate entropy (Kondepudi and Prigogine, 1998) and therefore a lower photovoltaic
efficiency. However, recent experiments show that there is no change in luminescence
when moving from electrical injection, through Voc to photovoltaic behaviour (EkinsDaukes et al, 2003a and 2003b). This implies that there is no change in quasi-Fermi level
under the reversal of current and is inconsistent with the prediction that a suppressed QW
quasi-Fermi level in the dark will lead to an inflated quasi-Fermi level under illumination.
Such behaviour is essentially in conflict with the second law of thermodynamics, so

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

ch-02

2/6/07

66

7:36 PM

Page 66

N. J. EKINS-DAUKES

requires an alternative explanation. We can speculate that it may be inappropriate to


define quasi-Fermi levels over distances as short as a QW (Settler and Lundstrom, 1994;
Corkish and Honsberg, 1997). Also much of the QW luminescence stems from the
recombination of excitons, whose population has a strongly non-thermal momentum
distribution (Koch et al, 2002) and leads to a breakdown of Kirchoffs law relating
absorptivity and emissivity, on which efficiency limits derived through the principle of
detailed balance rely (Arajo and Marti, 1994). Nevertheless, one does not abandon quasiFermi levels lightly, as the analysis of photovoltaic efficiency becomes difficult without the
convenience of quasi-thermal equilibrium (Wrfel, 1995).
Such speculation, however, is only hinting at possible mechanisms for the 20meV
quasi-Fermi level suppression, and a much larger effect is required for a practical
intermediate band device. It has also been pointed out that the MQW is not expected to
be the best material to demonstrate an intermediate band (Luque et al, 2001), on account
of the freedom of movement for carriers in the plane of the well, which promotes
rapid phonon-mediated relaxation of carriers. A quantum dot solar cell with discrete
levels may be better in this respect (Marti et al, 2000), but nevertheless the same concern
applies over whether, in practice, an intermediate band can be established.
The conclusion is therefore that plenty of mechanisms exist on paper that predict a
global efficiency enhancement in an MQW device, but that experimental work
should proceed to establish what properties can realistically be exploited in QW and other
low dimensional materials. For example, a detailed investigation of electroluminescence
from AlGaAs/GaAs double QW samples composed of a wide and narrow QW
(Kluftinger et al, 2000 and 2001; Kluftinger, 2000) showed a quasi-Fermi level
suppression in the wide well, but not in the narrow well. The reason for this remains
elusive and may be the key to understanding the origin of the 20meV quasi-Fermi level
suppression.

2.6 CONCLUSION
Over the last 15 years a considerable quantity of work has been performed on the QW
solar cell and this marks one of the first attempts to apply low dimensional structures to
photovoltaics. The ability to engineer the device band gap has led to practical applications
in multi-junction solar cells and thermophotovoltaics, while the strain-balance technique
broadens the range of materials whose properties can be exploited in MQW or
superlattice structures. To date only the band gap has been deliberately engineered in QW
solar cell structures, but it is likely that other properties, such as the optical transitions and
phonon modes, will also be controlled in order to produce highly efficient devices in the
future.

ACKNOWLEDGEMENTS
Extensive discussions with Keith Barnham, Massimo Mazzer and Jenny Nelson are
gratefully acknowledged. David Bushnell kindly provided recent experimental data and
Martin Green and Peter Wrfel are thanked for useful discussions concerning efficiency
limits. The Japan Society for the Promotion of Science (JSPS) and Masafumi Yamaguchi
are thanked for their respective financial and scientific support.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

ch-02

2/6/07

7:36 PM

Page 67

Quantum Well Solar Cells

67

AUTHOR CONTACT DETAILS


N. J. Ekins-Daukes, School of Physics, University of Sydney, NSW 2006, Australia.
Tel: +61 2 9036 9259; Fax: +61 2 9351 7726; e-mail: ned@physics.usyd.edu.au

NOTE
1 The contamination is most likely to be due to C which is inevitably incorporated into AlGaAs during MOVPE growth on
account of the strong AlC bond.

REFERENCES
Abbott, P., Rohr, C., Connolly, J. P., Ballard, I., Barnham, K., Ginige, R., Corbett, B., Clarke, G., Bland, S. W. and Mazzer, M. (2002)
A comparative study of bulk InGaAs and InGaAs/InGaAs straincompensated quantum well cells for thermophotovoltaic
applications, proceedings of 29th IEEE PV Specialists Conference, IEEE, New Orleans, US, May, pp10581061
Anderson, N. G. (1995) Ideal theory of quantumwell solar cells, J. Appl. Phys., vol 78, no 3, pp18501861
Anderson, N. (2002) On quantum well solar cell efficiencies, Physica, vol E14, no 12, pp126131
Arajo, G. and Marti, A. (1994) Absolute limiting efficiencies for photovoltaic energy conversion, Solar Energy Materials and
Solar Cells, no 33, pp213240
Arajo, G. L. and Marti, A. (1995) Electroluminescence coupling in 24 multiple quantum well diodes and solar cells, Appl.
Phys. Lett, vol 66, no 7, pp894895
Arajo, G., Marti, A., Ragay, F. and Wolter, J. (1994) Efficiency of multiple quantum well solar cells, proceedings of 12th
European Photovoltaic Solar Energy Conference, H. S. Stephens & Associates, Bedford, UK, pp14811484
Ballard, I., Barnham, K. W. J., Connolly, J. P., Nelson, J., Rohr, C., Roberts, C., Roberts, J. and Button, C. (2001) The effect of
temperature on the efficiency of multi quantum well cells for solar and thermophotovoltaic applications, proceedings of
17th European Conference and Exhibition on PV Solar Energy, WIP and ETA, Munich, Germany and Florence, Italy, pp4144
Ballard, I., Barnham, K. W. J., Nelson, J., Connolly, J. P., Roberts, C., Roberts, J. S. and Pate, M. A. (1998) The effect of
temperature on the efficiency of multiquantum well solar cells, proceedings of 2nd World PV Energy Conversion
Conference European Commission, Ispra, Italy and Vienna, Austria, July, pp36243626
Barnes, J. (1994) An experimental and theoretical study of GaAs/InGaAs quantum well solar cells and carrier escape from
QWs, Ph.D. thesis, Imperial College, University of London, available from www.sc.ic.ac.uk/~q_pv/
Barnes, J., Ali, T., Barnham, K. W. J., Nelson, J. and Tsui, E. S. M. (1994) Gallium arsenide/indium gallium arsenide
multiquantum well solar cells, proceedings of 12th European PV Solar Energy Conference, H. S. Stephens & Associates,
Bedford, UK, pp13741377
Barnes, J., Nelson, J., Barnham, K., Roberts, J., Pate, M., Grey, R., Dosanjh, S., Mazzer, M. and Ghiraldo, F. (1996)
Characterization of GaAs/InGaAs QWs using photocurrent spectroscopy, J. Appl. Phys., no 79, pp77757779
Barnham, K. (1992) Solar Cell, UK Patent Application PCT/GB92/01913, EC Patent Application 92921571.3, US Patent
Application 211850, Australian Patent Application 27905/92
Barnham, K. W. J. and Duggan, G. (1990) A new approach to highefficiency multiband-gap solar cells, J. Appl. Phys., no 67,
pp34903493
Barnham, K., and Vvedensky, D. (2001) Low Dimensional Semiconductor Structures: Fundamentals and Device Applications,
Cambridge University Press, Cambridge, UK
Barnham, K. W. J., Braun, B., Nelson, J., Paxman, M., Button, C., Roberts, J. S. and Foxon, C. T. (1991), Shortcircuit current
and energy efficiency enhancement in a low dimensional structure photovoltaic device, Appl. Phys. Lett., no 59, pp135137
Barnham, K. W. J., Connolly, J., Griffin, P., Haarpaintner, G., Nelson, J., Tsui, E., Zachariou, A., Osborne, J., Button, C., Hill, G.,
Hopkinson, M., Pate, M., Roberts, J. S. and Foxon, T. (1996) Voltage enhancement in quantum well solar cells, J. Appl.
Phys., vol 80, no 2, pp12011206

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

ch-02

2/6/07

7:36 PM

68

Page 68

N. J. EKINS-DAUKES

Barnham, K., Marques, J. L., Hassard, J. and OBrien, P. (2000) Quantum dot concentrator and thermodynamic model for the
global redshift, Appl. Phys. Lett., vol 76 no 9, pp11971199
Barnham, K., Ballard, I., Connolly, J., Ekins-Daukes, N., Kluftinger, B., Nelson, J. and Rohr, C. (2002) Quantum well solar cells,
Physica, vol E14, no 12, pp2736
Bastard, G. (1988) Wave Mechanics Applied to Semiconductor Heterostructures, Les Editions de Physique, les Ulis
Bedair, S., Roberts, J., Jung, D., Moody, B., El-Masry, N. and Katsuyama, T. (2000) GaAsP/InGaAsN strained layer
superlattices for solar cell applications, proceedings of 28th IEEE PV Specialists Conference, IEEE, US, pp12691272
Bremner, S., Corkish, R. and Honsberg, C. (1999) Detailed balance efficiency limits with quasi-Fermi level variations, IEEE
Trans. Elect. Dev., vol 46, no 10, pp19321939
Brown, A. and Green, M. (2003) Intermediate band solar cell with many bands: Ideal performance, J. Appl. Phys., vol 94,
no 9, pp61506158
Bushnell, D. (2002) Optimisation of strain compensated multi-quantum-well solar cells, Ph.D. thesis, Imperial College,
University of London, available from www.sc.ic.ac.uk/~q_pv/
Bushnell, D. B., Barnham, K. W. J., Connolly, J. P., Ekins-Daukes, N. J., Airey, R., Hill, G. and Roberts, J. S. (2002)
Lighttrapping structures for multiquantum well solar cells, proceedings of 29th IEEE PV Specialists Conference, IEEE, New
Orleans, US, May, p1035
Bushnell, D., Ekins-Daukes, N., Barnham, K., Connolly, J., Roberts, J., Hill, G., Airey, R. and Mazzer, M. (2003a) Shortcircuit
current enhancement in Bragg stack multi-quantum-well solar cells for multi-junction space applications, Solar Energy
Materials and Solar Cells, vol 75, no 12, pp299305
Bushnell, D. B., Barnham, K. W. J., Connolly, J. P., Mazzer, M., Ekins-Daukes, N. J., Roberts, J. S., Hill, G., Airey, R. and Nasi, L.
(2003b) Effect of barrier composition and well number on the dark current of quantum well solar cells, proceedings of 3rd
World Conference on Photovoltaic Energy Conversion, Osaka, Japan, May
Chaffin, R. (1987) Quantum Well Multi-Junction Photovoltaic Cell, US Patent 4,688,068
Chaffin, R., Osbourn, G., Dawson and L., Biefeld, R. (1984) Strained superlattice, quantum well, multi-junction photovoltaic
cells, proceedings of 17th IEEE PV Specialists Conference, IEEE, US, pp743746
Connolly, J. (1997) Modelling and optimising GaAs/AlGaAs mqw solar cells, Ph.D. thesis, Imperial College, University of
London, available from www.sc.ic.ac.uk/~q_pv/
Connolly, J. P. and Rohr, C. (2003) Quantum well cells for thermophotovoltaics, Semiconductor Science and Technology, vol
18, no 5, p216
Connolly, J. P., Barnham, K. W. J., Nelson, J., Roberts, C., Pate, M. and Roberts, J. S. (1998) Short circuit current
enhancement in GaAs/AlGaAs MQW solar cells, proceedings of 2nd World PV Energy Conversion Conf. European
Commission, Ispra, Italy, pp36313634
Connolly, J. P., Nelson, J., Barnham, K. and Ballard, I. (2000) Multiple quantum well solar simulator, proceedings of 28th IEEE
PV Specialists Conference, Anchorage, IEEE, US, pp13041307
Connolly, J. P., Nelson, J., Ballard, I., Barnham, K. W. J., Rohr, C., Button, C., Roberts, J. and Foxon, T. (2001) Modelling multi
quantum well solar cell efficiency, proceedings of 17th European Conference and Exhibition on PV Solar Energy, WIP, Munich,
Germany
Corkish, R. and Green, M. (1993) Recombination of carriers in quantum well solar cells, proceedings of 23rd PV Specialists
Conference, IEEE, New York, pp675689
Corkish, R. and Honsberg, C. (1997) Dark currents in double-heterostructure and quantumwell solar cells, proceedings of
26th IEEE PV Specialists Conference, pp955960
Coutts, T. J. and Fitzgerald, M. C. (1998) Thermophotovoltaics, Scientific American, vol 279, no 3, pp9095
Dimroth, F., Schubert, U. and Bett, A. (2000) 25.5% efficient GaInP/GaInAs tandem solar cells grown on GaAs substrates,
IEEE Elect. Dev. Lett., vol 21, no 5, pp209211

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

ch-02

2/6/07

7:36 PM

Page 69

Quantum Well Solar Cells

69

Dingle, R., Gossard, A. and Wiegmann, W. (1975) Direct observation of superlattice formation in a semiconductor
heterostructure, Phys. Rev. Lett., vol 34, no 21, pp13271330
Ekins-Daukes, N. (2000) An investigation into the efficiency of strained and strain-balanced quantum well solar cells, Ph.D.
thesis, Imperial College, University of London, available from www.sc.ic.ac.uk/~q_pv/
Ekins-Daukes, N., Mazzer, M., Barnes, J., Ballard, I., Barnham, K., Holmes, S., Grey, R., Roberts, J., Pate M. and Hill, G. (1998)
New approaches in GaAs/InGaAs quantum well solar cells, proceedings of 2nd World Conf. and Exhibition on Photovoltaic
Solar Energy Conversion, European Commission, Ispra, Italy, p3627
Ekins-Daukes, N., Barnham, K., Connolly, J., Roberts, J., Clark, J., Hill, G. and Mazzer, M. (1999) Strain-balanced
GaAsP/InGaAs quantum well solar cells, Appl. Phys. Lett., vol 75, no 26, pp41954197
Ekins-Daukes, N., Bushnell, D., Zhang, J., Barnham, K. and Mazzer, M. (2000) Strain-balanced materials for highefficiency
solar cells, proceedings of 28th IEEE PV Specialists Conference, IEEE, US, pp12731276
Ekins-Daukes, N., Barnes, J., Barnham, K., Connolly, J., Mazzer, M., Clark, J., Grey, R., Hill, G., Pate and M., Roberts, J. (2001)
Strained and strain-balanced quantum well devices for high efficiency tandem solar cells, Solar Energy Materials and
Solar Cells, no 68, pp7187
Ekins-Daukes, N., Kawaguchi, K. and Zhang, J. (2002) Strain-balanced criteria for multiple quantum well structures and its
signature in X-ray rocking curves, Crystal Growth & Design, vol 2, no 4, pp287292
Ekins-Daukes, N., Ballard, I., Calder, C., Barnham, K., Hill, G. and Roberts, J., (2003a) Photovoltaic efficiency enhancement
through thermal up-conversion, Appl. Phys. Lett., vol 82, no 12, pp19741976
Ekins-Daukes, N., Calder, C., Ballard, I., Barnham, K., Nelson, J., Hill, G. and Roberts, J. (2003b) Quantum well and bulk quasiFermi
level behaviour under illuminated conditions, proceedings of 3rd World Conference on Photovoltaic Energy Conversion, Osaka,
Japan, May
Fatemi, N. S., Pollard, H. E., Hou, H. Q. and Sharps, P. R. (2000) Solar array trades between very high-efficency multi-junction
and Si space solar cells, proceedings of 28th IEEE PV Specialists Conference, IEEE, US, pp1083804
Freundlich, A. (1998) Strained Quantum Well Photovoltaic Energy Converter, US Patent 5,851,310
Freundlich, A. (2000) Quantum Well Thermophotovoltaic Convertor, US Patent 6,150,604
Freundlich, A. (2002) Multi-Quantum Well Tandem Solar Cell, cont. US Patent 6,372,980
Freundlich, A. and Serdiukova, I. (1998) Multi-quantum well tandem solar cells with efficiencies exceeding 30% AM0,
proceedings of 2nd World Conf. and Exhibition on Photovoltaic Solar Energy Conversion, European Commission, Ispra, Italy,
Vienna, Austria, July, p3707
Freundlich, A., Rossignol, V., Vilela, M., Renaud, P., Bensaoula, A. and Medelci, N. (1994) InPbased quantum well solar cells
grown by chemical beam epitaxy, proceedings of 1st World Conference and Exhibition on Photovoltaic Solar Energy
Conversion, IEEE, US, pp18861889
Friedman, D., Geisz, J., Kurtz S. and Olson, J. (1998) 1ev solar cells with GaInNAs active layer, J. Crystal Growth, no 195,
pp409415
Gevaux, D. G., Green, A. M., Phillips, C. C., Vurgaftman, I., Bewley, W. W., Felix, C. L., Meyer, J. R., Lee, H. and Martinelli,
R. U. (2001) GaInSb/InAs/AlGaAsSb W quantum-well light emitting diodes, Appl. Phys. Lett., vol 79, no 25,
pp40734075
Green, M. (2000) Prospects for photovoltaic efficiency enhancement using low-dimensional structures, Nanotechnology, vol
11, no 4, pp401405
Griffin, P. R., Barnes, J., Barnham, K. W. J., Haarpainter, G., Mazzer, M. Zanotti-Fregonara, C., Grunbaum, E., Olson, C., Rohr,
C., David, J. P. R., Roberts, J. S., Grey, R. and Pate, M. A. (1996) Effect of strain relaxation on forward bias dark currents
in GaAs/InGaAs multiquantum well pin diodes, J. Appl. Phys. no 80, pp58155820
Griffin, P., Ballard, I., Barnham, K., Nelson, J., Zachariou, A., Epler, J., Hill, G., Button, C. and Pate, M. (1997) Advantages of
quantum well solar cells for TPV, in T. J. Coutts, C. S. Allman and J. P. Benner (eds) Thermophotovoltaic Generation of

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

ch-02

2/6/07

7:36 PM

70

Page 70

N. J. EKINS-DAUKES

Electricity: Third NREL Conference, vol 401 of AIP Conference Proceedings, American Institute of Physics, Woodbury,
New York, Colorado Springs, Colorado, US, May, pp411422
Griffin, P., Ballard, I., Barnham, K., Nelson, J., Zachariou, A., Epler, J., Hill, G., Button, C. and Pate, M. (1998) The application
quantum well solar cells to thermophotovoltaics, Solar Energy Materials and Solar Cells, no 50, pp213219
Hepp, M. (1998) Light trapping in quantum well solar cells grown on patterned substrate, Masters thesis, Imperial College
of Science, Technology and Medicine, University of London., available from www.sc.ic.ac.uk/~q_pv/
Hoffman, R., Fatemi, N., Stan, M., Jenkins P. and Weizer, V. (1998) High efficiency InGaAs on GaAs devices for monolithic
multi-junction solar cell applications, proceedings of 2nd World Conf. and Exhibition on Photovoltaic Solar Energy
Conversion, European Commission, Ispra, Italy, pp36043608
Honsberg, C., Bremner, S. and Corkish, R. (2002) Design tradeoffs and rules for multiple energy level solar cells, Physica, vol
E14, no 12, pp136141
Kettemann, S. and Guillemoles, J. (1995) Limiting efficiency of LDS solar cells, proceedings of 13th European Photovoltaic
Solar Energy Conference, Stephens, Bedford, UK, pp119121
Kettemann, S. and Guillemoles, J. (2002) Thermoelectric field effects in lowdimensional structure solar cells, Physica, vol
E14, no 12, pp101106
Kluftinger, B. (2000) An experimental study of the quasiFermi level separations in quantum well solar cells, Ph.D. thesis,
Imperial College, University of London, available from www.sc.ic.ac.uk/~q_pv/
Kluftinger, B., Barnham, K., Nelson, J., Foxon, T. and Cheng, T. (2000) Temperature dependent study of the quasi-Fermi level
separation in double quantum well pin structures, Microelectronic Engineering, no 5152, pp265274
Kluftinger, B., Barnham, K., Nelson, J., Foxon, T. and Cheng, T. (2001) Temperature dependent study of radiative losses in
double quantum well solar cells, Sol. Energy Mater. Sol. Cells, no 66, pp510509
Koch, S., Meier, T., Hoyer, W. and Kira, M. (2002) Theory of the optical properties of semiconductor nanostructures, Physica,
vol E14, no 12, pp4552
Kojima, N., Goetzberger, O., Ohshita, Y. and Yamaguchi, M. (2000) Approach to the carbonbased superlattice structure solar
cell, proceedings of 28th IEEE PV Specialists Conference, IEEE, US, pp873875
Kondepudi, D. and Prigogine, I. (1998) Modern Thermodynamics, John Wiley & Sons
Kurtz, S., Faine, P. and Olson, J. (1990) Modeling of 2junction, series-connected tandem solar cells using top cell thickness as
an adjustable parameter, J. Appl. Phys., vol 68, no 4, pp18901895
Levine, B. (1993) Quantum well infrared photodetectors, J. Appl. Phys., vol 74, no 8, R1R81
Li, N. Y., Sharps, P. R., Stan, M., Newman, F., Hills, J. S., Hou, H. Q., Gee, J. M. and Aiken, D. J. (2000) Development of 1.25
eV InGaAsN for triple junction solar cells, proceedings of 28th IEEE PV Specialists Conference, IEEE, US, pp986989
Luque, A. and Marti, A. (1997a) Entropy production in photovoltaic conversion, Phys. Rev. B., vol 55, no 11, pp69946999
Luque, A. and Marti, A. (1997b) Increasing the efficiency of ideal solar cells by photon induced transitions at intermediate
levels, Phys. Rev. Lett., vol 78, no 26, pp50145017
Luque, A., Marti, A. and Cuadra, L. (2001), Thermodynamic consistency of subband-gap absorbing solar cell proposals, IEEE
Trans. Electron Devices, vol 48, no 9, pp21182124
Marti, A., Cuadra, L. and Luque, A. (2000) Quantum dot intermediate band solar cell, proceedings of 28th IEEE PV
Specialists Conference, IEEE, US, pp940943
Matthews, J. W. and Blakeslee, A. E. (1974) Defects in epitaxial multilayers (I), J. Crystal Growth, no 27, pp118125
Mazzer, M., Grnbaum, E., Barnham, K., Barnes, J., Griffin, P., Holt, D., Hutchinson, J., Norman, A., David, J., Roberts, J. and
Grey, R. (1996) Study of misfit dislocations by EBIC, CL and HRTEM in GaAs/InGaAs lattice strained multi-quantum well
pin solar cells, Materials Science and Engineering, no B42, pp4351
McFarlane, S. C., Barnes, J., Barnham, K. W. J., Tsui, E. S. M., Button, C. and Roberts, J. S. (1999) Space charge effects in
carrier escape from single quantum well structures, vol 86, no 9, pp51095115

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

ch-02

2/6/07

7:36 PM

Page 71

Quantum Well Solar Cells

71

Monier, C., Freundlich, A. and Vilela, M. (1999) Oscillator strength of excitons in InGaAs/GaAs QWs in the presence of a large
electric field, vol 85, no 5, pp27132718
Nagashima, T., Okumura, K., Murata, K. and Yamaguchi, M. (2003) A germanium backcontact type cell for
thermophotovoltaic applications, proceedings of 3rd World Conference on Photovoltaic Energy Conversion, Osaka; Japan,
May
Nelson, J. (1995) Quantum well structures for photovoltaic energy conversion, in M. Francome and J. Vossen (eds) Physics
of Thin Films, Academic Press, New York, vol 21, pp311368
Nelson, J. (2001) Quantum well solar cells, in M. D. Archer and R. Hill (eds) Clean Electricity from Photovoltaics, Imperial
College Press, London, pp7187
Nelson, J. (2003) The Physics of Solar Cells, Imperial College Press, London
Nelson, J., Paxman, M., Barnham, K. W. J., Roberts, J. S. and Button, C. (1993) Steady state carrier escape from single
QWs, IEEE Journal of Quantum Electronics, no 29, pp14601467
Nelson, J., Kluftinger, B., Tsui, E., Barnham, K., Roberts J. and Button, C. (1995) QuasiFermi level separation in quantum well
solar cells, proceedings of 13th European Photovoltaic Solar Energy Conference, H. S. Stephens & Associates, Bedford,
UK, pp150153
Nelson, J., Barnes, J., Ekins-Daukes, N., Kluftinger, B., Tsui, E., Barnham, K. W. J., Foxon, T., Cheng, T. and Roberts, J. S.
(1997) Observation of suppressed radiative recombination in single quantum well pin photodiodes, no 82, pp62406246
Nelson, J., Ballard, J., Barnham, K. W. J., Connolly, J., Roberts, J. S. and Pate, M. (1999) Effect of QW location on quantum
well solar cell dark currents, vol 86, no 10, pp58985905
Okada, Y., Seki, S., Takeda, T. and Kawabe, M. (2002) Control of dark currents in multi-quantum well solar cells fabricated by
atomic H assisted molecular beam epitaxy, J. Cryst. Growth, no 237, pp15151518
Osbourne, J. (1994) InGaP/GaAs quantum well solar cells, Masters thesis, Imperial College, University of London
Paxman, M. (1992) An experimental study of GaAs/AlGaAs multiple quantum well solar cells, Ph.D. thesis, Imperial College,
University of London
Paxman, M., Nelson, J., Braun, B., Connolly, J., Barnham, K. W. J., Foxon, C. T. and Roberts, J. S. (1993) Modeling the
spectral response of the quantum well solar cell, no 74, pp614621
Peng, R., Mazzer, M. and Barnham, K. (2003) Efficiency enhancement of ideal photovoltaic solar cells by photonic excitations
in multiintermediate band structures, Appl. Phys. Lett., vol 83, no 4, pp770772
Ragay, F., Wolter, J., Marti, A. and Arajo, G. (1994a) Experimental analysis of GaAsInGaAs MQW solar cells, proceedings of
1st World Conference and Exhibition on Photovoltaic Solar Energy Conversion, IEEE, US, p1754
Ragay, F., Wolter, J., Marti, A. and Araujo, G. (1994b) Experimental analysis of the efficiency of MQW solar cells,
proceedings of 12th European Photovoltaic Solar Energy Conference, H. S. Stephens & Associates, Bedford, UK,
pp14291433
Raisky, O. Y., Wang, W. B., Alfano, R. R., Reynolds, Jr., C.L. Stampone, D. V. and Focht, M. W. (1998) In1-xGaxAs1-yPY /InP
multiple quantum well solar cell structures, J. Appl. Phys., vol 84, no 10, pp57905794
Raisky, O. Y., Wang, W. B., Alfano, R. R. and Reynolds, Jr., C.L. (2001) Carrier screening effects in photoluminescence
spectra of InGaAsP/InP multiple quantum well photovoltaic structures, Appl. Phys. Lett., vol 79, no 3, pp430432
Renaud, P., Vilela, M., Freundlich, A., Bensaoula, A. and Medelci, N. (1994) Modeling p-i-n multi quantum well n solar cells: A
contribution for a near optimum design, proceedings of 1st World Conference and Exhibition on Photovoltaic Solar Energy
Conversion, IEEE, US, pp17871790
Rimada, J.C. and Hernandez, L. (2001a) Modelling of ideal AlGaAs quantum well solar cells, Microelectronics, vol 32, no 9,
pp719723
Rimada, J.C. and Hernandez, L. (2001b) A new approach to ideal AlGaAs MQW solar cells, Mod. Phys. Lett., vol B15, no
1719, pp778781

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

ch-02

2/6/07

7:36 PM

72

Page 72

N. J. EKINS-DAUKES

Rohr, C. (2000) InGaAsP quantum well cells for thermophotovoltaic applications, Ph.D. thesis, Imperial College, University of
London
Rohr, C., Connolly, J. P., Barnham, K. W. J., Griffin, P. R., Nelson, J., Ballard, I., Zachariou, A., Button, C. and Clark, J. (1998)
InGaAsP quantum well cells for thermophotovoltaics, proceedings of 2nd World Conf. and Exhibition on Photovoltaic Solar
Energy Conversion, Ispra, Italy and Vienna, Austria, July 1998, European Commission, pp230233
Rohr, C., Connolly, J. P., Barnham, K. W. J., Griffin, P. R., Nelson, J., Ballard, I., Button, C. and Clark, J. (1999) Optimisation of
InGaAsP quantum well cells for hybrid solarthermophotovoltaic applications, in T. J. Coutts, J. P. Benner and C. S. Allman
(eds) Thermophotovoltaic Generation of Electricity: Fourth NREL Conf., vol 460 of AIP Conference Proceedings, Denver,
Colorado, US, October 1998, American Institute of Physics, Woodbury, New York, pp8392
Rohr, C., Barnham, K. W. J., Connolly, J. P., Nelson, J., Button, C. and Clark, J. (2000a) Potential of InGaAsP quantum well
cells for thermophotovoltaics, proceedings of 26th International Symposium on Compound Semiconductors, no 166 in
Institute of Physics Conference Series, Berlin, Germany, August 1999, Institute of Physics Publishing, Bristoland,
Philadelphia, US, pp423426
Rohr, C., Connolly, J., Barnham, K. W., Mazzer, M., Button, C. and Clark, J. (2000b) Strain-balanced In0.62Ga0.38As/In0.47Ga0.53As
(InP) quantum well cell for thermophotovoltaics, proceedings of 28th IEEE Photovoltaic Specialists Conference, IEEE, US,
pp12341237
Rohr, C., Abbott, P., Ballard, I., Connolly, J. P., Barnham, K. W., Nasi, L., Ferrari, C., Lazzarini, L., Mazzer, M. and Roberts, J.
(2002a) Straincompensated InGaAs/InGaAs quantum well cell with 2m band-edge, in T. J. Coutts, G. Guazzoni and
J. Luther (eds) The Fifth Conference on Thermophotovoltaic Generation of Electricity. Vol. 653 of AIP Conference
Proceedings, American Institute of Physics, Woodbury, New York, pp344353
Rohr, C., Connolly, J. P., Ekins-Daukes, N., Abbott, P., Ballard, I., Barnham, K. W., Mazzer, M. and Button, C. (2002b)
InGaAs/InGaAs strain-compensated quantum well cells for thermophotovoltaic applications, Physica E: Low-dimensional
Systems and Nanostructures, vol 14, no 12, pp158161
Serdiukova, I., Newman, F., Aguillar, L., Vilela, M. F., Monier, C. and Ferundlich, A. (1997) Strained In0.53Ga0.47As/InxGa1-xAs
(x>0.6) multiquantum well thermophotovoltaic converters, proceedings of 26th IEEE PV Specialists Conference, IEEE, US,
pp963966
Serdiukova, I., Monier, C., Vilela, M. and Freundlich, A. (1999) Critical built-in electric field for an optimum carrier collection in
multi-quantum well pin diodes, Appl. Phys. Lett., vol 74, no 19, pp28122814
Settler, M. and Lundstrom, M. (1994) A detailed investigation of hetero junction transport using a rigorous solution to the
Boltzmann equation, IEEE Trans. Elec. Dev., vol 41, no 4, pp592600
Shockley, W. and Queisser, H. (1961) Detailed balance limit of efficiency of pn junction solar cells, J. Appl. Phys., vol 32,
no 3, pp510519
Steinke, L. (1996) Light trapping in MQW solar cells, Masters thesis, Imperial college, University of London, available at
www.sc.ic.ac.uk/~q_pv/
Stradling, R. and Klipstein, P. (1989) Growth and Characterization of Semiconductors, Adam Hilger, UK
Suzuki, Y., Kikuchi, T., Kawabe, M. and Okada, Y. (1999) Atomic hydrogen assisted molecular beam epitaxy for the fabrication
of multi-quantum well solar cells, J. Appl. Phys., vol 86, no 10, pp58585861
Takamoto, T., Ikeda, E., Kurita, H., Tanabe, T., Tanaka, S., Matsubara, H., Mine, Y., Takagishi, S. and Yamaguchi, M. (1997)
InGaP/GaAs and InGaAs mechanically stacked triplejunction solar cells, proceedings of 26th PV Specialists Conference,
IEEE, pp10311034
Takamoto, T., Agui, T., Ikeda, E. and Kurita, H. (2000) High efficiency InGaP/InGaAs tandem solar cells on Ge substrates,
proceedings of 28th IEEE PV Specialists Conference, IEEE, US, pp976981
Takamoto, T., Agui, T., Ikeda, E. and Kurita, H. (2001) High efficiency InGaP/In0.01Ga0.99As tandem solar cells lattice matched to
Ge substrates, Solar Energy Materials and Solar Cells, no 66, pp511516

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

ch-02

2/6/07

7:36 PM

Page 73

Quantum Well Solar Cells

73

Tang, P., Pullin, M., Phillips, C. and Stradling, R. (1995) Photoluminescence studies of NIPI superlattices in InSb and InAs
suppression of Auger recombination due to type-II potentials, Semicond. Sci. Technol., vol 10, no 4, pp476482
Tibbits, T., Ballard, I., Barnham, K., Bushnell, D., Ekins-Daukes, N., Airy, R., Hill, G. and Roberts, J. (2003) The potential for
strain-balanced quantum well solar cells in terrestrial concentrator applications, proceedings of 3rd World Conference on
Photovoltaic Energy Conversion, Osaka, Japan, May
Tobin, S., Vernon, S., Bajgar, C., Woktczuk, S., Melloch, M., Keshavarzi, A., Stellwag, T., Venkatensan, S., Lundstrom, M. and
Emery, K. (1990) Assessment of MOCVD and MBE grown GaAs for high efficiency solar cell applications, IEEE Trans.
Electron Devices, vol 37, no 2, pp469477
Varonides, A. and Berger, A. (1997) Proceedings of 15th European Photovoltaic Solar Energy Conference, H. S. Stephens &
Associates, Bedford, UK, pp17121715
Vegard, L. (1921) Die Konstitution der Mischkristalle und die Raumfllung der Atome, Z. Phys., no 5, pp1726
Walters, R., Summers, G., Messenger, S., Freundlich, A., Monier, C. and Newman, F. (2000a) Radiation hard multiquantum
well InP/InAsP solar cells for space applications, Prog. Photovolt. Res., vol 8, no 3, pp349354
Walters, R., Summers, G., Messenger, S., Romero, M., Arajo, D., Garcia, R., Freundlich, A., Newman, F. and Vilela, M.
(2000b) Electron beam induced current and cathodoluminescence study of proton irradiated InAsP/InP quantum well solar
cells, proceedings of 28th IEEE PV Specialists Conference, IEEE, US, pp13121316
Walters, R., Summers, G., Messenger, S., Romero, M., AlJassim, M., Garcia, R., Arajo, D., Freundlich, A., Newman, F. and
Vilela, M. (2001) Electron beam induced current and cathodoluminescence study of proton irradiated InAsP/InP
quantumwell solar cells, J. Appl. Phys., vol 90, no 6, pp28402846
Weisbuch, C. and Vinter, B. (1991) Quantum Semiconductor Structures: Fundamentals and Applications, Academic Press
Wrfel, P. (1995) Is an illuminated semiconductor far from thermodynamicequilibrium, Sol. Energy Mater. Sol. Cells, vol 38,
no 14, pp2328
Wrfel, P. (2003) Improvement of solar cell efficiencies by impurity transitions, Sol. Energy Mater. Sol. Cells, vol 79, no 1,
pp153161
Yablonovitch, E. and Cody, G. (1982) Intensity enhancement in textured optical sheets for solar cells, IEEE Trans. Electron
Devices, ED29, pp300305
Yamaguchi, M. and Luque, A. (1999) High efficiency and high concentration in photovoltaics, IEEE Trans. Elec. Dev., vol 46,
no 10, pp21392144
Yang, M. J., Yamaguchi, M., Takamoto, T., Ikeda, E., Kurita, H. and Ohmori, M. (1997) Photoluminescence analysis of InGaP
top cells for high-efficiency multi-junction solar cells, Solar Energy Materials and Solar Cells, no 45, pp331339
Zachariou, A. (1996) An experimental study of InP/InGaAs quantum well solar cells, Ph.D. thesis, Imperial College of Science,
Technology and Medicine, University of London, available from www.sc.ic.ac.uk/~q_pv/
Zachariou, A., Barnham, K. W. J., Griffin, P., Nelson, J., Button, C. C., Hopkinson, M., Pate, M. and Epler, J. (1996) A new
approach to pdoping and the observation of efficiency enhancement in InP/InGaAs quantum well solar cells, proceedings
of 25th IEEE PV Specialists Conference, IEEE, US, pp113116
Zachariou, A., Barnes, J., Barnham, K. W. J., Nelson, J., Tsui, E. S. M., Epler, J. and Pate, M. (1998a) A carrier escape study
from InP/InGaAs single quantum well solar cells, J. Appl. Phys., vol 83, no 2, pp877881
Zachariou, A., Barnham, K. W. J., Griffin, P., Nelson, J., Osborne, J., Hopkinson, M. and Pate, M. (1998b) GaInP/GaAs
quantum well solar cells, proceedings of 2nd World Conf. and Exhibition on Photovoltaic Solar Energy Conversion,
European Commission, Ispra, Italy, pp223226

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 4573

ch-03

2/6/07

8:21 PM

Page 74

Recent Progress of Organic


Photovoltaics
Sam-Shajing Sun

Abstract
This article briefly reviews and summarizes recent key developments in organic and
polymeric photovoltaic materials and devices. Specifically, the article briefly reviews the
four major stages and/or types of the organic and polymeric photovoltaic developments
from single layer Schottky junction cell, to donor/acceptor bilayer Tang cell, to
donor/acceptor blend heterojunction cell to donor/acceptor bicontinuous phase
separated nano-structured cell. The current relatively low photoelectric power
conversion efficiencies (less than 6 per cent) of purely organic and polymeric
photovoltaic materials and devices are attributed mainly to the heavy photon loss, the
exciton loss and carrier loss. However, high efficiency can be achieved as these
losses can be addressed via optimizations in both space and energy domains as
discussed. Organic versus inorganic photovoltaics is also briefly compared.

Keywords organic and polymeric photovoltaics; plastic solar cells; organic and inorganic
semiconductors; band theory; excitons; light harvesting; portable power

3.1 INTRODUCTION
While inorganic crystalline-based solar cell technology has become relatively mature
over 30 per cent photoelectric power conversion efficiencies have been demonstrated,1-4
and much higher theoretical efficiencies have been predicted3 this technology
suffers from a relatively high manufacturing cost and a current severe shortage of the
feedstock materials. As the need for renewable and solar energy technologies is rapidly
growing, alternative materials or technologies that could reduce the solar cell
manufacturing costs and have sufficient feedstock supplies become very attractive and
critical. Organic or polymeric photovoltaic materials and technology fall right into this
category.57
In comparison to the traditional inorganic solar cell, recently developed organic and
polymeric conjugated semiconducting materials appear very promising for photovoltaic
applications for several reasons:

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 7498

ch-03

2/6/07

8:21 PM

Page 75

Recent Progress of Organic Photovoltaics

75

their lightweight, flexible shape, versatile materials synthesis and device fabrications,

and low cost on large-scale industrial production;


almost continuous tunability of the materials energy levels and gaps via molecular

design, synthesis and processing;


ultra-fast optoelectronic response and charge carrier generation at organic

donor/acceptor interfaces (this makes organic photovoltaic materials also attractive


for developing potential fast photodetectors); and
integrability into other products such as textiles that can be used for fabricating solar
cell tents (see Figure 3.1) and clothing, packaging systems, consumption goods and
future all-plastic optoelectronic devices.57

The Sun
Future Camping/Hiking

Plastic Solar Cell Tent

FIGURE 3.1 Lightweight and flexible thin film plastic solar cells are very attractive for camping tents or
any mobile units.

3.2 ORGANIC VERSUS INORGANIC SEMICONDUCTORS


In organic conjugated semiconductor materials, typically the valence shell electrons are
responsible for the optoelectronic properties.57 When the material is at its lowest ground
state, the highest occupied molecular orbital (HOMO) typically refers to bonding orbital, and
the lowest unoccupied molecular orbital (LUMO) typically refers to * anti-bonding orbital. In
typical organic semiconductors, including most organic crystalline semiconductors, the
spatial and therefore the electronic overlap or coupling of orbitals are much poorer
compared to their inorganic semiconductor counterparts. Therefore traditional long-range
delocalized conduction band (CB) and valence band (VB) can hardly be formed in organics due
to this poor frontier orbital overlap, particularly the poor inter-molecular overlap.7 Specifically,
poor inter-molecular orbital overlap results in relatively small charge delocalization length or
small frontier orbital (i.e., HOMOs and LUMOs) band size (BS)8 (l, the extent of charge
delocalization) that would limit or suppress the exciton radius (r). Since the exciton Coulombic
binding energy E is inversely proportional to the exciton radius (E=K/r, as shown in Figure
3.2), for most organic or polymeric materials where small radius and strongly correlated
Frenkel type of exciton are typically generated9 and the binding energy is mostly larger than

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 7498

2/6/07

76

8:21 PM

Page 76

S-S. SUN

the
room
temperature
thermal
energy
(the
top
diamond
curve
in
Figure 3.2), in other words a secondary force is needed to dissociate the photogenerated
exciton. This is called the secondary photo carrier generation process (or excitonic
mechanism) and will be discussed in the following section. In fact, the initial sunlight
harvesting steps of photosynthesis in natural plants usually follow the secondary process
more closely.10 However, as shown in Figure 3.2, VB to CB band-to-band-like photo carrier
generation (also called the primary photo carrier generation process or SSH model8) is also
possible if the orbital overlap is really good, such that the exciton radius is very large, or the
dielectric constant of the material is very large so the exciton Coulombic binding energy is
below the thermal energy level. A Wannier type of exciton curve affiliated with large dielectric
constant materials such as silicon crystal is also shown (as triangles) in Figure 3.2. Due to the
typical small band size (BS), the excitation energy gap Eg in organic or polymeric materials only
represents the energy difference between the relatively localized LUMO and HOMO, not the
traditional delocalized VB and CB. When a photon with energy equal to or over Eg excite the
organic or polymeric molecule, an electron first transfers from the HOMO to the LUMO and
then quickly relaxes with the hole to form a Frenkel type exciton, also called polaron-exciton
in organics.8 Depending on the dielectric constant of the materials, the Frenkel type exciton
typically has a size of less than 2nm with binding energies larger than 0.1eV.

'Frenkel' Exciton Curve

kT

'Wannier' Exciton Curve

100

10
Coulombic
Energy E (eV)

ch-03

0.1

0.01

0.001
0

10

20

30

40

Exciton Radius r (nm)


The diamond curve (top) is obtained from an organic exciton. The triangle curve (bottom) is obtained from
an inorganic exciton.

FIGURE 3.2 Scheme of exciton Coulombic binding energy (in log scale) versus exciton radius.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 7498

50

ch-03

2/6/07

8:21 PM

Page 77

Recent Progress of Organic Photovoltaics

77

As discussed above, one key difference between organic and inorganic


semiconductors appears to be the extent of the spatial (and therefore the electronic)
overlap of the frontier orbitals, or the band size (BS), as BS is directly affected by the
extent of orbital spatial overlap or coupling. Recent experimental data on the charge
mobility of an organic crystal, rubrene, has confirmed this argument. Rubrene is an
organic molecule that can self-assemble to form a highly ordered crystalline structure
such that its frontier orbitals are spatially well overlapped between adjacent molecules, so
the CB and VB band-like states might be formed, and charge mobility up to 20cm2/Vs has
been observed in a rubrene.11 On the mechanisms of photo carrier generation, it is
possible the exciton is generated with its binding energy below the thermal energy line,
so that a photon plus room temperature may instantaneously generate a free electron
(also called negative polaron) at CB and a free hole (also called positive polaron) at VB, in
other words a band-to-band or primary photo carrier generation mechanism applies.

3.3 ORGANIC/POLYMERIC SOLAR CELL DEVELOPMENTS


3.3.1 SINGLE LAYER SCHOTTKY TYPE SOLAR CELLS
The earliest photoconductivity observation in organic materials was reported in
Anthracene in 1907.12 Later, some organic dyes and biological materials, such as
methylene blue, carotenes, chlorophylls and related porphyrins or phthalocyanines were
also found to exhibit photoconductive or very weak photovoltaic effects,1314 with best
photoelectric power conversion efficiency of 0.05 per cent obtained on a
chromium/chlorophyll-a/mercury cell.15 Conjugated polymers such as polyacetylene,16
polythiophenes,17 and poly-p-phenylenevinylenes (PPVs)18 were also investigated for
photovoltaic effects, with best power conversion efficiency of 0.1 per cent achieved under
white light illumination,18 and a best open-circuit voltage (Voc) of up to 1.7 volt was
reported using calcium as the small work function electrode.19 The low photoelectric
conversion efficiency can be attributed mainly to the Frenkel exciton model. As shown in
Figure 3.3, a single layer organic cell is composed of an organic semiconductor
sandwiched between two different metal electrodes with different work functions. The
small work function electrode (SWFE) typically acts as the negative electrode to collect
photogenerated electrons, and the large work function electrode (LWFE) acts as the
positive electrode to collect photogenerated holes. At least one electrode is very thin and
semi transparent, so the light can pass through. As Figure 3.3 shows, when an energymatched photon strikes most part of the organic layer (assumed defect free), only a
strongly bound Frenkel exciton is generated, which typically decays radiatively or nonradiatively back to the ground state without contributing to the carrier generation (Figure
3.3b). This explains the very low photoelectric efficiency of single layer organic PV cells.
However, for those few excitons that were generated at or diffused into the Schottky
junction area near the organic/metal interface as shown in Figure 3.3, the exciton can be
dissociated into an electron and a hole by a strong field formed from the orbital bending
or depletion region in the Schottky junction area.15, 20 Additionally, frontier energy level
offsets due to the presence of molecular oxygen21 and certain impurity or structure defect
sites (such as carbonyl groups)22 also contribute to the charged carrier generations.
However, within the typical exciton lifetime of pico- to nano-seconds, most Frenkel

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 7498

ch-03

2/6/07

8:21 PM

78

Page 78

S-S. SUN

excitons can only travel 550nm (also called average exciton diffusion length, AEDL),
much less than the thickness of most PV materials films, which are typically over 100nm;
most excitons are therefore wasted. This exciton loss is a major loss mechanism. Those
early cells are also called Schottky cells or defect cells due to the carriers mainly being
generated in the Schottky junction or defect sites. These single layer solar cells may be
called first generation cells.

Schottky
Junction

Vacuum Level

SWFE

LWFE

Schottky Junction

hv
LUMO

hv

ex

em
ex

E
SWFE

LWFE
HOMO

(a)

(b)

FIGURE 3.3 First generation single layer type organic photovoltaic cell or Schottky cell with (a) device
structure and (b) energy diagram

3.3.2 DONOR/ACCEPTOR BILAYER HETEROJUNCTION CELL OR TANG CELL


From the spatial (geometry) structure point of view, the second generation solar cells were
of the donor/acceptor bilayer structure type as shown in Figure 3.4. This bilayer
heterojunction type cell was first demonstrated in 1986 with a surprising 1 per cent power
conversion efficiency under AM 2.23 Since then, many other donor/acceptor (D-A) bilayer
systems have been investigated extensively,56 including, for instance, molecule-fullerene,24
polymer-fullerene2527 and polymer-polymer D-A bilayer systems,2829 with a best 1 per cent
power conversion efficiency achieved in a PPV/C60 bilayer cell.27 It was also found that the
average exciton diffusion length (AEDL) in PPV is about 9nm.27 As shown in Figure 3.4a,
once a photogenerated Frenkel exciton in either donor or acceptor layer diffuses to the
middle donor/acceptor interface, charge separation occurs whereby the electrons transfer
to or remain in the acceptor layer and holes transfer to or remain in the donor layer. The
charge carrier is now generated at such an artificially created donor/acceptor interface and
the frontier orbital energy offset between the donor and acceptor has become the key
driving force for exciton dissociation. Due to both the electrode-induced internal field and
chemical potential driving forces,3031 the electrons and holes than diffuse or hop to their
respective electrodes much quicker than in single layered cells. The likelihood of carrier

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 7498

ch-03

2/6/07

8:21 PM

Page 79

Recent Progress of Organic Photovoltaics

79

recombination is also relatively small due to electrons and holes now moving in two
separate layers or domains. However, a major limiting factor for the double layer cell is still
the relatively thick donor or acceptor layer (typically over 100nm) versus the relatively short
average exciton diffusion length (550nm), in other words many photogenerated excitons
decay before they reach the D-A interface (Figures 3.4a and b). On the other hand, if the

D/A Interface

Vacuum Level

SWFE

LWFE

D/A Interface

hv

D-LUMO

hv

ex

A-LUMO

em
ex

Donor

Acceptor

em

ex

D-HOMO

A-HOMO

(a)

(b)

FIGURE 3.4 Second generation donor/acceptor bilayer type organic photovoltaic cell or Tang cell with (a)
device structure and (b) energy diagram

N N
N M N

O
n

PPV
n

Phthalocyanines

O
MDMO-PPV

MEH-PPV

O
S
O

RO-PPV-10

n
R
PTh

PBZT (D)

FIGURE 3.5 Representative organic/polymeric electron donors (p-type semiconductors)

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 7498

ch-03

2/6/07

8:21 PM

80

Page 80

S-S. SUN

R N

O
R

C60

Perylenes
R
SO2

OR

NC

RO

PCBM

CN-PPV

BBL

SF-PPV

FIGURE 3.6 Representative organic/polymeric electron acceptors (n-type semiconductors)

vacuum
2

vacuum
Donor (hole acceptor)

MDMO MEH
PPV
-PPV -PPV
2.7 eV
-

Acceptor (electron acceptor)

C10O
-PPV

-2
2.6 eV

LUMOs

Ca

-2.83 eV -2.80 eV
-2.91 eV

PCBM C
60

3.53 eV 3.53 eV 3.55 eV

SF-PPV

P3HT P3OT P3DDT


3.75 eV

CN-PPV

3.83 eV

3.43 eV

3.60 eV

4
Al
4.3 eV

ITO
4.7 eV

PSS-PEDOT
5.2 eV

Au
5.1 eV

5.17 eV 5.20 eV

5.20 eV
5.31 eV

-5

4.90 eV
5.22eV

5.29 eV

HOMOs

5.80 eV 5.89 eV

-6

6.10 eV 6.10 eV

FIGURE 3.7 Frontier orbital levels of representative polymeric electron donors and acceptors; the work
functions of several representative electrodes are also shown on the sides

film thickness is too thin, this results in severe photon losses. Figure 3.5 shows mostly
used or studied organic donors. Figure 3.6 shows mostly used organic acceptors, and
Figure 3.7 shows the frontier HOMO/LUMO levels of some commonly used conjugated

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 7498

ch-03

2/6/07

8:21 PM

Page 81

Recent Progress of Organic Photovoltaics

81

semiconducting polymers. The donor/acceptor (or p/n) double layer solar cells may be
called second generation organic cells.

3.3.3 DONOR/ACCEPTOR BLEND BULK HETEROJUNCTION SOLAR CELLS


Since the donor/acceptor interface is critical for organic exciton dissociation, and the
average Frenkel exciton diffusion length (AEDL) in most organic semiconductors is
between 5 and 50nm, it is therefore desirable that D-A interfaces are within the AEDL
dimension everywhere in the bulk of the materials. The donor/acceptor
blended/composite or bulk heterojunction organic and polymeric solar cells (shown in
Figure 3.5) are thus a logical approach. These may be called third generation
organic cells.3240 These cells were fabricated by intimately blending a donor material
with an acceptor material mechanically. In this way, the donor/acceptor interface
(and so the exciton dissociation sites) would be located everywhere in the bulk, so that it
would be more convenient for an exciton generated anywhere in the bulk to reach a
nearby donor/acceptor interface and be dissociated into carriers. For instance, it was
found that a cell with a D-i-A tri-layer structure, where D is a donor layer, A is an
acceptor layer and i represents a donor/acceptor blend layer, had nearly doubled
photoelectric power conversion efficiency over that of a corresponding D-A bilayer cell
under similar conditions.32 So far, numerous bulk heterojunction cells using PPV or
polythiophene derivatives (MEH-PPV, MDMO-PPV, P3HT and P3OT, for example) as the
donor and CNPPV or fullerene derivatives (mostly fullerenes such as PCBM) as the
acceptor have been intensely studied,3340 and near unity photo-induced charge
separation (also called internal quantum efficiency) and between 1 and 6 per cent
photoelectric power conversion efficiencies have been reported under different
conditions.3340 For instance, a 4.4 per cent power conversion efficiency (under NREL
calibrated AM 1.5 and 1 sun condition) was recently described on a PCBM/P3HT
binary cell with carefully controlled thin film processing and device fabrications.40
The higher efficiencies of D/A blend cells over the previous D-A double layered cells can
be attributed mainly to the reduction of exciton loss due to a larger donor/acceptor
interface, and also to the photon loss as the films now can be made thicker to
harvest more light. However, even if the charge carrier generation may be efficient, the
carrier transport to the electrodes still appears not so easy or might even be worse
compared to the bilayer cell in other words a major problem is carrier loss. This is
due to the fact that the two phases may not be really bicontinuous between the two
electrodes (Figure 3.5a), in other words a carrier transport pathway may be interrupted
easily by the other phase. This would cause carriers to recombine more frequently.
Additionally, if the donor and acceptor are in direct contact with both electrodes, carrier
recombination at the organic/electrodes interface would be severe, and carrier
collection efficiency at electrodes would be poor. One interesting approach was to
fabricate a D/A bilayer first and then enable donor and acceptor to partially diffuse into
each other to form a D-D/a-d/A-A gradient type cell structure:41 this would increase the
D/A interface and at the same time still keep the D/A spatial asymmetry between the
two electrodes.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 7498

ch-03

2/6/07

82

8:21 PM

Page 82

S-S. SUN

LWFE

SWFE

Vacuum

D-LUMO
A-LUMO
SWFE
LWFE

(a)

D-HOMO
A-HOMO

(b)

FIGURE 3.8 Third generation donor/acceptor blend type organic photovoltaic cells or bulk heterojunction
cells with (a) device structure and (b) energy diagram

3.3.4 DONOR/ACCEPTOR BICONTINUOUS PHASE SEPARATED


AND NANO-STRUCTURE ORDERED SOLAR CELL
If the carrier transport in a donor/acceptor blend system is poor due to transport pathways
disorder, then a donor/acceptor bicontinuous phase separated and nano-structure
ordered solar cell as illustrated in Figure 3.6 become desirable.4245 Since this structure
was proposed in a conjugated block copolymer system,42, 43 a number of other approaches
such as honeycomb- style patterned inorganic n-type semiconductors with p-type
polymers,46 acceptor-type carbon nano-tubes with donor-type polymers,47 and n-type
semiconducting nano-rods with p-type polymers48 have been pursued in order to realize
it. In this solar cell type, a donor/acceptor bicontinuous phase separated nano-structure
can be in the form of ordered column or cylinder nano-structure or honeycomb pattern
sandwiched and perpendicularly oriented between the two electrodes, each column or
cylinder cross-section diameter controlled to be within the average exciton diffusion
length of most organic semiconductors (550nm, Figure 3.6a).42, 43 In comparison to the
third generation donor/acceptor blend system, the excitons now still easily reach the
donor/acceptor interfaces, particularly in the direction perpendicular to the column (Figure
3.6a), yet the carriers now have a continuous or uninterrupted transportation pathway
toward their respective electrodes along the direction of the column. In this way the active
layer thickness (or lengths of the columns) can be much greater than the average exciton
diffusion length, so more photons can be captured. Discotic-shaped small molecules
stacked in p/n bicontinuous column style crystalline structures are also desirable. While
one molecular phase stacked crystalline structure has already been reported,49, 50 both
donor and acceptor bicontinuous stacked columnar structures have yet to be

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 7498

ch-03

2/6/07

8:21 PM

Page 83

Recent Progress of Organic Photovoltaics

83

demonstrated. Most importantly, these kind of bicontinuous ordered nano-structures may


be applied to all organic D/A, inorganic p/n, or any hybrid inorganic/organic p/n binary type
solar cells, as the two common features are that the nano-domain junctions dissociate
excitons while the continued/uninterrupted channels provide smooth transport pathways
for the carriers. These may be called fourth generation organic cells.

A
SWFE

LWFE

Vacuum
D/A Interface

D-LUMO

A-LUMO

D-HOMO

A-HOMO

LWFE-Femi

SWFE-Femi

(b)

(a)

FIGURE 3.9 Fourth generation donor/acceptor bicontinuous phase separated and nano-structured organic
photovoltaic cell with (a) device structure and (b) energy diagram

3.3.5 TANDEM STRUCTURED CELLS


In inorganic solar cells, one well-known approach to minimize the solar photon loss has
been to fabricate a cell with multiple serially connected and parallel stacked sub-cells
having different energy gaps (also called energy gap graded tandem structure cells).14,
51, 52
The advantages of the tandem structure include: 1) increased solar photon capture
due to both increased overall cell thickness and energy gap grading, and 2) increased
open-circuit voltage as a result of Voc serial summation. It is obvious that the optical
excitation of the cell should match the energies of the photons to be harvested, and the
frontier orbital energy offsets between the donor and acceptor must also be just right in
order to dissociate the exciton most efficiently and at the same time minimize the charge
recombination.44, 45, 53, 54 For this reason, optimizations at energy domain are also critical.
Since sunlight radiation spans a wide range from UV all the way to IR, and in each cell
either donor or acceptor has only one energy gap and can only capture a very narrow
energy range of photons, a tandem-style stacked and serially connected cell structure is
therefore desirable, with the energy gaps graded among the stacked cells with an energy
range spanning the whole solar spectrum, so that most of the sunlight can be captured.51, 52
Since the open-circuit voltage of each cell is correlated closely to the donor HOMO and
acceptor LUMO,55 it is critical to connect the cells in a serial manner so that photovoltage

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 7498

ch-03

2/6/07

84

8:21 PM

Page 84

S-S. SUN

can be added up. In this way, large solar cell voltages together with a high power
conversion efficiency can be achieved.8 For instance, over 30 per cent power conversion
efficiencies have been reported in inorganic triple-junction tandem-type solar cell.14 For
organics, a 5.7 per cent power conversion efficiency (under 1 sun AM 1.5 condition) and
an open-circuit voltage of 1.2 volts was reported in a tandem-type PV device containing
two serially connected and stacked sub-cell units.56 In this device, though both sub-cells
are composed of copper phthalocyanine (CuPc) as the donor and C60 as the acceptor, the
front sub-cell was tuned (via optical engineering) to absorb mainly the 650nm wavelength
low energy light, while the back sub-cell mainly captures the 450nm wavelength high
energy photons due to the optical match issue.56 In an ideal tandem cell, however, it is
desirable that the front sub-cells absorb high energy photons while the rear sub-cells
capture low energy light.2, 3, 51, 52 In this way, the largest-gapped cell at the front would
capture highest energy photons first, but allow lower energy photons to pass through to
the lower-gapped cells behind, where the lower energy photos can be captured by lower
energy gap sub-cells, and so on. Even if some excitons in the front large-gapped cells do
not dissociate and relax to emit a smaller energy photon, that photon can be captured by
the next lower-gapped cells. However, current density needs to be balanced between subcells.13, 51, 52 As a whole, as the cell could capture most sunlight, it should appear as a dark
color.7

3.4 ORGANIC SOLAR CELL FABRICATIONS


In a typical organic or polymeric solar cell device, as shown in Figure 3.10, the organic or
polymeric semiconductor layer (called the active layer) is typically sandwiched between a
transparent conducting electrode (TCE) (for example Indian-Tin-Oxide or ITO coated glass
as the LWFE, bottom) and a metal electrode (for example aluminum as the SWFE, top). It
has recently been found that a poly(ethylene dioxythiophene):polystyrene sulfonic acid
layer (PSS-PEDOT, chemical structure shown in Figure 3.11) greatly facilitates the hole
transfer between the active layer and the ITO electrode, and a thin layer LiF greatly
facilitates the electron transfer between the active layer and the metal electrode as will be
further elaborated below. For small organic molecules, typically high vacuum (at least
10-6 Torr pressure) vapor deposition and occasionally solution crystallization protocols are
used to grow thin films on the TCE substrate, followed by the vacuum deposition of the
metal electrode on top of the photovoltaic active layer. For polymers, typically solution
spin coating (small devices), inkjet printing, or roll-to-roll printing (large sized sheets)
protocols can be used. Solution processing generally offers advantages of low cost and
convenience on large-scale industrial productions.5, 6
The PSS-PEDOT layer is coated between the ITO glass electrode and the active layer
in order to optimize the hole transfer between the active layer and the ITO electrode.5, 6
The conductivity of PSS-PEDOT can reach 80S/cm if produced by electro-polymerization,
and can be as low as 0.03S/cm if produced by chemical polymerization. This is believed
to be due to different composition.57 Sheet resistance of PSS-PEDOT as low as
350500 ohm/square was reported.58 It became an extremely attractive material when it
was found that the performance and stability of polymer light emitting diodes (LEDs)
could be improved by inserting this material between the polymer active layer and ITO

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 7498

ch-03

2/6/07

8:21 PM

Page 85

Recent Progress of Organic Photovoltaics

85

SWFE, e.g., Al, Ca


LiF
Photovoltaic Active Materials
PSS-PEDOT

Transparent LWFE, e.g., ITO


Transparent substrate, e.g., glass

Light
FIGURE 3.10 General scheme of an organic/polymer solar cell with buffering layers

O
S
+

S
O

+
S

S
O

S
O

O
OH
O
O S OO S OO S O

FIGURE 3.11 Chemical structure of PSS-PEDOT

glass as a buffer layer.59, 60 PSS-PEDOT is available in an aqueous dispersion, which can


form uniform, transparent, conductive film by spin coating. The most important physical
properties for its application in devices are the high work function (-5.2eV) and the smooth
surface. The improvement of the device performance after PSS-PEDOT layer application
may be attributed to a number of factors in either spatial or energy regimes. In the spatial
regime, for instance, commercial ITO glass surfaces have been found to be very rough
and the conductivities were area sensitive.61 The PSS-PEDOT layer was believed to help
smooth both the surface roughness and the conductivities at different spots. In the
energy regime, it is believed that the work function of PSS-PEDOT lies between the work
function of the ITO (-4.7eV) and the HOMO levels of most organic donor materials. This
intermediate level would facilitate the hole transfer between the polymer and the ITO. It is

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 7498

ch-03

2/6/07

86

8:21 PM

Page 86

S-S. SUN

also possible that PSS-PEDOT simply prevents the acceptor from being in direct contact
with the ITO, so charge recombination at the ITO interface is minimized. However,
PSS-PEDOT layer thickness should be optimized since too thin a layer may yield many pin
holes, while thicker layers cause larger increase in the series resistance.61
Insertion of a thin layer of lithium fluoride (LiF) between the active layer and the SWFE
metal electrode has been shown to improve electron injection in some organic LEDs.62
The LiF layer is typically very thin (< 1nm), since thicker layers are found to be detrimental
to electron injection. For instance, 3 thickness of LiF was found to be optimal for both
the fill factor and the current density in one system.62 While the exact causes were not
very clear, the improved device optoelectronic performance of incorporating LiF in some
cases may be attributed to, for instance, the LUMO of the LiF lying between the acceptor
LUMO and work function of the metal, or the LiF-modified work function of the metal
electrode due to the dipolar nature of LiF layer. Both would facilitate the electron transfer.
Another cause might be that the LiF layer prevents the donor from being in direct contact
with the metal, or prevents the chemical reactions between the organic active layer and
the metal, or reduces the serial resistance of the interface between the metal and the
active layer.61
The equivalent circuit of an organic/polymeric solar cell J-V curves can be
approximated by:63
(1)
where J0 is the saturation dark current density, q is the elementary charge, n is the diode
ideality factor, V is the applied voltage of the cell, RS and RP are the serial and parallel
resistivity, Jsc is the short-circuit photocurrent density (A/cm2), k is the Boltzman constant,
and T is the absolute temperature. For a pin-hole free organic active layer, the parallel
(shunt) resistivity was approximately the same for many organic photodiodes (typically
over 1k), but the total serial resistivity (RS) may vary depending on materials and device
fabrications. In one device, for instance, Rs decreased from 10 with no LiF layer to 4
for LiF thickness between 3 and 9 thick and increased to 5 for LiF 1215 thick for one
cell.63 In comparison to inorganic solar cells, organics typically have very large RS due to
the typically very poor orbital overlaps.
When characterizing and comparing organic solar cells, like in inorganic solar cells, the
open-circuit voltage (Voc), short-circuit current density (Jsc), the fill factor (FF), and the
overall photoelectric power conversion efficiencies () are critical parameters. Like in
inorganic solar cells, the open-circuit voltage (Voc) can be experimentally obtained from the
currentvoltage curves of an illuminated device when the current is zero. Because of the
special charge generation and separation mechanisms in organic devices, the efficiencylimiting factors are therefore distinct from those in conventional inorganic solar cells (for
example silicon p-n junction solar cells). For instance, while the maximum photovoltages
(Voc) achievable in silicon cells are generally limited to the magnitude of the built-in
potential, it is common to observe experimentally Voc greater than the built-in potential in
organic-based photovoltaic devices.55 It is believed the Voc of an organic donor/acceptor
binary cell is closely correlated to the HOMO of the donor and the LUMO of the

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 7498

ch-03

2/6/07

8:21 PM

Page 87

Recent Progress of Organic Photovoltaics

87

acceptor.55, 64 The short-circuit current density (Jsc) is the photocurrent density value under
zero applied bias. It is a result of the photo-induced charge separation (or exciton
dissociation) and the charge transport driven by the internal field and the chemical
potentials of the material.
The photoelectric power conversion efficiency (PCE or PCE) is the ratio of cell
maximum output electrical power (Pout) over the total radiated input optical power (Pin):

(2)
where a fill factor (FF) is related to the percentage of maximum electrical power that can
be extracted from the solar cell. The fill factor for devices is described as

(3)

where Jm and Vm are the values for the current density and voltage at maximum power
Pout/max = (JV)max.
To achieve a high FF factor in organic solar cells, it is desirable that the parallel
resistivity is large to prevent short or leakage current, while the serial resistivity is small to
enable a large forward current. The series resistivity simply adds up from all series
resistive contributions in the device, including both carrier transport in materials and
interface charge transfer. One key problem in organics is the very high series resistance
due to poor molecular frontier orbital overlaps, so charges have to hop between
molecules instead of diffuse in bands. However, molecular self-assembly may solve this
problem in the future.

3.5 ORGANIC SOLAR CELL OPTIMIZATIONS


The overall photoelectric power conversion efficiency of an organic/polymeric solar cell is
determined by at least the following five critical steps:
1
2
3
4
5

photon capture and exciton generation;


exciton diffusion to the donor/acceptor interface;
exciton dissociation and/or carrier generation at donor/acceptor interface;
carrier diffusion to the respective electrodes; and
carrier collection by the respected electrodes.

For all currently reported organic/polymeric photovoltaic materials and devices, none of
the above mentioned five steps have been optimized. For instance, in the first step, the
photon loss is heavy due to most organic semiconductors having too high energy gaps
(typically larger than 2eV), meaning that only a very small fraction of the sunlight could be
captured. In the second step, the exciton loss is severe unless a phase separated and
ordered nano-structure can be materialized. In the third step, both the exciton loss and
carrier loss are severe unless the frontier orbital energy offsets are at their optimal value

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 7498

ch-03

2/6/07

88

8:21 PM

Page 88

S-S. SUN

so that the exciton dissociation is maximized and at the same time the charge
recombination is minimized. In the fourth step, the carrier loss is notoriously severe and
well known in most organic/polymeric semiconductors due to the typically poor orbital
overlaps, so that sizable bands are difficult to form, and carrier transport is mainly through
inter- and intra-molecular hopping mechanisms. In the last step, the carrier loss may be
severe if the energy levels or chemical states between the organic semiconductors and
the electrodes are not optimized. It is thus not surprising that the power conversion
efficiencies of most currently reported organic or polymeric solar cells are relatively small
in comparison to their inorganic counterparts. However, there is plenty of room to improve,
and the optimizations should be done in both energy and spatial domains.

3.5.1 OPTIMIZATIONS IN THE SPATIAL DOMAIN


As mentioned earlier, in the spatial domain, a donor/acceptor bicontinous phase
separated and nano-structure ordered morphology appears ideal in order to minimize
both the exciton loss and the carrier loss.4245 As mentioned in section 3.4, there are
several approaches being pursued that could potentially realize this structure. In the cases
of carbon nano-tubes and semiconducting nano-rods, the main challenges are to fabricate
cells with uniformly distributed and well-aligned tubes/rods perpendicular to the horizontal
conducting substrates. In the cases of any techniques using block copolymers, the main
challenge lies in both synthetic chemistry and block copolymer processing.
Block copolymer solid melts are well known to exhibit behavior similar to conventional
amphiphilic systems such as lipidwater mixtures, soap, and surfactant solutions.65, 66 The
covalent bond connection between different blocks imposes severe constraints on
possible equilibrium states, this results in unique supra-molecular nano-domain structures
such as lamellae (LAM), hexagonally (HEX) packed cylinders or columns, spheres packed
on a body-centered cubic lattice (BCC), hexagonally perforated layers (HPL) and at least
two bicontinuous phases: the ordered bicontinuous double diamond phase (OBDD) and
the gyroid phase.6567 The morphology of block copolymers is affected by chemical
composition, block size, temperature, processing and other factors. Clearly, the block
copolymer approach to photovoltaic function offers some intrinsic advantages and has
attracted a number of research efforts.4245 For instance, block copolymers containing a
conjugated PPV donor block with a non-conjugated polystyrene acceptor block derivatized
with fullerenes have been investigated, and phase separations at nano-scale were indeed
observed in some cases.68, 69 However, the carrier loss problem in the non-conjugated
polystyrene phase would be a problem compared to the conjugated main chain
systems. On the other hand, when a conjugated donor block was connected directly to a
conjugated acceptor block to form a direct p-n type conjugated diblock copolymer, while
energy transfer from higher gap block to lower gap block was observed, no charge
separated states (which are critical for photovoltaic functions) were detected.70 A DBABtype of block copolymer and its potential tertiary supra-molecular nano-structure was
therefore designed (Figures 3.123.14),4244 where D is a electron conjugated donor
block, A is a electron conjugated acceptor block, and B is a non-conjugated and flexible
bridge unit. Additionally, the flexibility of the bridge unit enables the rigid donor and
acceptor conjugated blocks more easily to self-assemble, phase separate and become

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 7498

ch-03

2/6/07

8:21 PM

Page 89

Recent Progress of Organic Photovoltaics

89

less susceptible to distortion of the main chain conjugation (Figures 3.12 and 3.13). Since
both donor and acceptor blocks are electron conjugated chains, if they are selfassembled in planes perpendicular to the molecular plane like a stacking morphology
well known in a typical conjugated system (Figure 3.13),7173 good carrier transport in
both donor and acceptor phases would become feasible.

Source: S. Sun, Photovoltaic Devices Based on a Novel Block Copolymer, US patent publication # 20040099307

FIGURE 3.12 Scheme of a DBAB- type block copolymer primary structure

D
A

Source: S. Sun, Photovoltaic Devices Based on a Novel Block Copolymer, US patent publication # 20040099307

FIGURE 3.13 Scheme of a potential DBAB- type block copolymer secondary structure

While the DBAB- block copolymer backbone structure may be called primary structure
(Figure 3.12), the conjugated chain orbital closely stacked and ordered morphology may
be called secondary structure (Figure 3.13). This secondary structure has been known to
possess dramatically enhanced charge carrier mobility as demonstrated in, for instance,

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 7498

ch-03

2/6/07

90

8:21 PM

Page 90

S-S. SUN

crystalline orbital stacked molecular rubrene,10 self-assembled regioregular


polythiophenes,72 or template aligned poly-p-phenylenevinylenes.73 Finally, through the
adjustment of block size, block derivatization, and thin film processing protocols, a tertiary
structure (Figure 3.14) could be obtained, where a bicontinuous columnar (or HEX) type
of morphology of the donor and acceptor blocks vertically aligned on top of the substrate
and sandwiched between two electrodes may be formed.6 Even better, a thin donor layer
could be inserted between the ITO and the active HEX layer (preventing the acceptor from
being in direct contact with the ITO), and a thin acceptor layer could be inserted between
the metal electrode and the active layer (preventing the donor from being in direct contact
with metal) (Figure 3.9a).42 The terminal donor and acceptor layers not only prevent charge
easy recombination at the electrodes, they also enable a desired asymmetry and favorable
chemical potential gradient for asymmetric (selective) carrier diffusion and collection at
respective electrodes.30, 31 Since the diameter of each donor or acceptor block column can
be conveniently controlled via design, synthesis and processing to be within the average
exciton diffusion length (AEDL) of 550nm, every photo-induced exciton will be in
convenient reach of a donor/acceptor interface along the direction perpendicular to the
column. At the same time, photogenerated charge carriers can diffuse more smoothly to
their respective electrodes via a truly bicontinuous block copolymer columnar
morphology. Additionally, with appropriate adjustment of donor and acceptor block sizes
and their substituents, energy levels, gaps, and tandem device structures, it is expected
that the photon loss, the exciton loss, and the carrier loss (including charge
recombination) issues can all be addressed and optimized simultaneously in one such
block copolymer photovoltaic device. To examine the feasibility of this design,42, 43 a series
of DBAB- type block copolymers has been developed.44, 45, 7480 Preliminary experimental
results have indeed shown the photovoltaic properties of the DBAB- block copolymer are
much better than the corresponding D/A blend cells.44, 45, 79, 80

D
A
D

D
A
D
A

A
D
A

'HEX' Columnar Morphology


Source: S. Sun, Solar Energy Materials and Solar Cells, 79, pp257264 (2003)

FIGURE 3.14 Scheme of a potential DBAB- type block copolymer tertiary structure

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 7498

ch-03

2/6/07

8:21 PM

Page 91

Recent Progress of Organic Photovoltaics

91

3.5.2 OPTIMIZATIONS IN THE ENERGY DOMAIN


Photon loss, exciton loss and carrier loss are all critically related to the frontier orbital levels
of the donor, the acceptor and the work functions of the two electrodes. In sunlight
harvesting applications, the solar radiation spectrum is very broad, from UV all the way to IR,
with maximum photon flux ranging between 1.3 and 2.0eV on the surface of the Earth (air
mass 1.5) and between 1.8 and 3.0eV in outer space (air mass zero).56 For optical
telecommunications and signal processing, an optical excitation energy gap of 0.8eV (for
1.55 m IR signal) is needed. Energy excitation gaps (energy difference between the HOMO
and LUMO) in both the donor and the acceptor should be fine tuned to match the photon
energy, as both can absorb photons and incur charge separation at the donor/acceptor
interface. The questions are: 1) how to capture as much solar photons as possible while each
material (donor or acceptor) has typically one energy gap that can only capture a very narrow
energy range of photons; and 2) what would be the optimal frontier orbital level offsets
between the donor and the acceptor, which is the key driving force for exciton dissociation.
For the first question, the best answer would be a tandem-type cell structure with
energy gap grading as has been discussed extensively in the literature.14, 49, 50 For instance,
over 30 per cent power conversion efficiencies have been reported in inorganic tandemtype solar cells,14 and a 5.7 per cent power conversion efficiency was reported in a two
sub-cell stacked tandem-type organic device.56 However, it is obvious the more cells with
more energy gaps, the smaller the photon losses.
For the second question, the current widely cited view is that the frontier orbital
energy offset between the donor and the acceptor should be no less than the exciton
binding energy EB (in other words the minimum energy needed to overcome the electric
Coulomb forces and dissociate the Frenkel type exciton into a separate or free electron
at acceptor LUMO and a free hole at donor HOMO).81 Indeed when the LUMO energy
offset is too small, photo-induced charge separation appears to become inefficient.82 On
the other hand, if the energy offset is too large, the Marcus inverted region would slow
down charge separation,8389 and thermal ground state charge separation without photo
excitation may also occur.7 These are not desirable for light harvesting applications. A
large energy offset also reduces open-circuit voltages.55 Therefore an analysis of optimal
donor/acceptor energy offset is very critical and necessary44, 45, 51, 52 and is briefly summarized
here.
For solar cell purposes, the photo-induced charge separated state (Figure 3.15, after
steps 3 or 7) is the desired starting point. However, the exciton charge separation is also
competing with exciton decay (Figure 3.15, steps 2 and 6). The ratio of charge separation
rate constant (ks) versus exciton decay rate constant (kd) can be defined as exciton
quenching parameter (EQP, mathematically represented as Yeq) as:

YeqX = ksX / kdX

(4)

Here X=D (donor) or A (acceptor). The parameter Yeq reflects to a certain degree the
efficiency of exciton
>charge conversion. It was experimentally observed that the charge
separation could be orders of magnitude faster then the exciton decay in MEHPPV/fullerene binary systems.33, 34 Secondly, the charge separations (steps 3 and 7) are

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 7498

ch-03

2/6/07

92

8:21 PM

Page 92

S-S. SUN

also competing with charge recombination (step 4). The ratio of charge separation rate
constant (ks) over charge recombination rate constant (kr) can therefore be defined as
recombination quenching parameter (RQP, mathematically represented as Yrq):

YrqX = ksX / kr

(5)

For any light harvesting applications, such as solar cell applications, it is desirable that
both Yeq and Yrq parameters are as large as possible.
From semi-classical electron transfer theory, when LUMO orbital energy offset E is
set as variable (Figure 3.15),8389 and based on measured, calculated and estimated
parameters of RO-PPV/SF-PPV-I donor/acceptor pairs developed recently,42, 45 a plot of
normalized YeqD, kr, and YrqD versus E is shown in Figure 3.16.51, 52 As Figure 3.16 shows,
when E varies, kr, YeqD and YrqD all exhibit their own maximum values. For the RO-PPV/SFPPV-I pair, the fastest photo-induced charge separation occurs when the RO-PPV/SF-PPVI LUMO offset (E, driving force) equals the sum of charge separation reorganization
energy and the donor exciton Coulombic potential energy.51, 52 Also, the fastest charge
recombination occurs at a LUMO offset far away from the optimum charge separation
offset as well as the actual RO-PPV/SF-PPV-I offset. Therefore, the charge recombination
in the RO-PPV/SF-PPV-I pair does not seem to be of a major concern as long as the LUMO
offset is nearby the EeqD. Figure 3.16 also shows the recombination quenching parameter
YrqD (ksD/kr) does not reach its maximum until a positive energy offset. At this positive
energy offset, the photo-induced charge separation might be too slow to be attractive for
efficient photovoltaic function; therefore, the positive ErqD value appears not critical in
this particular case. It is desirable that through careful molecular design, the ErqD is

Vacuum
3

D-LUMO

EA

A-LUMO
IP

1
4
6

D-HOMO

A-HOMO
D/A Interface
FIGURE 3.15 Scheme of molecular frontier orbitals and photo-induced electron transfer as well as Dexter
energy transfer processes in a donor/acceptor binary light harvesting system

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 7498

2/6/07

8:21 PM

Page 93

Recent Progress of Organic Photovoltaics

Kr

ET Rates/Rate Ratios (Normalized)

ch-03

3.5

2.5

1.5

Yeq(D)

0.5

93

Yrq(D)

0.5

1.5

Donor/Acceptor Frontier Orbital Energy Offset (E/eV)


Source: S. Sun, Sol. Energy Mat. Sol. Cells, 85, 261-267 (2005)

FIGURE 3.16 Donor RO-PPV exciton quenching parameter (Yeq(D)=ks(D)/kd(D), middle solid curve), charge
recombination rate constant (Kr, left long dashed curve), and charge recombination quenching parameter
(Yrq(D)=ks(D)/kr(D), right short dashed curve) versus LUMO offset of RO-PPV/SF-PPV-I pair

coincident with or close to EeqD, and that ErD is far away from EeqD. Similar results can
also be deducted for the acceptor and for the donor/acceptor pair.44, 52
This donor/acceptor charge separation model can readily be used in the impurity or
defect photo doping cases (including the single layer Schottky junction cell), as the
impurities or structural defects can be either donor or acceptor type in the energy domain.
The model can also be used in ground state chemical doping cases.7 Chemical doping
typically refers to a mobile or free charge carrier generation phenomenon when one
material is mixed or in direct contact with a second material at room temperature without
the intentional use of strong external excitation forces such as light or high heat. This is
sometimes also called ground state charge separation. However, there are two different
situations involved. In the first situation, the acceptor LUMO is lower then the donor
HOMO. The energy offset E between the D-HOMO and A-LUMO therefore acts as a key
driving force (in addition to other potential driving forces such as thermal force kT) for the
electron transfer from the donor HOMO directly to the acceptor LUMO. In this case, the
most efficient or optimal charge separation would occur when the energy offset E plus
the thermal force kT equals the charge separation reorganization energy and the attractive
Coulomb force. Even at absolute zero temperature when kT=0, the electron transfer still
occurs due to the presence of the driving E. However, in the second scenario, the
acceptor LUMO might be the same or a little higher then the donor HOMO and the
thermal driving force kT becomes critical. In this case, the most efficient electron transfer
would occur when the driving kT is equal to the charge separation reorganization energy
plus the Coulomb attractive force plus the D-HOMO/A-LUMO energy offset increase.
Because of the kT driving force involved, this process actually is not really ground state
electron transfer, though it may occur at room temperature. Thermally induced chemical
doping mechanism may also be used to explain static electrical charge generation when
two different materials are rubbed against each other, as the friction at the interface might

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 7498

ch-03

2/6/07

94

8:21 PM

Page 94

S-S. SUN

generate enough thermal energy to incur electron transfer. In chemical doping cases,
since the dopant is typically a minority component and it traps one charge carrier, the
opposite charged carrier at non-dopant majority phase then becomes the mobile or free
charge carrier and can transport at the majority component phase.

3.6 CONCLUSIONS AND FUTURE PERSPECTIVES


The current relatively small (less than 6 per cent) photoelectric power conversion
efficiencies of organic and polymeric photovoltaic devices can be attributed mainly to the
severe photon loss, exciton loss and carrier loss due to materials improper frontier
energy levels and gaps, energy offsets between the donors and the acceptors, poor
material morphologies, and cell structures and fabrications that are not optimized.
However, there are plenty of areas where improvements might be made. Optimizations in
both spatial and energy/time domains should be pursued simultaneously in order to
achieve high efficiency organic and polymeric photovoltaic materials and devices.
In the spatial domain, the ideal structure or morphology of a single cell appears to be
a donor/acceptor nano-phase separated and bicontinuous columnar type morphology,
where the diameter of each column phase should be within the average exciton diffusion
length (AEDL) of 550nm depending on the materials involved, and with the columns
aligned perpendicular to the electrode planes. Among the several approaches mentioned,
a DBAB- type of block copolymer and its potential self-assembled columnar tertiary
supra-molecular nano-structure has been proposed and preliminarily examined
experimentally with promising results. In this system, along the carrier transport direction
which is perpendicular to the two electrode surfaces, it is bicontinuous of the two
phases, so both positive and negative charges have smooth transport pathways. Yet in the
direction parallel to the electrode planes, it has donor/acceptor phase separated
morphology, and each phase diameter is within the AEDL.
In the energy/time domain, first the optical excitation energy gaps in each donor and
acceptor phase should match the intended photon energy. For sunlight harvesting, an
energy gap graded tandem-type cell structure is desirable to minimize photon losses.
Second, the optimal donor/acceptor energy offset that is the critical driving force for photoinduced charge separation should be identified and materialized. Specifically, in an electron
transfer dynamic regime, there exists an optimal donor/acceptor LUMO (or HOMO) level
offset where exciton-charge conversion is most efficient (or exciton quenching parameter
EQP reaches its maximum), and another optimal LUMO (or HOMO) level offset can be
identified where charge recombination is relatively slow compared to charge separation (or
recombination quenching parameter RQP become largest). Though the cell efficiency is
mainly affected by the EQP, the molecules should be designed and developed such that
the maximum RQP is close to or coincides with maximum EQP. There also exists a third
energy offset where the charge recombination becomes most severe. The molecules
should be designed and developed such that this worst charge recombination is far away
from maximum EQP. These orbital offset values are critically important in molecular
structure and energy level fine tuning. Finally, since sunlight is a very broad radiation with
photon energy ranging from UV all the way to IR, a tandem-style serially connected and
parallel stacked cell structure with energy gaps gradually descending from UV to IR along

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 7498

ch-03

2/6/07

8:21 PM

Page 95

Recent Progress of Organic Photovoltaics

95

the light radiation direction appears ideal. This would enable a broad capture of most solar
photons and the summation of induced photovoltages.
With optimizations in both space and energy/time domains, it is expected that high
efficiency photoelectric power conversion efficiency organic light harvesting systems,
including organic photovoltaic cells, photo detectors or any artificial photo-charge
synthesizers/converters, can be realized. The dream of renewable, clean, inexpensive,
portable and low cost energy supply can be a reality.

ACKNOWLEDGEMENTS
The author would like to thank all those involved in the research cited in the literature and
for the research/educational grant supports from a number of funding agencies including
NASA, the Air Force Office of Scientific Research, the National Science Foundation, the
Department of Education (Title III award) and the Dozoretz foundation.

AUTHOR CONTACT DETAILS


Sam-Shajing Sun, Center for Research & Education in Advanced Materials and Chemistry Department, Norfolk
State University, 700 Park Avenue, Norfolk, VA 23504, US
Tel: 757-823-2993; Fax: 757-823-9054; e-mail: ssun@nsu.edu

REFERENCES
1 M. D. Archer and R. Hill (eds) (2001) Clean Electricity From Photovoltaics, Imperial College Press, London
2 A. Luque and S. Hegedus (eds) (2003) Handbook of Photovoltaic Science and Engineering, John Wiley, Chichester, UK
3 M. A. Green (2003) Third Generation Photovoltaics: Advanced Solar Energy Conversion, Springer, Berlin
4 A. Hepp, S. Bailey and R. Raffaelle (2005) Inorganic photovoltaic materials and devices: Past, present, and future in S.
Sun and N. S. Sariciftci (eds) Organic Photovoltaics: Mechanisms, Materials and Devices, CRC Press, Boca Raton, Florida,
US, p19
5 C. Brabec, V. Dyakonov, J. Parisi, and N. Sariciftci (2003) Organic Photovoltaics: Concepts and Realization, Springer, Berlin
6 S. Sun and N. S. Sariciftci (eds) (2005) Organic Photovoltaics: Mechanisms, Materials and Devices, CRC Press, Boca
Raton, Florida, US
7 S. Sun (2006) Organic and polymeric solar cells, in H. S. Nalwa (ed) Handbook of Organic Electronics and Photonics,
American Scientific Publishers, Los Angeles, California, US, Vol 2, Chapter 23
8 A. J. Heeger (2000) Semiconducting and metallic polymers: the fourth generation of polymeric materials, Nobel Lecture
in Chemistry.
9 Y. I. Frenkel (1931) Phys. Rev., no 37, p1276
10 R. Blankenship (2005) Natural organic photovoltaic solar energy transduction, in S. Sun and N. S. Sariciftci (eds) Organic
Photovoltaics: Mechanisms, Materials and Devices, CRC Press, Boca Raton, Florida, US, p37
11 V. Podzorov, E. Menard, A. Borissov, V. Kiryukhin, J. A. Rogers and M. E. Gershenson (2004), Phys. Rev. Lett., 93,
086602.
12 Pochettino (1906) Acad. Lincei Rendiconti, no 15, pp355363
13 R. Bube (1960) Photoconductivity of Solids, Wiley, New York
14 S. Anthoe (2002) Rom. Rep. Phys., no 53, p427
15 C. W. Tang and A. C. Albrecht (1975) J. Chem. Phys., no 62, p2139

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 7498

ch-03

2/6/07

96

8:21 PM

Page 96

S-S. SUN

16 B. R. Weinberger, M. Akhtar and S. Gau (1982) Synth. Metals, no 4, p187


17 S. Glenis, G. Tourillon and F. Garnier (1986) Thin Solid Films, no 139, p221
18 S. Karg, W. Riess, V. Dyakonov and M. Schwoerer (1993) Synth. Metals, no 54, p427
19 R. Marks, J. Halls, D. Bradley, R. Friend and A. Holmes (1994) J. Phys.: Cond. Matter, no 6, p1379
20 H. Antoniadis, B. Hsieh, M. Abkowitz, S. Jenekhe and M. Stolka (1994) Synth. Metals, no 62, p265
21 E. Frankevich et al (1992) Phys. Rev. B, no 46, p9320
22 L. Rothberg et al (1996) Synth. Metals, no 80, p41
23 C. W. Tang (1986) Appl. Phys. Lett. no 48, pp183185
24 Y. Yamashita, W. Takashima and K. Kaneto (1993) Jpn. J. Appl. Phys., no 32, p1017
25 N. Sariciftci et al (1993) Appl. Phys. Lett., no 62, p585
26 S. Morita, A. Zakhidov and K. Yoshino (1993) Jpn. J. Appl. Phys., no 32, p873
27 J. Halls et al (1996) Appl. Phys. Lett., no 68, p3120
28 K. Tada et al (1999) Synth. Metals, no 102, p982
29 M. Granstrom, K. Petritsch, A. Arias, A. Lux, M. Andersson and R. Friend (1998) Nature, no 395, p257
30 B. Gregg (2005) Coulomb forces in excitonic solar cells, in S. Sun and N. S. Sariciftci (eds) Organic Photovoltaics:
Mechanisms, Materials and Devices, CRC Press, Boca Raton, Florida, US, p139
31 B. A. Gregg (2003) J. Phys. Chem. B, no 107, pp46884698
32 M. Hiramoto, H. Fujiwara and M. Yokoyama (1991) Appl. Phys. Lett., no 58, pp10621064
33 N. S. Sariciftci, L. Smilowitz, A. J. Heeger and F. Wudl (1992) Science, no 258, p1474
34 B. Kraabel, J. Hummelen, D. Vacar, D. Moses, N. Sariciftci, A. Heeger and F. Wudl (1996) J. Chem. Phys., no 104,
pp42674273
35 G. Yu, J. Gao, J. Hummelen, F. Wudl and A. Heeger (1995) Science, no 270, pp17891791
36 J. Halls et al (1995) Nature, no 376, p498
37 L. S. Roman, M. Anderson, T. Yohannes and O. Inganas (1997) Adv. Mater., no 9, p1164
38 H. Hoppe and N. S. Sariciftci (2004) J. Mater. Res., no 19, pp19241945
39

H. Hoppe and N. Sariciftci (2005) Bulk heterojunction solar cells, in S. Sun and N. S. Sariciftci (eds) Organic
Photovoltaics: Mechanisms, Materials and Devices, CRC Press, Boca Raton, Florida, US, p217

40 G. Li, V. Shrotriya, J. Huang, Y. Yao, T. Moriarty, K. Emery and Y. Yang (2005) Nature Materials, no 4, pp864868
41 M. Drees, R. Davis and R. Heflin (2005) Polymer-fullerene concentration gradient photovoltaic devices by thermally
controlled interdiffusion, in S. Sun and N. S. Sariciftci (eds) Organic Photovoltaics: Mechanisms, Materials and Devices,
CRC Press, Boca Raton, Florida, US, (2005) p559
42 S. Sun, Photovoltaic Devices Based on a Novel Block Copolymer, US patent publication no 20040099307 (application
dated 14 November 2002)
43 S. Sun (2003) Sol. Energy Mat. Sol. Cells, no 79, pp257264
44 S. Sun, Z. Fan, Y. Wang and J. Haliburton (2005) J. Mater. Sci., no 40, pp14291443
45 C. Zhang, S. Choi, J. Haliburton, R. Li, T. Cleveland, S. Sun, A. Ledbetter, C. Bonner (2006) Macromolecules, no 39,
pp4317-4326.
46 K. Coakley and M. McGehee (2004) Chem. Mater., no 16, pp45334542
47 M. Jin and L. Dai (2005) Vertically aligned carbon nanotubes for organic photovoltaic devices, in S. Sun and
N. S. Sariciftci (eds) Organic Photovoltaics: Mechanisms, Materials and Devices, CRC Press, Boca Raton, Florida, US,
p579
48 Y. Kang, N. Park and D. Kim (2005) Appl. Phys. Lett., no 86, 113101.
49 L. Schmidt-Mende A. Fechtenktter, K. Mllen, E. Moons, R. H. Friend and J. D. MacKenzie (2001) Science, no 293,
p1119

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 7498

ch-03

2/6/07

8:21 PM

Page 97

Recent Progress of Organic Photovoltaics

97

50 B. Kippelen, S. Yoo, J. Haddock, B. Domercq, S. Barlow, B. Minch, W. Xia, S. Marder and N. Armstrong (2005) Liquid
crystal approaches to organic photovoltaics, in S. Sun and N. S. Sariciftci (eds) Organic Photovoltaics: Mechanisms,
Materials and Devices, CRC Press, Boca Raton, Florida, US, p271
51 G. V. Tsarenkov (1975) Sov. Phys. Semicond., vol 9, no 2, pp166171
52 G. Sassi (1989) J. Appl. Phys., no 65, pp322328
53 S. Sun (2005) Sol. Energy Mat. Sol. Cells, no 85, pp261267
54 S. Sun (2005) Mater. Sci. Eng. B., vol 116, no 3, pp251256
55 C. J. Brabec, A. Cravino, D. Meissner, N. S. Sariciftci, T. Fromherz, M. Minse, L. Sanchez and J. C. Hummelen (2001)
Adv. Funct. Mater., no 11, pp374380
56 J. Xue, S. Uchida, B. Rand and S. R. Forrest (2005) Appl. Phys. Lett., no 86, p5757
57 X. Crispin, S. Marciniak, W. Osikowicz, G. Zotti, A. W. D. van der Gon, F. Louwet, M. Fahlman, L. Groenendaal, F. De
Schryver and W. R. Salaneck (2003) Journal of Polymer Science Part B Polymer Physics, no 41, pp25612583
58 A. J. Mkinen, I. G. Hill, R. Shashidhar, N. Nikolov and Z. H. Kafafi (2001) Applied Physics Letters, no 79, pp557559
59 Y. Cao, G. Yu, C. Zhang, R. Menon and A. J. Heeger (1997) Synthetic Metals, no 87, pp171174
60 S. A. Carter, M. Angelopoulos, S. Karg, P.J. Brock and J.C. Scott (1997) Applied Physics Letters, no 70, pp20672069
61 A. Djuristic and C-Y. Kwong (2005) in S. Sun and N. S. Sariciftci (eds) Organic Photovoltaics: Mechanisms, Materials and
Devices, CRC Press, Boca Raton, Florida, US, p453
62 W. J. H. van Gennip, J. K. J. van Duren, P. C. Thne, R. A. J. Janssen and J. W. Niemantsverdriet (2002) J. Chem. Phys.,
no 117, pp50315035
63 P. Lane and Z. Kafafi (2005) in S. Sun and N. S. Sariciftci (eds) Organic Photovoltaics: Mechanisms, Materials and
Devices, CRC Press, Boca Raton, Florida, US, p49
64 Y. Gao (2005) in S. Sun and N. S. Sariciftci (eds) Organic Photovoltaics: Mechanisms, Materials and Devices, CRC Press,
Boca Raton, Florida, US, p421
65 N. Hadjichristidis, S. Pispas and G. Floudas (eds) (2003) Block Copolymers: Synthetic Strategies, Physical Properties, and
Applications, John Wiley & Sons, Inc., New York
66 M Lazzari, and M. Lopez-Quintela (2003) Adv. Mater., no 15, pp15841594
67 F. S. Bates and G. H. Fredrickson (1990) Ann. Rev. Phys. Chem., no 41, pp525557
68 B. de Boer, U. Stalmach, P. F. van Hutten, C. Melzer, V. V. Krasnikov and G. Hadziioannou (2001) Polymer, vol 42, no 21,
pp90979109
69 R. Segalman, C. Brochon and G. Hadziioannou (2005) Solar cells based on diblock copolymers: a PPV donor block and a
fullerene derivatized acceptor block, in S. Sun and N. S. Sariciftci (eds) Organic Photovoltaics: Mechanisms, Materials
and Devices, CRC Press, Boca Raton, Florida, US, p403
70 X. L. Chen and S. A. Jenekhe (1996) Macromolecules, no 29, p6189
71 T. A. Skotheim, R. L. Elsenbaumer and J. R. Reynolds (eds) (1998) Handbook of Conducting Polymers, 2nd ed, Marcel
Dekker, New York
72 Z. Bao, A. Dodabalapur and A. J. Lovinger (1996) Appl. Phys. Lett., no 69, p4108
73 T. Nguyen, J. Wu, V. Doan, B. Schwartz and S. H. Tolbert (2000) Science, no 288, pp652656
74 S. Sun, Z. Fan, Y. Wang, C. Taft, J. Haliburton and S. Maaref (2002) in D. Fichou and K. Kafafi (eds) Organic Photovoltaics
II, SPIE, Bellingham, WA, US, vol 4465, pp121-128
75 S. Sun, Z. Fan, Y. Wang, J. Haliburton, C. Taft, K. Seo and C. Bonner (2003) Syn. Met., no 137, pp883884
76 S. Sun (2004) in P. Lane and Z. Kafafi (eds) Organic Photovoltaics IV, SPIE, Bellingham, WA, US, vol 5215, pp195205
77 S. Sun (2004) in P. Lane and Z. Kafafi (eds) Organic Photovoltaics V, SPIE, Bellingham, WA, US, vol 5520, pp126135
78 S. Thomas, C. Zhang and S. Sun (2005) J. Poly. Sci. (A) Polym. Chem., vol 43, pp42804287
79 S. Sun, Z. Fan, Y. Wang, K. Winston and C. E. Bonner (2005) Mater. Sci. Eng. B., vol 116, no 3, pp279282

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 7498

ch-03

2/6/07

98

8:21 PM

Page 98

S-S. SUN

80 S. Sun, C. Zhang, J. Haliburton, A. Ledbetter, C. Bonner, M. Drees and N. Sariciftci (2005) in P. Lane and Z. Kafafi (eds)
Organic Photovoltaics VI, SPIE, Bellingham, WA, US, vol 5938, pp59380D1-8
81 M. Knupfer (2003) Appl. Phys. A, no 77, pp623626
82 S. Sensfuss and M. Al-Ibrahim (2005) in S. Sun and N. S. Sariciftci (eds) Organic Photovoltaics: Mechanisms, Materials
and Devices, CRC Press, Boca Raton, Florida, US, p529
83 R. A. Marcus (1993) Rev. Mod. Phys., no 65, pp599610
84 V. Balzani (ed) (2000) Electron Transfer in Chemistry, Wiley-VCH, New York
85 E. Peeters, P. Hal, J. Knol, C. Brabec, N. Sariciftci, J. Hummelen and R. Janssen (2000) J. Phy. Chem. B, no 104,
pp1017410190
86 E. Neuteboom, S. J. Meskers, P. van Hal, J. van Duren, E. Meijer, R. Janssen, H. Dupin, G. Pourtois, J. Cornil, R.
Lazzaroni, J. Brdas and D. Beljonne (2003) J. Am. Chem. Soc., no 125, pp86258638
87 J. Cornil, V. Lemaur, M.C. Steel H. Dupin, A. Burquel, D. Beljonne and J. L. Brdas (2005) Electronic structure of organic
photovoltaic materials: modeling of exciton-dissociation and charge-recombination processes, in S. Sun and N. S.
Sariciftci (eds) Organic Photovoltaics: Mechanisms, Materials and Devices, CRC Press, Boca Raton, Florida, US,
pp161182
88 J. Miller, L. Calcaterra and G. Closs (1984) J. Am. Chem. Soc., no 106, pp30473049
89 A. Weller (1982) Photoinduced electron transfer in solution, Phys. Chem., no 133, pp9398

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 7498

ch-04

2/7/07

11:51 AM

Page 99

Thermal and Material


Characterization of Immersed
Heat Exchangers for Solar
Domestic Hot Water
Jane H. Davidson, Susan C. Mantell and Lorraine F. Francis

Abstract
This chapter provides an overview of methods to characterize the thermal performance
of heat exchangers immersed in thermal storage vessels intended for solar domestic
water heating systems and to assess the mechanical durability and scaling potential of
polymers for this application. Recent research at the University of Minnesota is
summarized and recommendations are made for future research to further the
development of polymer heat exchangers and piping.

Keywords heat exchangers; domestic hot water; integral collector storage; polymers; natural
convection; nylon; heat transfer

4.1 INTRODUCTION
Immersed heat exchangers are common in European combistore water storage tanks for
combined domestic hot water and space heating, but in the US, low temperature solar
thermal systems are used almost exclusively for domestic hot water and immersed heat
exchangers are atypical. This situation is expected to change as lower cost and simplified
solar domestic hot water (SDHW) systems are developed for the US market. The recent
focus of US efforts to reduce system cost is an indirect integral collector storage (ICS)
system (Figure 4.1). An immersed heat exchanger is required to discharge the
unpressurized solar storage. During discharge, the domestic pressurized water flows
through the heat exchanger and is delivered to the conventional water heater. Because
the load-side flow is driven by the domestic water pressure, no mechanical pump is
required. Recent research and development for this system, as well as for systems
intended for cold climates, addresses the barriers and opportunities to shift from copper
and glass components to integrated systems manufactured using mass production
techniques, such as those associated with polymeric materials. Significant cost savings

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

ch-04

2/7/07

11:51 AM

100

Page 100

J. H. DAVIDSON, S. C. MANTELL AND L. F. FRANCIS

GLOSSARY OF TERMS USED IN THIS PAPER


a
A
C
cp
D
E
Er
f
g
h
i
k
Ksp
L
mj
n
NuD
Nu*D
NTU
o
P
Pr
.
Q
r
Ri
Ro
Rw
RaD
RaD
s
S
t
t0
Tf,in
Tf,out
Ts
Ts,0
Tw
U
UA
Vr
Vs
.
V
w

ionic activity, units depend on constituent


heat transfer surface area of each heat exchanger tube, m2
empirical constant
specific heat of water, J/kgK
outer diameter of heat exchanger tube
delivered energy, J
ratio of energy delivered by divided storage to energy delivered by undivided storage
subscript used to indicate heat exchange fluid
gravitational constant, m/s2
convective heat transfer coefficient, W/m2K
subscript indicating inside of tube or inner diameter
thermal conductivity, W/mK
solubility product for crystalline CaCO3, moles/l
length of heat exchanger tube (per compartment), m
heat exchanger mass flow rate, kg/s
number of heat exchanger tubes
Nusselt number, hD/kf
average Nusselt number with initial storage temperature as reference temperature
number of transfer units
subscript indicating outside of tube
tube pitch, or center to center spacing, m
Prandtl number, /
heat transfer, W
radial distance from the tube center, m
convective resistance of forced flow through the heat exchanger, K/W
convective resistance of natural convection on the outer or storage side of the heat
exchanger, K/W
conductive resistance across the heat exchanger wall, K/W
Rayleigh number, gD3(T)/()
initial Rayleigh number, gD3(Tw Ts,0)/()
subscript indicating storage side medium
supersaturation of water with respect to calcium carbonate,
time, s
time scale, s
water temperature at the inlet of the heat exchanger, K
water temperature at the outlet of heat exchanger, K
average storage temperature, K
initial storage temperature, K
temperature of the outer tube wall, K
velocity scale, m/s
overall heat transfer coefficient-area product, W/K
ratio of hot water volumetric output to storage volume
volume of storage fluid, m3
tube-side volumetric flow rate, m3/s
subscript used to indicate tube surface
thermal diffusivity, m2/s

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

ch-04

2/7/07

11:51 AM

Page 101

Thermal and Material Characterization of Immersed Heat Exchangers

coefficient of volumetric thermal expansion, K1


indicates temperature difference
heat exchanger effectiveness
dimensionless variable, (cp)f/(cp)s.
positive roots of the Bessel function
kinematic viscosity, m2/s
dimensionless temperature
dimensionless tank averaged temperature
dimensionless time
geometric parameter

Yo

Bessel function of the second kind

101

FIGURE 4.1 Conceptual drawing of an indirect integral solar collector storage (ICS) system with a load-side
immersed heat exchanger for solar domestic water heating

are anticipated using multi-component injection molding or extrusion, and integration of


components and fittings. Additional savings are expected due to weight reduction, which
translates to reductions in shipping, and installation costs. The status of material selection
for the glazing, collector enclosure and heat exchanger were first described in volume 15
of Advances in Solar Energy (Davidson et al, 2002). The present paper summarizes the
results of our continuing study of heat transfer and long-term durability and reliability of
polymers for immersed heat exchangers. In section 4.2, a summary of the operation of
immersed heat exchangers is presented with an emphasis on study and characterization
of natural convection heat transfer. In section 4.3, the durability of candidate polymers in
hot chlorinated water is discussed and recent data are presented. Section 4.4 summarizes
the state of knowledge regarding the formation of calcium carbonate, usually called scale,
on polymeric surfaces.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

ch-04

2/7/07

102

11:51 AM

Page 102

J. H. DAVIDSON, S. C. MANTELL AND L. F. FRANCIS

4.2 THERMAL CHARACTERIZATION AND DESIGN


4.2.1 THERMAL PERFORMANCE
Immersed heat exchangers may be used to charge or discharge water storage tanks. Heat
transfer between the heat exchange surface and the storage fluid is via natural
convection. Thus the heat exchanger should be located to take advantage of the greatest
difference between the temperature of the storage fluid and the temperature of the fluid
flowing through the heat exchanger (Figure 4.2). To discharge the tank, the heat exchanger
should be located near the top of the storage tank where the storage temperature is
presumably the warmest. The cold plumes formed in the boundary layer at the heat
exchange surface will move downward in the tank. Conversely, to charge the storage tank,
the heat exchanger should be placed at the bottom of the tank. These arrangements
ensure that the storage vessel can be completely charged or discharged.

(a)

(b)

FIGURE 4.2 Conceptual sketch of an immersed heat exchanger in a vertical thermal storage tank; the heat
exchanger is shown here in two positions depending on the intended mode of operation: a) top mounted heat
exchanger intended to discharge the tank, and b) bottom mounted heat exchanger intended to charge the tank

The thermal performance of the heat exchanger is described by the overall heat
transfer coefficient-area product (UA), which is the inverse of the sum of the convective
and conductive thermal resistances:
(1)
In equation (1), Ri is the thermal resistance of the forced convection flow through the heat
exchanger, Rw is the conductive resistance across the heat exchanger wall and Ro is the
natural convection thermal resistance on the outer storage-side of the heat exchanger.
The effectiveness of an immersed heat exchanger is related to the UA by:
(2)

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

ch-04

2/7/07

11:51 AM

Page 103

Thermal and Material Characterization of Immersed Heat Exchangers

103

The subscript s denotes the storage fluid and the subscript f denotes the heat
exchange fluid. The bracketed term on the right hand side of equation (2) is referred to as
the number of transfer units (NTU) of the heat exchanger. For metal heat exchangers, the
conductive resistance across the tube wall (Rw) may be neglected (unless there is
substantial scaling due to use in hard water or if double wall tubes are used). If the tube
is plastic, thin-walled tubes with a relatively high ratio of outer diameter-to-wall thickness,
termed the standard diameter ratio (SDR), are desirable to minimize Rw (Liu et al, 2000;
Davidson et al, 2002). The relative magnitude of the thermal resistances depends on many
factors including the flow rate and flow regime (laminar or turbulent) through the heat
exchanger, the geometry of the combined heat exchanger and storage tank, and the
temperature difference between the outer heat exchange surface and the storage fluid
(the state of charge). The natural convection heat transfer may limit the rate of heat
transfer, especially if the internal flow is turbulent. In this case, it is very important to
accurately characterize Ro:
(3)
h o Ao

where ho is the natural convection heat transfer coefficient.


Determination of the ho of heat exchangers immersed in solar storage tanks has been
the subject of numerous experimental and numerical studies (Feiereison et al, 1982;
Farrington and Bingham, 1986 and 1987; Reindl, 1992; Reindl et al, 1992a and 1992b; Liu
et al, 2003, 2004 and 2005; Su and Davidson, 2005). The problem poses unique
challenges compared to the extensively studied problem of natural convection to bodies
in an unbounded fluid. First, the process is transient. During both charge and discharge,
the temperature difference that drives natural convection to the heat exchanger
decreases in magnitude as charge or discharge procedure progresses. Thus in the
expected situation where the natural convection poses a significant thermal resistance
compared to wall conduction and forced convection heat transfer, the overall heat transfer
coefficient U, and, thus, effectiveness and NTU decrease with time. Second, the spatial
distribution of the storage temperature may change during the discharge process and
depends on the specific heat exchanger/tank geometry. Thus a generalized choice for the
appropriate temperature difference to describe natural convection has been elusive.
The starting point for most efforts to quantify the natural convection heat transfer
coefficient has been the conventional empirical correlation of Nusselt and Rayleigh
numbers:
(4)
where:
(5)

(6)

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

ch-04

2/7/07

104

11:51 AM

Page 104

J. H. DAVIDSON, S. C. MANTELL AND L. F. FRANCIS

The characteristic length scale in expressions (5) and (6) is the outer diameter D of an
individual tube in a tube bundle or coil. The empirical constant C and exponent m should
depend only on the geometry of the heat exchanger/tank and the range of Rayleigh
number. The real challenge in developing correlations for this application is measurement
and identification of an appropriate characteristic temperature difference between the
heat exchange surface and the storage fluid, T = Tw Ts. Early studies of smooth and
finned coiled and bayonet heat exchangers (Feiereison et al, 1982; Farrington, 1986;
Farrington and Bingham, 1986 and 1987) did not measure the heat exchanger surface
temperature but rather used various definitions of the log-mean temperature difference to
correlate data. The storage fluid was assumed to be fully mixed. The correlations were
generally unsuccessful in the sense that the constant C was temperature dependent.
Natural convection to a single tube (Liu et al, 2003) and several tube bundle heat
exchangers (Liu et al, 2003 and 2004) immersed in a thin rectangular storage tank with a
height to width aspect ratio of 9.3:1 and inclined at 30 degrees to the horizontal was
measured at the University of Minnesota. The tube wall temperature was measured
directly. When the storage fluid was initially isothermal, it was possible to correlate the
data using the average storage temperature. When the storage fluid was initially stratified,
it was necessary to use the local storage fluid temperature surrounding the heat
exchanger tubes to define Ts and to obtain statistically significant correlations of the
measured data. For a single tube, the data were best correlated by:
NuD = (0.675 0.001)RaD0.25, for 106 RaD 108.

(7)

For a bundle of eight tubes with pitch-to-diameter (P/D) = 2.43:


NuD = (0.728 0.002)RaD0.25, 105 RaD 107.

(8)

For bundles of 240 tubes with 1.5 P/D 3.3:


NuD = (2.45 0.03)RaD0.188, 102 RaD 104.

(9)

Three important findings emerged from these studies. First, comparison of equations
(7)(9) to the correlations developed for a horizontal tube in an unbounded fluid (Morgan,
1975) proved that the presence of the storage tank enhanced heat transfer. The heat
transfer enhancement is attributed to the buoyancy driven flow in the tank. For a single
tube, the correlation predicts average heat transfer coefficients ~20 per cent larger than
those for a heated tube in an unbounded fluid. Second, fluid motion in the storage fluid
was sufficient to maintain a fully mixed tank for an initially isothermal storage tank. If the
storage fluid was initially stratified, operation of the heat exchanger destroyed
stratification relatively quickly. Third, for tube bundles, the pitch to diameter ratio (P/D) had
no statistically significant impact on the overall heat transfer for 1.5 P/D 3.3. This result
implies that it is possible to pack the heat exchanger tubes in close proximity to each
other without substantially decreasing overall heat transfer. For the same heat exchange
area, multiple tubes reduce the pressure drop across the heat exchanger.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

ch-04

2/7/07

11:51 AM

Page 105

Thermal and Material Characterization of Immersed Heat Exchangers

105

Use of equations (7)(9) for design of heat exchangers is difficult because it requires
knowledge of the temporal behavior of the storage fluid temperature during the discharge
process. An alternate approach that relies solely on the initial storage temperature was
first suggested by Reindl et al (1992b). From a scale analysis of the problem, a semiempirical expression for the time-dependent Nusselt number based on the initial Rayleigh
number of an initially isothermal tank was proposed. Su and Davidson (2005) extended the
scale analysis to consider discharge of a storage tank with tube bundle heat exchangers
and presented the results in a more generalized dimensionless form.
Su and Davidson (2005) divide the process into four temporal periods referred to as
conduction, quasi-steady, fluctuating and decay. General formulations for the transient
Nusselt number and the volume averaged water temperature in the enclosure were
developed for each period. Three dimensional CFD results illustrate the flow and
temperature fields within the storage tank (Figure 4.3). At the initiation of the discharge, heat
transfer is dominated by conduction. The convective flow is only apparent near the heat
exchanger and the buoyancy-induced velocity is very low. The dimensionless temperature
in the tank during the conduction period may be determined analytically and is given by:
(10)
where n are the positive roots of the Bessel function of the second kind,Y0 (n) = 0, and
r is the dimensionless radial distance measured from the heat exchanger. The volume
averaged storage temperature is defined as:
(11)

where Ts is the average storage temperature at a specified time, and Ts,0 is the fully-mixed
storage temperature at the start of the charge or discharge process (at t = 0). The average
Nusselt number is expressed as:
(12)
The asterisk on the Nusselt number denotes that it is defined in terms of the initial, and
presumed known, difference between the initial surface temperature of the heat exchanger
and the initial temperature of the storage fluid, i.e. T = Tw Ts,0. The conduction period
. The dimensionless time, , is given by:
ends relatively quickly at = Pr
s
(13)
where t0 is the ratio of the convective length scale and a velocity scale for natural
convection. The natural convection velocity scale U is given by:
(14)
Thus the time scale is:
(15)

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

ch-04

2/7/07

106

11:51 AM

Page 106

J. H. DAVIDSON, S. C. MANTELL AND L. F. FRANCIS

FIGURE 4.3 Three dimensional streamlines (top) and isotherms (bottom) for a rectangular storage with a single
immersed tube; T = |Tw To| = 20K (T0 = 353K and Tw = 333K) corresponding to Ra*D = 2.345 107 during
(a) conduction ( = 0.029, t = 0.014s); (b) conduction ( = 1.953, t = 0.955s); (c) quasi-steady ( = 11.778,
t = 5.76s); (d) fluctuating ( = 111.78, t = 54.66s); (e) decay ( = 2441, t = 1194s); and (f) decay ( = 65,710,
t = 32,315s)

Once the conduction period ends, heat transfer becomes dominated increasingly by
natural convection. Initially, the presence of the storage container does not impact the
flow field or heat transfer to the heat exchanger. Consequently, the convective heat
transfer is described by the average Nusselt number for a body immersed in an
unbounded fluid, given in the form of equation (4). This quasi-steady period endures until

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

ch-04

2/7/07

11:51 AM

Page 107

Thermal and Material Characterization of Immersed Heat Exchangers

107

the flow in the enclosure begins to disrupt the boundary layer of the heat exchanger. The
duration depends on the geometry of the tank, the position of the heat exchanger within
the tank and the magnitude of the convective velocity given in equation (14).
At the end of the quasi-steady period, the flow field and heat transfer are influenced
by the presence of the tank. CFD results show there is a short period of intense mixing of
the storage fluid and intermittent disruption and growth of the boundary layer at the heat
exchanger surface. The Nusselt number fluctuates around the quasi-steady value. The
duration of the fluctuating period is not well defined except by observation of numerical
or measured data.
The last period and most important one in terms of overall system performance is the
decay period during which over 90 per cent of the energy is removed (discharge) or added
(charge) to the storage tank. It is during this period that all measured data have been
obtained. The driving temperature difference between the heat exchanger surface and the
storage fluid decreases exponentially. The transient Nusselt number is expressed as:
(16)
where the dimensionless transient temperature difference is given by:

(17)
The geometric parameter is specified by the heat exchanger/storage configuration.
It equals the volume of the storage fluid divided by the product of the heat transfer
surface area of the heat exchanger and the characteristic length scale for natural
convection heat transfer. For N tubes, each of diameter D and outer heat transfer surface
area Ao:
(18)
A generalized expression for the transient Nusselt number is:

(19)
The strength of equation (19) is that it requires only the initial storage temperature and the
heat exchanger geometry and operating conditions, from which Ra*D and can be
determined, to determine heat transfer during the entire decay period. The values of C and
m must be determined empirically and depend only on the geometry. Su and Davidson (2005)
used the data of Liu et al (2003, 2004 and 2005) for single- and multiple-tube heat exchangers
to test the usefulness of equation (19). The data for the single-tube heat exchanger in an
initially isothermal storage were successfully correlated with m = 0.25 and C = 0.50. For the
240-tube heat exchanger with P/D = 3.3, the figures were m = 0.188 and C = 2.6.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

ch-04

2/7/07

108

11:51 AM

Page 108

J. H. DAVIDSON, S. C. MANTELL AND L. F. FRANCIS

In situations where storage fluid is stratified at the initiation of the charge or discharge
and stratification is maintained by the use of manifolds or baffles, equation (19) is not
applicable. If the fluid is stratified, the average tank temperature does not represent the
thermal environment of the heat exchanger. In this case, use of the initial Rayleigh number
in equation (19) will underestimate the natural convection film coefficient at the beginning
of the charge or discharge process and overestimate it as the storage tank reaches a full
or depleted state of charge, respectively. In this case, correlations of the form of equation
(4) may be used but it will be necessary to characterize the temperature of the storage
fluid surrounding the heat exchanger during use.
Once the UA of the heat exchanger has been estimated, the transient behavior of the
heat exchanger/storage tank can be predicted. The transient storage temperature Ts(t) and
delivered outlet water temperature Tf,out(t) can be estimated from energy balances on the
heat exchanger and storage. Assuming the inlet fluid temperature and mass flow rate
) are fixed and the storage fluid remains fully mixed
through the heat exchanger (m
f
throughout the charge or discharge process, the transient storage temperature and heat
exchanger outlet temperature are given by:
(20)

(21)
The parameter = (cp)f/(cp) is approximately equal to the heat exchanger
effectiveness when the heat exchange and storage fluids are the same fluid. The volume
ratio Vr equals the volume of fluid passed through the heat exchanger divided by the
ft/f /Vs.
storage volume: Vr=m

4.2.2 DESIGNS FOR IMPROVED PERFORMANCE


Efforts to improve the thermal performance of thermal stores with immersed heat
exchangers have focused on three concepts: shrouds or baffles to direct the buoyant flow
that develops at the heat exchanger, stratification manifolds, and partitioned storage
vessels. Several European combistores include a shroud around the immersed heat
exchanger connected to a stratification pipe to prevent the natural convective plumes that
develop in the boundary layer of the coil from mixing with the surrounding fluid. Work to
characterize these systems is reported by the International Energy Agency (IEA) Task 26
(Drck, 2002; Weiss, 2003).
This review focuses on the use of a divided or partitioned storage vessel to improve
performance. The divided storage concept is an adaptation of the use of multiple storage
tanks connected in series first suggested by Bejan (1982). Analyses of the concept are
presented by Sekulic and Krane (1992a and 1992b) for two tanks in series each with an
immersed heat exchanger and by Boies and Homan (2004 and 2005) and Boies et al
(2005) for multiple storage tanks. Mather et al (2002) proposed using multiple tanks for
large solar thermal storage systems with the view that a series of small storage tanks
should be more economical and more practical than the single large (greater than 2000
liter) storage tank used in a combistore. For smaller SDHW systems, dividing a single tank

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

ch-04

2/7/07

11:51 AM

Page 109

Thermal and Material Characterization of Immersed Heat Exchangers

109

into two or more storage compartments is practical and has the advantage of reduced
thermal losses to the ambient.
At the University of Minnesota, we have conducted experiments with a single storage
tank divided into two compartments (Davidson et al, 2005). The dual compartment
storage tank, illustrated in Figure 4.4 for discharge of the store, provides a staged heating
process. The inlet of the heat exchanger is in the first compartment. The heat exchange
fluid entering the second compartment is warmer than that entering the first
compartment and thus the storage fluid in compartment two remains warmer for longer
periods of time. Consequently, for much of the discharge process, the heat exchanger
outlet temperature is higher than it would be using the same heat exchanger in an
undivided storage. The advantage of the approach is that the enhanced heat transfer
attributed to convective motion in the storage (Liu et al, 2003, 2004 and 2005) is
maintained in each compartment. Boies and Homan (2004) point out that the benefit of
this concept can be realized in either of two ways. For a fixed total storage volume, a
higher energy delivery rate is maintained for a longer period, which means more hot water
is produced. Alternately, the same energy output can be provided by a smaller storage
volume.

FIGURE 4.4 Sketch of a divided indirect thermal storage vessel with two storage compartments; the heat
exchanger is positioned for discharge

The advantage of the divided storage depends on the effectiveness, or equivalently


the NTU, of the immersed heat exchanger, and the size of the thermal store relative to the
hot water draw (Vr). Too low an NTU will limit the energy transfer rate and diminish the
advantage of the divided storage. On the other hand, increasing the NTU is advantageous
only up to a certain value. At high NTU, heat transfer is limited by the effective
temperature difference between the heat exchange and storage fluids. The effects of NTU
and Vr on cumulative delivered energy of a storage tank with two compartments are
illustrated in Figure 4.5 for an idealized situation in which the heat exchanger NTU is

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

2/7/07

11:51 AM

110

Page 110

J. H. DAVIDSON, S. C. MANTELL AND L. F. FRANCIS

1.18
1.16

NTU = 12

1.14

Energy Ratio (Er)

ch-04

NTU = 6

1.12
1.10

NTU = 4

1.08
Equal outlet
temperatures

1.06
1.04

NTU = 2

1.02
NTU = 1

1.00
0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Volume Ratio (Vr)

FIGURE 4.5 Predicted ratio of delivered energy of a divided storage and an undivided storage as a function
of the dimensionless output volume and NTU. The divided storage has two compartments. The dashed line
indicates the volume ratio at which the heat exchanger outlet temperatures of the two storage configurations
are equal

constant during discharge and thermal losses are neglected. Here the ratio of energy
delivered by a divided storage to that delivered by an undivided storage, Er, is plotted
versus Vr for 1 NTU 12. The NTU values shown are those of the undivided storage. For
the divided storage, the NTU is assumed to be divided equally between the two storage
compartments. The energy ratio increases as overall NTU is increased from 1 to 10, with
little change from NTU = 10 to 12. For a given NTU, the energy ratio depends on the
volume ratio, or extent of depletion of the storage. The maximum cumulative delivered
energy is at Vr  0.85. For example, at NTU = 10, Er is 1.15. The dashed line on the plot
indicates the value of Vr for which the outlet temperatures of the undivided and divided
storage are equal. The region to the left of the dashed line, where the divided storage
delivers higher temperature water than the undivided storage, represents the desirable
operating range. The only case for which there is no advantage to the divided storage
system is if the storage is too small to meet the hot water load and the stored energy is
completely discharged during a hot water draw. When the storage unit is completely
discharged, Er = 1.
The measured performance of a divided storage vessel is shown in Figure 4.6. These
data were obtained for a 126 liter rectangular storage vessel with a copper heat exchanger
with a total surface area of 2.38m2 immersed in the storage fluid. Experiments were
conducted with one and two compartments. In the divided storage vessel, the compartments
were separated with a 2.54cm thick polystyrene sheet (R-value = 0.881K-m2/W). Figure 4.6
is a plot of Er versus Vr for nominal NTUs of 2.5 and 7. With NTU = 7, the divided storage
delivers 11 per cent more energy than the undivided storage at Vr = 0.8 (i.e. when 100
liters of hot water or 55 per cent of the stored energy has been delivered). Higher outlet
temperatures are sustained by the divided storage until Vr = 1.2. At this point, the divided

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

2/7/07

11:51 AM

Page 111

Thermal and Material Characterization of Immersed Heat Exchangers

111

1.14
1.12

Energy Ratio (Er)

ch-04

NTU = 7

1.10
1.08
1.06
1.04

NTU = 2.5

1.02
1.00
0.0

0.5

1.0

1.5

2.0

2.5

3.0

Volume Ratio (Vr)


FIGURE 4.6 Measured ratio of delivered energy of a divided and undivided storage as a function of ratio of
the volume of delivered hot water to the volume of storage fluid for nominal heat exchanger NTU of 2.5 and 7

storage provides 8 per cent more energy than the undivided storage and the storage is 74
per cent depleted. With NTU = 2.5, the divided storage delivers 5 per cent more energy
at Vr = 0.8. At Vr = 1.3, when the outlet temperatures of the divided and undivided storage
are equal, the divided storage provides 4.6 per cent more energy and the stored energy
is 65 per cent depleted. The measured advantage of the divided storage is slightly lower
than that predicted because the NTU of the heat exchanger decreases as the tank is
discharged. However, even with a simple partition, the divided storage vessel provides the
same benefit as that predicted for multiple storage tanks. The concept is expected to be
easy to implement in solar systems at minimal or no extra cost.

4.2.3 SUMMARY
Substantial progress has been made in recent years to understand and characterize the
thermal/fluid processes in storage vessels with immersed heat exchangers. Natural
convection heat transfer correlations have been developed for tube bundles and the
generalized approach to express measured data in the form of equation (19) holds
promise for fully mixed storage vessels. The empirical constant in this expression will
depend on the combined heat exchanger/tank design and must be determined from
experiment or CFD. The computational problem requires substantial computational
resources. It is anticipated that most developers and designers will choose the
experimental approach. Once the constant is determined, the transient charge or
discharge process can be predicted using only the initial conditions. If baffles or other
stratification devices are shown to be effective, equation (4) may be used. In this case, the
key to predicting the natural convection heat transfer is to determine the thermal
environment to which the heat exchanger is exposed.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

ch-04

2/7/07

112

11:51 AM

Page 112

J. H. DAVIDSON, S. C. MANTELL AND L. F. FRANCIS

4.3 MECHANICAL CHARACTERIZATION OF POLYMERS


Traditionally, immersed heat exchangers are copper. The switch from copper to polymer
materials requires careful consideration of the compatibility of the material with the
working fluids. Polymer materials must maintain mechanical properties in the working
fluid over their target lifetimes. Although thin walled tubes with a high SDR are desirable
for maximizing heat transfer (see section 4.2), tube wall thickness must be sufficient to
ensure structural integrity throughout the lifetime of the heat exchanger. The traditional
approach to evaluating long-term performance of polymers has been to consider burst
strength. However, the creep compliance is equally important, because material
deformation may exceed acceptable limits long before the polymer component ruptures
(Wu et al, 2004). In the past, we have focused on a number of potential materials. Of
these, we selected polysulfone (PSU), polybutylene (PB) and polyamide 6,6 (PA66) for
additional study. PSU was selected because it is currently used in aqueous environments
for pipe fittings and water heater dip tubes. PB was selected because it is used outside
the US for hot water distribution and in the US as an internal liner tank in a commercially
available plastic electric water heater. In addition, PB represents polyolefins, which use
additives to protect them from degradation in water. Polyamide 6,6 was selected because
it is a high strength material. In addition it is known to absorb water and this property is
of interest in terms of understanding the mechanisms for degradation of mechanical
properties and scaling (discussed in section 4). Cross-linked polyethylene (PEX) was not
selected for further study because PEX tubes are designed for a 40-year lifetime of
exposure to hot chlorinated water (following ASTM F2023, 2000).
In potable water, chlorine and pH combine to create an oxidative environment,
characterized by the oxidative reduction potential (ORP), that can chemically attack a
polymer, resulting in permanent loss of mechanical strength and increase in creep
compliance. Water absorption and hydrolysis can also diminish polymer mechanical
properties. The mechanism of polymer degradation will depend on the polymer molecular
structure. Thus both polymer morphology data and mechanical properties are required to
understand the relationship between degradation and loss of mechanical performance.

4.3.1 BACKGROUND ON POLYMER DEGRADATION AND PRIOR STUDIES


The molecular structures of PSU, PB and PA66 differ and predispose each polymer to
different degradation mechanisms in hot potable water (Scott, 1999). Of the three
polymers, PSU is the most resistant to molecular degradation because the polymer chain
is comprised of aromatic rings and strong carbon, sulfur and oxygen bonds within the
polymer backbone. Polybutylene, a polyolefin comprised of carbon and hydrogen atoms,
will degrade by oxidation: a hydrogen atom is easily abstracted from a tertiary carbon
within the polymer chain. The lifetime of PB in an oxidative environment is extended by
including an antioxidant. Polybutylene performance will not degrade significantly until the
antioxidant has been depleted. Polyamide 6,6 is a semi-crystalline polyamide that has
good mechanical properties in the air but is prone to water absorption and hydrolysis. The
absorbed water acts as a plasticizing agent in PA66 and increases polymer chain mobility
resulting in a reduction in tensile strength (Aharoni, 1997). The properties may be restored
if the water is removed. However, if PA66 remains in water for longer periods, hydrolysis

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

ch-04

2/7/07

11:51 AM

Page 113

Thermal and Material Characterization of Immersed Heat Exchangers

113

is likely. The PA66 polymer chain includes an amide group which can by hydrolyzed in
water, leading to chain scission and permanent damage.
The effect of hot water on the molecular structure of polyolefins was analyzed using
infrared spectroscopy, size exclusion chromatography, thermal analysis and visual
observations (Gedde et al, 1994). Cross-linked polyethylene (PEX), polybutene-1,
polypropylene and medium-density polyethylene were studied. These researchers found
that there is an incubation period during which there is a decrease in the antioxidant
concentration in a polyolefin. The incubation period ends when the antioxidants are
completely depleted, leaving the material with an extremely limited lifetime. Oxidation
was found to be the main mechanism responsible for polymer degradation. The study,
although limited to polyolefins, established a method for predicting the lifetime of a
polymer based on antioxidant depletion.
Prior work addressing the effect of a hot potable water environment on mechanical
properties of polymer tubing has focused on polymer strength. The National Sanitary
Foundation (NSF) and the American Society for Testing and Materials (ASTM) have both
developed testing protocols, NSF standard P171 (1999) and ASTM standard F2023 (2000),
to evaluate the burst strength of polymer tubing in hot chlorinated water. In these tests, the
tube rupture must be a result of the oxidative environment as indicated by a failure in which
there are many cracks emanating from the interior tube wall and extending to the tube
surface (stage III failure). The water chlorine level ranges between 3 to 5ppm and pH level
ranges from 6 to 8 such that an ORP of 825mV is maintained throughout the test. (Note that
the ORP of typical tap water ranges from 400 to 550mV (ASTM, 2000).) Tests are performed
over several combinations of pressure and temperature so as to extrapolate the burst
strength of the tubing after 10 to 40 years exposure to hot chlorinated water. Test results
are limited to the particular tube geometries (diameter to thickness ratio) tested. These
standards provide a methodology to predict polymer lifetime based on a material strength
limit, the hydrostatic burst strength. Material deformation (in other words creep compliance)
is not evaluated during testing and is not considered in predicting material lifetime.
Bradley et al (2000) used a similar test to determine the effects of temperature and
chlorine levels on the tensile strength and weight of polymer tubes. Polymers were tested
in conditions of 28C, 60C and 90C with 5ppm chlorine and also at 90C with 0ppm
chlorine. The strength of the materials when immersed in a hot chlorinated environment
for 1500 hours was evaluated under (1) no load and (2) continuous loading conditions. The
harshest condition, 90C and 5ppm chlorine, was very detrimental to PA66, causing it to
lose 70 per cent of its weight and 95 per cent of its tensile strength. Reducing the
temperature to 60C from 90C caused the degradation of PA66 to slow dramatically. PSU,
polyphenylsulfone, polyphenylsulfide and polyvinylidene fluoride exhibited good
resistance to chlorine exposure at high temperature with little weight or tensile strength
loss. The creep compliance was not evaluated. Limited creep data in oxidative
environments for PB and PA66 have been published by researchers at the University of
Minnesota (Walter et al, 2003; Wu et al, 2004). Because the test data were obtained for
only one environmental condition (825mV ORP and 82C), the effects of changes in ORP
and temperature could not be evaluated. No tests of chemical or physical degradation
(such as molecular weight or SEM) were performed.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

ch-04

2/7/07

114

11:51 AM

Page 114

J. H. DAVIDSON, S. C. MANTELL AND L. F. FRANCIS

4.3.2 EVALUATION OF POLYMER MATERIALS: PSU, PB, AND PA66


Recent work at the University of Minnesota has focused on creep of PSU, PB and PA66 in
an oxidative aqueous environment (Freeman, 2005; Freeman et al, 2005). Other indicators
of polymer degradation, tensile strength, molecular weight and surface morphology
(through SEM images) were also evaluated. A review of this work is provided in this
section.
Creep and tensile strength were measured in a series of experiments in which the
polymers were submerged in water at controlled temperature and ORP levels. Specimens
were either injection molded (PSU and PB) or cut from extruded sheets (PA66). The test
environments consisted of combinations of two temperature settings, 60C and 80C, and
two ORP settings, 550mV and 825mV, for a total of four test environments. Data were also
obtained in air at room temperature. Creep compliance was measured with specimens
that were continuously loaded at stress levels of approximately 11MPa (PSU), 10.5MPa
(PB), and 7MPa (PA66). The stress levels were chosen to give measurable creep over the
test duration and to ensure creep data were in the linear range for each material. Creep
was measured for specimens exposed for 300 hours. Measurements of tensile strength
and molecular weight were obtained for specimens that had been exposed to the full
range of test conditions for 300 hours and the most aggressive condition (80C and
825mV) for up to 1100 hours.
Measurement of molecular weight of the PSU, PB and PA66 required different
techniques. Gel permeation chromatography (GPC) was used to analyze the molecular
weight of PSU. As recommended by the material supplier, the molecular degradation of
PB was determined using melt flow index tests according to ASTM D1238 (ASTM, 2001).
A higher melt flow index indicates a lower molecular weight. Changes in molecular weight
of PA66 were based on the measurement of relative viscosity (adaptation of ASTM D789
(ASTM, 2004)). A lower viscosity indicates material degradation. Scanning electron
microscopy was used to observe the morphology of the unexposed and exposed polymer
surfaces after 300 hours exposure.
For the relatively short duration of these experiments, both PSU and PB were stable
in hot chlorinated water. After 300 hours the measured creep compliance of PSU in air at
room temperature ranged from 0.2 to 0.7GPa1, comparable to the manufacturers
reported compliance of approximately 0.4GPa1. The creep compliance in all water
environments ranged from 0.3 to 1.2GPa1. The data for air and water are indistinguishable
based on the experimental uncertainty ( 0.4GPa1). The average tensile strength of all
specimens tested (air and hot water environments) is 77 2MPa, comparable to the
manufacturers reported value of 70MPa (Solvay Advanced Polymers, 2002). Strain at
failure is relatively low (57 per cent) for exposed and unexposed specimens. The GPC
analysis and SEM images support the stable nature of this polymer. The Mw of unexposed
specimens ranged from 57,000 g/mole to 67,000 g/mole. The Mw of the exposed
specimens ranged from 44,000 g/mole to 63,000 g/mole. SEM images taken of exposed
PSU specimens are featureless and provide no evidence of surface degradation.
The creep compliance of PB in air ranged from 0.3 to 0.6GPa1, while the creep
compliance in water ranged from 0.5 to 1.7GPa1 (after 300 hours). Even though
specimens exposed to water show larger creep compliance than in air, the difference is

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

2/7/07

11:51 AM

Page 115

Thermal and Material Characterization of Immersed Heat Exchangers

115

within the experimental uncertainty ( 0.3GPa1). The strength of the unexposed


specimens and those exposed to the potable water environments is 50 20MPa. The
strength data are comparable to the manufacturers reported value of 36.5MPa (Basell
Polyolefins, 2004). Additional tests indicated that the large variation in the tensile strength
can be reduced significantly if the injection flow direction is consistent among specimens
(Freeman, 2004). Strain at failure data were also comparable between unexposed and
exposed specimens, ranging from 2 to 6 per cent for unexposed specimens and from 1.9
to 6.9 per cent for specimens exposed to 830mV, 80C. The melt flow index of the
unexposed specimens and those exposed to chlorinated water for 250 hours were within
the instrument uncertainty: the average melt flow index of the unexposed material is
0.60g/10 min and that of the exposed specimens is 0.63 0.05g/10 min. SEM images of
the unexposed and exposed PB specimens are similar and do not indicate degradation of
the material.
The effect of exposure to hot, chlorinated water on PA66 is dramatic. The loss in
mechanical properties of PA66 is evident in both the creep and tensile strength/strain at
failure data. Creep compliance is plotted in Figure 4.7. The creep compliance of
specimens exposed to water is greater than that in air for all conditions. The creep
compliance in air ranged from 0.2 to 1.0GPa-1 after 160 hours. The creep compliance in
water ranged from 6.5 to 14.5GPa1 (after 230340 hours). The average tensile strength of
the unexposed material is 94 1MPa, comparable to the manufacturer reported strength
of 85MPa (DuPont, 2004). After 24 hours of exposure to any of the hot potable water
conditions, the strength is reduced more than half, to approximately 45MPa, regardless of
temperature or ORP (Figure 4.8). This decrease in tensile strength is attributed primarily
to the absorption of water. Approximately 15 per cent of the strength is not recovered by
drying, indicating some molecular degradation had occurred. Strain at failure for
unexposed specimens ranges from 26 to 30 per cent. For specimens exposed to 830mV,
100

Creep compliance (1/GPa)

ch-04

Air, 24C nominal


570 50 mV, 27 1C
550 20 mV, 60 1C
830 20 mV, 60 1C
540 50 mV, 80 2C
830 20 mV, 80 1C

10

0.1
0.1

10

100

1000

Time (hours)
FIGURE 4.7 Creep compliance of PA66 in air and exposed to hot chlorinated water

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

2/7/07

11:51 AM

116

Page 116

J. H. DAVIDSON, S. C. MANTELL AND L. F. FRANCIS

Tensile strength (MPa) G

100

No exposure
550 20 mV, 60 1C
830 20 mV, 60 1C
830 20 mV, 80 8C

80

uncertainty, stress

60

40

20

uncertainty, strain
250
Strain at failure (%)

ch-04

200
150
100
50
0
0

50

100

150

200

250

300

Time (hours)
FIGURE 4.8 Tensile strength (top) and strain (bottom) of PA66 before and after exposure to hot
chlorinated water

80C conditions, strain at failure varies depending on the exposure time (Figure 4.8).
Considerable elongation occurs for samples aged from 2 to 5 hours (strain at failure during
this period was as high as 200 per cent). After 12 hours exposure the strain at failure
ranges from 35 to 45 per cent. This variation in elongation over time may be associated
with initial water absorption followed by hydrolysis. Although strain at failure is similar for
unexposed and exposed specimens (exposure time greater than 10 hours), the exposed
material has consistently lower strength and reduced stiffness.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

ch-04

2/7/07

11:51 AM

Page 117

Thermal and Material Characterization of Immersed Heat Exchangers

117

There is no difference in the relative viscosity of unexposed and exposed PA66


specimens: the average relative viscosities of the unexposed PA66 specimens and those
specimens exposed to 80C water with ORP levels of 540 and 800mV for 240 hours are
50.4 and 50.3 0.6, respectively. Degradation of the surface may not be detected
because of the large volume of bulk material required for relative viscosity tests. SEM
images of the PA66, however, show significant surface degradation. The surface of the
unexposed PA66 is smooth with no visible defects (Figure 4.9a). Surface cracks, ranging
from 1m to 10m in length, are visible on the surface of all specimens exposed to water
(Figure 4.9b). Although the physical breakdown of polymer chains is not indicated by the
relative viscosity, these images, along with the creep and tensile test data, provide
convincing evidence of material degradation.

(a)

(b)

FIGURE 4.9 SEM images of PA66: (a) native unexposed surface and (b) after 292 hours exposure at 830
40mV, 59 2C. The surface of the unexposed PA66 is smooth with no visible defects. Surface cracks, ranging
from 1m to 10m in length, are visible on the surface of all specimens exposed to water

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

ch-04

2/7/07

118

11:51 AM

Page 118

J. H. DAVIDSON, S. C. MANTELL AND L. F. FRANCIS

4.3.3 SUMMARY
These results highlight the importance of designing heat exchanger components based
on data from specimens that have been exposed to an aqueous, oxidative environment.
The challenge is to elicit degradation within the test duration. It is difficult to accelerate
the test sufficiently using a traditional ORP control approach.
The data presented here show that for up to 1100 hours exposure in hot potable
chlorinated water, PSU and PB (with antioxidant additives) maintain their mechanical
properties, while PA66 degrades significantly. These test results are consistent with
the expected degradation mechanisms for each polymer. PSU has a rigid backbone
with strong chemical bonds and was expected to be resistant to oxidation. PSU tubing
is commercially available in limited sizes. Manufacturing process development is
required to cost-effectively produce small diameter, thin-walled PSU tubes required for
heat exchanger applications. The PB included an antioxidant package, such that
degradation of the polymer is delayed until the antioxidant is depleted. The combination
of test environments and durations studied were not sufficient to deplete the
antioxidant. Further investigation of the relationship between depletion of the
antioxidant additives in polyolefins is recommended to determine the expected life of
these materials. Polyamide 6,6 is not recommended for this application as test results
indicate substantial degradation of the material in relatively short times. The polymer
readily absorbs water and the polymer chain includes an amide group which can be
hydrolyzed.

4.4 SCALING OF CANDIDATE POLYMERS


4.4.1 BACKGROUND
Scaling is the formation of a hard mineral deposit on a surface that is in contact with
untreated water containing dissolved minerals (in other words hard water). Scale deposits
are composed mostly of calcium carbonate, and this type of scale is the focus here.
Scaling is pervasive on heated surfaces because the solubility of calcium carbonate
decreases with increasing temperature and the rates of the deposition processes are
enhanced at elevated temperature. In solar collectors (Vliet and Baker, 1998; Baker, 2000;
Baker and Vliet, 2001; Baker and Vliet, 2003) and heat exchangers (Epstein, 1986; Marner
and Suitor, 1987) scaling can be a serious problem. Scale may reduce heat transfer rates
due to the additional conductive resistance across the heat exchanger wall. In addition,
the scale build-up increases pressure drop across the heat exchanger. In extreme cases,
the deposit may completely block the flow of water. Until recently, literature reports have
focused on scaling of metal tubes (see, for example, Branch and Muller-Steinhagen, 1991;
Khan et al, 1996; Budair et al, 1998). With the increased use of polymer tubes, more
attention has been paid to scaling of polymers. The focus of this section is scaling of
candidate polymer tubes and more broadly the formation of calcium carbonate on
polymer surfaces.
Scaling of a heated surface involves several steps, as described more completely
elsewhere (Cowan and Weintritt, 1976; Hasson, 1981; Knudsen, 1981; Ferguson, 1984;
Bott, 1988; Karabelas, 2002). First, the surface (in this case a polymer) comes into contact

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

ch-04

2/7/07

11:51 AM

Page 119

Thermal and Material Characterization of Immersed Heat Exchangers

119

with water that is supersaturated with respect to calcium carbonate. The supersaturation (S)
is commonly given by:

(22)
where aCa2+ and aCO32 are the activities of calcium and carbonate ions, respectively and Ksp
is the solubility product for crystalline CaCO3 (e.g. calcite) (Snoeyink and Jenkins, 1980;
Stumm and Morgan, 1996). When S is greater than 1, the solution is supersaturated and
the formation of solid calcium carbonate is possible. A higher value of S indicates a
greater thermodynamic driving force for scaling. To calculate S, several chemical reactions
that control species concentrations must be taken into account along with the
temperature, which influences the reaction rates as well as the solubility product. Next,
solid calcium carbonate forms on the polymer surface by nucleation and growth
(Nancollas, 1979; Mullin, 1993). During the initial nucleation stage, the chemical and
physical properties of the polymer surface, as well as the supersaturation and the
temperature, are expected to play a role. During growth, ions from the water are
transported to the surface and the calcium carbonate scale layer grows. Growth can be
controlled by transport to the surface or reaction at the surface. When transport controls,
the flow rate (e.g., Reynolds number) enters in as another important variable in scaling, as
flow impacts the mass transfer to the surface. Lastly, the scale can be removed,
depending on the flow conditions and the adhesion of the scale to the tube. Hence at
least four variables impact the scaling process: the composition of the water, the
temperature, the surface composition and structure, and the flow condition.

4.4.2 SCALING OF POLYMERS


Various configurations have been used to study scaling of polymers. Experiments have
been performed with 1) flow through heated tubes, 2) isothermal flow through tubes, and
3) in stagnant water under isothermal conditions. Some researchers have focused solely
on the behavior of polymers while others have included polymers as one of several
materials under study.
Very few studies have tackled scaling of polymer tubes using conditions that simulate
the application of a polymer in a heat exchanger or solar absorber. An early exploration of
heat exchangers with Teflon tubing explored scaling using a boiler test and data on the
thermal performance of the heat exchanger (Githens et al, 1965). Results of the boiler test
suggested that the smooth surface of the Teflon reduced adherence of the scale under
the agitation from boiling. The scale also had less of an impact on the heat transfer
performance of the Teflon-based heat exchanger compared with that of the one
containing metal tubes; this result is due in part to the similarity in the thermal
conductivity between the polymer and the scale.
More recently, Wang and co-workers at the University of Minnesota (Wang et al, 2005)
studied scaling of a variety of candidate polymer tubes in a tube-in-shell heat exchanger.
In this research commercial tubes of candidate polymers were studied using conditions

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

ch-04

2/7/07

120

11:51 AM

Page 120

J. H. DAVIDSON, S. C. MANTELL AND L. F. FRANCIS

representative of solar water heating applications. Tubes were oriented vertically in a


tube-in-shell geometry with 60C shell side temperature. Tap water with its composition
adjusted by the addition of calcium nitrate, sodium bicarbonate and potassium hydroxide
was used as the inlet water. The inlet water was heated as it traveled down the length of
the 0.89m tubes. While the diameters of the tubes varied (see Table 4.1), the average flow
rate was kept constant and all flow was in the laminar regime. Since the experiment was
carried out for a fixed time, different quantities of water passed through each tube. There
were two phases of the study: (i) a pretreatment phase with a lower supersaturation inlet
water (S = 2, [Ca2+]t = 2 x 103 M, pH = 8) and (ii) an accelerated phase with a higher
supersaturation inlet water (S = 8, [Ca2+]t = 4 x 103 M, pH = 9). Scale deposition was
monitored periodically by removing sections of the tubes and characterizing the deposit
by chemical analysis, scanning electron microscopy (SEM) and X-ray diffraction (XRD).
All materials in the study formed a calcium-rich scale on their surfaces. Figure 4.10
shows representative SEM micrographs of the scale layer deposited on several of the
materials investigated in the study. A layer of new material of a distinctly different
morphology from the native polymer is apparent. Small particles (~100nm) appear in the
layer. Chemical analysis data showed that the scale contained calcium with minor
quantities of iron and phosphorus. However, the only crystalline phases found by XRD
were CaCO3 polymorphs (calcite and aragonite), indicating that there may be some
amorphous iron- and phosphorus-containing phases present. In addition, the scale
formed on Teflon and a high temperature nylon (HTN) was amorphous or nano-crystalline.
Table 4.1 provides a summary of the final chemical analysis results of the study.
Scale accumulated on all polymer tubes in quantities and at rates that were
comparable to those on copper, with two exceptions more scale was found on
polyamide 6,6 and less on HTN. The greater amount on PA66 was linked to the moisture
absorption by the polymer, which is an order of magnitude more than the other polymers
in the study, and to hydrolysis, which creates anionic groups capable of binding Ca cations

TABLE 4.1 Scale accumulation and scaling rate for tubes in a tube-in-shell heat exchanger1 as
determined by chemical analysis
MATERIAL

Polyamide 6,6, PA66


High Temperature
Nylon, HTN
Polybutylene, PB
Polypropylene, PP
Teflon
Copper

EXTERNAL
DIAMETER
(mm)
3.25

WALL
THICKNESS
(mm)
0.41

VOLUME
OF WATER2 (L)
1158

SCALE
ACCUMULATED
CaCO3 (g /m2)
1.275

5.89
2.54
2.18
1.52
4.57

6.6 x 107

2.19

2535

0.082

8.3 x 109

0.76
1.07
1.48
1.78

478
303
327
143

0.172
0.202
0.173
0.395

8.9 x 108
1.0 x 107
8.9 x 108
2.0 x 107

Scale accumulation and rate calculated assuming that all calcium was present as CaCO3.

Volume of water passed through each tube in 540 hours during accelerated scaling phase.

Source: Wang et al (2005)

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

SCALING RATE
(g /m2s)

ch-04

2/7/07

11:51 AM

Page 121

Thermal and Material Characterization of Immersed Heat Exchangers

121

Source: Wang et al (2005)

FIGURE 4.10 SEM micrographs of scale formed on (a) polyamide 6,6, (b) polypropylene, (c) polybutylene
and (d) copper tubes in a tube-in-shell heat exchanger. SEM micrographs were taken at the end of 540 hours of
accelerated scale testing. A layer of new material of a distinctly different morphology from the native polymer
is apparent

(Aizenberg et al, 1999). The lesser scaling rate for the HTN was attributed to the tubes
larger wall thickness, which prevented the interior surface temperature from reaching as
high a temperature as the other materials. For this water chemistry, the temperature had
little effect on the supersaturation, but the lower temperature in the case of HTN resulted
in slower kinetics of deposition. These conclusions were consistent with a surfacereaction limited model for scaling that took into account differences in temperatures due
to the differing wall thicknesses and thermal conductivities of the materials. From these
experimental and modeling results, several of the candidate polymers for solar thermal
applications (polypropylene (PP), PB, HTN and Teflon) had very similar scaling tendencies.
Scaling studies have also been carried out using flow-through systems under
isothermal conditions. Andritsos and co-workers (Andritsos et al, 1996 and 1997) studied
scale formation at room temperature under turbulent flow conditions. Their experimental
system used a syringe pump to boost the pH, and hence the supersaturation, in-line. This
design allowed them to achieve high supersaturation without the risk of precipitation of
calcium carbonate particles in the bulk liquid. Their results showed a marked increase in
scale deposition rate with temperature and flow velocity, consistent with work by Hasson

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

ch-04

2/7/07

122

11:51 AM

Page 122

J. H. DAVIDSON, S. C. MANTELL AND L. F. FRANCIS

et al (1968). Interestingly, the scaling rate was found to undergo a dramatic increase at a
critical value of supersaturation (S = 8). The authors speculate that below the critical
supersaturation the formation of scale is by a surface-controlled mechanism and above
the critical value a diffusion-controlled mechanism dominates. Likewise, the scale
accumulation with time depended on the supersaturation; at low S, the scale
accumulated slowly in the early stage, with an apparent induction period, but at high S,
the scale quantity increased linearly with time. The researchers compared the
morphologies of the scale deposits on steel, copper and Teflon and found that the tube
material had little effect. The tenacity of the scale was less on Teflon, however.
Using a similar isothermal flow-through system, Sanft et al (2006) carried out
accelerated scaling experiments at room temperature on tubes of cross-linked
polyethylene (PEX), a polypropylene copolymer (PP-r) and copper. Two tests were run
under identical conditions using distilled water with its composition adjusted (S = 7.8,
[Ca2+]t = 3 x 103 M, pH = 9). The accumulation of scale was slow at early times and then
increased, similar to the results of Andritsos et al (1996). There was large variability in the
data, however, which was attributed to the randomness of the nucleation process. Within
this limitation, scaling tendencies of the materials in the study were similar. SEM
micrographs revealed isolated calcite and vaterite particles, as shown in Figure 4.11.
Coverage is incomplete, indicative of the early nucleation-dominated stage of scaling.
Another approach is to expose polymers to supersaturated water without flow. In one
such study, Roques and Girou (1974) noted a connection between the type of material used
for the walls of their reactor, which contained supersaturated water, and the formation of
calcium carbonate on the walls. The induction time for this heterogeneous nucleation on
the walls was decreased in the following order: poly(vinyl chloride), poly(methyl
methacylate), glass and stainless steel. In another study, nucleation times were compared

Source: Sanft et al (2006)

FIGURE 4.11 SEM micrographs of scale formed on PEX tubes after (a) 5 hours and (b) 7.5 hours of
exposure to flowing hard water at room temperature. After exposure, isolated calcite and vaterite particles are
visible. Coverage is incomplete, indicative of the early nucleation-dominated stage of scaling

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

ch-04

2/7/07

11:51 AM

Page 123

Thermal and Material Characterization of Immersed Heat Exchangers

123

for various substrate materials that were immersed in hard water kept at constant
temperature (Ben Amor et al, 2004). The nucleation time was determined by monitoring the
pH and the calcium ion content of the water. The researchers observed that under low
hardness conditions the nucleation times were much different, decreasing from polyamide
to polyvinyl chloride to chrome to steel, consistent with Roques and Girou. However, when
harder water was used, the nucleation times were all shorter and roughly the same. This
result concurs with evidence from flow-through experiments that the nature of the material
has the greatest impact when scaling conditions are not severe and nucleation dominates.
A series of reports document the formation of calcium carbonate on the surfaces of
fine polymer particles circulating in supersaturated water under isothermal conditions
(Dalas, et al, 1988, 1999 and 2000; Kanakis and Dalas, 2000; Dousi et al, 2003). In these
studies, polymer particles are placed in a constant concentration reactor in which a drop
in pH associated with calcium carbonate formation triggers the addition of reactant
solutions, allowing for quantitative analysis of deposition. The results of studies of
calcium carbonate formation on sulfonated polystyrene and polystyrene divinylbenzene
(Dalas et al, 1988), carboxylated poly(vinyl chloride) copolymer (Dalas et al, 1999) and
cellulose (Dalas et al, 2000) polymers parallel earlier work on calcium carbonate formation
on calcite seeds (Nancollas and Reddy, 1971; Reddy and Nancollas, 1971; Kazmierczak
and Tomson, 1982), supporting the finding that the induction time decreases and the
calcium carbonate formation (nucleation and growth) rate increases as the
supersaturation of the water is increased. These studies generally point to a surfacecontrolled or surface diffusion-controlled mechanism and reveal some of the effects of
polymer chemistry. An important chemical factor is the ability of the polymer to bind
calcium ions. For example, calcium ions can bind to the carboxylate (C=O) functional
group (Dousi et al, 2003). With this binding ability, the local supersaturation at the surface
is high and so the induction time drops. This research on polymer particles gives good
guidance on the selection of candidate polymers for solar water heating systems:
polymers lacking active functional groups, such as PEX, PP-r, and PP, have the greatest
promise in terms of minimizing scaling.
In recent work at the University of Minnesota, Wang (2004) and McGill (2005) studied
the formation of calcium carbonate on polymers in isothermal cells containing hard water.
They immersed thin sheets of the candidate polymers, including those studied by Wang et
al (2005), as well as polysulfone and polyphenylsulfone, in hard water. Calcium carbonate
deposition was quantified by chemical analysis of the deposit. Conditions were chosen to
study nucleation, and hence the randomness of the process led to some variability in the
data. These researchers found the scaling tendencies of the candidate polymers to be the
same within the experimental error. Experiments were conducted at room temperature and
approximately 50C using water that was supersaturated to approximately the same degree
at the two temperatures. A significantly greater amount of scale formed during the elevated
temperature test, showing the importance of kinetics in enhancing the rate of deposition.

4.4.3 SUMMARY
While progress has been made in recent years, more research is needed to understand
and control the scaling of polymers. Results to date indicate all surfaces are prone to

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

ch-04

2/7/07

124

11:51 AM

Page 124

J. H. DAVIDSON, S. C. MANTELL AND L. F. FRANCIS

scale, but the scaling rate may be different if the conditions ensure that nucleation or
surface reaction control the deposition. Scale formation on polymers is linked in part to
the chemical structure of the material. For example, PA66 appears to form scale more
readily due to its susceptibility to water absorption and hydrolysis (Wang et al, 2005). Yet
unknown is the importance of the physical structure, including surface roughness, which
has been shown to influence the scaling of metals (Keysar et al, 1994), and stress
concentration, which is documented as trouble causing in calcification of biomedical
polymers in the body (Banas and Baier, 2000). In general, scaling is enhanced for water
with higher supersaturation. Scaling rate also increases with temperature, an effect due,
to some extent, to an increase in supersaturation and perhaps to a greater extent to the
acceleration of the kinetics of deposition process. The flow conditions become critical
when the scale grows by ion transport to the surface scaling rate increases with flow
rate. This phenomenon is well documented for turbulent flow (Hasson et al, 1968), but
less so for laminar flow, which is characteristic of flow through narrow polymer tubes
such as might be used in heat exchangers. The shear removal of scale from polymer
tubes has not been investigated extensively. Here, the difference in scale adherence,
which will likely be related to the chemical and physical structure of the polymer, may be
important. Lastly, the field of scaling research on polymers as well as metals is in need of
innovation in characterization tools. Some progress has been made on this front with
several reports of new in situ methods: microscopic imaging (Kim et al, 2002), rotating
disc electrode monitoring (for metals) (Chen et al, 2005), attenuated total reflectance
infrared sensing (Smith et al, 2004), quartz crystal microbalance monitoring (Kohler et al,
2001), optical fiber sensing (Lyons et al, 2001), and monitoring of laser reflectance and
scattering (Euvrard et al, 2004; Sanft, 2005).

4.5 CONCLUSION
Thermal characterization of heat exchangers immersed in solar water storage vessels for
domestic hot water systems has been the subject of study for nearly 20 years. Yet
predicting thermal performance remains a challenge because the transient natural
convection heat transfer coefficients for such heat exchangers depend on the geometry
of the heat exchanger/tank combination, the state of charge and the extent of thermal
stratification. Based on recent work summarized in this paper, we recommend equation
(19) to predict the transient natural convection Nusselt number during either charge or
discharge. The advantage of this transient formulation is that only the initial Rayleigh
number and an empirical constant C are required. The empirical constant must be
determined for each heat exchanger/tank combination. The correlation is valid for fully
mixed storage tanks. In most cases, charge or discharge of the tank with an immersed
heat exchanger maintains a fully mixed tank and is likely to destroy existing stratification.
In situations where stratification of the storage fluid is maintained by the use of manifolds
or other devices, equation (19) is not applicable. In this case, correlations of the form of
equation (4) may be used and it will be necessary to characterize the temperature of the
storage fluid surrounding the heat exchanger during use. Either a validated transient zonal
model of the heat exchanger/tank such as that used in TRNSYS or computational fluid
dynamics models will be needed to predict the temperature distribution in the tank.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

ch-04

2/7/07

11:51 AM

Page 125

Thermal and Material Characterization of Immersed Heat Exchangers

125

As manufacturers and system integrators consider the use of polymer materials for
immersed heat exchangers, durability and reliability of the heat exchanger must be
considered. A critical aspect is the long-term mechanical stability of materials in water at
elevated temperature. In addition, chlorinated water may create an oxidative environment in
which certain polymers are chemically attacked, resulting in permanent loss of mechanical
strength and increase in creep compliance. Water absorption and hydrolysis can also
diminish polymer mechanical properties. The mechanism of polymer degradation will depend
on the polymer molecular structure. Based on testing of polysulfone, (PSU), polybutylene (PB)
and polyamide 6,6 (PA66), we do not recommend the use of polyamide 6,6 for this
application. Data for polybutylene is inconclusive. Although the material did not degrade after
1100 hours of exposure to hot, chlorinated water, PB is expected to fail once the antioxidant
additives have been depleted. Further investigation of the relationship between depletion of
the antioxidant additives in polyolefins is recommended to determine the expected life of
these materials. Polysulfone is a promising material as it is stable in the intended environment
and does not require the use of antioxidants.
All heat exchangers operated in an open loop with water supersaturated with respect
to calcium carbonate or other inverse soluble salts are prone to scale. Measurement of
scaling in a number of candidate polymer tubes indicates the scaling rate for polymers is
of the same order of magnitude as that for copper. Additional research is needed to
assess scale adhesion on polymers and to develop polymers that are resistant to scale.

ACKNOWLEDGEMENTS
The authors appreciate the collaboration of Dr. Jay Burch at the US National Renewable Energy
Laboratory, Professor F. A. Kulacki and graduate students at the University of Minnesota, and
Professor Kelly Homan at the University of Missouri-Rolla. We gratefully acknowledge the
financial support of the National Renewable Energy Laboratory, the US Department of Energy
and the University of Minnesota Initiative for Renewable Energy and the Environment.

AUTHOR CONTACT DETAILS


Dr. Jane H. Davidson (corresponding author), Department of Mechanical Engineering, University of Minnesota,
111 Church St., S.E., Minneapolis, MN 55455
e-mail: david020@umn.edu; jhd@me.umn.edu
Dr. Susan C. Mantell, Department of Mechanical Engineering, University of Minnesota, 111 Church St., S.E.,
Minneapolis, MN 55455
e-mail: smantell@me.umn.edu
Dr. Lorraine F. Francis, Department of Chemical Engineering and Materials Science, University of Minnesota,
421 Washington Ave., S.E., Minneapolis, MN 55455

REFERENCES
Aharoni, S. M. (1997) n-Nylons: Their Synthesis, Structure and Properties, John Wiley & Sons, New York
Aizenberg, J., Black, A. J. and Whitesides, G. M. (1999) Oriented growth of calcite controlled by self-assembled monolayers
on functionalized alkanethiols supported on gold and silver, J. Am. Chem. Soc., vol 121, pp45004509
Andritsos, N., Kontopoulou, M., Karabelas, A. J. and Koutsoukos, P. G. (1996) Calcium carbonate deposit formation under
isothermal conditions, Can. J. Chem. Eng., vol 74, p911

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

ch-04

2/7/07

11:51 AM

126

Page 126

J. H. DAVIDSON, S. C. MANTELL AND L. F. FRANCIS

Andritsos, N., Karabelas, A. J. and Koutsoukos, P. G. (1997) Morphology and structure of CaCO3 scale layers formed under
isothermal flow conditions, Langmuir, vol 13, pp28732879
ASTM (2000) F 2023-00: Standard test method for evaluating oxidative resistance of crosslinked polyethylene (PEX) tubing
and systems to hot chlorinated water, in Annual Book of ASTM Standards, American Society of Testing and Materials,
West Conshohocken, PA, US
ASTM (2001) D 1238: Standard test method for melt flow rates of thermoplastics by extrusion plastometer, in Annual Book
of ASTM Standards, American Society of Testing and Materials, West Conshohocken, PA, US
ASTM (2004) D789-04: Standard test method for determination of relative viscosity of polyamide (PA), in Annual Book of
ASTM Standards, American Society of Testing and Materials, West Conshohocken, PA, US
Baker, D. K. (2000) An investigation of calcium carbonate scaling rates based on experiments and modeling, Ph.D. Thesis in
Mechanical Engineering, University of Texas, Austin, US
Baker, D. K. and Vliet, G. C. (2001) Designing solar hot water systems for scaling environments, Journal of Solar Energy
Engineering, Transactions of the ASME, vol 123, no 1, pp4347
Baker, D. and Vliet, G. (2003) Identifying and reducing scaling problems in solar hot water systems, Journal of Solar Energy
Engineering, Transactions of the ASME, vol 125, pp6166
Banas, M. D. and Baier, R. E. (2000) Accelerated mineralization of prosthetic heart valves, Mol. Cryst. and Liq. Crystl., vol
354, pp249267
Basell Polyolefins (2004) Product Data Sheet: Polybutylene-1, PB 4137 Grey
Bejan, A. (1982) Entropy Generation Through Heat and Fluid Flow, John Wiley & Sons, New York, p165
Ben Amor, M., Zgolli, D., Tlili, M. M. and Manzola, A. S. (2004) Influence of water hardness, substrate nature and
temperature on heterogeneous calcium carbonate nucleation, Desalination, vol 166, pp7984
Bradley, S. W., El-Hibri, J., Bersted, B. H. and Bradley, W. L. (2000) A study of the effect of chlorinated water on engineering
thermoplastics at elevated temperatures, presented at ANTEC, Orlando, US
Branch, C. A. and Muller-Steinhagen, H. M. (1991) Influence of scaling on the performance of shell-and-tube heat
exchangers, Heat Transfer Engineering, vol 12, pp3745
Boies, A. M. and Homan, K. O. (2004) Improving discharge characteristics of indirect integral collector storage systems with
multi-element storage, ASME International Solar Energy Conference, Portland, OR, US, June 2004, CDROM
Boies, A. M. and Homan, K. O. (2005) System performance of an indirect collector storage preheater in series with a
conventional electric resistance water heater, 2005 ISES Solar World Congress, Orlando, FL, US, August 2005, ASME,
paper ISEC2005-76200, CDROM
Boies, A. M., Homan, K. O., Davidson, J. H. and Liu, W. (2005) A variable effectiveness model for indirect thermal storage
devices, 2005 ASME Summer Heat Transfer Conference, San Francisco, CA, US, July 2005, ASME, paper HT2005-72711,
CDROM
Bott, T. R. (1988) Crystallisation fouling Basic science and models, Fouling Science and Technology, vol 145, pp251260
Budair, M. O., Khan, M. S., Zubair, S. M., Sheikh, A. K. and Quddus, A. (1998) CaCO3 scaling in AISI 316 stainless steel
tubes Effect of thermal and hydraulic parameters on the induction time and growth rate, Heat & Mass Transfer, vol 34,
pp163170
Chen, T., Neville, A., and Yuan, M. (2005) Calcium carbonate scale formation Assessing the initial stages of precipitation
and deposition, J. Petroleum Science and Engineering, vol 46, pp185194
Cowan, J. C. and Weintritt, D. J. (1976) Water-Formed Scale Deposits, Gulf Publishing Co., Houston, TX, US
Dalas, E., Kallitsis, J. and Koutsoukos, P. G. (1988) The crystallization of calcium carbonate on polymeric substrates,
J. Crystal Growth, vol 89, pp287294
Dalas, E., Klepetsanis, P. and Koutsoukos, P. G. (1999) The overgrowth of calcium carbonate on poly(vinyl chloride)-co-vinyl
acetate-co-maleic acid, Langmuir, vol 15, pp8322327

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

ch-04

2/7/07

11:51 AM

Page 127

Thermal and Material Characterization of Immersed Heat Exchangers

127

Dalas, E., Klepetsanis, P. G. and Koutsoukos, P. G. (2000) Calcium carbonate deposition on cellulose, J. Colloid. Inter. Sci.,
vol 224, pp5662
Davidson, J. H., Mantell, S. C. and Jorgensen, G. (2002) Status of the development of polymeric solar water heating
systems, in D.Y. Goswami (ed) Advances in Solar Energy, American Solar Energy Society, vol 15, pp149186
Davidson, J. H., Ragoonanan, V. and Homan, K. (2005) Divided storage in an indirect integral collector storage with immersed
heat exchanger, Proceedings of the 2005 Solar World Congress, Orlando, FL, US
Dousi, E., Kallitsis, J., Chrissanthopoulos, A., Mangood, A. H. and Dalas, E. (2003) Calcite overgrowth on carboxylated
polymers, Journal of Crystal Growth, vol 253, pp496503
Drck, H. (2002) Influence of different combistore concepts on the overall system performance, IEA-SHC Task 26, Industry
Workshop, Oslo, Norway, pp3946
DuPont (2004) Product Data Sheet: Zytel 103 HSL NC010
Epstein, N. (1986), Fouling of heat exchangers, in J. W. Palen (ed.) Heat Exchanger Source Book, Hemisphere Publishing
Corporation, New York, pp677697
Euvrard, M., Membrey, F., Filiatre, C. and Foissy, A. (2004) Crystallization of calcium carbonate at a solid/liquid interface
examined by reflectance of a laser beam, Journal of Crystal Growth, vol 265, pp322330
Farrington, R. B. and Bingham, C. E. (1986) Testing and analysis of immersed heat exchangers, SERI/TR-253-2866, National
Renewable Energy Laboratory, Golden, Colorado, US
Farrington, R. B. and Bingham, C. E. (1987) Testing and analysis of load-side immersed heat exchangers for solar domestic
hot water systems, SERI/TR-254-3094, National Renewable Energy Laboratory, Golden, Colorado, US
Feiereison, T. J., Klein, S. A., Duffie, J. A. and Beckman, W. A. (1982) Heat transfer from immersed coils, Paper No 82
WA/SOL-18, American Society of Mechanical Engineers, New York
Ferguson, R. J. (1984) A kinetic model for calcium carbonate deposition, Materials Performance, vol 23, pp2534
Freeman, A. (2004) Mechanical performance of polysulfone, polybutylene, and nylon 6,6 in hot, chlorinated water, MS
Thesis in Mechanical Engineering, University of Minnesota, Minneapolis, US
Freeman, A., Mantell, S. C. and Davidson, J. H. (2005) Mechanical performance of polysulfone, polybutylene, and polyamide
6,6 in hot chlorinated water, Solar Energy, vol 79, pp624637.
Gedde, U. W., Viebke, J., Leijstrom, H. and Ifwarson, M. (1994) Long-term properties of hot-water polyolefin pipes A
review, Polymer Engineering and Science, vol 34, no 24, pp17731787
Githens, R. E., Minor, W. R. and Tomsic, V. J. (1965) Flexible tube heat exchangers, Chem. Eng. Prog., vol 61, no 7, pp5562
Hasson, D. (1981); Precipitation fouling: A review, in E. F. C. Somerscals and J. G. Knudsen (eds) Fouling of Heat Transfer
Equipment, Hemisphere Publishing Corporation, New York, pp527568
Hasson, D., Avriel, M., Resnick, W., Rozenman T. and Windreich, S. (1968) Mechanism of calcium carbonate scale deposition
on heat-transfer surfaces, I&EC Fundamentals, vol 7, no 1, pp5965
Kanakis, J. and Dalas, E. (2000) The crystallization of vaterite on fiber, Journal of Crystal Growth, vol 219, pp277282
Karabelas, A. J. (2002) Scale formation in tubular heat exchangers Research priorities, International Journal of Thermal
Science, vol 41, pp682692
Kazmierczak, T. F. and Tomson, M. B. (1982) Crystal growth of calcium carbonate: A controlled composition kinetic study,
J. Phys. Chem., vol 86, pp103107
Keysar, S., Semiat, R., Hasson, D. M. and Yahalom, J. (1994) Effect of surface roughness on the morphology of calcite
crystallizing on mild steel, J. Colloid and Interface Sci., vol 162, p311
Khan, M. S., Zubair, S., Budair, M., Sheikh A. and Quddus, A. (1996) Fouling resistance model for prediction of CaCO3 scaling
in AISI 316 tubes, Heat and Mass Transfer, vol 32, pp12
Kim, W. T., Bai, C. and Cho, Y. I. (2002) A study of CaCO3 fouling with a microscopic imaging technique, International Journal
of Heat and Mass Transfer, vol 45, pp597607

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

ch-04

2/7/07

11:51 AM

128

Page 128

J. H. DAVIDSON, S. C. MANTELL AND L. F. FRANCIS

Knudsen, J. G. (1981) Cooling water fouling A brief review, American Society of Mechanical Engineers, Heat Transfer
Division, vol 17, pp2938
Kohler, N., Courbin, G. and Ropital, F. (2001) Static and dynamic evaluation of calcium carbonate scale formation and
inhibition, European Formation Damage Control Conference, Expanding Horizons, Society of Petroleum Engineers, The
Hague, The Netherlands
Liu, W., Davidson, J. H., and Mantell, S. C. (2000) Analysis of polymer heat exchangers for solar water heating: A case
study, Journal of Solar Energy Engineering, Transactions of the ASME, vol 122, no 2, pp8491
Liu, W., Davidson, J. H., Mantell, S. C. and Kulacki, F. A. (2003) Natural convection from a horizontal tube heat exchanger
immersed in a tilted enclosure, Journal of Solar Energy Engineering, Transactions of the ASME, vol 125, no 1, pp6775
Liu, W., Davidson, J. H. and Kulacki, F. A. (2004) Natural convection from a tube bundle in a thin inclined enclosure, Journal
of Solar Energy Engineering, Transactions of the ASME, vol 126, no 2, pp702709
Liu, W., Davidson, J. H., and Kulacki, F. A. (2005) Thermal characterization of prototypical integral collector storage systems
with immersed heat exchangers, Journal of Solar Energy Engineering, Transactions of the ASME, vol 127, no 1, pp2128
Lyons, W. B., Ewald, H., Flanagan, C., Lochmann, S. and Lewis, E. (2001) A neural networks based approach for determining
fouling of multipoint optical fibre sensors in water systems, Meas. Sci. Technol., vol 12, pp958965
Mather, D., Hollands, K. G. T. and Wright, J. L. (2002) Single- and multi-tank energy storage for solar heating systems:
fundamentals, Solar Energy, vol 73, no 1, pp313
Marner, W. J. and Suitor, J. W. (1987) Fouling with Convective Heat Transfer, John Wiley & Sons, New York
McGill, K. N. (2005) Calcium carbonate scaling of polymers in isothermal, stagnant water, Mechanical Engineering,
University of Minnesota, Minneapolis, MN, US
Morgan, V. T. (1975) The overall convective heat transfer from smooth circular cylinders, Advances in Heat Transfer, vol 11,
pp199264
Mullin, J. W. (1993) Crystallization, Butterworth-Heinemann, Oxford, UK
Nancollas, G. H. (1979) The growth of crystals in solution, Advances in Colloid and Interface Science, vol 10, pp215252
Nancollas, G. H. and Reddy, M. M. (1971) The crystallization of calcium carbonate: II Calcite growth mechanism, J. Colloid
and Interface Sci., vol 36, pp824
NSF (1999) NSF Protocol P171: Chlorine resistance of plastic piping materials, National Sanitary Foundation International,
Ann Arbor, MI, US
Reddy, M. M. and Nancollas, G. H. (1971) The crystallization of calcium carbonate: I. Isotopic exchanger and kinetics,
J. Colloid and Interface Science, vol 36, p166
Reindl, D. T. (1992) Source driven transient natural convection in enclosures, Ph.D. Thesis in Mechanical Engineering,
University of Wisconsin, Madison, US
Reindl, D. T., Beckman, W. A. and Mitchell, J. W. (1992a) Transient natural convection from a vertical flat plate in a rectangular
enclosure, Proceedings, ASME 28th National Heat Transfer Conference and Exhibition, San Diego, CA, US, pp9198
Reindl, D. T., Beckman, W. A. and Mitchell, J. W. (1992b) Transient natural convection in enclosures with application to solar
thermal storage tanks, Journal of Solar Energy Engineering, Transactions of the ASME, vol 114, pp175181
Roques, H. and Girou, A. (1974) Kinetics of the formation conditions of carbonate tartars, Water Research, vol 8, p907
Sanft, P. M. (2005) Development of an apparatus for in situ analysis of calcium carbonate formation on polymer surfaces,
MS Thesis in Chemical Engineering and Materials Science, University of Minnesota, Minneapolis, US
Sanft, P., Francis, L. F. and Davidson, J. H. (2006) Calcium carbonate formation on cross-linked polyethylene (PEX) and
polypropylene random copolymer (PP-r), Journal of Solar Energy Engineering, Transactions of the ASME, vol 128, no 2,
pp251254
Scott, G. (1999) Polymers and the Environment, The Royal Society of Chemistry, Cambridge, UK

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

ch-04

2/7/07

11:51 AM

Page 129

Thermal and Material Characterization of Immersed Heat Exchangers

129

Sekulic, D. P. and Krane, R. J. (1992a) Use of multiple storage elements to improve the second law efficiency of a thermal
storage system. Part I: Analysis of the storage process, in A. Valero and G. Tsatsaronis (eds) ECOS 92 International
Symposium on Efficiency, Cost, Optimization, ASME, New York, pp6166
Sekulic, D. P. and Krane, R. J. (1992b) Use of multiple storage elements to improve the second law efficiency of a thermal
storage system. Part II: Completion of the analysis and presentation of results, in A. Valero and G. Tsatsaronis (eds) ECOS
92 International Symposium on Efficiency, Cost, Optimization, ASME, New York, pp6772
Smith, J. K., Yuan, M., Lopez, T. H., Means, M. and Przybylinski, J. L. (2004) Real time and in-situ detection of calcium
carbonate scale in a west Texas oil field, Society of Petroleum Engineers Production and Facilities, vol 19, pp9499
Snoeyink, V. L. and Jenkins, D. (1980) Water Chemistry, John Wiley & Sons, New York
Solvay Advanced Polymers (2002) Product Data Sheet: P-1700 NT11, NT06 & CL2611, Alpharetta, GA, US
Stumm, W. and Morgan, J. J. (1996) Aquatic Chemistry: Chemical Equilibria and Rates in Natural Waters, John Wiley & Sons,
New York
Su, Y. and Davidson, J. H. (2005) Natural convective flow and heat transfer in a collector storage with an immersed heat
exchanger: Numerical study, Journal of Solar Energy Engineering, Transactions of ASME, vol 127, no 3, pp324332
Vliet, G. C. and Baker, D. K. (1998) Designing solar hot water systems for scaling environments, Proceedings of the 1998
American Solar Energy Society Annual Conference, Albuquerque, NM, US, pp307317
Walter, D., Mantell, S. C. and Davidson, J. (2003) Performance of polymer tubes in hot chlorinated water, SOLAR 2003:
Conference Proceedings, American Solar Energy Society, pp7782
Wang, Y. (2004) Experimental study of calcium carbonate scaling on polymer tubes and films, MS Thesis in Chemical
Engineering and Materials Science, University of Minnesota, Minneapolis, US
Wang, Y., Davidson, J. H. and Francis, L. F. (2005) Scaling in polymer tubes and interpretation for use in solar water heating
systems, Journal of Solar Energy Engineering, Transactions of the ASME, vol 127, pp314
Weiss, W. (2003) Solar Heating Systems for Houses: A Design Handbook for Solar Combisystems, International Energy
Agency, London and James and James, London
Wu, C., Mantell, S. C. and Davidson, J. H. (2004) Polymers for domestic solar hot water: Long-term performance of PB and
nylon 6,6 in hot water, Journal of Solar Energy Engineering, Transactions of the ASME, vol 126, no 1, pp581586

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 99129

ch-05

2/7/07

2:25 PM

Page 130

Photocatalytic Detoxification
of Water with Solar Energy
Sixto Malato, Julin Blanco, Diego C. Alarcn, Manuel
I. Maldonado, Pilar Fernndez-Ibez and Wolfgang Gernjak

Abstract
During the past ten years there has been tremendous research and development in the
area of photocatalysis. One of the major applications of this technology is the
degradation of organic pollutants in water, by what are called advanced oxidation
processes (AOPs). This paper reviews the use of sunlight to produce OH radicals for
AOPs. The pilot plant-scale systems necessary for performing solar photocatalytic
experiments, including the basic pilot plant components and the fundamental
parameters related to solar photocatalytic reactions, are described. The paper also
summarizes most of the recent research related to solar photocatalytic degradation of
water contaminants, and how it could significantly contribute to the treatment of very
persistent toxic compounds. It further describes the possibility of using the solarilluminated photo-Fenton reaction to extend the horizon of solar photocatalysis
applications. Various solar reactors for photocatalytic water treatment based mainly on
non-concentrating collectors erected during the last few years are also described in
detail in the last part of this review.

Keywords photocatalysis; non-concentrating collectors; reaction rate; mineralization; electrons;


degradation; collector efficiency

5.1 INTRODUCTION
The main causes of surface and groundwater contamination are industrial discharges (even
in small amounts), excessive use of pesticides, fertilizers (agro-chemicals) and domestic
waste landfills. Wastewater treatment is based on various mechanical, biological, physical
and chemical processes. In fact, it is a combination of many operations, such as filtration,
flocculation, sterilization or chemical oxidation of organic pollutants. After filtration and
elimination of particles in suspension, the ideal process is biological treatment (natural
decontamination). Unfortunately, some organic pollutants, classified as bio-recalcitrant, are
not biodegradable. Advanced oxidation processes (AOPs) may be used for

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 131

Photocatalytic Detoxification of Water with Solar Energy

131

decontamination of water containing these compounds (Andreozzi et al, 1999; Gogate and
Pandit, 2004a; Legrini et al, 1993; Pera-Titus et al, 2004). These methods rely on the
formation of highly reactive chemical species which degrade even the most recalcitrant
molecules into biodegradable compounds. Although there are different reacting systems,
all of them are characterized by the same chemical feature: production of OH radicals
(OH), which are able to oxidize and mineralize almost any organic molecule, yielding CO2
and inorganic ions. Rate constants (kOH, r = kOH [OH] C) for most reactions involving
hydroxyl radicals in aqueous solution are usually in the order of 106 to 109M1s1.
They are also characterized by their not very selective attack, which is a useful
attribute for wastewater treatment and solution of pollution problems. The versatility of
AOPs is also enhanced by the fact that there are different ways of producing hydroxyl
radicals, facilitating compliance with the specific treatment requirements. Methods based
on UV, H2O2/UV, O3/UV and H2O2/O3/UV combinations use photolysis of H2O2 and ozone to
produce the hydroxyl radicals. Other methods, like heterogeneous photocatalysis and
homogeneous photo-Fenton, are based on the use of a wide band-gap semiconductor
and the addition of H2O2 to Fe2+ salts, respectively, and irradiation with UV-VIS light. Since
1976, photocatalytic detoxification has been discussed in the literature as an alternative
method for cleaning up polluted water (Carey et al, 1976). Todays renewed interest
(Pignatello, 1992) in the photo-assisted Fenton process, discovered by Fenton in the 19th
century, is underlined by a significant number of studies devoted to wastewater treatment
(Gogate and Pandit, 2004b). Both these processes are of special interest since sunlight
can be used for them. The main disadvantage of AOPs is their high cost (expensive
reactants such as H2O2 and UV generation). Future applications of these processes could
therefore be improved through the use of catalysis and solar energy.
The heterogeneous solar photocatalytic detoxification process consists of making use
of the near-ultraviolet (UV) band of the solar spectrum (wavelength shorter than 380nm)
to photo-excite a semiconductor catalyst in the presence of oxygen (Goswami and Blake,
1996). Under these circumstances, oxidizing species, either bound hydroxyl radicals (OH)
or free holes, which attack oxidizable contaminants, are generated producing a
progressive break-up of molecules yielding CO2, H2O and diluted inorganic acids. The
most commonly used catalyst is the semiconductor TiO2, which is cheap, non-toxic and
abundant. The homogeneous solar photocatalytic detoxification process (photo-Fenton) is
based on the production of OH radicals by Fenton reagent (H2O2 added to Fe2+ salts). The
rate of degradation of organic pollutants with Fenton-like reagents is strongly accelerated
by irradiation with UV-VIS light. This is an extension of the Fenton process which can make
use of UV-VIS light irradiation at wavelengths over 300nm. Under these conditions, the
photolysis of Fe3+ complexes enables Fe2+ regeneration and Fenton reactions due to the
presence of H2O2 (Bauer et al, 1999; Sagawe et al, 2001).
Although these processes have been studied for at least two decades,
industrial/commercial applications, engineering systems and engineering design
methodologies have only been developed recently (Blanco and Malato, 2003; Goswami
et al, 1997 and 2003). This paper summarizes such work done during the last decade and
attempts to continue previous work published in Advances in Solar Energy on this subject
(Goswami, 1995).

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 132

132 S. MALATO et al

5.2 SOLAR COLLECTORS FOR PHOTOCHEMISTRY


5.2.1 CONCENTRATING COLLECTORS
Contrary to solar thermal processes, which collect large amounts of photons at any
wavelength to reach a specific temperature range, solar photochemical processes use
only high-energy short wavelength photons to promote photochemical reactions. Most
solar photochemical processes use UV or near-UV sunlight (300 to 400nm), but in some
photochemical synthesis processes, sunlight up to 500nm can be absorbed and photoFenton heterogeneous photocatalysis uses sunlight up to 580nm. Sunlight at wavelengths
over 600nm is normally not useful in any photochemical process. Nevertheless, the
specific hardware needed for solar photochemical applications has much in common with
those used for thermal applications. As a result, both photochemical systems and
reactors have followed conventional solar thermal collector designs, such as parabolic
troughs and non-concentrating collectors (Goswami, 1997). At this point, their design
begins to diverge, since in photochemical processes: (i) the fluid must be directly exposed
to the solar radiation and, therefore, the absorber must be transparent to the photons, and
(ii) temperature usually does not play a significant role, so no insulation is required.
The original solar photo-reactor designs (Goswami, 1995) for photochemical
applications were based on line-focus parabolic trough concentrators (PTCs). In part, this
was a logical extension of the historical emphasis on trough units for solar thermal
applications. Furthermore, PTC technology was relatively mature and existing hardware
could be easily modified for photochemical processes. The parabolic trough collector
consists of a structure that supports a reflective concentrating parabolic surface. This
structure has one or two motors controlled by a solar tracking system on one or two axes
respectively which keep the collector aperture plane perpendicular to the solar rays. In this
situation, all the solar radiation available on the aperture plane is reflected and
concentrated on the absorber tube that is located at the geometric focal line of the
parabolic trough. The first outdoor engineering-scale reactor developed (in the US) was a
converted solar thermal parabolic trough collector in which the absorber/glazing tube
combination had been replaced by a simple Pyrex glass tube through which contaminated
water could flow. It was designed and built jointly by the Solar Energy Research Institute
(SERI), now known as the National Renewable Energy Laboratory (NREL), and Sandia
National Laboratories at the end of the 1980s. This system was installed at the National
Solar Thermal Test Facility at the Sandia National Laboratories in Albuquerque, New
Mexico (US) in 1989 (Pacheco et al, 1990; Blake, 1994). The facility was made up of a total
of six aligned parabolic trough collectors with single-axis solar tracking, with an aperture
of 2.13m and a length of 36.4m, for a total of 465m2 aperture area. The collector
concentrated the sunlight about 50 times on the photo-reactor (Anderson et al, 1991;
Pacheco and Tyner, 1990; Prairie et al, 1992).
Immediately afterwards, in the 1990s, a similar facility was designed and built at the
Plataforma Solar de Almera. The first engineering-scale solar photochemical facility for
water detoxification in Europe was developed by Centro de Investigaciones
Medioambientales, Energticas y Tecnolgicas (CIEMAT) (Minero et al, 1993), using 12
two-axis solar tracking PTCs, each having a total of 32 mirrors in 4 parallel parabolas with
a collecting area of 32m2. Typical overall optical efficiencies obtained in PTCs were around

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 133

Photocatalytic Detoxification of Water with Solar Energy

133

50 per cent (Curc et al, 1996), mainly due to:


the solar tracking error, attributable to the control system (s);
the module construction (C);
the reflectivity (which is a function of the wavelength) of the aluminized surface of

the parabolic mirrors (R,). As this surface is outdoors it gets dirty and can be
damaged, requiring that the reflectivity should be measured periodically; and
the spectral transmissivity of the absorber tube (T,).
Figure 5.1 shows the path of direct radiation (ID) until it arrives inside the absorber tube.
It must reach the surface and be reflected (part is lost due to R,) in the right direction
(here affected by s) by the real mirror surface (C), before penetrating (T,) in the tube.
Furthermore, the parabolic trough concentration factor must also be considered (ratio of
surface area of the parabola capturing the radiation and surface area of the tube,
Sp/ST). Therefore the effective photon flux corresponding to the direct UV inside the
absorber (ID,E) is:

I D ,E = f I D ,

Sp
Sr

,c ,R ,Tl

(1)

Global Radiation is also collected by the PTCs, but only that small fraction which falls
directy on the transparent absorber tube without intervention of the collector and is only
affected by the transmissivity of the glass, T,:

I G E = f I G ,T

(2)

where IG,E is the effective photon flux corresponding to global UV inside the absorber and
therefore the total radiation reaching a PTC is the sum of both components.

ID
IG
ID
IG

ID
ID,E
IG,E

FIGURE 5.1 The various loss factors () affecting the photon flux (I) inside a PTC photo-reactor

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 134

134 S. MALATO et al

Both these facilities, which were followed by others, were based on parabolic trough
collectors with hundreds of square meters of collector surface and can be considered the
starting point of solar photocatalytic technology development (Goswami, 1995). Parabolic
trough collectors make efficient use of direct solar radiation and, as an additional
advantage, the thermal energy collected from the concentrated radiation could
simultaneously be used for other applications. The reactor is small, while receiving a large
amount of energy per volume unit. The flow is turbulent and volatile compounds do not
evaporate, making handling and control of the liquid to be treated simple and inexpensive.
The main disadvantages are that the collectors (i) use only direct radiation, (ii) are
expensive and (iii) have low optical and quantum efficiencies (at least for TiO2 applications)
(Alfano et al, 2000).

5.2.2 NON-CONCENTRATING COLLECTORS


One-sun (non-concentrating) collectors are, in principle, cheaper than PTCs, as they have
no moving parts or solar tracking devices (Wyness et al, 1994a and 1994b). They do not
concentrate radiation, so efficiency is not reduced by factors associated with
concentration and solar tracking. Manufacturing costs are cheaper because their
components are simpler, which also means easy, low-cost maintenance. Nonconcentrating collector support structures are also easier and cheaper to install and the
surface required for their installation is smaller because, since they are static, they do not
project shadows on the others. They are able to utilize the diffuse as well as the direct
portion of the solar UV-A. It should be noted that with AM (air mass) at 1.5, the diffuse and
direct portion of the solar radiation reaching the surface of the earth are almost the same.
This means that a light concentrating system, for example a PTC, can in principle only
employ half of the solar radiation available in this particular spectral region. An extensive
effort in the design of small non-tracking collectors has resulted in the testing of several
different non-concentrating solar reactors:
Free-falling film. The process fluid falls slowly over a tilted plate which faces the

sun and is open to the atmosphere (Goslich et al, 1997a), with a catalyst supported in
a permanent matrix on the reactor surface (Guillard et al, 2003) or suspended or
dissolved in the fluid (photo-Fenton applications) (Gernjak et al, 2003). Two large pilotscale fixed-bed photocatalytic reactors were tested in Australia (Feitz et al, 2000).
One (6.5 x 0.5m, total volume 200L) was based on a TiO2-coated (5g m2) woven
glass fiber mesh supported on an inclined corrugated support. The other was based
on a packed bed configuration (1 x 2m2) containing TiO2-coated raschig rings (6.5cm
deep). The packed bed configuration processed a similar concentration of phenol
seven times faster under similar solar light conditions.
Pressurized flat plate. This consists of two plates between which the fluid
circulates using a separating wall (Dillert et al, 1999a). A kind of this nonconcentrating reactor is the double skin sheet reactor (DSSR), which consists of a flat
and transparent structured box made of Plexiglas (Well et al, 1997), with channels
and walls for water circulation. The structure of the reactor is schematically drawn in
Figure 5.2. The suspension containing the model pollutant and the photocatalyst is

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 135

Photocatalytic Detoxification of Water with Solar Energy

135

pumped through these channels. The Plexiglas used to manufacture the double skin
sheets has high transmittance in the UV-A portion of the solar spectrum below
400nm. A similar system (based on a double skin acrylic panel) was developed by the
Solar Energy and Energy Conversion Laboratory (University of Florida) for the
treatment of BTEX-contaminated groundwater (Goswami, 1995). A detailed
description and discussion of the experiments can be found in Srinivasan et al (1997
and 1998).
Solar ponds. These are small, shallow on-site pond reactors (Bedford et al, 1994;
Gimnez et al, 1999). The flat reactor resides in a cylindrical tank in which several
perforated tubes lie at the bottom, through which air circulates, bubbling the solution
to be treated. The system operates in a discontinuous mode and the catalyst is
maintained in suspension by using the air stream. Air also supplies the oxygen
necessary for the pollutants oxidation.
Based on all of the above, the main advantages and disadvantages of the different
technologies for solar photocatalytic applications can be summarized.

FIGURE 5.2 Structural schematic view of a DSSR reactor showing water flowing in the transparent box.
The channels are shown connected in series. The channels may also be connected in parallel to reduce
pressure drop

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 136

136 S. MALATO et al

Advantages of non-concentrating systems:


1 they can make use of both direct and diffuse solar radiation;
2 they are simpler systems, with lower investment cost and fewer maintenance
requirements;
3 water may not heat up significantly if no special measures are taken to that effect; and
4 they have both high optical efficiency and high quantum efficiency since there is a
lower recombination of e/h+, given that the photonic density is lower than in a
concentrating system.
Disadvantages of non-concentrating systems:
1 construction problems associated with weather resistance, chemical inertness and
ultraviolet transmission;
2 they usually work with laminar flow, which can cause mass transfer problems in
photocatalysis;
3 reactants vaporization; and
4 reactants contamination.
Advantages of concentrating systems:
1 they have a noticeably smaller reactor tube area (a shorter loop in which to confine,
control and handle the water);
2 more practical use of a supported catalyst;
3 turbulent flow (favors mass transfer and avoids possible catalyst sedimentation
problems); and
4 there is no evaporation of volatile compounds.
Disadvantages of concentrating systems:
1 they can only make use of direct solar radiation;
2 usually expensive;
3 lower efficiency, both optical and quantum, the second arising from a higher
recombination of e/h+ than in non-concentrating systems; and
4 there can be problems with water overheating.

5.2.3 COMPOUND PARABOLIC CONCENTRATORS (CPCs)


The non-concentrating devices referred to above are of the flat plate type, even when tubes
are directly used, since they are just laid side by side. But as mentioned, there is a category
of low concentration collectors called Compound Parabolic Concentrators (CPCs), which
are used in thermal applications both with fins in non-evacuated concentrators and with
tubes in vacuum collectors. These are an interesting option for thermal applications since
they combine characteristics of parabolic concentrators and static flat systems. They
concentrate solar radiation but they conserve the properties of the flat plate collectors,

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 137

Photocatalytic Detoxification of Water with Solar Energy

137

being static and collecting diffuse radiation. Thus they also constitute a good option for
solar photochemical applications (May et al, 1991). CPCs are static collectors with a
reflective surface designed to be ideal in the sense of non-imaging optics and can be
designed for any given reactor shape. When the absorber has another shape a more
precise way of describing them would be CPC-type collectors. CPCs were invented in the
1960s (Welford and Winston, 1978) to achieve solar concentration with static devices
(Collares-Pereira, 1995), since they were able to concentrate on the receiver all the radiation
that arrives within the collectors angle of acceptance. They do this by illuminating the
complete perimeter of the receiver, rather than just the front of it, as in conventional flat
plates or in the case of tubes laid side by side. These concentrating devices have ideal
optics, thus maintaining both the advantages of the PTCs and static systems. The
concentration factor (CCPC) of a two dimensional CPC collector is given by Equation 3:

CCPC =

1
sin a

(3)

The normal values for the semi-angle of acceptance (a) for photocatalytic applications are
going to be between 60 and 90 degrees. This wide angle of acceptance allows the receiver
to collect both direct and a large part of the diffuse light (1/C of it), with the additional
advantage of decreasing errors of both the reflective surface and receiver tube alignment,
which become important for achieving a low cost photo-reactor. A special case is that in
which a = 90, whereby CCPC = 1 (non-concentrating solar system) and each CPC curve is an
ordinary involute (Figure 5.3a). When this occurs, all the UV radiation that reaches the aperture
area of the CPC (direct and diffuse) can be collected and redirected to the reactor. If the CPC
is designed for an acceptance angle of +90 to 90, all incident solar diffuse radiation can be
collected. The light reflected by the CPC is distributed all around the tubular receiver (Figure
5.3b) so that almost the entire circumference of the receiver tube is illuminated and the light
incident on the photo-reactor is the same as would impinge on a flat plate. The advantages
of CPCs can be summarized as follows. Just as in the parabolic trough collector, in CPC-type
collectors, the photo-reactor is tubular so that the water can be pumped easily. CPC devices
for photocatalytic applications are generally fabricated with aluminum reflectors and the
structure is usually made of a simple frame which, in turn, forms the support for connecting
the glass tubes that make up the photo-reactor (Romero et al, 1999a; Blanco et al, 2000).
CPCs have the advantage of both technologies (PTCs and non-concentrating collectors) and
none of the disadvantages, apparently making them the best option for photocatalytic
processes based on the use of solar radiation (Ajona and Vidal, 2000; Blanco et al, 2000;
Jubran et al, 1999; Malato et al, 2002a and 2003a; Robert et al, 1999). Advantages include:
1 they can make highly efficient use of both direct and diffuse solar radiation without
the need for solar tracking (Blanco et al, 1999);
2 there is no evaporation of possible volatile compounds and water does not heat up;
3 they have high optical efficiency, since they make use of almost all the available
radiation, and high quantum efficiency, as they do not receive a concentrated flow of
photons; and
4 flow is turbulent inside the tube reactor.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 138

138 S. MALATO et al

FIGURE 5.3 Schematic drawing of CPCs

5.3 FUNDAMENTAL PARAMETERS IN SOLAR PHOTOCATALYSIS


5.3.1 INITIAL CONTAMINANT CONCENTRATION INFLUENCE
As oxidation proceeds, less and less of the surface of the TiO2 particles is covered as the
contaminant is decomposed. Evidently, at total decomposition, the rate of degradation is
zero and a decreased photocatalytic rate is to be expected with increasing illumination
time. Most authors agree that, with minor variations, the expression for the rate of photomineralization of organic substrates with irradiated TiO2 follows the LangmuirHinshelwood (L-H) law for the same saturation-type kinetic behavior in any of the four
possible situations:
1 the reaction takes place between two adsorbed substances;
2 the reaction occurs between a radical in the solution and the adsorbed substrate;
3 the reaction takes place between the radical linked to the surface and the substrate
in the solution; or
4 the reaction occurs with both species in solution.
In all cases, the expression is similar to the L-H model. From kinetic studies only, it is
therefore not possible to find out whether the process takes place on surfaces or in
solution. Although the L-H isotherm has been useful in modeling the process, it is
generally agreed that both rate constants and orders are only apparent (Cunningham and

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 139

Photocatalytic Detoxification of Water with Solar Energy

139

Sedlak, 1996; Minero, 1999). They serve to describe the rate of degradation, and may be
used for reactor optimization, but they have no physical meaning and may not be used to
identify surface processes. Thus while not a useful tool for describing the active species
involved in oxidation, engineers and solar designers seem to have a common
understanding of the usefulness of the unmodified L-H model.
For L-H standard data treatment it is therefore assumed that the reaction occurs on the
surface, which is also the assumption most widely accepted as possible. Under these
conditions, two extreme situations are defined to illustrate the adsorption on the catalyst
surface: (1) substrate and water compete for the active catalyst sites, and (2) the reactant and
the solvent are adsorbed on the surface without competing for the same active catalyst sites.
According to the L-H model, the reaction rate (r) is proportional to the fraction of surface
covered by the substrate (x). In each case the following expression can be obtained:

r=-

kr KC
dC
= krx =
dt
1 + KC + K sCs

(4)

k KC
dC
= kr x = r
dt
1 + KC

(5)

r=-

where kr is the reaction rate constant, K is the reactant adsorption constant, C is the
concentration at any time, KS is the solvent adsorption constant and CS is its concentration
(in water CS 55.5 M). As CS >> C and CS remains practically constant, the part of the
catalyst covered by water is unalterable over the whole range of C and the previous
equations can be integrated.
Using an L-H model, graphics similar to those depicted in Figure 5.4 may be obtained
from the experimental data and from the linearization of the previous equations. The
effect of the initial concentration on the degradation rate is shown in Figure 5.4; due to
the saturation produced on the semiconductor surface as the concentration of the
reactant increases, it reaches a point at which the rate becomes steady.
It should be emphasized that photo-decomposition gives rise to intermediates, which
could also be adsorbed competitively on the surface of the catalyst. The concentration of
these intermediates varies throughout the reaction up to their mineralization and thus
Equation 6 may also take the following form:

r0

r01

t1/2

FIGURE 5.4 Graphics related to the adjustment of data to an L-H type kinetic model

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

2/7/07

2:25 PM

Page 140

140 S. MALATO et al

r=

kr KC

(6)

1 + KC + K iCi (i = 1, n )
i =1

where i is the number of intermediates formed during degradation (the solvent is also
included in the summation).
An understanding of the reaction rates and how the reaction rate is influenced by
different parameters is important for the design and optimization of an industrial system.
The L-H reaction rate constants are useful for comparing the reaction rate under different
experimental conditions. Once the reaction constants kr and K have been evaluated, the
disappearance of reactant can be estimated if all other factors are held constant. Due to
this, a series of tests at different initial substrate concentrations must be performed to
demonstrate whether the experimental results could be adjusted with this model. The
concentration range must be wide enough to allow correct fit of the L-H linearization. This
means from the lowest concentration at which the initial rate could be determined until
the limit where the relationship between initial reaction and initial concentration remains
constant (see Figure 5.4).
Since hydroxyl radicals react non-selectively, numerous intermediates are formed en
route to complete mineralization at different concentrations. Because of this, all tests
should be carried out using TOC (total organic carbon) as a crucial parameter (instead of
concentration of parent compound, C), because the photocatalytic treatment must
destroy not only the initial contaminant, but any other organic compound as well. The
results shown in Figure 5.5 are examples of the experiments carried out with mixtures of
different commercial pesticides. It is possible to see that mineralization, once begun,
maintains the same slope until at least 6070 per cent of the initial TOC has been
degraded.

L1

ch-05

Q, kJ L1
FIGURE 5.5 Pesticide decomposition at different initial concentrations. Maximum rate as a function of
TOCmax is shown in the inserted graphic

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 141

Photocatalytic Detoxification of Water with Solar Energy

141

As the reaction is not expected to follow simple models like first- or zero-order
kinetics, overall reaction rate constants cannot be calculated. The complexity of the
results, of course, is caused by the fact that the TOC is a sum parameter often including
many products which undergo manifold reactions. One parameter is proposed in order to
obtain a practical point of comparison for various experiments: the maximum gradient of
the degradation curve, which is the slope of the tangent at the inflection point (rQ,0). This
has the unit of a zero-order rate constant (mg/kJ instead of mg/min) and therefore appears
to be easy to handle. Furthermore this gradient can be roughly considered as the initial
rate of the mineralization reaction, because it is preceded by a period of nearly constant
TOC level. The parameter rQ,0 is referred to as maximum rate. In the graphic insert in
Figure 5.5, it may be observed that the initial rate is steady from 2030mg of TOC per liter.
At this concentration, saturation occurs and the reaction rate becomes constant.
Once the optimum initial concentration is known, a model for predicting plant behavior
is necessary. This model must allow calculation of the area of solar collectors required for
treating water contaminated with different amounts of pesticides. As commented before,
the L-H model is not a perfect explanation of the mechanism of the photocatalytic process,
but the behavior of the reaction rate versus reactant concentration can very often be
adjusted to a mathematical expression with it. In the present case, instead of using the LH model (r = kKC/(1+KC)) directly, the use of an alternative model is preferred for fitting
experimental data in large solar photocatalytic plants, using an approximate kinetic solution
of the general photocatalytic kinetic system, which has the analytical form of an L-H
equation. With these considerations, the rate of TOC disappearance is given by Equation 7
(analogous to the L-H model but without its original significance).

rQ ,0 =

1 [TOC ]max
2 + 3 [TOC ]max

(7)

The experimental results shown in Figure 5.5 have been used to calculate the constants
(i). By inversion of Equation 7 these constants can be calculated from the intercept and
the slope of the line of fit (Equation 8), which is shown in the inset in Figure 5.5.

1 3 2
1
=
+
; 3 = 1.67mg -1 kJ; 2 = 5 .07kJ L-1
1
rQ ,0 1 1 [TOC ]max 1

(8)

Using these values, experimental results and the corresponding lines of fit are shown in
Figure 5.6. The lines of fit were drawn with Equation 9, using the constants reported
previously.

[TOC ]max
1
2 ln
+ 3 [TOC ]max - [TOC ]
1
[TOC ]

) = Q

UV

(9)

The experimental results agree reasonably well with the model proposed and the constants
calculated. This equation allows TOC degradation to be predicted as a function of initial TOC

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

2/7/07

2:25 PM

Page 142

142 S. MALATO et al

L1

ch-05

Q, kJ L1
FIGURE 5.6 Application of the proposed kinetic model for mineralization of a pesticide mixture. The inset
shows the fit of Equation 9

and available radiation, and the reverse, incident energy on the reactor necessary to reach a
specific degree of mineralization. Therefore, useful design equations may be obtained with
an L-H type model, in spite of not fitting the heterogeneous photocatalytic reaction
mechanism. For now, these equations must be obtained at pilot plant size; however, they
will be useful for larger plants if the same type of collector is used.

5.3.2 RADIANT FLUX


Since 1990 there has been a clarification of the kind of solar technology which should be
involved in detoxification (Dillert et al, 1999a; Alfano et al, 2000; Mukherje and Ray, 1999).
The question posed was whether it is necessary to concentrate radiation for photocatalysis
and if a non-concentrating collector can be as efficient as concentrating ones. Initially, it
was thought that non-concentrating collectors were the ideal alternative and, in fact, the
first large pilot plants operated with them. However, their high cost and the fact that they
can only operate with direct solar radiation (which implies their location in highly insolated
areas) led researchers to consider the alternative of static non-concentrating collectors.
The reason for using one-sun systems for water treatment is firmly based on two factors:
(1) the high percentage of UV photons in the diffuse component of solar radiation (see
Figure 5.7), and (2) the low order dependence of reaction rates on light intensity.
As commented before, the first non-concentrating collectors for detoxification were
tested a few years after the first tests with parabolic troughs. Early engineering-scale tests
with solar systems were based on parabolic troughs with concentration factors in the
range of 10X, but immediately a second generation of non-concentrating collectors
started being tested on thin films, flat plates and CPCs. Figure 5.8 presents degradation
tests comparing a CPC and a two-axis parabolic trough. Both systems were operated in
parallel at Plataforma Solar de Almera (Spain), at the same catalyst concentration
(200mg L1 Degussa P-25), at the same dichloroacetic acid initial concentration (5mM) and

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 143

Photocatalytic Detoxification of Water with Solar Energy

143

FIGURE 5.7 Average UV direct and UV global (direct + diffuse) irradiance for each month of the year at
Plataforma Solar de Almera, Spain (5 years average)

FIGURE 5.8 Overall dichloroacetic acid degradation rate (as TOC disappearance reaction rate) comparison
for a concentrating (PTC) and a non-concentrating (CPC) collector system. The performance ratio (CPC/PTC)
and the mean value of global UV-radiation during the experimental periods are also shown

performing one test near the 15th of each month during one year. The tests were
performed on perfect, sunny days to permit the parabolic trough to operate all day from
sunrise to sunset. The ratio encountered between both systems working in analogous
conditions for degradation is higher than 5 in favor of the CPC, and this ratio would be
even higher taking cloudy periods into account.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 144

144 S. MALATO et al

It has been experimentally shown that above a certain flux of UV photons, the reaction
rate changes from one to a half-order dependence on intensity. Most authors believe that
the transition from r = f (I1.0) to r = f (I0.5) occurs due to the excess of photogenerated
species (e, h+ and OH). A very simple explanation could be the following (based on the
first stages of the process). The first stages considered are: (i) formation of electron/hole
pairs (Equation 10), (ii) recombination of the pairs (Equation 11), and (iii) oxidation of a
reactant R (Equation 12).
(10)
(11)

(12)

From these reactions, the concentration of holes is:

(13)

where I is the intensity of incident radiation. If it is considered that [e] [h], then in
stationary state:
(14)

When I is very high, a large number of holes and electrons are generated and therefore
kR[h]2 ko[h]R:
(15)
As the reaction rate depends on the amount of hydroxyl radicals present, and these are
generated in the holes, then r0.5 when I is high. Under these conditions, the quantum
yield diminishes because of the high rate of recombination of e/h+ pairs formed. In the
same manner, when I is small, the inverse is true, kR[h]2 ko[h]R:
(16)
At higher radiation intensities, another transition from r = f (I0.5) to r = f (I0) is produced.
At this moment, mass transfer limits the photocatalytic reaction, therefore the rate
is constant although the radiation increases. There may be several factors that limit
the mass transfer, such as the lack of electron scavengers (O2), organic molecules in the
proximity of the TiO2 surface and/or excess of products occupying active centers of the

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 145

Photocatalytic Detoxification of Water with Solar Energy

145

catalyst. These phenomena appear more frequently when working with a supported
catalyst (Pozzo et al, 1997) and/or at a low agitation level. This implies a low catalyst
surface in contact with the liquid and smaller turbulence, which diminishes the contact of
reactants with the catalyst and the diffusion of products from the proximity of the catalyst
to the liquid.
These effects may be appreciably attenuated if some reagent that reduces the
importance of the electron/hole recombination is added. When the electrons are trapped,
recombination of e/h+ is impeded. Either way, addition of oxidants can improve the
efficiency of the process at high illumination intensities. Moreover, this type of compound
can increase the quantum yield even at low irradiation levels due to their strong oxidizing
character. The use of inorganic peroxides has been demonstrated to remarkably enhance
the rate of degradation of different organic contaminants because they trap the photogenerated electrons more efficiently than O2 (Chen et al, 1998; Doong and Chang, 1997;
Malato et al, 1998 and 2000a; Poulios et al, 1998). Hydrogen peroxide is the obvious
candidate and it has been tested with a large number of compounds. Also it is a very
commonly used chemical and so is inexpensive. Being an electron acceptor, hydrogen
peroxide can be a beneficial oxidizing agent because it can react with conduction band
electrons to generate hydroxyl radicals which are required for the photo-mineralization of
organic pollutants (Equation 17):
(17)
The effect depends on H2O2 concentration, generally showing in the optimum range, but
that strength usually depends on the H2O2/contaminant molar ratio. An optimal molar ratio
of between 10 and 100 has been found by different authors (Akmehmet and Ynel., 1996;
Bellobono et al, 1994; Harada et al, 1990; Hofstadler et al, 1994; Poulios et al, 1998).
Inhibition could be explained in terms of TiO2 surface modification by H2O2 adsorption,
scavenging of photo-produced holes and reaction with hydroxyl radicals (Equations 18
and 19). This produces additional problems when working with complex mixtures,
because the molar concentration is usually different for each compound in the mixture.
Moreover, for the treatment of real wastewater, the concentration of each product is
unpredictable.
(18)
(19)
Peroxydisulfate can be a beneficial oxidizing agent in photocatalytical detoxification
because SO4 is formed from the oxidant compound by reaction with the semiconductor

). The peroxydisulfate ion accepts an electron and


photogenerated electrons (eCB
dissociates (Equation 20). The sulfate radical proceeds through the reactions shown in
Equations 21 and 22. In addition, the strongly oxidizing SO4 (Eo = 2.6V) can directly
participate in the degradation processes. The improvement in the mineralization reaction
rate, the relatively low consumption of this reactant and the apparent non-relation

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 146

146 S. MALATO et al

between molar ratio S2O82/contaminant and reaction rate shows peroxydisulfate as a good
choice. It must be mentioned here that in many highly toxic wastewaters where
degradation of organic pollutants is the major concern, the increase of salt content by
addition of an inorganic anion to enhance the organic degradation rate may be justified.
(20)
(21)
(22)
Other oxidants have been used in photocatalysis for reducing solar (or artificial) UV
exposure time: ClO3, BrO3, IO4 and HSO5. Nevertheless, these additives are very
expensive compared to hydrogen peroxide and peroxydisulfate, and their application
would dramatically increase treatment cost. Even more importantly, they do not dissociate
into harmless products (Br and I), and hundreds of milligrams per liter of these anions
are undesirable in water.

5.3.3 QUANTUM YIELD


In photochemistry, quantum yield is used to evaluate the results obtained and compare
different experimental conditions. The quantum yield of a photochemical reaction is
defined with regard to the number of reacting molecules and the number of photons
absorbed ( = n Na1). A photocatalytic heterogeneous system is made up of a
suspended solid (TiO2), a gas (O2) in bubbles and/or dissolved, and an aqueous solution of
a multitude of compounds (initial substrate, intermediates, H+, anions). It is very difficult
to determine the amount of photons absorbed by the catalyst from the behavior of the
radiation incident on a suspension such as this. In order to calculate this, if so desired,
one would have to: 1) evaluate the light absorption of a very complex mixture, which,
moreover, changes its composition throughout the reaction, 2) from this basis, determine
the photon flux that arrives at each particle of the catalyst to photo-activate it, and 3)
estimate the photons absorbed and dispersed. And it seems that this, for the moment, is
a difficult undertaking (Serpone et al, 1996). Remember that in heterogeneous catalysis
the reaction rate is usually expressed as a function of the grams of catalyst. In
photocatalysis, it should include the number of active centers as well as the surface area
of the catalyst. But as a consequence of the above comments, the number of active
centers is unknown and the surface of the catalyst exposed to light is undetermined.
This is therefore simplified by considering only the radiation of a certain wavelength
incident inside the reactor for the calculation of Na. The value obtained from this is called
the estimated quantum yield: E. No distinction is made between the photons
corresponding to each wavelength, assuming that all of them have the same effect on the
surface of the catalyst. In all cases, this simplification is accepted as valid by a multitude
of authors and widely used in the literature. Consequently, the reported quantum yields
have sometimes been reported as lower limits not allowing for scattered light. A simple
means of assessing process efficiencies for equal absorption of photons is therefore

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 147

Photocatalytic Detoxification of Water with Solar Energy

147

desirable in heterogeneous photocatalysis. The initial photo-conversion of phenol has


been proposed as the standard process and Degussa P-25 titania as the standard
photocatalyst (Serpone and Emeline, 2002). The choice of phenol was dictated by the
recognition that the molecular structure of phenol is present in many organic pollutants
and, like many of them, is essentially degraded by oxidation rather than reduction.

(23)

In Equation 23, r is called relative photonic efficiency (Serpone and Emeline, 2002). When
the reaction rate for the test substances and phenol (secondary actinometer) are obtained
under identical experimental conditions, there is no need to measure the photon flux. The
use of relative photonic efficiency renders comparison of process efficiencies between
studies carried out in different laboratories or pilot plants possible because r is basically
independent of the fundamental photocatalysis parameters (light intensity, reactor
geometry and TiO2 concentration for a given catalyst). However, it depends on the initial
concentration of substrate and on temperature. In any case, based on initial rates of
degradation, r illustrates only one aspect of photo-degradation and is also useful for
comparing different photocatalyst materials for water treatment purposes. The
mineralization rate (measured by TOC analysis) is also included because efficiencies
based on the disappearance of organic carbon (r ,TOC, Equation 24) provide more practical
information:

r, TOC

rateof disapearanceof substrateTOC


rateof disappearance of TOC from phenol

(24)

5.4 FACTORS AFFECTING SOLAR PHOTOCATALYSIS


5.4.1 INFLUENCE OF OXYGEN
In semiconductor photocatalysis for water purification, the pollutants are usually organic
and, therefore, the overall process can be summarized by Equation 25. Given the reaction
stoichiometry of this equation, there is no photo-mineralization unless O2 is present. The
literature provides a consensus regarding the influence of oxygen (Herrmann, 1999). It is
necessary for complete mineralization and does not seem to be competitive with other
reactives during the adsorption on TiO2 since the places where oxidation takes place are
different from those of reduction. The O2 avoids the recombination of e/h+ and photoactivated oxygen (O2) also reacts directly:
(25)

The concentration of oxygen also affects the reaction rate, which is faster when the partial
pressure of oxygen (pO2) in the atmosphere in contact with the water increases. In any
case, it seems that the difference between using air (pO2 = 0.21atm) or pure oxygen

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 148

148 S. MALATO et al

(pO2 = 1atm) is not drastic. In an industrial plant it would be purely a matter of economy
of design. Injection of pure O2 becomes necessary in once-through plants at low flow
rates. At high flow rates or with recirculation, the addition of oxygen is not always
necessary since the illumination time per pass is short. The water again recovers the
oxygen consumed when it reaches the tank (open to the atmosphere and stirred).

5.4.2 INFLUENCE OF CATALYST CONCENTRATION


Whether in static, slurry or dynamic flow photo-reactors, the initial reaction rates were found
to be directly proportional to catalyst mass. This indicates a truly heterogeneous catalytic
regime. However, above a certain value, the reaction rate levels off and becomes
independent of catalyst mass. This limit depends on the geometry and working conditions
of the photo-reactor and is for a specific amount of TiO2 in which all of the particles (in other
words all the surface) exposed are totally illuminated. When catalyst concentration is very
high, after traveling a certain distance on an optical path, turbidity impedes further
penetration of light in the reactor. In any given application, this optimum catalyst mass must
be found in order to avoid excess catalyst and ensure total absorption of efficient photons.
There are a number of studies in the literature discussing the influence of catalyst
concentration on process efficiency (Cassano and Alfano, 2000). The results are very
different, but from all of them it may be deduced that radiation incident on the reactor and
path length inside the reactor are fundamental in determining the optimum catalyst
concentration. If the lamp is outside, but the path length is several centimeters long
(similar to a reactor illuminated by solar radiation), the appropriate catalyst concentration
is several hundreds of milligrams per liter. In this case, it is very clear that the optimum
rate is attained at lower catalyst concentrations when the photo-reactor diameter is
increased.
As one important factor related to the photo-reactor design is its diameter, it seems
obvious that a uniform flow must be maintained at all times in the reactor, since irregular
flows cause non-uniform residence times that can lower efficiency compared to the ideal
(Blake et al, 1997). In the case of the heterogeneous process with TiO2 in suspension,
sedimentation and deposition of the catalyst along the hydraulic circuit must also be
avoided, so turbulent flow in the reactor is guaranteed (Malato et al, 2003b). Turbulent
flow makes pressure loss an important parameter that can condition design, especially in
the case of an industrial plant with long reactor tube lengths. For these reasons diameters
of less than 2025mm are not feasible. On the other hand, diameters over 5060mm are
considered impractical. Furthermore, every photo-reactor design must guarantee that all
the useful incoming photons are used and do not escape without having intercepted a
particle in the reactor. Although TiO2 suspensions absorb solar photons with less than
390400nm wavelengths, there is also a strong light scattering effect due to particles.
Both effects must be considered in determining the optimum catalyst load as a function
of light path length in the photo-reactor. After many experiments with different photoreactors, the optimum TiO2 concentration obtained with sunlight has been found to be
several hundreds of milligrams per liter (Dillert et al, 1999a; Gimnez et al, 1999; Goslich
et al, 1997a; Guillard et al, 1999; Minero et al, 1996a) and the ideal diameter of the photoreactor for use with sunlight must be in the range of 25 to 50mm.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 149

Photocatalytic Detoxification of Water with Solar Energy

149

Fe3+ (related species and organic complexes) absorbs solar photons as a function of
its absorptivity (Mazellier et al, 1997). This effect must be considered when determining
the optimum load as a function of light path length in the photo-reactor. These
calculations must be experimentally demonstrated because they are strongly affected by
different complexes formed by Iron (III) during the photo-Fenton process. The optimum
concentrations of 0.20.5mM were obtained after many experiments with different photoreactors under sunlight (Fallmann et al, 1999; Gernjak et al, 2003; Malato et al, 2002b).

5.4.3 TEMPERATURE INFLUENCE


Because of photonic activation, photocatalytic systems do not require heating and
operate at room temperature. The true activation energy is nil, while the apparent
activation energy is often very low (a few kJ/mol) in the medium temperature range
(2080C). However, at very low temperatures (400C), activity decreases and activation
energy becomes positive. The decrease in temperature favors reactants adsorption,
which is a spontaneous exothermic phenomenon, but also favors adsorption of the final
reaction products, desorption of which tends to be the rate-limiting step. By contrast, at
high temperatures (>7080C) for various types of photocatalytic reactions, the activity
decreases and the apparent activation energy becomes negative. When temperature
increases above 80C, nearing the boiling point of water, the exothermic adsorption of
reactants is disfavored and this tends to become the rate-limiting step.
In addition to these mechanical effects, other consequences of plant engineering
must be considered. At high temperatures the oxygen concentration in water decreases.
Also materials used at high temperatures must be stable at those temperatures.
Consequently, the optimum temperature is generally between 20 and 80C. The absence
of a need for heating is attractive for photocatalytic reactions carried out in aqueous
media and in particular for environmental purposes (photocatalytic water purification).
In the case of photo-Fenton treatment, reaction rates increase with temperature (Lee
and Yoon, 2004; Sagawe et al, 2001). As a consequence this process seems especially
interesting for wastewater streams leaving an industrial process at elevated
temperatures. Additionally, the application of heat exchangers or insulation can be an
attractive plant design adaptation (Sagawe et al, 2001).

5.5 SOLAR UV PHOTOCATALYTIC DEGRADATION OF CONTAMINANTS


Research activities devoted to environmental protection have developed relatively quickly
as a consequence of the special attention paid to the environment by international social,
political and legislative authorities, leading in some cases to very severe regulations
(Angelakis et al, 1999). The fulfillment of strict quality standards is especially called for in
the case of those toxic substances which affect the biological sphere and prevent
activation of biological degradation processes. These non-biodegradable contaminants
are an accumulative problem with unpredictable consequences for the mid-term future
(Hayo, 1996). The destruction of toxic pollutants such as biologically recalcitrant
compounds must be handled by non-biological technologies. Photocatalysis aims at the
mineralization of the contaminants into carbon dioxide, water and inorganics. Up to now,
practical applications of solar technologies have been studied and developed most

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 150

150 S. MALATO et al

intensively for heterogeneous TiO2 photocatalysis and homogeneous photo-Fenton. In


this context, treatment of industrial wastewater seems to be one of the most promising
fields of application of solar detoxification, despite inefficient production of hydroxyl
radicals and slow kinetics which may limit economic feasibility. There is no general rule
here at all each case is completely different (Malato et al, 2003a). Consequently,
preliminary research is always required to assess potential pollutant treatments and
optimize the best option for any specific problem on nearly a case-by-case basis. Solar
photocatalytic degradation technology might be feasible for the treatment of wastewater
containing hazardous contaminants at medium or low pollutant concentrations when
treatment plants are impossible. The technology is dependent on the energy flux, as is the
associated investment contingent on the collector surface. Reasonable orders of
magnitude for inflow into typical treatment plants would be in the range of from several
dozen up to a few hundred m3 per day (Bekblet et al, 1996; Freudenhammer et al, 1997;
Goswami et al, 1997; Malato et al, 1996, 2000b; Funken et al, 2001).
In general, the types of compounds which have been degraded include alkanes,
haloalkanes, aliphatic alcohols, carboxylic acids, alkenes, aromatics, haloaromatics,
polymers, surfactants, herbicides, pesticides and dyes. Equation 26 generally holds true
for an organic compound of general formula CnHmOp:

(26)
In the case of organic compounds containing halogens, Equation 27 shows how the
corresponding halide is formed:

(27)

Under photocatalytic oxidative conditions, sulfur is recovered as sulfate in sulfurcontaining compounds according to Equation 28:

(28)

In photo-degradation, transformation of the parent organic compound is desirable in order


to eliminate its toxicity and persistence, but the principal objective is to mineralize all
pollutants. The effectiveness of degradation is not demonstrated only because the entire
initial compound is decomposed. Furthermore, the stoichiometry proposed for the
general reactions (Equations 2628) has to be demonstrated in each case by a correct
mass balance. Reactants and products might be lost (evaporation, adsorption on reactor
components), which introduces uncertainty in the results. The mineralization rate is
determined by monitoring inorganic compounds, such as CO2, Cl, SO42, NO3 and PO43.
When organics decompose, a stoichiometric increase in the concentration of inorganic
anions is produced in the water treated and there is very often an increase in the

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 151

Photocatalytic Detoxification of Water with Solar Energy

151

concentration of hydrogen ions (decrease in pH). For this reason, the analysis of these two
products of the reaction is of interest for the final mass balance. However, the decrease
in pH is not a very reliable parameter of this balance, except in some cases, because it is
influenced by other processes which take place in the medium: the effect of the TiO2
suspension, the formation of CO2 and intermediates.
The oxidation of carbon atoms into CO2 is relatively easy. It is, however, in general,
markedly slower than the de-aromatization of the molecule. Until now, the absence of
total mineralization has been observed only in the case of s-triazine herbicides, for which
the final product obtained was essentially 1,2,5-triazine-2,4,6, trihydroxy (cyanuric acid),
which is, fortunately, not toxic (Minero et al, 1996b). This is due to the strong stability of
the triazine nucleus, which resists most methods of oxidation. For chlorinated molecules,
Cl ions are easily released into the solution. Nitrogen-containing molecules are
mineralized into NH4+ and mostly NO3. Ammonium ions are relatively stable and the
proportion depends mainly on the initial oxidation degree of nitrogen and on the
irradiation time. Organo-phosphorous pesticides produce phosphate ions. However,
phosphate ions in the pH range used remain adsorbed on TiO2. This strong adsorption
partially inhibits the reaction rate, though it remains acceptable. Until now, the analyses of
aliphatic fragments resulting from the degradation of the aromatic ring have only revealed
formate and acetate ions. Other aliphatics (presumably acids, diacids and hydroxylated
compounds) are very difficult to separate from water and to analyze. Formate and acetate
ions are relatively stable, which in part explains why total mineralization takes much longer
than de-aromatization (Franch et al, 2002). Information concerning degradation of
contaminants at pilot plant-scale using solar collectors is available in several recently
published reviews (Alfano et al, 2000; Bahnemann et al, 2004; Dillert et al, 1999a; Gogate
and Pandit, 2004a; Konstantinou and Albanis, 2003; Malato et al, 2002a; Malato and
Agera, 2003; Pera-Titus et al, 2004).

5.6 EVALUATION OF SOLAR UV RADIATION


Solar ultraviolet radiation is an essential parameter for the correct evaluation of data
obtained during photocatalytic experiments in a solar water decontamination pilot plant.
The kinetic constants of photocatalytic processes can be obtained by plotting substrate
concentration as a function of three different variables: time, incident radiation inside the
reactor and photonic flux absorbed by the catalyst. The complexity of obtaining these
constants, as well as their applicability, varies depending on the procedure. When the
photonic flux absorbed by the catalyst is used as an independent variable, extrapolation
of the results is better. However, many parameters must be known for determining
photonic flux absorbed by the catalyst (including incident photons passing through the
reactor without interacting with the catalyst, directions of light scattering and size
distribution of the TiO2 particles suspended in the liquid), making it almost impracticable
in large size solar photo-reactors (Cassano and Alfano, 2000; Curc et al, 1996; Sagawe
et al, 2003a and 2003b).
Use of the experimental time as the calculation unit could give rise to
misinterpretation of results, because the reactor consists of illuminated and nonilluminated elements. Large experimental reactors must be as versatile as possible and

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 152

152 S. MALATO et al

require a great deal of instrumentation, thus substantially increasing the non-illuminated


volume. With use of residence time, in other words the time the water was irradiated, the
conclusions would be erroneous too. This is because when time is the independent
variable, the differences in the incident radiation in the reactor during an experiment are
not taken into account.
Goslich et al (1997b) have proposed a very useful mathematical approach for the
treatment of such data obtained in real solar experiments using a relationship between
experimental time (t), plant volume (Vt), collector surface (Ar) and the radiant power density
(UVG = WUVm2) measured by a specific UV radiometer. They suggested using a so-called
mass area ratio Si of the organic pollutant i. This is the overall mass of pollutant per
surface element of the reactor: Si = CiVtAr1. As radiation data are collected continuously,

it is very easy to calculate the average incident radiation on the collector surface (UVG,n)
for each period of t, and therefore calculate the accumulated energy during this period,
Euv,ges. It is reasonable to divide the mass area ratio of the pollutant by the accumulated
energy, yielding a value for an entity called efficiency: Eff = Si (Euv,ges)1. Its dimension is
mass per unit energy. Instead of using Euv,ges and Eff, we have proposed using QUV and rQ
(Malato et al, 2000c). QUV fits better with the term radiant energy (Q) recommended by
IUPAC (Braslavsky and Houk, 1988) and makes it possible to use C (reactant
concentration) in the plots, which is easier to understand than Si. Besides, rQ is better than
Eff to define a parameter which is not dimensionless. At the same time, we think that
symbol rQ (-dC/ dQUV) is more consistent with r (dC/dt), which is commonly used in the
scientific community.
Consequently, the amount of energy collected by the reactor (per unit of volume) from
the beginning of the treatment until each sample is collected may be found by:

QUV,n = QUV,n -1 + Dt n UV G,n

Ar
Vt

(29)

Dt n = tn - t n -1
where tn is the experimental time of each sample and QUV,n is the accumulated energy (per
unit of volume, kJ L1) incident on the reactor for each sample taken during the
experiment. This equation comprises all those proposed by Goslich et al (1997a),
substituting the parameter used by these authors (Euv,ges) by QUV,n.
Figure 5.9 shows the improvement obtained using this equation to calculate the
reaction rate in a two-day photocatalytic degradation experiment with a model compound.
Obviously, UV power changes during the day, and clouds during the first day make this
variation still more noticeable, but with Equation 29, the data for both days can still be
combined and compared with other photocatalytic experiments. Consequently, with QUV,
the initial reaction rate (rQ,0 = -dC/dQUV; C1(QUV) = C0 - kapQUV) is expressed in terms of
pesticide degraded per kJ of UV incident on the collector surface. A similar treatment
could be done for first order kinetics. As Goslich et al have remarked, if rQ (in their case
Eff) is known, collector efficiency is already included through the use of incident surface
radiation, since different rQ, with the same substance and different solar collectors, means
collector efficiency is different.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

2/7/07

2:25 PM

Page 153

Photocatalytic Detoxification of Water with Solar Energy

153

C, mg.L1

C, mg.L1

UVG, W.m2

ch-05

QUV, kJ.L1
FIGURE 5.9 Plots of pesticide concentration as a function of experiment time (top) and accumulated
energy (bottom). Solar UV power throughout the experiment is also shown

As commented by Wolfrum and Turchi (1992), this procedure for deducing rQ is exact only
if the concentration of reactives is constant along the entire reactor (in other words if the
system is at steady state). Solar detoxification pilot plants are frequently operated in a
recirculating batch mode as depicted in Figure 5.10. In this scheme, the fluid is
continuously pumped between a reactor zone and a tank in which no reaction occurs until
the desired degradation is achieved (Bahnemann et al, 2004). Since these are dynamic
systems, the photo-reactor (see Figure 5.10) outlet concentration (C1) is not exactly the
same as the mixed tank outlet concentration (C2). In a large field system, the amount of
conversion each time the mixture passes through the reactor is noticeable (Goswami,
1995). As the relatively clean water in the reactor is mixed with the dirty water in the batch
tank, the water sent from it to the reactor has an increasingly lower concentration. Because
the rate usually decreases with concentration, the overall rate in the reactor responds
likewise. Thus unless properly accounted for, the presence of the tank will alter the
perceived performance of the photo-reactor. Two solutions are available to solve this
problem. The first solution is to use a very high flow rate to achieve low conversion each
time through. This high flow rate must allow more than 1 per cent conversion per pass
(C1(t) C2(t)) to be avoided. We have selected this 1 per cent because it is lower than the
error associated with any analytical chemical method applied for Ci analysis. The second
solution is that the concentration in the reactor outlet at time t, C1(t) is determined from the
inlet concentration (which does not change with time) and the reaction kinetics. However,
because the system is a transient process, the normal steady-state plug flow reactor

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 154

154 S. MALATO et al

equation cannot be used to model the photo-reactor. Because the inlet concentration
changes with time, C2(t) is defined by what went into the reactor one residence time prior
to t, C2(t-), and the kinetics. The batch tank could still be modeled as a well-mixed tank.
Solving these equations is more difficult than the low-conversion-per-pass case and a
numerical routine is required to fit the data from the batch test. This routine also takes into
account the volume of the piping between the reactor outlet and the batch tank. This
methodology can be used to numerically solve for the concentration profile in any batch
process, even though this method is only necessary when the recirculation flow is not high
enough. As also pointed out by Mehos et al (1992), one must account for the mixing in the
dark tank only if substantial conversion per pass is obtained (C2 > C1).

FIGURE 5.10 Schematic of the pilot plant batch operation

5.7 INSTALLED SOLAR PHOTOCATALYTIC TREATMENT PLANTS


Despite its obvious potential for the detoxification of polluted water, there has been very
little commercial or industrial use of photocatalysis as a technology to date. Several years
ago, and according to a review by Goswami (1997), the published literature showed only
two engineering-scale demonstrations for groundwater treatment in the US and one
industrial wastewater treatment in Spain (at Plataforma Solar de Almera). But recently
more installations have been erected, mainly based on non-concentrating collectors.
In laboratory and bench-scale experiments Dillert et al have treated biologically
pretreated industrial wastewaters from the Volkswagen AG factories in Wolfsburg
(Germany) and Taubat (Brazil). The results of the experiments, which were performed
using the DSSR mentioned in section 5.2.2, were so promising that a pilot plant was
installed in the Wolfsburg factory during the summer of 1998 (Dillert et al, 1999b). The
flowsheet of a more recent version of this pilot plant, which was installed in 2000, has
been recently published by Bockelmann et al (2004).
This pilot plant, which is operated in a recycle batch mode, consists of 12 double-skin
sheet photo-reactors (manufactured by Solacryl) with a total irradiated area of 27.6m2.
These reactors are connected by a recirculation pipe with a tank (approx. 1.1m3 volume).
The water (500L) coming from the biological treatment plant is pumped into the tank and
mixed with a catalyst slurry. The solar photocatalytic treatment starts after the tank has
been filled. The suspension (total volume approx. 1m3) recycles between the tank and the
DSSRs for 8 to 11 hours during the daytime. The flow rate can be varied between 830 and
1300L per hour in the reactor sheets. After the desired treatment time the suspension is
pumped out of the reactors into the tank. The photocatalyst is allowed to settle during the

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 155

Photocatalytic Detoxification of Water with Solar Energy

155

night. After this sedimentation period the supernatant liquid (500L) is pumped out of the
tank and the treatment cycle can be started again by filling the tank with a new batch of
wastewater. The initial values of the TOC were determined to vary between 26.9 and
7.3mg L1. During the solar-catalytic treatment a decrease in the concentration of the
organic pollutants, determined as COD (chemical oxygen demand) and TOC, was
observed in most cases. More than 50 per cent of the organic pollutants initially present
in the mixed water inside the pilot plant could thus be degraded within 8 to 11 hours of
illumination. Naturally, the total mass of the degraded contaminants was found to depend
on the initial pollutant concentration, the time of illumination and, in particular, on the solar
UV energy flux density.
In 1997 Freudenhammer et al reported their results from a pilot study using thin film
fixed bed reactors (TFFBRs), which was performed in various Mediterranean countries and
showed that biologically pretreated textile wastewater can be cleaned by solar
photocatalysis with a maximum degradation rate of 3g COD m2 per hour (Freudenhammer
et al, 1997). It was concluded that photocatalysis should be a suitable technology as the final
stage of purification of biologically or physically pretreated wastewater in particularly sunrich areas. However, it was also pointed out that there is a strong demand for cheap
photocatalysts of higher activity for industrial applications of solar wastewater treatment on
a larger scale to be competitive with treatment methods already established on the market.
Based on these results, a pilot plant, financed by the European Commission, has been
built at the site of a textile factory in Tunisia (Menzel Temime). The TFFBR was chosen
because previous studies showed sufficient degradation rates with the selected textile
wastewater in combination with its simple, low cost construction and the low energy
consumption. However the possibility of using suspended catalysts has also been
considered, to integrate results obtained with suspended catalysts showing, in some
cases, a higher efficiency than the fixed system,. The pilot plant and the flow chart have
been published recently (Bahnemann et al, 2004).
Two TFFBRs of width 2.5m and length 10m, corresponding to a total illuminated area of
50m2, were built in concrete and are oriented to the south with an inclination angle of 20.
The reactors can be operated in parallel or as a cascade flow and in a continuous or a
recycling mode, depending on the reaction kinetics. The pumps are designed for a
maximum flow of 3m3 per hour. The plant is connected to the sewerage system of the textile
factory for continuous and long-term experiments. Two sequencing batch reactors (SBRs)
and a membrane aeration system are connected to the TFFBR for pre- and post-treatment,
each with a total volume of 15m3. An automatic injection system of chemical compounds
can be used for pH regulation or other necessities. The plant is designed to operate the two
reactors independently. This allows a comparison of different catalysts, fixation techniques,
hydrodynamics and effluents/model compounds under identical conditions. The plant can
be operated with suspended and fixed catalysts. A sedimentation tank is connected to the
reactors (via a storage tank) to separate and recycle the catalyst. Instrumentation and control
instruments are installed and connected to a computer system to allow automatic
operation: pumps and valves control flow rates, tank levels, UV-A intensity, dissolved
oxygen, pH and conductivity. Experiments are currently being conducted to test the
operation conditions of this newly installed facility (Bousselmi et al, 2004).

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 156

156 S. MALATO et al

Under the Solar Detoxification Technology for the Treatment of Industrial NonBiodegradable Persistent Chlorinated Water Contaminants (SOLARDETOX) project, a
consortium coordinated by Plataforma Solar de Almera, Spain, has been formed in
Europe for the development and marketing of solar detoxification treatments for
recalcitrant water contaminants. The main project objective was the technical and
economic optimization of real solar detoxification applications. The main innovations were
in the engineering, there being no specific developments in the solar detoxification
technology itself, since it was derived from the already existing solar thermal technology
with only minor modifications. This general objective was divided into several parts:
assessment of the performance and best working conditions for newly developed

catalyst powders compared with the efficiencies of commercial products (Degussa


P25) in the treatment of chlorinated solvent compounds;
assessment of degradation mechanisms, photonic efficiencies, formation of sideproducts and kinetic modeling of the process from degradation data;
development of a reactor with highly UV-transmissive glass in the solar UV range;
improvement of the solar collector with a highly efficient UV-reflective surface; and
demonstration of the technical and economic feasibility of achieving a well-defined
system under real conditions; optimization of the design and standardization of the
components (collectors, tubes, catalyst, control and operating systems) in order to
minimize costs of production; erection and operation through the formation of a
consolidated industrial and institutional consortium.

The main goal (financed by the EC-DGXII (European Commission Directorate General XII)
through the Brite Euram III Program, 19972000) was to develop a commercial nonconcentrating solar detoxification system using compound parabolic collector (CPC)
technology, with a concentration ratio of one. The beauty of the solar CPC system (as
commented in section 5.2.3) is its intrinsic simplicity while remaining cost-effective and
easy to use it requires low capital investment. Field demonstration was intended to
identify any pre- or post-processing requirements, potential operating problems, and
capital and operating costs.
Based on accumulated experience in pilot plant design, construction and testing (Blanco
et al, 2000), a full-size demonstration plant was erected at the facilities of HIDROCEN (Madrid,
Spain). This plant was designed to treat 1m3 of water contaminated with 100m2 of collector
aperture area (see Figure 5.11). The main plant characteristics are: i) 2 rows of 21 collectors
each; ii) total collector (tilted 40, local latitude) aperture area = 98m2; iii) total loop volume =
675L; iv) total plant volume = 975L; v) 200mg/L TiO2 slurry; and vi) 1.5 x 1.5m collectors with
sixteen 29.2mm internal diameter (ID) tubes. The CPC reflector is made of a highly reflective
anodized aluminum sheet held by a galvanized frame supporting 16 parallel 1.5m-long tubes,
each with an appropriate connector for the adjacent tube. A complete module is formed by
a series of collectors connected in a row. The final prototype plant consists of EW oriented
parallel rows of 21 collectors each. The structure was slightly tilted (1 per cent) in the same
direction to drain rain water and avoid its accumulation in the CPC troughs. Final system
design is completely modular, with the collectors attached in series using HDPE (high density

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 157

Photocatalytic Detoxification of Water with Solar Energy

157

polyethylene) quick connections between glass reactor tube absorbers. Water flows
simultaneously through all parallel tubes and there is no limit to the number of collector
components modules may have. Water enters and exits rows through two manifolds at
opposite ends. As this plant is to be a demonstration of what a commercial plant would be
like, operation is fully automatic and maintenance requirements are minimal. Among the
electronic control devices the most innovative system is a solar UV-A sensor that integrates
solar UV with time during the treatment. This sensor is connected to a programmable logic
controller (PLC) and once sufficient energy (see Equation 29) for finishing the treatment has
been achieved (based on preliminary testing for plant design according to the specific
contaminated water to be treated), the PLC stops the main pump and advises the operator
that the treatment has been completed. The PLC also receives other data signals (flow rate,
tank level, temperature, etc) for controlling system pumps and valves.

2m3

2m3
0.1m

0.2m3

0.3m

100m2
800L

~5

0m

FIGURE 5.11 Schematic diagram of solar detoxification demonstration plant constructed in SOLARDETOX
project at HIDROCEN, Madrid

Since late 1999, this plant has been fully functional and its operation is easily described.
Wastewater enters the system from a 2m3 wastewater storage tank while detoxified
water from a previous run returns to the catalyst sedimentation tank. When the system is
completely full (0.8m3), a pump recirculates through the collectors and a small tank
(0.1m3). The concentrated TiO2 slurry and the air necessary for the reaction are injected
into the circuit. Once the water is detoxified, the entire volume passes to the

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 158

158 S. MALATO et al

sedimentation tank and the system is filled with more wastewater for another batch.
Meanwhile, the detoxified water and the TiO2 in the sedimentation tank undergo pH
adjustment to provoke fast sedimentation of the catalyst (FernndezIbez et al, 2003).
The concentrated catalyst slurry is transferred from the bottom of the tank to another
smaller tank from which the catalyst enters the photo-reactor. The supernatant is removed
through an outlet in the side almost at the bottom of the sedimentation tank and enters
another tank (0.3m3) where the small fraction of the initial catalyst contained in the
supernatant (<7mg L1) is removed by micro-filtration with a suitable membrane before
disposal of water. The membrane has two outlets, one for clean water and one for the
concentrated slurry. The concentrated slurry is recirculated through the membrane until
there are several grams of catalyst per liter. Then this is added to the TiO2 injection tank
and also reused.
More recently (2004) a new CPC-based plant has been installed. This project focuses
on problems that originate in intensive greenhouse agriculture, a sector that has been
growing exponentially in recent years in the Mediterranean basin. There are currently over
200,000 hectares of greenhouses, most of them in EU countries. This type of agriculture
requires up to 200 times more pesticides than conventional agriculture. The
environmental problems are one of the greatest disadvantages for the development of
this economic sector. One of these problems is the uncontrolled dumping of plastic
pesticide containers, which usually still contain residues. This poses a serious risk of
pollution of soil and groundwater. In the area of El Ejido, a town in the province of Almera
in southern Spain, intensive agriculture in 400km2 of greenhouses consumes
approximately 2 million plastic bottles of pesticide per year. So far, these empty plastic
bottles have simply been discarded. Although the amount of product remaining in each
bottle after use is minimal, the numbers are so high that they become a danger to the
environment (poisoning of fauna and flora, not only on land, but also at sea by bottles
carried out to sea by flooding; contamination of water supply by ground filtration). The
solution is to selectively collect these containers for recycling. But before the plastic can
be recycled, it must be washed and the water used for this becomes polluted by the
pesticides. This water must be treated before it is discharged. Therefore, the
development of a simple, clean treatment process for this water in the place where it is
produced is necessary. It is in the detoxification of this water that solar photocatalysis
suggests itself as a very promising process, made even more so by the availability of yearlong strong sunlight as a cheap energy source in this region. Thus the availability of
sunlight and the lack of other alternatives justify the application of a new technology that
has not yet been evaluated on an industrial scale. The ALBAIDA Company and CIEMAT
(Spain) has presented a project entitled Environmental Collection and Recycling of Plastic
Pesticide Bottles using Advance Oxidation Process, Driven by Solar Energy to the
European LIFEENVIRONMENT program, which was approved and began in October
2001. The definitive plant has been in operation since 2004.
The plastic container recycling starts with shredding and then industrial washing of the
shredded plastic, which produces water polluted with highly toxic persistent compounds
(pesticides). The shredded plastic is sent to a series of three baths of the same size (see
Figure 5.12). This is where the separation of the plastic and the wash water would begin.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 159

Photocatalytic Detoxification of Water with Solar Energy

159

The wash water would be drained from each of the baths at floor level and flow to a
storage tank. This water is considered a hazardous toxic product as it carries the dissolved
organic matter that was in the washed pesticide containers. As the water is continuously
recycled and reused, the contaminants must be treated. It is important that care be used
in selecting the containers, since it is very common for the grower to throw away any kind
of used or useless container, while this specific plant is exclusively for plastic.

10mg/L

mg/L
1g/L

CPCs FIELD
(150m2)

FIGURE 5.12 Schematic diagram of the container washing and photocatalytic water treatment
plant concept

Before entering the solar field, the contaminated water is treated physically and
chemically. The water still contains solid waste, which is eliminated by a rough screen. It
then flows to another tank where reactants needed for the photocatalytic reaction are
added. Once the sample is prepared, it enters the solar field of several photo-reactors
connected in series, through which the water flows from one to another. When the water
leaves the solar field, it is tested to see if the process has considerably diminished the
TOC concentration in the treatment water. If not, it is sent back through the solar field
again to undergo the solar detoxification process in the collectors again. When the TOC
measurement is sufficient, the water is reused, re-entering the washing unit in the
container storage bay. When the water has recirculated four or five times, it can be filtered
by an active carbon filter, where any substance resistant to the solar process would be
retained, and discharged into a storage pool for later reuse. The plant was constructed to
treat 1.6m3 of contaminated water with 150m2 of collector aperture area. The main
plant characteristics are: i) 4 rows of 14 collectors each; ii) total collector (tilted 37,
local latitude) aperture area = 150m2; iii) total loop volume = 1060L; iv) total plant
volume = 1600L; v) photo-Fenton using Fe (II) 1mM; and vi) collectors with twenty

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 160

160 S. MALATO et al

29.2mm ID glass tubes. The final prototype plant consists of EW oriented parallel rows
of 14 collectors each. Final system design is completely modular. Collectors are attached
in series using polypropylene quick connections between glass reactor tube absorbers.
Water flows simultaneously through all four rows and in series through all tubes of each
row. Operation is fully automatic and maintenance requirements are minimal.

(150m )

FIGURE 5.13 Schematic diagram of solar detoxification demonstration plant constructed by ALBAIDA at
La Mojonera, Almera, Spain

5.8 PHOTOCATALYTIC DETOXIFICATION OF WATER WITH SOLAR


ENERGY: OUTLOOK FOR THE FUTURE
The proposed technology could be applicable to different organic hazardous contaminants,
such as pesticides, solvents, detergents and a variety of industrial chemicals, all capable of
substantial contamination of the environment due to their persistent toxicity. Nevertheless,
while the process efficiency can be considered linearly dependent on the energy flux, only 5
per cent of the whole solar spectrum is available for the TiO2 band gap. A realistic assumption
of solar collector efficiency of 75 per cent and 1 per cent for the catalyst (Romero et al, 1999b)
means that 0.04 per cent of the original solar photons are efficiently used in the process.
From the standpoint of solar collecting technology, this is a rather inefficient process even
taking into consideration a high added value application (Parent et al, 1996). Solar AOPs have
the advantage over other AOPs of using sunlight, which has the important characteristic of
being an environmentally friendly technology. TiO2 is a cheap photo-stable catalyst, and the
process may run at ambient temperature and pressure. Additionally the oxidant, molecular
oxygen (O2) is the mildest. Therefore, in principle, the process involves a mild catalyst working
under mild conditions with mild oxidants. However, as concentration and number of
contaminants increase, the process becomes more complicated and challenging, with
problems such as slow kinetics, low photo-efficiency and unpredictable mechanisms

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 161

Photocatalytic Detoxification of Water with Solar Energy

161

needing to be solved. It is clear that naked TiO2 needs extra help to undertake practical
applications and this may cause it to lose some of the charm of its easy operation.
Two basic lines of R&D (increasing quantum yield) have been working on modifying
catalyst structure and composition and by adding electron acceptors (Herrmann, 1999). A
third approach has focused on finding new catalysts able to work with band gaps which
better overlap the solar spectrum. There have been many attempts within the first and
third approach, such as improving specific surface, doping and deposition with metal ions
and oxides (Blake, 2001). Successful innovative catalyst compositions have been
developed, but they have not been used in large-size plants because no cheap solution
has yet been developed. Our experience in testing at large solar facilities and with
different contaminants qualifies the use of electron acceptors as the most versatile way
of improving reaction rates (for the moment), opening the opportunity to extend the use
of heterogeneous photocatalysis to complicated wastewater (Malato et al, 2002a).
Photo-Fenton treatment presents an interesting alternative to titanium dioxide
photocatalysis, because when the reaction rates of these AOPs are compared, photoFenton shows usually higher reaction rates and is less susceptible to inner filter effects
by the compounds of the wastewater itself (Fallmann et al, 1999; Malato et al, 2002b).
Both issues are mainly due to the fact that iron complexes can absorb light to higher
wavelengths (up to 580nm, Bauer et al, 1999) depending on their composition and on the
homogeneous nature of the process. On the other hand, commonly mentioned
disadvantages of the photo-Fenton treatment are the low pH at which the process is
performed and the need to remove the iron after the process. Therefore its applicability
will be positively influenced by possibilities of water reuse where reactives employed
could be re-valorized, as in fertilizer in irrigation.

FIGURE 5.14 Schematic of a solar treatment coupled with an aerobic biological treatment

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 162

162 S. MALATO et al

Contaminant treatment, in its strictest meaning, is the complete mineralization (TOC = 0) of


the contaminants, but photocatalytic processes only make sense for hazardous nonbiodegradable pollutants. When feasible, biological treatment is the cheapest treatment and
also the most compatible with the environment. Therefore biologically recalcitrant
compounds could be treated with photocatalytic technologies until biodegradability is
achieved, later transferring the water to a conventional biological plant. Such a combination
reduces treatment time and optimizes overall economics, since the solar detoxification
system can be significantly smaller. Solar photocatalytic post-treatment of an effluent from an
anaerobic treatment has also been described by Zaidi et al (1995). Due to the kinetic
mechanism (see Equations 4 and 5), the first part of the photocatalytic process is the
quickest. Therefore the use of AOPs as a pretreatment step can be justified if the
intermediates resulting from the reaction (more oxidized compounds as carboxylic acids and
alcohols) are readily degraded by micro-organisms (Sarria et al, 2003; Gogate and Pandit,
2004b). The feasibility of such a photocatalytic-biological process combination should be
assessed, because it could provide an important cost reduction by reducing the size of the
necessary solar collector field. It must be taken into account that, as with most solar systems,
economics of the water detoxification systems are dominated by their capital cost. A recent
project (A Coupled Advanced OxidationBiological Process for Recycling of Industrial
Wastewater Containing Persistent Organic Contaminants, CADOX), sponsored by the
European Commission and with the participation of nine EU partners, is focused in this way
(www.psa.es/webeng /projects/cadox/index.html). This project attempts to demonstrate that
the treatment cost of water containing persistent contaminants can be drastically reduced.
Another way to decrease AOP operating costs is by determining the waters toxicity at
different stages of AOP treatment, using different micro-organisms. In this case,
biocompatibility with the environment can be stated. Toxicity testing of the
photocatalytically treated wastewater is therefore necessary, particularly when incomplete
degradation is planned. Recently, the use of acute toxicity bioassays (Fernndez-Alba et al,
2002) has meant an important improvement in the evaluation of AOPs because of their
reproducibility, adequate format for quick analysis, short analysis time and well-defined
analytical protocols (Fernndez-Alba et al, 2001). Assessment of the contaminants effect
involves summarizing data on the effects of the chemical on representative organisms and
using these data to predict a no-effect concentration on a specific niche. Toxicity of a
chemical is usually expressed as the effective concentration of the material that would
produce a specific effect in 50 per cent of a large population of test species (EC50). The no
observed effect concentration (NOEC) is the concentration immediately below the lowest
level eliciting any type of toxicological response in the study. All these analytical
performance findings make the use of acute bioassays very attractive for AOPs evaluation.
Numerous bioassay procedures are now available (Tothill and Turner, 1996); however, if we
consider that toxicity is a biological response, the values obtained by a single toxicity assay
may be an insufficient measure of the adverse biological impact. Consequently, a battery
of assays is recommended to be applied to adequately assess toxicity, and careful
selection is essential. A high degree of confidence in the detoxification assessment is
achieved when two or more different bioassay representatives of different taxonomic
groups point in the same direction (Malato et al, 2003b).

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 163

Photocatalytic Detoxification of Water with Solar Energy

163

ACKNOWLEDGEMENTS
The authors wish to thank the European Commission (Research DG) for its financial
assistance within the Energy, Environment and Sustainable Development Program
(Contract No EVK1CT200200122, CADOX Project) and EU-DG Research, Confirming
the International Role of Community Research for Development (Contract No
ICA4CT200210001, SOLWATER Project). They also wish to thank Mrs. Deborah
Fuldauer for correcting the English. Mr. Gernjak wishes to thank the Austrian Academy of
Sciences for financial support through a DOC grant.

AUTHOR CONTACT DETAILS


Sixto Malato, (corresponding author), Plataforma Solar de Almera (CIEMAT), Carretera Sens, km 4, 04200
Tabernas, (Almera), Spain
Tel: 34-950 387940; Fax: 34-950365015; e-mail: sixto.malato@psa.es

REFERENCES
Ajona, J. A. and Vidal, A. (2000) The use of CPC collectors for detoxification of contaminated water: Design, construction
and preliminary results, Solar Energy, no 681, pp109120
Akmehmet, Y. and Ynel, Y. (1996) Photocatalytic degradation of organic contaminants in semiconductor suspensions with
added H2O2, J. Environ. Sci. Health A, vol 31, no 1, pp123138
Alfano, O. M., Bahnemann, D., Cassano, A. E., Dillert, D. and Goslich, R. (2000) Photocatalysis in water environments using
artificial and solar light, Catalysis Today, no 58, pp199230
Anderson, J. V., Link, H., Bohn, M. and Gupta, B. (1991) Development of US solar detoxification technology: An introduction,
Solar En. Mat., no 24, pp538549
Andreozzi, R., Caprio, V., Insola, A. and Martota, R. (1999) Advanced oxidation processes (AOP) for water purification and
recovery, Catalysis Today, no 53, pp5159
Angelakis, A. N., Marekos, M. H. F., Bontoux, L. and Asano, T. (1999) The status of wastewater reuse practice in the
Mediterranean basin: Need for guidelines, Wat. Res., vol 33, no 10, pp22012217
Bahnemann, D. (2004) Photocatalytic Water Treatment: Solar Energy Applications, Solar Energy, no 77, pp445459
Bauer, R., Waldner, G., Fallmann, H., Hager, S., Klare, M., Krutzler, T., Malato, S. and Maletzky, P. (1999) The photo-Fenton
reaction and the TiO2/UV process for waste water treatment Novel developments, Catalysis Today, no 53, pp131144
Bedford, J., Klausner, J. F., Goswami, D. Y. and Schanze, K. S. (1994) Performance of nonconcentrating solar photocatalytic
oxidation reactors part II: Shallow pond configuration, Journal of Solar Energy Engineering, vol 116, no 1, pp813
Bekblet, M., Lindner, M., Weichgrebe, D. and Bahnemann, D. W. (1996) Photocatalytic detoxification with the thin-film fixedbed reactor (TFFBR): Clean-up of highly polluted landfill effluents using a novel TiO2-photocatalyst, Solar Energy, no 56,
pp455469
Bellobono, I. R., Carrara, A., Barni, B. and Gazzotti, A. (1994) Laboratory- and pilot-plant-scale photodegradation of
chloroaliphatics in aqueous solution by photocatalytic membranes inmovilizing titanium dioxide, J. Photochem. Photobiol.
A: Chem., no 84, pp8390
Blake, D. M. (1994) Solar Processes for the Destruction of Hazardous Chemicals. Alternative Fuels and the Environment, Lewis
Publishers, Boca Raton, FL, US
Blake, D. M. (2001) Bibliography of Work on the Photocatalytic Removal of Hazardous Compounds from Water and Air,
National Technical Information Service, US Dept. of Commerce, Springfield, VA, US

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 164

164 S. MALATO et al

Blake, D. M., Magrini, K., Wolfrum, E. and May, E. K. (1997) Material issues in solar detoxification of air and water, in Optical
Materials Technology for Energy Efficiency and Solar Energy Conversion XV, The Int. Society for Optical Engineering, pp154162
Blanco, J. and Malato, S. (2003) Solar Detoxification, UNESCO Publishing, France
Blanco, J., Malato, S., Maldonado, M.I., Fallmann, H., Krutzler, T. and Bauer, R. (1999) Techno-economical assessment of
solar detoxification systems with compound parabolic collectors, J. Phys., IV, Pr 3, pp259264
Blanco, J., Malato, S., Fernndez, P., Vidal, A., Morales, A., Trincado, P., de Oliveira, J.C., Minero, C., Musci, M., Casalle, C.,
Brunotte, M., Tratzky, S., Dischinger, N., Funken, K-H., Sattler, C., Vincent, M., Collares-Pereira, M., Mendes, J. F. and
Rangel, C. M. (2000) Compound parabolic concentrator technology development to commercial solar detoxification
applications, Solar Energy, no 67, pp317330
Bousselmi, L., Geissen, S. U. and Schroeder, H. (2004) Textile wastewater tratment and reuse by solar photocatalysis: results
from a pilot plant in Tunisia, Wat. Sci. Technol., vol 49, no 4, pp331337
Braslavsky, S. E. and Houk, K. N. (1988) Glossary of terms used in photochemistry, Pure Applied Chem., no 60,
pp10551106
Carey, J. H., Lawrence, J. and Tosine, H. M. (1976) Photodechlorination of PCBs in the presence of TiO2 in aqueous
suspensions, Bull. Environ. Contamination & Toxicol., vol 16, no 6, pp697701
Cassano, A. E. and Alfano, O. E. (2000) Reaction engineering of suspended solids heterogeneous photocatalytic reactors,
Catal. Today, no 58, pp167197
Chen, T., Doong, R. and Lei, W. (1998) Photocatalytic degradation of paration in aqueous TiO2 dispersion, the effect of
hydrogen peroxide and light intensity, Wat. Sci. Technol., vol 37, no 8, pp187194
Collares-Pereira, M. (1995) The relevance of non-imaging optics for solar energy: A review, Revista de Engenharia Tcnica, 3
Cunningham, J. and Sedlak, J. (1996) Kinetic studies of depollution process in TiO2 slurries: Interdependences of adsorption
and UV intensity, Catalysis Today, no 29, pp309315
Curc, D., Malato, S., Blanco, J. and Gimnez, J. (1996) Photocatalysis and radiation absorption in a solar plant, Solar Energy
Materials and Solar Cells, no 44, pp199217
Dillert, R., Cassano, A.E., Goslich, R. and Bahnemann, D. (1999a) Large scale studies in solar catalytic wastewater
treatment, Catalysis Today, no 54, pp267282
Dillert, R., Vollmer, S., Gross, E., Schober, M., Bahnemann, D., Wienefeld, J., Pahlmann, K., Schmedding, T., Arntz, H. J. and
Sager, G. (1999b) Solar-catalytic treatment of an industrial wastewater, Z. Phys. Chem., no 213, pp141147
Doong, R. and Chang, W. (1997) Photoassisted titanium dioxide mediated degradation of organophosphorus pesticides by
hydrogen peroxide, J. Photochem. Photobiol., A: Chem., no 107, pp239244
Fallmann, H., Krutzler, T., Bauer, R., Malato, S. and Blanco, J. (1999) Applicability of the photo-Fenton method for treating
water containing pesticides, Catalysis Today, no 54, pp309319
Feitz, A. J., Boyden, B. H. and Waite, T. D. (2000) Evaluation of two solar pilot scale fixed-bed photocatalytic reactors, Water
Res., no 34, pp39273932
Fernndez-Alba, A. R., Hernando, L., Daz, G. and Chisti, Y. (2001) Toxicity of pesticides in wastewater: a comparative
assessment of rapid bioassays, Anal. Chim. Acta, no 426, pp289301
Fernndez-Alba, A. R., Hernando, D., Agera, A., Cceres, J. and Malato, S. (2002) Toxicity assays: a way for evaluating
AOPs efficiency, Wat. Res., no 36, pp42554262
Fernndez-Ibez, P., Blanco, J., Malato, S. and de las Nieves, F. J. (2003) Application of the colloidal stability of TiO2 particles
for recovery and reuse in solar photocatalysis, Wat. Res., no 37, pp31803188
Franch, M., Aylln, J.A., Peral, J. and Dom`enech, X. (2002) Photocatalytic degradation of short-chain organic diacids,
Catalysis Today, no 76, pp221235
Freudenhammer, H., Bahnemann, D., Bousselmi, L, Geissen, S. U., Ghrabi, A., Saleh, F., Si-Salah, A., Siemon, U. und
Vogelpohl, A. (1997) Detoxification and recycling of wastewater by solar-catalytic treatment, Wat. Sci. Tech., no 35,
pp149156

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 165

Photocatalytic Detoxification of Water with Solar Energy

165

Funken, K-H., Sattler, C., Milow, B., de Oliveira, L, Blanco, J., Fernndez, P., Malato, S., Brunotte, M., Dischinger, N., Tratzky,
S., Musci, M. and de Oliveira, J. C. (2001) A comparison of prototype compound parabolic collectorreactors (CPC) on
the road to SOLARDETOX technology, Wat. Sci. Technol., vol 44, no 5, pp271278
Gernjak, W., Krutzler, T., Glaser, A., Malato, S., Cceres, J., Bauer, R. and Fernndez-Alba, A. R. (2003) Photo-Fenton
treatment of water containing natural phenolic pollutants, Chemosphere, no 50, pp7178
Gimnez, J., Curc, D. and Queral, M. A. (1999) Photocatalytic treatment of phenol and 2,4-dichlorophenol in a solar plant in
the way to scaling-up, Catalysis Today, no 54, pp229244
Gogate, P. R. and Pandit, A. B. (2004a) A review of comparative technologies for wastewater treatment. I: Oxidation
technologies at ambient conditions, Adv. Environ. Res., no 8, pp501551
Gogate, P. R. and Pandit, A. B. (2004b) A review of comparative technologies for wastewater treatment. II: Hybrid methods,
Adv. Environ. Res., no 8, pp553597
Goslich, R., Dillert, R. and Bahnemann, D. (1997a) Solar water treatment: Principles and reactors, Wat. Sci. Tech., vol 35, no
4, pp137148
Goslich, R., Bahnemann, D., Schumacher, H. W., Benz, V. and Mller, M. (1997b) Photocatalytic solar wastewater treatment:
New reactor concepts, proceedings of 8th Int. Symp. Solar Thermal Concentrating Technologies, Kln, Germany,
October, pp13371353
Goswami, D. Y. (1995) Engineering of solar photocatalytic detoxification and disinfection processes. Advances in solar
energy, Am. Solar En. Soc., pp165209
Goswami, D. Y. (1997) A review of engineering developments of aqueous phase solar photocatalytic detoxification and
disinfection processes, J. Sol. En. Eng., no 119, pp101107
Goswami, D. Y. and Blake, D. M. (1996) Cleaning up with sunshine, Mech. Engin., August, pp5659
Goswami, D. Y., Sharma, S. K., Mathur, G. D. and Jotshi, C. K. (1997) Techno-economic analysis of solar detoxification
systems, J. Solar En. Eng., no 119, pp108113
Goswami, D. Y., Vijayaraghavan, S., Lu, S. and Tamm, G. (2003) New and emerging developments in solar energy, Solar
Energy, no 76, pp3343
Guillard, C., Disdier, J., Herrmann, J. M., Lehaut, C., Chopin, T., Malato, S. and Blanco, J. (1999) Comparison of various titania
samples of industrial origin in the solar photocatalytic detoxification of water containing 4-chlorophenol, Catalysis Today,
no 54, pp217228
Guillard, C., Disdier, J., Monnet, Ch., Dussaud, J., Malato, S., Blanco, J., Maldonado, M. I. and Herrmann, J. M. (2003) Solar
efficiency of a new deposited titania photocatalyst: Chlorophenol, pesticide and dye removal applications, Appl. Catal. B:
Environ., no 46, pp319332
Hayo, M. G. (1996) Assessing the impact of pesticides on the environment, Agric. Ecosys. Environ., no 60, pp8196
Harada K., Hisanaga, T. and Tanaka, K. (1990) Photocatalytic degradation of organophosphorous insecticides in aqueous
semiconductor suspensions, Wat. Res., vol 24, no 11, pp14151417
Herrmann, J. M. (1999) Heterogeneous photocatalysis: Fundamentals and applications to the removal of various types of
aqueous pollutants, Catalysis Today, no 53, pp115129
Hofstadler, K., Bauer, R., Novalic, S. and Heisler, G. (1994) New reactor design for photocatalytic wastewater treatment with
TiO2 inmobilized on fused silica glass fibers: Photomineralization of 4-chlorophenol, Environ. Sci. Technol., vol 28, no 4, pp
670674
Jubran, B. A., Ismail, A. F. and Pervez, T. (1999) Prediction of the performance of a solar water detoxification system under
Malaysian climatic conditions, Energy Convers. Manage., no 41, pp112
Konstantinou, I. K. and Albanis, T. A. (2003) Photocatalytic transformation of pesticides in aqueous titanium dioxide
suspensions using artificial and solar light: Intermediates and degradation pathways, Appl. Catal. B: Environ., no 42,
pp319335

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 166

166 S. MALATO et al

Lee, C. and Yoon, J. (2004) Temperature dependence of hydroxyl radical formation in the hv/Fe3+/H2O2 and Fe3+/H2O2
systems, Chemosphere, no 56, pp923934
Legrini O., Oliveros, E. and Braun, A. M. (1993) Photochemical processes for water treatment, Chem. Rev., no 93, pp671698
Malato, S. and Agera, A. (2003) Detoxification of pesticides in water using solar photocatalysis, In Pesticide
Decontamination and Detoxification, ACS Symposium Series 863, Am. Chem. Soc., Washington DC, pp113126
Malato, S., Richter, C., Blanco, J. and Vincent, M. (1996) Photocatalytic degradation of industrial residual waters, Solar
Energy, vol 56, no 5, pp401410
Malato, S., Blanco, J., Richter, C., Braun, B. and Maldonado, M. I. (1998) Enhancement of the rate of solar photocatalytic
mineralization of organic pollutants by inorganic species, Appl. Catal. B: Environ., no 17, pp347356
Malato, S., Blanco, J., Maldonado, M. I., Fernndez-Ibez, P. and Campos, A. (2000a) Optimizing solar photocatalytic
mineralization of pesticides by adding inorganic oxidising species: Application to the recycling of pesticide containers,
Appl. Catal. B: Environ., no 28, pp163174
Malato, S., Blanco, J., Richter, C. and Maldonado, M. I. (2000b) Optimization of pre-industrial solar photocatalytic
mineralization of commercial pesticides; application to pesticide container recycling, Appl. Catal. B: Environ., no 25,
pp3138
Malato, S., Blanco, J., Richter, C., Fernndez, P. and Maldonado, M. I. (2000c) Solar photocatalytic mineralization of
commercial pesticides: Oxamyl, Solar En. Mat. Solar Cells, no 64, pp114
Malato, S., Blanco, J., Vidal, A. and Richter, C. (2002a) Photocatalysis with solar energy at a pilot-plant scale: An overview,
Appl. Catal. B: Environ., no 37, pp115
Malato, S., Blanco, J., Cceres, J., Fernndez-Alba, A. R., Agera, A. and Rodrguez, A. (2002b) Photocatalytic treatment of
water-soluble pesticides by photo-Fenton and TiO2 using solar energy, Catalysis Today, no 76, pp209220
Malato, S., Blanco, J., Vidal, A., Alarcn, D., Maldonado, M. I., Cceres, J. and Gernjak, W. (2003a) Applied studies in solar
photocatalytic detoxification: An overview, Solar Energy, no 75, pp329336
Malato, S., Blanco, J., Campos, A., Cceres, J., Guillard, C., Herrmann, J. M. and Fernndez-Alba, A. R., (2003b) Effect of
operating parameters on the testing of new industrial titania catalysts at solar pilot plant scale, Appl. Catal. B: Environ.,
no 42, pp349357
Malato, S., Cceres, J., Fernndez-Alba, A. R., Piedra, L., Hernando, M. D., Agera, A. and Vial, J. (2003c) Photocatalytic
treatment of diuron by solar photocatalysis: Evaluation of main intermediates and toxicity, Env. Sci. Technol., no 37,
pp25162524
May, E. K., Gee, R., Wickham, D. T., Lafloon, L. A. and Wright, J. D. (1991) Design and fabrication of a prototype solar
receiver/reactors for the solar detoxification of contaminated water, NREL Report, Industrial Solar Technology Corp.,
Golden, Colorado, US
Mazellier, P., Jirkovsky, J. and Bolte, M. (1997) Degradation of diuron photoinduced by iron III in aqueous solution, Pestic.
Sci., no 49, pp259267
Mehos, M., Turchi, C., Pacheco, J., Bogel, A. J., Merrill, T. and Stanley, R. (1992) Pilot-scale study of the solar detoxification
of VOC-contaminated groundwater, National Renewable Energy Laboratory, Golden, Colorado, US
Minero, C. (1999) Kinetic analysis of photoinduced reactions at the water semiconductor interface, Catalysis Today, no 54,
pp205216
Minero, C., Pelizzetti, E., Malato, S. and Blanco, J. (1993) Large solar plant photocatalytic water decontamination:
Degradation of pentachlorophenol, Chemosphere, vol 26, no 12, pp21032119
Minero, C., Pelizzetti, E., Malato, S. and Blanco, J. (1996a) Large solar plant photocatalytic water decontamination: Effect of
operational parameters, Solar Energy, no 56, pp421428
Minero, C., Pelizzetti, E., Malato, S. and Blanco, J. (1996b) Large solar plant photocatalytic water decontamination:
degradation of atrazine, Solar Energy, no 56, pp411419

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 167

Photocatalytic Detoxification of Water with Solar Energy

167

Mukherje, P. S. and Ray, A. K. (1999) Major challenges in the design of a large-scale photocatalytic reactor for water
treatment, Chem. Eng. Technol., no 22, pp253260
Pacheco, J. E. and Tyner, C.E. (1990) Enhancement of processes for solar photocatalytic detoxification of water, Solar
Engineering, no 1990, pp163169
Pacheco, J., Prairie, M., Evans, L. and Yellowhorse, L. (1990) Engineering-scale experiments of solar photocatalytic oxidation
of trichloroethylene, proceedings of 25th Intersociety Energy Conversion Engineering Conference, Reno, Nevada, US
Parent, Y., Blake, D., Magrini-Bair, K., Lyons, C., Turchi, C., Watt, A., Wolfrum, E. and Prairie, M. (1996) Solar photocatalytic
process for the purification of water: State of development and barriers to commercialization, Solar Energy, vol 56, no 5,
pp429438
Pera-Titus, M., Garca-Molina, V., Baos, M. A., Gimnez, J. and Esplugas, S. (2004) Degradation of chlorophenols by means
of advanced oxidation proceses: A general review, Appl. Catal. B: Environ., no 47, pp219256
Pignatello, J. J. (1992) Dark and photoassisted Fe3+ catalyzed degradation of chlorophenoxy herbicides by hydrogen
peroxide, Environ. Sci. Technol., no 26, pp944951
Poulios, I., Kositzi, M. and Kouras, A. (1998) Photocatalytic decomposition of triclopyr over aqueous semiconductor
suspensions, J. Photochem. Photobiol. A: Chem., no 115, pp175183
Pozzo, R. L., Baltans, M. A. and Cassano, A. E. (1997) Supported titanium oxide as photocatalyst in water decontamination:
State of the art, Catal. Today, no 39, pp219231
Prairie, M. R., Pacheco, J. E. and Evans, L. R. (1992) Solar detoxification of water containing chlorinated solvents and heavy
metals via TiO2 photocatalysis, Solar Engineering, no 1992, pp18
Robert, D., Piscopo, A., Heintz, O. and Weber, V. (1999) Photocatalytic detoxification with TiO2 supported on glass-fibre by
using artificial and natural light, Catalysis Today, no 54, pp291296
Romero, M., Vidal, A., Senhaji, F. and El-Hraiki, A. (1999a) Application of a CPC-based photocatalytic reactor to detoxification
of washing waters from agricultural industries, J. Phys., vol IV, no 9, pp283288
Romero, M., Blanco, J., Snchez, B., Vidal, A., Malato, S., Cardona, A. and Garca, E. (1999b) Solar photocatalytic
degradation of water and air pollutants: Challenges and perspectives, Solar Energy, vol 66 no 2, pp169182
Sagawe, G., Lehnard, A., Lbber, M. and Bahnemann, D. (2001) The insulated solar Fenton hybrid process: Fundamental
investigations, Helv. Chem. Acta, no 86, pp37423758
Sagawe, G., Brandi, R. J., Bahnemann, D. and Cassano, A. E. (2003a) Photocatalytic reactors for treating water pollution with
solar illumination. I: A simplified analysis of batch reactors, Chem. Eng. Sci., no 58, pp25872599
Sagawe, G., Brandi, R. J., Bahnemann, D. and Cassano, A. E. (2003b) Photocatalytic reactor for treating water pollution with
solar illumination. II: A simplified analysis for flow reactors, Chem. Eng. Sci., no 58, pp26012615
Sarria, V., Kenfack, S., Guillod, O. and Pulgarn, C. (2003) An innovative coupled solar-biological system at field pilot scale for
the treatment of biorecalcitrant compounds, J. Photochem. Photobiol. A: Chem., no 159, pp8999
Serpone, N. and Emeline, A. V. (2002) Suggested terms and definitions in photocatalysis and radiocatalysis, Int. J.
Photoenergy, no 4, pp91131
Serpone, N., Sauv, G., Koch, R., Tahiri, H., Pichat, P., Piccini, P., Pelizzetti, E. and Hidaka, H. (1996) Standardization protocol
of process efficiencies and activation parameters in heterogeneous photocatalysis: Relative photonic efficiencies, J.
Photochem. Photobiol. A: Chem., no94, pp191203
Srinivasan, M., Klausner, J. F., Jotshi, C. K. and Goswami, D. Y. (1997) Solar photocatalytic treatment of BTEX contaminated
groundwater, proceedings of the ISES 1997 Solar World Congress, August 1997
Srinivasan, M., Goswami, D. Y., Klausner, J. F. and Jotshi, C. K. (1998) Design, construction and testing of a solar
photocatalytic treatment facility for BTEX contaminated groundwater. Solar engineering 1998, proceedings of the ASME
Intern. Solar Energy, Albuquerque, NM, US, pp299304

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

ch-05

2/7/07

2:25 PM

Page 168

168 S. MALATO et al

Tothill, I. E. and Turner, A. P. F. (1996) Developments in bioassay methods for toxicity testing in water treatment, Trends Anal.
Chem., vol 15, no 5, pp178186
Welford, W. T. and Winston, R. (1978) The Optics of Non-Imaging Concentrators: Light and Solar Energy, Academic Press Inc.,
New York
Well, M., Dillert, R. H. G., Bahnemann, D. W., Benz, V. W. and Mueller, M. A. (1997) A novel nonconcentrating reactor for
solar water detoxification, J. Sol. En. Eng., no 119, pp114119
Wolfrum, E. I. and Turchi, S. (1992) Comments on reactor dynamics in the evaluation of photocatalytic oxidation kinetics, J.
Catal., no 136, pp626628
Wyness, P., Klausner, J. F., Goswami, D. Y. and Schanze, K. S. (1994a) Performance of nonconcentrating solar photocatalytic
oxidation reactors, Part I, J. Solar Energy Engineering, no 116, pp27
Wyness, P., Klausner, J. F., Goswami, D. Y. and Schance, K. S. (1994b) Performance of non-concentrating solar photocatalytic
oxidation reactors, Part II: Shallow pond configuration, J. Sol. En Eng., no 116, pp813
Zaidi, A. H., Goswami, D. Y. and Wilkie, A. C. (1995) Solar photocatalytic post-treatment of anaerobically digested distillery
effluent, proceedings of Am. Solar Energy Soc. Annual Conf., Minneapolis, MN, US, July, pp5156

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 130168

Ch-06

2/6/07

7:06 PM

Page 169

Solar-Hydrogen:
A Solid-State Chemistry
Perspective
J. Nowotny, T. Bak, L. R. Sheppard and C. C. Sorrell

Abstract
The present work considers the solid-state chemistry-related issues relevant to the
photo-electrochemical generation of hydrogen from water using solar energy. The focus
is on the properties of semiconducting photo-electrodes, which comprise the critical
component for increasing the efficiency of the conversion of solar energy into chemical
energy. The most important parameters of the photo-electrodes, which are essential for
the conversion and to be modified to this end, are the band gap and the Fermi energy.
The relationships between these properties and the materials properties that can be
experimentally determined conveniently (electrical conductivity, thermoelectric power
and work function) are outlined.
The present paper brings together the concepts of photo-electrochemistry with the
concepts of defect chemistry and solid-state electrochemistry. It is considered that the
Fermi energy (the chemical potential of electrons) at the surface of the photo-electrode
is the key quantity in the assessment of the reactivity of photo-electrode materials with
water. Consequently, the desired reactivity between the photo-electrode and water may
be achieved through appropriate modification of this quantity during processing or
subsequent treatment of the photo-electrode material. Further, the optimal photoelectrode for water photolysis with high solar energy conversion efficiency should
exhibit a band gap of ~2eV and be resistant to corrosion in water. Finally, the reactivity
between the photo-electrode and water must be considered in terms of both collective
properties, such as electronic structure, and local properties related to photocatalytic
active surface sites at which the photocatalytic reaction between water and the photoelectrode takes place. The leading candidate for such photo-electrodes is titania (TiO2-x),
which exhibits outstanding resistance to corrosion and photo-corrosion in aqueous
environments.
The optimal reactivity of titania with water, leading to maximal solar energy
conversion efficiencies, can be achieved through maximization of solar energy
absorption and minimization of energy losses due to recombination and charge transport.
This may be achieved through the modification of defect disorder and related properties
including charge transport, charge separation and electronic structure. Progress in
research on the determination of the relevant properties of titania essential to the

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

2/6/07

7:06 PM

170

Page 170

J. NOWOTNY, T. BAK, L. R. SHEPPARD AND C. C. SORRELL

performance of photo-electrodes, including defect chemistry and related electroactivity, is discussed. Significance, benefits and advantages of the solar-hydrogen
technology and cost-related estimates are briefly considered.

Keywords solar-hydrogen; water photolysis; titanium dioxide; photo-electrode; charge transfer; defect
chemistry; defect disorder; photo-electrochemistry; solar energy conversion

6.1 INTRODUCTION
6.1.1 GLOBAL WARMING
Increasing environmental pollution is becoming a vital global concern (Ooki, 1998). Figure
6.1, showing the increase in CO2 emissions over the period 17002000, provides a
dramatic illustration of the extent of the problem (Neftel et al, 1985; Friedli et al, 1986;
Russ, 1994). In order to reverse the global warming associated with CO2 emissions, it will
be imperative to reduce emissions of greenhouse gases, which are formed as a result of
the burning of fossil fuels. Therefore there is an urgent need to increase the production
of energy which is environmentally clean. This may be achieved through the development
of novel materials required for the conversion of renewable energy, such as solar energy,
into other types of energy, such as chemical energy (for example hydrogen). It is believed
that an understanding of materials interfaces is essential for the development of such
materials through interface engineering (Hirano et al, 1999).
Centre for Materials Research in Energy Conversion, UNSW

CO2 CONCENTRATION [ppm]

Ch-06

340

ATMOSPHERIC MEASUREMENTS
ICE PROBING

320

300

280
1800

1850

1900

1950

2000

YEAR
Sources: Neftel et al (1985), Friedli et al (1986) and Russ (1994)

FIGURE 6.1 Emission of carbon dioxide over the period 17002000 according to direct atmospheric
measurements and ice probing

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

Page 171

Solar-Hydrogen: A Solid-State Chemistry Perspective

171

GLOSSARY OF TERMS USED IN THIS PAPER


A
An, Ap
A
CPD
D
e
e
[e]
E
EA
EC
ED
EF
Eg
Eo
EV
fn(E)
fp(E)
h
h
[h]
Ho
Ir
[i]
J
k
K
n
Nn, Np
p
p(O2)
R
S
So
T
x
z

min
i
n
p
(e)
o(e)

Effective concentration of acceptors [atomic ratio]


Kinetic constants related to scattering of electrons and electron holes respectively
Singly ionized acceptor-type defects
Contact potential difference [V]
Singly ionized donor-type defects
Elementary charge [1.602 x 1019C]
Quasi-free electron
Concentration of electrons [atomic ratio]
Energy [eV]
Acceptor energy level [eV]
Energy of the bottom of the conduction band [eV]
Donor energy level [eV]
Fermi energy [eV]
Band gap width energy [eV]
Reference level corresponding to E=0 [eV]
Energy of the top of the valence band [eV]
Function describing the energy distribution of electrons
Function describing the energy distribution of electron holes
Plancks constant [6.626 1034J s]
Quasi-free electron hole
Concentration of electron holes [atomic ratio]
Standard free enthalpy [kJ mol1]
Incidence of solar irradiance [W m2]
Concentration of ionic charge carriers [m3]
Flux density (amount of some quantity flowing across a given area often unit area
perpendicular to the flow per unit time, e.g. number of particles) [m2s1]
Boltzmanns constant [8.6167 105eV K1; 1.3807 x 1023J K1]
Equilibrium constant
Concentration of electrons [m3]
Density of states for electrons and electron holes respectively [m3]
Concentration of electron holes [m3]
Oxygen partial pressure [Pa]
Universal gas constant [8.3144J mol1 K1]
Thermoelectric power [VK1]
Standard free entropy [kJ K mol1]
Absolute temperature [K]
Deviation from stoichiometry (in chemical formulae)
Valency
Difference
Electrical conductivity [1 cm1]
Electrical conductivity at the n-p transition [1 cm1]
Mobility of ionic charge carriers [cm2 V1 s1]
Mobility of electrons [cm2 V1 s1]
Mobility of electron holes [cm2V1s1]
Chemical potential of electrons [eV]
Standard chemical potential of electrons [eV]

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

2/6/07

172

R
s

[ ]

7:06 PM

Page 172

J. NOWOTNY, T. BAK, L. R. SHEPPARD AND C. C. SORRELL

Frequency [Hz]
Work function [eV]
Work function of reference electrode [eV]
Work function component related to the surface charge-induced potential barrier
[eV]
Thermovoltage (electrical potential) [V]
Concentration [atomic ratio]

The symbols for point defects are according to the Kroger-Vink notation (Kroger, 1974)

The protection of the environment is an international responsibility. However, since the


authors of the present work are based in Australia, the present paper will be focused on
the specific environmental issues of this country. According to recent reports by the
United Nations (United Nations Environmental Programme, 2002), Australia belongs to the
group of carbon-emitting countries that substantially exceed the target level established
by the Kyoto protocol, as shown in Figure 6.2. There is a growing awareness that
Australia, which is probably the cleanest inhabited continent on Earth, is not immune to
the consequences of global warming.

Centre for Materials Research in Energy Conversion, UNSW

ISS
ION

480

TU

AL

EM

460

440

T
RGE

AC

TOTAL CO2 EMISSION


IN MILLION TONNES

Ch-06

A
TO T
KYO

420

1990

1995

2000

2005

2010

YEAR
FIGURE 6.2 Emission of carbon dioxide in Australia over the period 19902000 and the emission level
according to the Kyoto target, according to the United Nations Environmental Programme, 2002

The list of problems related to pollution is long and has resulted in the understanding
that a reduction in greenhouse gas emissions must be addressed immediately.
Consequently, there is an increasing need to scale down the combustion of fossil fuels

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

Page 173

Solar-Hydrogen: A Solid-State Chemistry Perspective

173

and to develop alternative sources of energy that do not result in the emission of
greenhouse, pollutant or toxic gases and that are, therefore, environmentally friendly.
While there is a widespread perception that hydrogen is such form of energy, one must
ask the question: Is this perception correct?

6.1.2 IS HYDROGEN ENVIRONMENTALLY SAFE?


In general, increasing levels of urban pollution are largely attributed to the exhaust from
gasoline-powered cars. This has generated the belief that the introduction of hydrogenpowered cars will remedy this problem since the exhaust from such vehicles consists of
harmless water. However, it is important to consider that the use of hydrogen as a fuel
comprises two sources of emissions: those derived from its combustion and those derived
from its generation. While the former clearly involves the emission of H2O, the latter
depends on the method used to generate hydrogen in the first place. For example,
electrolysis of water using electricity generated from fossil-fuel-based power plants is
associated with CO2 emissions. The most common method of hydrogen generation,
methane reforming, also generates CO2 emissions when the methane molecule is split.
Consequently, it is essential that hydrogen is considered environmentally safe only when
there are no carbon emissions generated during both combustion and generation. In effect,
only hydrogen generated from sources of renewable energy are clean and environmentally
safe. Therefore, it is essential that efforts to introduce hydrogen-powered vehicles, already
at the experimental stage, should be accompanied by research to develop technologies to
generate environmentally clean hydrogen, such as solar, wind and tidal power.
While hydrogen generated from renewable energy is expected to be environmentally
safe, a concern is growing that unintended disposal of hydrogen to the atmosphere, due
to losses, may lead to unexpected consequences (Tromp et al, 2003; Prather, 2003). Such
emissions, and subsequent hydrogen oxidation in air, may lead to moistening and cooling
of the lower stratosphere and a decrease of stratospheric ozone. This hydrogen emission
has been projected to be of the order of 10 per cent or more. Therefore the evaluation of
the impact of hydrogen emission on climate changes is required.

6.1.3 HYDROGEN FUEL OF THE FUTURE


There is a growing consensus that hydrogen has the potential to supplement and possibly
replace fossil fuels for the production of energy by 20102020 (Thomas et al, 1998). It is
estimated that the potential market for hydrogen as a fuel is comparable to the present
markets for coal and gasoline combined. Therefore the availability of hydrogen will have
a major impact on the global energy scenario and the environment.
The major driving forces for the replacement of gasoline by hydrogen include:
Price: The price of hydrogen is expected to decrease while that of oil is expected to

increase.
Energy capacity: Hydrogen is the fuel with the highest energy capacity.
Amount of pollution: The absence of pollution during combustion (except nitrogen

oxides) in hydrogen-powered vehicles will lead to a significant reduction in urban


pollution.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

174

7:06 PM

Page 174

J. NOWOTNY, T. BAK, L. R. SHEPPARD AND C. C. SORRELL

Location of pollution: Even hydrogen produced from methane has an advantage in

that urban pollution will be reduced by removing power generation plants to less
populated areas.
Security of resources: Fossil fuel reserves are limited in quantity and they are
concentrated in politically uncertain regions.
Hydrogen generated from water using renewable sources of energy can resolve the issue
of the availability of a fuel that is 100 per cent environmentally safe.1 Consequently, the
development of hydrogen technologies using renewable sources of energy will reduce
the dominance of resource-rich nations in international energy markets. Such
technologies will provide these nations with increased energy security.
Substantial seed funding already has been allocated to research and development
specifically for solar-hydrogen technologies in the US, Japan and the European Union
(EU). These countries have recognized that it is in their best national interests not to be
left behind in the race to develop these technologies.
Recently, the Australian Government has entered the race to develop hydrogen-related
technologies by allocating funding through the Department of Industry, Tourism and
Resources to study the potential for an Australian hydrogen industry (Rand and Badwal,
2005) and to organize an international meeting on The Hydrogen Economy (1821 May
2003). The meeting aimed at mapping the future of hydrogen as an important potential
long-term source of energy for Australia and identifying Australias contribution to world
hydrogen developments (Hartley Management Group, 2003). The location of the meeting
site in Broome was not incidental since the Pacific coastline in this area is exposed to the
worlds largest tide (up to 11m) and so represents a potential for utilizing a substantial
amount of environmentally clean tidal energy.
The most critical issue in the development of hydrogen energy, and specifically the
technologies for the conversion of renewable energy into other forms of energy, such as
chemical energy (hydrogen), is the development of a special class of materials required
for efficient and clean conversion of energy. Development of these materials, which will
need to exhibit sophisticated functional properties, will require applying the most recent
progress in the science of materials interfaces and solid-state science. These issues were
addressed at the First International Conference on Materials for Hydrogen Energy held
in Sydney on 27 August 2004.

6.2 SOLAR-HYDROGEN
One of the most promising renewable energy technologies is the production of hydrogen
by water photolysis (Veziroglu, 1998 and 2000; Bockris et al, 1991; Bockris 1999; Bak
et al, 2002). The current predictions indicate that the production of hydrogen will
skyrocket by 2010 (Thomas et al, 1998). According to a recent comprehensive review the
method of photo-electrochemical water decomposition using solar energy is the most
promising method for the generation of hydrogen (Bak et al, 2002). The key functional
element in photo-electrochemical devices which allows the exploitation of this technology
is the photo-electrode (PE). However, there is a need to enhance the performance of this
element in order to achieve the level required for commercialization of the device

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

Page 175

Solar-Hydrogen: A Solid-State Chemistry Perspective

175

technology. Therefore the primary goal of research and development activities in many
research groups is the development of photo-electrodes with maximal efficiencies of
conversion of solar energy into chemical energy, viz hydrogen.

6.3 THE CONCEPT OF SOLAR-HYDROGEN GENERATION


The pioneers of solar-hydrogen include Fujishima et al (Fujishima and Honda, 1972;
Fujishima et al, 1975), who first reported photo-assisted water electrolysis using rutile as
the photo-anode of a photo-electrochemical cell (PEC). Both Fujishima and Honda also
showed for the first time that rutile, unlike other photo-sensitive materials, is resistant to
both corrosion and photo-corrosion in the aqueous environment.
The scientific foundation for this process was established mainly by Bockris (Bockris,
1980, 1999 and 2003; Bockris and Uosaki, 1976 and 1976a; Bockris et al, 1981), Gerischer
(1972) and Chandra (1985). In fact, Bockris first introduced the term hydrogen economy
back in the 1960s (Armstrong, 1996). His activity in this field encompasses a sustained
period of work spanning over four decades.
Fujishima et al (Fujishima and Honda, 1972; Fujishima et al, 1975) reported the
concept and performance of a PEC involving a TiO2-x single crystal as photo-anode and Pt
as cathode. They showed that solar irradiation of TiO2-x results in the evolution of oxygen
at the anode and hydrogen at the cathode. Their reports, which are cited extensively in
the authors recent review (Bak et al, 2002), resulted in an ongoing search for candidate
materials and design models for photo-electrodes. However, the lack of success in
increasing the electrochemical conversion efficiency (ECE) of TiO2-x above ~1 per cent
during the last 30 years has resulted in scepticism concerning the potential to increase
the ECE above ~10 per cent, which is the level that the US Department of Energy
(USDOE) considers to be the benchmark for commercialization (Service, 2002). However,
the recent report by (Khan et al, 2002), who obtained an increase in the ECE for
commercial titania to 8.5 per cent by exposing titania to a natural gas flame, revived
expectations of commercialization of this technology. Khan et al (2002) reported that
carbon incorporated into the TiO2-x lattice is responsible for the reduction of the band gap
from ~3eV to 2.32eV and an observed increase of ECE.
The concept of photo-electrochemical hydrogen generation is based on the splitting
of the water molecule on the surface of the photo-electrode using photo-energy.
Hydrogen is evolved at the cathode and oxygen produced at the photo-anode. Figures 6.3
and 6.4 show a simple photo-electrochemical device for water photolysis and its electrical
circuit respectively. Figure 6.5 shows the electrochemical chain of the photoelectrochemical device based on TiO2-x as the photo-anode. The experimental approaches
in the assessment of the electrochemical properties of the chain include the
determination of the effect of light on open cell voltage (electromotive force, EMF) and
currentvoltage characteristics. The latter can be used for evaluation of the energy
conversion efficiency (ECE).
As might be expected, ECE depends not only on the properties of the photo-electrode
but also on the construction of the PEC and its equivalent circuit. Although most of the
studies reported so far provide the data on ECE, they fail to provide the equivalent circuit.
This results in complications in the verification of the ECE data reported in the literature.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

176

7:06 PM

Page 176

J. NOWOTNY, T. BAK, L. R. SHEPPARD AND C. C. SORRELL

Centre for Materials Research in Energy Conversion, UNSW

O2
H2

H2O 2e + O2 + 2H

2H + 2e H2

PHOTO-ANODE

CATHODE
+

H2O H + OH

AQUEOUS ELECTROLYTE

FIGURE 6.3 Electrochemical cell showing the principle of electrochemical hydrogen generation through
water photolysis using solar energy and related reactions
R
e
h

EF
+

H /H2
EC
1.23 eV
Eg
O2 /H2O
EV

H2O + 2 h 2 H + O2
PHOTO-ANODE
(SEMICONDUCTOR)

H + e H2
AQUEOUS
ELECTROLYTE

CATHODE
(METAL)

FIGURE 6.4 Electrical circuit representing the photo-electrochemical device formed by a semiconducting
photo-anode and metallic cathode

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

Page 177

Solar-Hydrogen: A Solid-State Chemistry Perspective

177

Centre for Materials Research in Energy Conversion, UNSW

O2

h
1

2
e

Pt

H2

TiO2-x

H+

AQUEOUS SOLUTION

Pt

e
H2O + 2h 2H+ + O2

2H+ + 2e H2

R
V
FIGURE 6.5 The electrochemical chain of TiO2-x-based photo-electrochemical device and related
charge transfer

An essential part of the device for hydrogen generation using solar energy is the
semiconducting photo-electrode, which consists of an n-type semiconductor. Exposure to
light results in intrinsic ionization over the band gap, leading to the formation of an
electronhole pair:
(1)
The excess minority charge carriers (electron holes) gives rise to a photo-voltage, resulting
in splitting of the water molecule into hydrogen ions and gaseous oxygen:
(2)
Gaseous oxygen evolves at the photo-anode and the hydrogen ions migrate to the
cathode through the internal circuit (electrolyte), where the reduction of hydrogen ions to
gaseous hydrogen takes place:
(3)
The overall reaction of the photo-electrochemical process is as follows:
(4)
The preceding reaction takes place when:
the electromotive force (EMF) of the photo-electrochemical cell is 1.23V (in practice

this value is higher due to overvoltage); and

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

178

Page 178

J. NOWOTNY, T. BAK, L. R. SHEPPARD AND C. C. SORRELL

the energy of the photons is equal to or greater than the band gap of the

photo-electrode.
The economic feasibility of photo-electrochemical hydrogen generation technology
depends principally on meeting the following criteria:
the ECE must be 10 per cent, according to the USDOE (Service, 2002), where ECE

is defined as the ratio of the amount of the energy output, equal to the generated
chemical or electrical energy, to the energy input, equal to the solar energy striking
the photo-electrode;
the photo-electrode must be durable and resistant to corrosion and photo-corrosion
in aqueous environments;
the overall cost of the device and infrastructure must be at least comparable to the
value of the produced hydrogen, if not lower or substantially lower; and
the device should be either maintenance free or low maintenance.

6.4 MATERIALS PROPERTY REQUIREMENTS FOR PHOTO-ELECTRODES


The search for photo-electrodes that exhibit high ECEs has revealed two basic problems:
1 photo-anodes that possess good corrosion resistance to water, such as TiO2 (Bak
et al, 2002; Fujishima et al, 1999), exhibit low ECEs of 1 per cent, with the notable
exception of the work of Khan et al (2002); and
2 photo-anodes that exhibit high ECEs, such as valence semiconductors (El Zayat et al,
1998; Licht et al, 2000), where the ECE can rise to > 20 per cent, possess limited
corrosion and photo-corrosion resistance, resulting in short lifespans.
Reports on the development of PEs (Onishi et al, 1975; Mavroides et al, 1975; Memming,
1980; Soliman and Seguin, 1981) have concentrated on the effects of selected individual
issues on the ECE. Progress in this area requires the identification and optimization of all
of the materials properties that have an impact on the ECE, including defect chemistry
and related electronic structure, semiconducting properties, composition,
nonstoichiometry, and microstructure (Bak et al, 2002).
Most photo-electrodes that exhibit sustainable performance are fabricated from oxide
materials (Bak et al, 2002). However, another group of materials used for photo-electrodes
are valence semiconductors, such as GaAs (El Zayat et al, 1998; Licht et al, 2000).
Extensive analysis of the property requirements of photo-electrodes (Bak et al, 2002;
Nowotny et al, 2005) indicates that the following properties are essential (the physical
meanings of these properties have been outlined elsewhere):
Band gap: The band gap (the forbidden energy gap) is one of the most critical

properties of PEs. In the case of water photolysis, the band gap should remain
between 1.23eV (theoretical lower limit) and ~2eV (practical upper limit). The
electronic structure and related band gap of semiconductors based on
nonstoichiometric compounds, such as TiO2-x, are very sensitive to defect disorder

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

Page 179

Solar-Hydrogen: A Solid-State Chemistry Perspective

179

(Khan et al, 2002). On the other hand, in the case of single-crystal Si, used for
photovoltaic (PV) cells, the band gap is fixed at 1.1eV and it is a material property.
Therefore, the electronic structure of nonstoichiometric compounds may be modified
through the control of the defect chemistry.
Electrical resistance: Efficient charge transfer within the circuit of a
photo-electrochemical cell requires the electrical resistance to be at its minimal level.
Consequently, this requires the resistance of all elements of the circuit, including the
photo-electrode, to be minimized. The resistivity of the photo-electrode can be
minimized through the imposition of suitable defect disorder. Concurrently, the
materials engineering must be undertaken so as to retain or achieve the optimal band
gap.
Schottky barrier: The Schottky-type barrier is needed for effective charge separation
(prevention of recombination of the electron-hole pair). The consequence of this
separation is that electrons are transported into the n-type solid while electron holes
are transported toward the photo-electrode/electrolyte interface, resulting in the
formation of a p-type surface layer (Bak et al, 2002). In the case of nonstoichiometric
compounds used as photo-electrodes, the potential barrier may be formed either by
the imposition of a defect concentration gradient during processing or through the
effect of segregation leading to enrichment of the surface layer. These have the
effect of imposing a surface Fermi energy at a level different from that of the bulk
phase.
Microstructure: The properties of polycrystalline solids are well known to be very
sensitive to microstructure. The effects of the microstructure on properties may be
considered in terms of, inter alia, grain size and the concentration of grain
boundaries. The effect of the surface microstructure may be considered in terms of
the local properties of the linear defects formed at the intersections between the
grain boundaries and the external surface. So far, little is known of the effects of
these defects on the ECE. Further, the effect of the grain size (bulk/surface) may be
considered in terms of the interfacial energy and its impact on the local defect
chemistry and related charge transfer across the photo-electrode/electrolyte
interface.
Flat band potential: The flat band potential should be higher than the redox
potential of the H+/H2 couple. This potential may be engineered by the imposition of
specific surface-to-bulk composition gradient and related potential gradient. The flat
band potential was described by Bak et al (2002).
Helmholtz potential barrier: The Helmholtz potential barrier, which is formed at the
interface between the photo-electrode and the electrolyte, plays an important role in
retarding/accelerating the charge transfer within the electrochemical chain shown in
Figure 6.5 (Bak et al, 2002). Therefore its value is determined by the surface
compositions of the photo-electrode and the composition of the electrolyte.
Nonstoichiometry and related defect disorder: Although it is well known that
nonstoichiometry and point defects have a substantial impact on material properties,
so far little is known on the effect of defect disorder on electroactivity and
photo-sensitivity. Awareness is growing that properties of photo-electrodes based on

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

180

7:06 PM

Page 180

J. NOWOTNY, T. BAK, L. R. SHEPPARD AND C. C. SORRELL

nonstoichiometric compounds, such as TiO2-x, can be closely related to defect


disorder and the related electronic structure. Consequently, the desired properties of
the photo-electrodes may be achieved through the modification of the defect
disorder. The recent report by Khan (Khan et al, 2002) confirms that the defect
disorder has a significant impact on the electronic structure and related
photo-sensitivity of TiO2.
Corrosion resistance: Most photo-active materials are not resistant to corrosion in
aqueous environments, which results in the degradation of their properties. However,
little is known of effective corrosion inhibition in these materials. Consequently,
naturally occurring and stable oxides are the most promising materials for
photo-electrodes owing to their ability to meet this criterion. Since TiO2-x is among
the most stable of all oxides in aqueous environments at ~pH = 7 (Fujishima et al,
1999), it is the leading candidate.
Segregation-induced surface barrier: Segregation results in the formation of
concentration gradients in the surface layer of photo-electrodes that may affect
the charge transfer across the electrode/electrolyte interface. So far, knowledge
about segregation and its impact on the function of the photo-electrode is still
limited. However, the segregation-induced potential barrier in oxide materials may
be of the order of 1V (Nowotny, 1991), so its effect on charge transfer may be
crucial.
The list of the materials properties that are essential to achieve high-performance photoelectrodes is longer than that shown above. Some properties, such as band gap, electrical
resistance, flat band potential and microstructure are probably the most frequently used
in studies of photo-electrodes. The most extensive outline of these properties are
reported elsewhere (Nowotny et al, 2005). However, because there is no simple
relationship between these properties and their impact on the ECE, a multi-variant
approach is required in the development of photo-electrode materials rather than monovariant investigations.
In summary, there is a considerable experimental and theoretical body of evidence
that indicates that TiO2-x is the best candidate for highefficiency photo-electrode PEs.
However, there is a need to optimize the electronic structure in order to reduce the
effective band gap required for ionization. At the same time there is a need to address
other properties that have an impact on energy losses, such as the energy losses related
to recombination and the charge transfer.
Subsequent sections outline the issues related to the electronic structure of TiO2-x,
including the definition of basic quantities describing electronic structure and the
experimental approaches of their determination.

6.5 ELECTRONIC STRUCTURE


6.5.1 DEFINITION OF BASIC TERMS
The electronic structure of semiconducting solids may be described in terms of diagrams
of the density of states and represented in the form of plots of the number of electron
energy levels as a function of energy (Kofstad, 1972; Green, 1986).

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

Page 181

Solar-Hydrogen: A Solid-State Chemistry Perspective

181

The key energy quantities describing the electronic structure of solids and having an
impact on the photo-sensitivity include:

energy of the bottom of the conduction band, EC;


energy of the top of the valence band, EV;
energy levels of the donors and acceptors within the band gap, ED and EA; and
Fermi energy, EF.

It is well known that the width of the forbidden gap (Eg = EV EC) is the key property
controlling ECE (Bak et al, 2002; Chandra, 1985; Gerischer, 1977 and 1997; Seraphin,
1979). Another quantity that has a crucial impact on the reactivity of photo-electrodes is
the Fermi energy.
There is a close relationship between the Fermi energy and the lattice defect disorder
of semiconducting oxides. This involves point defects that form either donors, such as
oxygen vacancies and cation interstitials, or acceptors, such as cation vacancies. The
Fermi energy is determined by the location of the associated energy levels within the
band gap and the related density of states. Doping of metal oxides with aliovalent ions
results in the formation of either donors or acceptors located within the band gap. The
positions of these levels are determined by the incorporation mechanisms of the specific
ions into the lattice of the host semiconducting oxides. The densities of states are
controlled by the concentration of dopants within their solubility ranges.
The mobility of electronic charge carriers within the bands or mid-gap levels is
determined by the density of allowed states forming the bands and the donor and acceptor
states. Specifically, at high and low density of states, the charge carriers are quasi-free or
localized respectively. At low density of states, when these levels are isolated, the charge
transport occurs by the hopping mechanism. Increase of the density of states to very high
values may, in extreme cases, lead to metallic conduction (Sheppard, 2004).
Detailed analysis of the electronic structure of semiconducting solids is beyond the
scope of the present work, which is limited to a brief analysis of the following two
quantities:
1 band gap; and
2 Fermi energy.
These two quantities are selected for consideration for the following reasons:
They are the most important in the assessment of photo-sensitivity and reactivity at

the water/photo-electrode interface.


They may be assessed relatively easily by measurement of electrical properties

commonly available in materials science laboratories.

6.5.1.1 Band gap


As shown in Figure 6.6, the band gap Eg is the energy difference between the bottom of
the conduction band, EC, and the top of the valence band, EV. The band gap width is one

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

182

7:06 PM

Page 182

J. NOWOTNY, T. BAK, L. R. SHEPPARD AND C. C. SORRELL

of the most critical factors that control the properties of intrinsic semiconductivity in
general and photo-sensitivity in particular. The band gap is the energy difference through
which heat- and photon-induced ionization occur. This ionization results in the formation
of quasi-free charge carriers (electrons and electron holes) that take part in the charge
transfer within the electrode and its electroactivity (certain density of states is required for
the charge carriers to be quasi-free, alternatively they are isolated). In other words, only
the photons of the energy equal to or larger than that of the band gap may be absorbed
and be available for conversion of the electromagnetic energy into another form of energy,
such as electrical and chemical. The critical impact of Eg on photo-sensitivity and ECE has
been discussed previously (Bak et al, 2002).

Centre for Materials Research in Energy Conversion, UNSW

E0=0
O
1

EC

EF
Eg

S=-EF

EC

EF

O
O

EV
EV

DISTANCE FROM THE SURFACE


FIGURE 6.6 Band model of n-type semiconductor without surface charge (left) and involving surface
charge (right)

6.5.1.2 Fermi energy


Fermi energy (EF) is a specific parameter of the Fermi-Dirac statistics of electrons at which
50 per cent of states are ionized. These statistics describe the distribution of electrons
with respect to their energy levels (Kofstad, 1972):

fn ( E ) = 1 + exp ( Ei - E F ) / kT

-1

(5)

where fn(E) is the probability of occupation of the energy level Ei. The similar
expression for fp(E) describes the distribution of electron holes:
(6)

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

Page 183

Solar-Hydrogen: A Solid-State Chemistry Perspective

183

The EF may be referenced quantitatively by comparison to the energy of either the


bottom of the conduction band or the top of the valence band. The EF level may also be
referenced to the energy level where electrons are free of electrostatic interactions, which
is outside the solid surface. The energy difference between these two levels is termed the
work function. A schematic representation of the flat band model (free of surface charge)
for an n-type semiconductor, showing the EF and the work function , is shown on the left
side of Figure 6.6.
In solid-state physics, the EF is used to characterize semiconducting properties of
solids with respect to their ability to donate or accept electrons. The EF is also a key
quantity in the evaluation of the charge transfer accompanying chemical reactions, such
as catalysis. However, in chemistry, chemical equilibria are typically described in terms of
the chemical potentials of reactants and reaction products. In this case, the EF is
equivalent to the chemical potential of electrons, which describes the electroactivity of
the solids and so their ability either to accept or donate electrons during chemical
reaction. Knowledge of the value of the chemical potential is essential for quantitative
evaluation of the reactivity of solids whenever charge transfer takes place.
The reactivity (viz photo-reactivity) of solids involved in photo-induced processes is
no exception. Therefore, it is clear that the EF appropriate for photo-reaction of the photoelectrode with water must be optimized in order to achieve the desired photo-reactivity.
The semiconducting properties of the photo-electrode and thus the resultant chemical
potential are established during processing. Therefore it is useful to monitor the
chemical potential of electrons of the solid during processing. The desired chemical
potential of electrons can be established by conducting the processing in a controlled
manner.
The Fermi-Dirac statistics of electrons and electron holes, represented by Equations 5
and 6, can be approximated by the Maxwell-Boltzmann statistics when (Kofstad, 1972):
(7)
and
(8)
where EC and EV represent the energies of the bottom of the conduction band and the top
of the valence bands respectively, k is the Boltzmann constant, and T is the absolute
temperature. Then for a non-degenerated semiconductor there is a simple relationship
between the EF and the concentration of electronic charge carriers:

(9)

(10)
where n and p denote the concentrations of electrons and electron holes respectively, and
Nn and Np denote the densities of states of electrons and electron holes respectively.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

2/6/07

184

7:06 PM

Page 184

J. NOWOTNY, T. BAK, L. R. SHEPPARD AND C. C. SORRELL

Equations 9 and 10 may be used for the determination of the basic electronic structure
when the concentrations of the electronic carriers are known (or vice versa). In the case
of a flat-band structure, corresponding to semiconductors that are free of surface charge,
the value of EF is independent of the distance from the surface. Alternatively, the presence
of a surface charge, caused, for example, by the adsorption of charged species, results in
a difference in the Fermi energies of the surface and the bulk by the energy component
s, as shown on the right side of Figure 6.6.
There is a growing body of empirical data that indicate that the EF in the bulk is quite
different from that at the surface (Nowotny, 1991). This difference is due to segregationinduced concentration gradients at interfaces (surfaces and grain boundaries) and the
resultant electrical potential gradients (Adamczyk and Nowotny, 1986). These gradients
have a significant impact on the functional properties of solids, including photo-sensitivity.
Therefore quantitative analysis of surface versus bulk properties is a critical issue in the
interpretation of photo-induced effects. Accordingly, the imposition of specific surface
versus bulk properties during the processing of photo-sensitive materials is required for
the achievement of desired photo-sensitivity.
Centre for Materials Research in Energy Conversion, UNSW

ENERGY

Ch-06

PHOTON
PENETRATION
THICKNESS

(EF)n

EC
EF

(EF)p
EV

DISTANCE FROM THE SURFACE


Source: Gerischer (1977)

FIGURE 6.7 Effect of light in splitting of the Fermi energy level into two quasi-levels corresponding to
electrons and electron holes for an n-type semiconductor

The preceding discussion applies to solids in thermal electronic equilibrium. However,


when light is applied, then the system is in a dynamic state. Then EF may then be
considered in terms of EFs splitting into two quasi-Fermi levels, (EF)n and (EF)p, shown in
Figure 6.7 (Gerischer, 1977).

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

Page 185

Solar-Hydrogen: A Solid-State Chemistry Perspective

185

6.5.2 EFFECT OF LIGHT ON FERMI ENERGY


The light-induced split of the Fermi energy is the driving force for the charge transfer
within the photo-electrochemical cell. The split results from the change in the
concentration of electronic charge carriers within the valence band and the conduction
band as a result on the light-induced ionization. The quasi-Fermi energy levels will be
determined by increased concentrations of electrons and electron holes, due to
ionization, by the respective components n and p:

(11)

(12)
Therefore:
(13)

(14)
where EF is the change of the Fermi energy induced by light and no and po denote the
concentration of the electronic charge carriers before the light is applied.
The effect of light on (EF)n in an n-type semiconductor such as reduced TiO2 is
relatively small because the component no is relatively large compared to n. However,
the effect of light on (EF)p in the n-type material is substantial because the component p
is large compared to po, which, in the dark, corresponds to the minority charge carriers.
Consequently, application of the light results in a substantial increase of the oxidation
potential represented by (EF)p leading to the shift of the equilibria represented by
Equations 24 to the right. Then the rate of these reactions is determined by the light
intensity and the resulting (EF)p level.

6.5.3 BASIC RELATIONSHIPS


There have been an enormous number of reports on the application of electrical methods
to investigate the properties of materials, in particular their functional properties (Kofstad,
1972). The electrical properties of semiconducting solids are very sensitive to the
structure and chemical composition (including nonstoichiometry). The direct assessment
of the structure and composition of solids can be time-consuming, complicated and
expensive. The determination of these factors during processing at elevated temperatures
is much more difficult if at all possible. However, measurements of the electrical
properties at both room and elevated temperatures are more straightforward. The
disadvantage of the use of electrical properties for the assessment of structure and
composition is that these properties can be assessed only indirectly. However, there are
three substantial advantages to materials assessment through the measurement of
electrical properties:

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

186

Page 186

J. NOWOTNY, T. BAK, L. R. SHEPPARD AND C. C. SORRELL

1 they are extremely sensitive;


2 they require relatively uncomplicated analytical infrastructure; and
3 they can be used not only for assessment of structure and composition but also for
in situ monitoring of changes to them during processing.
The electrical properties used most frequently for the assessment and monitoring of
semiconducting solids include (Kofstad, 1972; Nowotny, 1997):

electrical conductivity;
thermoelectric power (Seebeck coefficient);
electrical conductivity and thermoelectric power measured simultaneously; and
work function.

The following sections discuss the relationships between these three electrical
properties, which can be determined experimentally, and the quantities describing the
electronic structure, Eg and EF.

6.5.3.1 Electrical conductivity


The electrical conductivity, , includes the conductivity components of all charge carriers
taking part in conduction (electrons, electron holes and ions):
(15)
where:
(16)
(17)

(18)
where e is the electron charge, is the mobility, n, p and [i] are the concentrations of
electrons, electron holes and ions respectively, z is the charge of ionic species, and l is
their number. Therefore:
(19)
The temperature dependence of the electrical conductivity can be used to determine the
electronic structure of a solid (Kofstad, 1972). Becker and Frederikse (1962) showed that
there is an explicit relationship between the Eg of a semiconductor that exhibits an np
transition, such as titania, and the temperature dependence of the minimal value of the
at the transition:
(20)

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

Page 187

Solar-Hydrogen: A Solid-State Chemistry Perspective

187

where min is at the np transition point, which then assumes a minimum. Solids,
particularly nonstoichiometric compounds, can be well defined when they are in a state
of gas/solid equilibrium. Consequently, well-defined data for the should be determined
at elevated temperatures and under controlled gas phase compositions. In the case of
oxide materials, oxygen is the most important component of the gas phase under which
the material is annealed or equilibrated since oxygen is a component of the lattice. Thus
its activity in the gas phase results in the imposition of its activity in the lattice and the
related defect disorder. The determination of the electrical conductivity of oxide materials
has been described elsewhere (Nowotny, 1997).
As seen in Equations 1619, the determination of the concentration of charge carriers
from the electrical conductivity data is possible when the mobility terms are known (or
vice versa). The evaluation of the mobility terms corresponding to electronic charge
carriers requires the application of more sophisticated experimental and theoretical
approaches (Kofstad, 1972). The electrical conductivity measurements may also be
applied for monitoring the gas/solid equilibration kinetics for a metal oxide/oxygen system
(Kofstad, 1972).
While the electrical conductivity is the most commonly measured electrical property,
its physical meaning is complicated because it incorporates both concentration and
mobility terms (Kofstad, 1972). The commonly assumed interpretation of the electrical
properties of materials is based on the assumption that changes in the electrical
conductivity are determined by the concentration term alone, while the mobility term
remains independent of concentration. Unfortunately, this assumption frequently is not
valid, which is the reason why studies based solely on electrical conductivity may lead to
a misleading picture of semiconducting properties.

6.5.3.2 Thermoelectric power


Although the experimental determination of thermoelectric power, S, is more complicated
than that of electrical conductivity, thermoelectric power is independent of the mobility
term and thus is determined solely by the concentration term (Nowotny, 1997). The
thermoelectric power is the temperature coefficient related to the electrical potential
difference (termed the Seebeck voltage or thermovoltage) generated along a
temperature gradient T imposed on a solid. Knowledge of both and T are required
for the determination of the thermoelectric power or thermopower (Nowotny, 1997):
(21)
Since S may vary with temperature, correct determination of the S requires measurement
of the thermovoltage at temperature gradients as small as possible. Assuming the
Maxwell-Boltzmann statistics, which may be applied at the limitations imposed by
Equations 7 and 8, the S may be related directly to the EF (Nowotny, 1997):
(22)

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

2/6/07

188

7:06 PM

Page 188

J. NOWOTNY, T. BAK, L. R. SHEPPARD AND C. C. SORRELL

(23)
where Sn and Sp denote the Seebeck coefficient components corresponding to electrons
and electron holes respectively, and An and Ap are kinetics constants related to the
scattering of electrons and electron holes respectively.
Experimental determination of Sn and Sp is straightforward when one type of charge
carrier is predominant. The separation of S into the two components Sn and Sp within the
np transition regime is more complicated. Methods for the determination of the
components of S in terms of different charge carriers are described elsewhere (Wagner,
1972).
Combination of Equations 9 and 10 with Equations 22 and 23 may be used to establish
a direct interrelationship between S, which is a measurable quantity, and the
concentration of electronic charge carriers, which are essential to the assessment of the
charge transfer.
Experimental determination of thermoelectric power at elevated temperatures and
under controlled gas phase compositions has been described previously (Nowotny, 1997).

6.5.3.3 Thermopower versus electrical conductivity


The simultaneous determination of both and S is a very useful experimental approach
in assessing the charge transport. Both and S may be used for the determination of
Centre for Materials Research in Energy Conversion, UNSW

min

eB

k ln10

lg 2

k ln
4e B - 1
2B -
e
k

2B

Sp

THERMOPOWER, S

Ch-06

Sn

log
Source: Jonker, 1968; Nowotny, 1997

FIGURE 6.8 The schematic Jonker plot of thermoelectric power vs log showing the critical parameters
(meanings of the symbols are explained in text)

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

Page 189

Solar-Hydrogen: A Solid-State Chemistry Perspective

189

semiconducting properties of metal oxides that exhibit measurable np transitions


according to the analysis proposed by Jonker (1968). The concept of this analysis is based
on determination of the key points of the Jonker plot, which consists of versus log S.
This plot, which exhibits a pear-like shape, is shown in Figure 6.8 and reveals that the
coordinate system used is described by the characteristic parameters min, B and D
(Jonker, 1968; Nowotny, 1991). Quantitatively, these parameters are defined in Equations
2426:

(24)
where:
(25)
and
(26)
where n and p are the mobilities of the electrons and electron holes respectively. The
parameters B and D, which can be determined from the Jonkers formalism, are directly
related to the band gap and the ratio of the mobility terms respectively.
Since the Jonker analysis allows determination of the mobilities, then Equation 19 may
be used for the determination of the concentration of charge carriers from the electrical
conductivity data. Also Equation 25 allows direct determination of the Eg when the
kinetics terms An and Ap are known.
A specific feature of the analysis described by Equations 2426 is that knowledge of
the p(O2) under which the oxide specimen is equilibrated is not required. However, it is
critical that both and S are determined in equilibrium under identical p(O2).
The simultaneous determination of both the and S of oxide semiconductors is
possible using a hightemperature Seebeck probe, which has been described previously
(Nowotny, 1991). A prototype of this instrument is available in the Centre for Materials
Research in Energy Conversion, University of New South Wales (UNSW).

6.5.3.4 Work function


According to the band model of semiconducting materials, as shown in Figure 6.6, the
work function is the energy difference between the reference energy level (E = 0) and
the Fermi energy level EF:
(27)
Therefore:
(28)
Consequently, the is a direct measure of the EF at the surface. Therefore
measurements may be used for monitoring changes in EF during processing. The most

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

190

Page 190

J. NOWOTNY, T. BAK, L. R. SHEPPARD AND C. C. SORRELL

widely applied method for the determination of work function changes of compounds at
elevated temperatures and under controlled gas phase environment is based on the
determination of the contact potential difference (CPD) (Nowotny, 1997):
(29)
where R is the work function of the reference material. Pt has been used for the
reference material in measurements of work function changes as a function of p(O2).
Therefore knowledge of the reference energy level allows determination of the EF at the
surface:
(30)
The changes in oxide materials at elevated temperatures (up to 1300K) in gas phases of
controlled oxygen activities using a high-temperature Kelvin probe have been described
previously (Nowotny, 1991). A prototype of this instrument is available in the Centre for
Materials Research in Energy Conversion, UNSW. This unique surface-sensitive tool may
be used for:
determination of the chemical potential of electrons; and
in situ monitoring of the changes in the chemical potential of electrons during the

processing of oxides at elevated temperatures enabling the establishment of


desired nonstoichiometry and related defect disorder through either imposition of
specific oxygen activities or incorporation of aliovalent ions (forming donors and
acceptors).

6.6 WHY TITANIA?


There are many candidates for photo-electrode materials for solar-hydrogen. While some
materials are known to demonstrate outstanding photo-sensitivity (Licht et al, 2000),
these are not suitable for commercial applications because they have limited resistance
to corrosion and photo-corrosion in aqueous environments. Consequently, it is essential
to consider all of the requirements that the photo-electrode must meet. TiO2-x has a unique
set of properties that make it the most promising candidate for photo-electrodes. The
relevant considerations are as follows:
TiO2-x is reactive with both light and water, which may be attributed to the ease

with which the Ti ions (Ti3+ and Ti4+) alter their valence (Kofstad, 1972; Matzke,
1981).
TiO2-x has excellent resistance to chemical and photochemical corrosion in aggressive
aqueous environments (Bak et al, 2002).
The properties of TiO2-x can be altered by varying the defect chemistry and associated
electronic structure through the introduction of aliovalent ions and the alteration of
the concentration of intrinsic defects.
TiO2-x is substantially less expensive than other photo-sensitive materials.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

Page 191

Solar-Hydrogen: A Solid-State Chemistry Perspective

191

The different processing technologies for TiO2-x production are relatively

straightforward and uncomplicated compared to those required for valence


semiconductors.
The economic feasibility of photo-electrochemical generation of solar-hydrogen requires
the energy conversion efficiency (ECE) to be increased from current levels of <1 per cent
to 10 per cent, while ensuring high durability. The main methods that have been used to
attempt to raise the ECE of TiO2-x consist of the following:
sensitization;
reduction of the band gap; and
reduction of internal energy losses.

6.6.1 SENSITIZATION
6.6.1.1 Dye sensitization
The function of dye sensitizers is to increase the absorption of solar energy through their
application as coatings to the external surface of the TiO2-x electrode (El Zayat et al, 1998;
Memming, 1980; ORegan and Gratzel, 1991; Bach et al, 1998). However, they have the
significant disadvantage that they have limited resistance to corrosion and photocorrosion in water compared to that of TiO2-x (Bak et al, 2002). The second problem is that
only the dye molecules in close proximity to the interface, corresponding to approximately
one monolayer, are efficient.2 This type of sensitizing has been already applied
commercially (AGO, 1999).

6.6.1.2 Metal sensitization


Fine particles and agglomerates of metals have been attached to the surfaces of TiO2-x
grains leading to an increase in photo-sensitivity (Reisfeld et al, 1988; Zhao et al, 1996).
Although this technology has resulted in limited increases in the photo-sensitivity, it has
been applied mainly to water treatment applications.

6.6.2 REDUCTION OF THE BAND GAP


Figures 6.9 and 6.10 show the solar spectrum in terms of the number of photons versus
photon energy and radiation energy versus wavelength respectively. The integral of the curve
in Figure 6.9, corresponding to the flux of photons, shows the regimes associated with
commercial TiO2-x [J(TiO2)] and reduced-band-gap TiO2-x [J(RBGT)]. The integral of the curve in
Figure 6.10 corresponds to the solar irradiance available for conversion for commercial TiO2-x
[Ir(TiO2)] and reduced-band-gap TiO2-x [Ir(RBGT)]. Accordingly, the RBGT TiO2-x exhibits
substantially enhanced photo-sensitivity compared to that of commercial TiO2-x.
The ramifications of these data have been discussed previously by the authors (Bak et
al, 2002). The recent work reported by Khan et al (2002) indicates that the formation of
point defects, attributed to carbon incorporation into TiO2-x, apparently results in the
formation of structural defects and the resultant modification of the band gap. This work
forcibly provides the message that defect chemistry is a crucial issue in the engineering
of the electronic structure of TiO2-x for the photolysis of water.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

2/6/07

Page 192

J. NOWOTNY, T. BAK, L. R. SHEPPARD AND C. C. SORRELL

NUMBER OF PHOTONS [s m eV ]

192

7:06 PM

Centre for Materials Research in Energy Conversion, UNSW

21

410

21

310

THEORETICAL ENERGY RANGE (Eg = 1.23eV)

21

210

J (RBGT)
J (TiO2)

21

110

TiO2 RANGE (Eg = 3eV)


1

PHOTON ENERGY [eV]


FIGURE 6.9 Solar energy spectrum in terms of the number of photons vs their energy, showing the flux
density for both commercial titania and reduced-band-gap titania denoted by J(TiO2) and J(RBGT) respectively

PHOTON ENEGY [eV]


2.0

1.0

0.5
Centre for Materials Research in Energy Conversion, UNSW

RADIATION ENEGY [kW/m nm]

5.0 3.0

1.23 eV

1.5

-2

2
1.0

1 - Ir (TiO2)
2 - Ir (RBGT)
1

0.5

0.0

VISIBLE

Ch-06

0.5

1.0

1.5

2.0

2.5

3.0

WAVELENGTH [m]
FIGURE 6.10 Solar energy spectrum in terms of the radiation energy vs wavelength where Ir(TiO2) and
Ir(RBGT) denote the incidence of solar irradiance for commercial TiO2 and reduced-band-gap TiO2 respectively

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

Page 193

Solar-Hydrogen: A Solid-State Chemistry Perspective

193

6.6.3 REDUCTION OF RECOMBINATION


Besides the measures applied above, sensitization could be enhanced by the reduction of
recombination through charge separation. This could be achieved by imposition of
chemicallyinduced internal electric field which results in the desired charge separation.

6.7 REDUCED-BAND-GAP TITANIA


The main disadvantage of commercial TiO2-x as a PE material is its relatively large band
gap (3.0eV), which results in a low ECE (Bak et al, 2002). The recent report by Khan et al
(2002) indicates that the band gap may be reduced through the effect of carbon on
TiO2x. However, the role of carbon in changing the electronic structure is not clear. The
explanation of the effect observed by Khan et al requires the following:
understanding of the effect of carbon on the chemistry of the surface and

near-to-surface layers of TiO2;


understanding of the effect of carbon incorporation into the TiO2 lattice on its defect

disorder;
understanding of the relationship between defect disorder and electronic

structure for both undoped TiO2-x, formed in a carbon-free environment, and


TiO2-x formed in natural gas flame according to the procedure applied by Khan
et al (2002);
understanding of the effect of carbon on other properties of titania that are essential
for ECE; and
understanding of the effect of aliovalent ions resulting in the reduction of band bap
of TiO2, such as Cr, Mo (Wilke and Brauer, 1999) and Pb (Rahman et al, 1999), on
other variables that are essential for ECE.
As is known, the reduction of the band gap of TiO2-x results in an increase of solar
energy absorption. Specifically, the reduction of the band gap from ~3eV to ~2eV
results in an increase of the energy spectrum available for absorption from ~7 per cent
to ~40 per cent (Sheppard, 2004). However, the incorporation of aliovalent ions into
TiO2-x, which are responsible for the reduction of the band gap, such as Cr, may also result
in increased energy losses through the modification of other properties, such as:
increased stability of the ionized surface oxygen species thus leading to increased

energy losses due to reduced charge transfer at the electrode/electrolyte interface


(Sharma et al, 1997);
reduction in the lifetime of the electron-hole pairs generated leading, in
consequence, to increase of energy losses due to recombination (Wilke and Brauer,
1999);
formation of a segregation-induced potential barrier in the surface layer leading to the
generation of a retarding effect on the charge transport (Nowotny, 1991);
increased ohmic resistance of the TiO2 photo-electrode (Sharma et al, 1997);
increased electrical potential barrier across the Helmholtz layer due to the enhanced
adsorption of dipoles;

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

194

Page 194

J. NOWOTNY, T. BAK, L. R. SHEPPARD AND C. C. SORRELL

formation of a bi-dimensional surface structure that exhibits extraordinary properties

that retard the charge transport across this structure (Nowotny, 1991); and
insignificant overlap of the electron wave functions of the dopants in rutile

(Tang, 1994).
The research and development programme on solar-hydrogen at the University of New
South Wales aims to engineer the electronic structure of TiO2-x and that of other Ti-based
compounds through the imposition of controlled defect chemistry. An important part of
this research is the derivation of defect diagrams and the establishment of the
relationships between defect disorder, electronic structure and the semiconducting
properties of undoped TiO2 as well as TiO2 doped with aliovalent ions. Preliminary
progress in this research has been reported recently by the authors (Nowotny et al, 1997
and 2005; Bak et al, 2002, 2003a and 2003b).

6.8 IMPACT OF DEFECT CHEMISTRY ON THE PROPERTIES OF TITANIA


6.8.1 BACKGROUND
The work of Khan et al (2002) suggests that there is a relationship between the defect
disorder of TiO2-x and its band gap. However, clarification of this observation requires
further studies in several relevant areas:
determination of the defect chemistry diagrams for TiO2-x;
preparation of TiO2-x with controlled defect disorder; and
use of the preceding to modify the band gap of TiO2-x.

The present work considers Ti-based compounds, mainly TiO2x and its solid solutions,
which are considered to be primary candidates for PEs. The following sections overview
several key properties of TiO2x that are essential for its performance as a photo-electrode:
nonstoichiometry and related defect chemistry;
charge transfer; and
segregation.

6.8.2 NONSTOICHIOMETRY
Titania is known to be an oxygen-deficient material with its nonstoichiometry being
determined by the apparent deficit of oxygen (Kofstad, 1972). Accordingly, the correct
formula of titanium dioxide is TiO2-x, where x is determined by the concentration of ionic
lattice defects. However, the present state of understanding is such that the specific value
of x is not limited solely to oxygen vacancies; it should be considered in terms of the
vacancies in both the titanium and oxygen sublattices as well as titanium interstitials.
Taking into account all these defects and also assuming the presence of donor- and
acceptor-type foreign ions (impurities and dopants) the formula of TiO2-x may be
expressed in Krger-Vink notation (Kroger, 1974) as follows:
(31)

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

Page 195

Solar-Hydrogen: A Solid-State Chemistry Perspective

195

where ATi and DTi denote singly ionized acceptor- and donor-type ions. The law of
conservation of the number of lattice sites requires that:
(32)
The lattice charge neutrality condition requires that:
(33)
where [e] and [h] denote the concentration of electrons and electron holes respectively.
Equation 33 represents the general charge neutrality condition. This condition assumes a
simple form when only the predominant defects are involved (Kofstad, 1972).

6.8.3 DEFECT DISORDER


It seems that defect chemistry is the key to understanding the photo-electrochemical
properties of oxide materials and, therefore, the materials of desired photo-sensitivity can
be obtained through controlled defect chemistry.
The theory of defect chemistry concerns a complete description of the structure
of point defects, such as oxygen vacancies and cation vacancies, and electronic
defects, such as electrons and electron holes. Defect chemistry explains in a quantitative
manner the relationship between nonstoichiometry and the concentration of defects.
Defect disorders are usually represented in terms of the concentration of defects as a
function of temperature or partial pressure of constituent atoms, such as oxygen. The
point defects can be considered in terms of defects equilibria that can be described by
equilibrium constants which are specific materials properties (Kroger, 1974).
The properties of nonstoichiometric compounds, including TiO2-x, are strongly affected
by the presence of point defects. Therefore knowledge of the relationship between defect
disorder and specific properties can be used to predict the properties of TiO2-x of
controlled defect chemistry. An extensive survey on the chemistry of point defects and
the impact of defect chemistry on the properties of binary metal oxides was provided by
Kofstad (1972). It appears that there is a close relationship between defect chemistry and
properties, including electronic structure, charge transfer and related electrical properties.
One of the most important defect-related electrical properties is electroactivity. This
property, which can be considered quantitatively in terms of the chemical potential of
electrons, has a significant impact on reactivity.
The crucial property of TiO2-x as a PE is its reactivity with water. In effect, this reactivity
is determined by the chemical potential of electrons. Accordingly, defect chemistry
considerations may be used to guide the processing procedures of TiO2-x with controlled
(desired) properties for specific applications, including photo-electrodes. However, this
process requires the defect chemistry to be described using correct defect diagrams.
An extensive overview of the defect chemistry of TiO2-x has been reported by Kofstad
(1972). Subsequent studies aimed at the verification of defect disorder models of TiO2-x have
shown that the reported defect diagrams are in disagreement not only in the absolute values
of the concentrations but also in the nature of the defects in both the titanium and oxygen

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

196

7:06 PM

Page 196

J. NOWOTNY, T. BAK, L. R. SHEPPARD AND C. C. SORRELL

sublattices (Yahia, 1963; Baumard and Tani, 1977; Balachandran and Eror, 1988; Son and Yu,
1996; Marucco et al, 1981; Blumenthal et al, 1967; Sawatari et al, 1982) that have been
taken into account. These disagreements are the reason why the authors of the
present work have undertaken extensive studies in the determination of the defect disorder
of titania.

6.8.3.1 Defect equilibria


Defects in TiO2-x may be considered in terms of defect equilibria. Assuming Krger-Vink
notation (Krger, 1974) the following equilibria should be considered:

(34)
(35)
(36)
(37)
(38)
(39)
It has recently been shown that prolonged oxidation results in oxygen incorporation into
the TiO2 leading to the formation of Ti vacancies (Nowotny et al, 2005):
(40)
The equilibrium constants for the preceding six reactions may be expressed according to
the following equations respectively:
(41)
(42)
(43)
(44)
(45)
(46)
where the square brackets denote the concentrations of defects, expressed in atomic
ratio. The preceding equilibrium constants can be related to the standard state
thermodynamic quantities entropy So and enthalpy Ho:

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

Page 197

Solar-Hydrogen: A Solid-State Chemistry Perspective

197

(47)

where both So and Ho are the specific materials properties. The thermodynamic
quantities related to the equilibrium constants have been reported elsewhere (Kofstad,
1972; Bak et al, 2003a).

6.8.3.2 Effect of oxygen partial pressure on the concentration of defects


According to Equations 4042, the concentrations of electronic charge carriers may be
expressed as a function of equilibrium p(O2) within the several p(O2) regimes, including:
a strongly reduced regime;
a reduced regime; and
and an oxidized regime.

A strongly reduced regime is characterized by the p(O2) exponent, 1/m = 1/4. This
exponent may be explained by the defect disorder based on Ti interstitials as the
predominant defects. Then simplified charge neutrality assumes the form:
(48)
The combination of Equations 42 and 48 leads to the following relation:

p(o2)

(49)

A reduced regime corresponds to the p(O2) exponent, 1/m = 1/4. This exponent is
consistent with the defect disorder based on doubly ionied oxygen vacancies as the
predominant defects. The simplified charge neutrality is:
(50)
The combination of Equations 41 and 50 results in:

p(o2)

(51)

An oxidized regime involves three sub-regimes including n-type, p-type and mixed
np sub-regimes in which the p(O2) exponent assumes the following respective values:
1/m = 1/4 (in the n-type regime);
1/4 <1/m <1/4 (in the np transition regime); and
1/m = 1/4 (in the p-type regime).

The oxygen vacancies in this regime are compensated by acceptor-type defects.


Therefore:

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

198

7:06 PM

Page 198

J. NOWOTNY, T. BAK, L. R. SHEPPARD AND C. C. SORRELL

(52)
where A denotes an effective concentration of acceptors, which includes the
concentration of extrinsic acceptor-type impurities, [A], donor-type defects, [D], as well
as Ti vacancies:
(53)
where A and D denote the singly ionized acceptors and donors respectively. While the
concentration of acceptor-type impurities is statistically higher than that of donors, it will
be shown below that the major component of the parameter A are Ti vacancies. It will also
be shown that the component A should be considered as quenched under usual
experimental conditions. Combining Equations 41 and 52 results in:

p(o2)

(54)

It has been recently shown that TiO2 may also exhibit p-type properties. Assuming that the
predominant defects in this regime are acceptor-type defects and oxygen vacancies, the
concentration of electronic defects is the following function of the p(O2):

p(o2)

(55)

The concentration of electrons, as well as all other defects, may be determined as a


function of p(O2) using the combination of Equations 4045 and 4751. Equations 33 and
4751 may be used to guide the selection of the processing conditions for TiO2-x in terms
of temperature, p(O2), and concentrations of both donors and acceptors in order to
achieve the desired chemical potential of electrons (e), which is required to achieve the
desired electroactivity of TiO2-x-based photo-electrodes:
(56)
where:
(57)
The experimental determination of the chemical potential of electrons at the surface and
in the bulk was discussed in section 6.5.

6.8.3.3 Defect diagrams


Figure 6.11 shows the effect of p(O2) on the defect concentration in undoped TiO2-x at
1273K where either a) the concentrations of both external acceptors (including titanium
vacancies) and external donors assume negligibly small values or b) the donors are
compensated by the acceptors (Bak et al, 2003a). As seen, the predominant electronic

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

2/6/07

7:06 PM

Page 199

Solar-Hydrogen: A Solid-State Chemistry Perspective

199

Centre for Materials Research in Energy Conversion, UNSW

TiO2x

1273 K, A = 0
2

VO

4+

Ti i

e
3+

Ti i

log [ ]

Ch-06

10

12

14

25

20

15

10

log p(O2) [p in Pa]


Source: Bak et al, 2003

FIGURE 6.11 Defect diagram showing the concentration of defects vs oxygen partial pressure for undoped
TiO2 at 1273K in absence of foreign elements forming donors or/and acceptors

defects are electrons within the entire range of p(O2). In the case where the mobility terms
are independent of p(O2), the slope of the log versus log p(O2) plot (where is electrical
conductivity) is the same as the slope of the log n versus log p(O2) plot. Accordingly,
reduction of undoped TiO2-x, through the reduction of p(O2) during equilibration, results in
an increase of both the concentration of electrons and the electrical conductivity. As seen
in Figure 6.11, the n-type conductivity of undoped TiO2-x at high and low p(O2) is
determined by oxygen vacancies and tri-valent Ti interstitials respectively. Similar defect
diagrams may be derived for acceptor- and donor-doped TiO2-x (Bak et al, 2003a and 2003c).
One should emphasize that in the temperature range usually applied for the
characterization of electrical properties (10001500K) the concentration of titanium

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

200

7:06 PM

Page 200

J. NOWOTNY, T. BAK, L. R. SHEPPARD AND C. C. SORRELL

vacancies does not assume the equilibrium concentration because the mobility of these
defects is too low (Baumard and Tani, 1977; Bak et al, 2003a and 2003c). This is also
the case for BaTiO3 (Nowotny and Rekas, 1994). In other words, titanium vacancies attain
equilibrium concentrations at substantially higher temperatures, typically above 2000K
(Bak et al, 2003c). At lower temperatures, the concentrations of titanium vacancies cannot
reach an equilibrium state, so their concentrations may vary depending on specific
processing conditions. This is the main reason why there is a substantial discrepancy
between the reported data for electrical properties and the related defect diagrams.
According to the derived defect diagrams, undoped TiO2-x exhibits n-type properties within
the entire range of p(O2) and does not exhibit an np transition (Bak et al, 2003). This is the
case when titanium vacancies are absent or present at very low concentrations. The fact that
the majority of the reported electrical conductivity data reveal minima indicates that titanium
vacancies are formed during either a) processing of the specimen or b) the experiment if
performed over a prolonged period of time (Bak et al, 2003c; Nowotny et al, 2005).
The defect diagrams for undoped TiO2-x may be considered to be divisible into two
p(O2) regimes, where the predominant ionic defects are:
1 tri-valent titanium interstitials under extremely reducing conditions; or
2 oxygen vacancies at intermediate and high p(O2) values.
Presence of aliovalent ions has a substantial impact on the electrical properties of TiO2-x.
Therefore the characterization of TiO2-x specimens should include the determination of the
concentrations of impurities that act as donors and acceptors. The defect diagrams reported
in this study (Bak et al, 2003) provide the first comprehensive representation of the defect
chemistry of TiO2-x containing both intrinsic and foreign defects, the latter including both
acceptor- and donor-type elements in TiO2-x, for a wide range of processing conditions.

6.8.4 CHARGE TRANSFER IN TITANIA


The charge transfer may be determined from both the concentration and mobility terms.
Comparative analysis of the concentration of defects (Bak et al, 2003) with the available
empirical electrical conductivity data (Nowotny et al, 1997 and 1998; Yahia, 1963;
Baumard and Tani, 1977; Balachandran and Eror, 1988; Son and Yu, 1996; Marucco et al,
1981; Blumenthal et al, 1967; Sawatari et al, 1982) allow the determination of the mobility
terms for electrons and electron holes (Bak et al, 2003c).
In summary, knowledge of defect diagrams and mobility terms may be used to guide
the procedures for the preparation of TiO2-x with desired electrical properties that can be
achieved through doping and imposition of desired p(O2). Specifically, n-type TiO2-x which
exhibits high conductivity could be achieved through donor doping and imposition of low
p(O2). Alternatively, p-type TiO2-x could be achieved through the incorporation of acceptors
at high p(O2).

6.8.5 CHARGE TRANSFER BETWEEN TITANIA AND WATER


The reactivity between solid surfaces such as TiO2-x and adsorbed species such as H2O
and the related charge transfer involve the following two components:

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

Page 201

Solar-Hydrogen: A Solid-State Chemistry Perspective

201

1 the component related to the collective electronic properties of the solid, as


represented by the Fermi energy; and
2 the component related to local electrostatic interactions between the adsorption
surface-active site and the adsorbed species.

Centre for Materials Research in Energy Conversion, UNSW

EC

Eact

Eact

EC

EF

EF
EV

EV

a)

b)

FIGURE 6.12 Effect of Fermi energy on the charge transfer between the surface of semiconducting solid
and adsorbed species forming either acceptors or donors

When the collective electronic properties control the reactivity, the charge transfer is
determined by the difference between the Fermi energy and the electron affinity or
ionization potential of the adsorbed species, as shown schematically in Figure 6.12. As
can be seen, the adsorbed molecules may be either donors or acceptors, depending on
the Fermi energy of the solid and the electronic structure of the adsorbed species. In this
case one should expect a linear dependence between the activation energy of the charge
transfer and the Fermi energy of the solid (Nowotny, 1991).
The preceding model does not apply when the local electrostatic interactions between
the adsorbed species and surface-active sites control the reactivity. These local interactions,
which are determined by the electronic structure of the specific surface site, are independent
of the Fermi energy which may be considered as a collective factor of the lattice.
A study by Kowalski et al (1980) indicates that the reduction of the TiO2-x surface
results in the formation of surface states that enhance the dissociation of water
molecules.
Hepel et al (1982) reported that photo-etching of TiO2 in acid solutions results in
dissolution of surface donor states, such as tri-valent Ti ions, and increase in the surface
area of the electrode. In consequence, the acid etching is expected to enhance the ECE
of the TiO2-based photo-electrode.

6.8.6 SEGREGATION IN TITANIA


An excess of interfacial energy is the driving force for segregation of certain lattice
defects, resulting in an enrichment of the surface in these defects. In consequence,

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

202

7:06 PM

Page 202

J. NOWOTNY, T. BAK, L. R. SHEPPARD AND C. C. SORRELL

TABLE 6.1 Effect of Cr on the energy conversion efficiency (ECE) of a photo-electrochemical cell involving
photo-anode made of Cr-doped TiO2
EFFECT ON SPECIFIC PROPERTY
Reduction of band gap
Increase of stability of chemisorbed
oxygen species
Reduction of the electron-hole pair lifetime
Formation of segregation-induced retarding
potential barrier

Increase of ohmic resistance


Increased retarding potential over Helmholtz layer
Formation of retarding potential over
segregation-induced low dimensional
interface structure
Procedure of Cr incorporation results in the
formation non-equilibrium concentration gradients
The effect of Cr on ECE
established experimentally

EFFECT ON ECE
Increase
Decrease

REFERENCE
Wilke and Brauer, 1999
Sharma et al, 1997

Decrease
Decrease

Wilke and Brauer, 1999


Up to 1V barrier in oxide
semiconductors:
Adamczyk
and Nowotny, 1986
Sharma et al, 1997
To be determined
To be verified

Decrease
Decrease
Decrease

Decrease

Decrease

The reported data


correspond to
non-equilibrium systems
Bak et al, 2002

the phenomenon of segregation leads to the formation of a concentration gradient


within the interface layer and a related potential barrier (Nowotny, 1991). This
potential barrier influences charge separation. Therefore, quantitative evaluation of the
segregation-induced concentration gradients is an essential part of the characterization of
photo-electrodes. Conversely, the phenomenon of segregation may be used as a
technological tool for the imposition of concentration gradients in order to engineer
the optimal Schottky-type potential barrier, which is required for charge separation. Ikeda
and Chiang (1991) have considered the effect of segregation of aliovalent ions (Nb5+
and Al3+) on the local defect chemistry and electrical properties of the surface layer of
TiO2-x. In this study, however, the conclusions about the segregation-induced surface
properties were based on electrical conductivity, which is a bulk property. The effect of
segregation on surface composition of titania was reported by Gulino et al (1995) for Sn,
Thevuthasan et al (1997) for Nb, Ruiz et al (2003 and 2004) for Cr and Nb, and Wang et al
(2004) for Y.

6.8.7 NANO-SIZED TITANIA


The effect of grain size on the properties of TiO2-x may be considered in terms of the
impact of the interfacial energy on the electronic structure. Fievet et al (1979) have shown
that there is a critical grain size below which the structure-related quantities, such as
lattice constant, may be affected substantially. For NiO, this critical grain size was
observed to be approximately 30nm. Consequently, ultrafine grain sizes can be expected

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

Page 203

Solar-Hydrogen: A Solid-State Chemistry Perspective

203

ENERGY
ATOM

MOLECULE
(2 ATOMS)

SMALL
CLUSTER
(10 ATOMS)

CLUSTER
PARTICLE
(~200 ATOMS)

SEMICONDUCTOR
(>200 ATOMS)

VACUUM

CONDUCTION
BAND

Eg

Eg

VALENCE
BAND

FIGURE 6.13 Effect of cluster and grain size on electronic structure and related band gap width according
to Hoffmann et al (1995)

to have an effect on the electronic structure and the band gap as well. This effect has
been considered by Hoffmann et al (1995) in terms of a quantum size effect, where finegrained materials exhibit larger Eg values than those of coarse-grained materials, as
shown in Figure 6.13. Therefore it is to be expected that a reduction in the grain size of
TiO2-x will result in a decrease in its photo-sensitivity.

6.9 COLLECTIVE AND LOCAL FACTOR


It is general practice for the performance of the photo-electrodes to be considered
principally in terms of collective properties, such as electronic structure, the concentration
of charge carriers, and the flat band potential required for charge separation. However, the
reactivity at the liquid/solid interface must be considered not only in terms of the
collective electrochemical properties, which are required for efficient charge transfer, but
also in terms of the specific surface sites at which the reactions take place.
The surface studies are in agreement that the defect-free surface of TiO2 is not
reactive with water and oxygen (Lo et al, 1978). Therefore the presence of defects is
required for the surface to enable the charge transfer between the adsorbed water
molecule and TiO2. Recent studies (Nowotny et al, 2005) have shown that:
defect disorder of TiO2 involves titanium vacancies, which form acceptor centres;

and
titanium vacancies are the surface active sites for water splitting.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

204

Page 204

J. NOWOTNY, T. BAK, L. R. SHEPPARD AND C. C. SORRELL

6.10 SPIN-OFF APPLICATIONS OF TITANIA


In addition to the use of TiO2-x as a photo-electrode for water photolysis, TiO2-x exhibits
many other useful properties related to its photo-sensitivity and reactivity with water.
While many other oxide materials have a band gap that is much more favourable to light
absorption, the increasing interest in TiO2-x results from its outstanding resistance to
corrosion and photo-corrosion in aqueous environments at pH ~7. Owing to this stability,
titania, unlike other materials, exhibits stable properties over periods of years.
This corrosion resistance has been established for commercial titania for which the
final technological process corresponds to annealing in air. Then titanias chemical
composition is close to stoichiometric composition and its formula is TiO2. Corrosion
resistance of reduced titania requires confirmation.
At present, the major spin-off applications that have been reported include:
Decontamination of water containing organic and inorganic pollutants (Fujishima

et al, 1999). The in status nascendi atomic oxygen formed on the surface of TiO2-x is
extremely reactive. Therefore even traces of TiO2-x present in water result in the rapid
oxidation of toxic and pollutant compounds.
Decontamination of water containing viruses and bacteria (Fujishima et al, 1999).
The in status nascendi atomic oxygen also is known to oxidize these species,
resulting in the sterilization of water.
Generation of electricity using so-called electric windows (Bach et al, 1998;
Nakato et al, 1982). Dye-sensitized TiO2-x, deposited on the surface of glass windows,
forms the PE component of a solar cell to produce electricity (AGO, 1999).
Inhibition of water condensation on ceramic tiles, glass and other building
materials (Fujishima et al, 1999). It has been observed that thin films of TiO2-x exhibit
self-cleaning due to their strongly oxidizing properties.
Anti-tumour effect on gastric cancers (Fujishima et al, 1999).
Anti-fogging coatings on glasses due to superhydrophilicity (Fujishima et al,
1999).
Coatings for enhanced cleanability of car bodies and monuments due to
hydrophilicity (Fujishima et al, 1999).

It is clear that there is a wide range of spin-off applications for TiO2-x with reduced band
gap and increased photo-sensitivity. Ultimately, it is likely that it has the potential to
replace silicon as a photovoltaic material in the future (Bak et al, 2003).

6.11 MULTIPHASE PHOTO-SENSITIVE SYSTEMS


Morisaki et al (1976) have reported a hybrid photo-electrode consisting of a silicon
photovoltaic cell, forming the underlayer, and an overlayer formed of TiO2-x. The
multiphase photo-sensitive electrochemical chain of this system is shown in Figure 6.14.
In principle, this type of stack electrode combines a silicon solar cell with a coating of
TiO2-x. The role of silicon is to provide the bias voltage required for high ECE and the role
of the TiO2-x is to provide corrosion protection of silicon and absorption of the solar
spectrum range, which is limited to that shown in Figures 6.9 and 6.10.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

Page 205

Solar-Hydrogen: A Solid-State Chemistry Perspective

205

Centre for Materials Research in Energy Conversion, UNSW

h
1
e

(1)

n-Si

p-Si

O2

H2

4
e

TiO2-x

e
(2)

H2 O + M X

(3)

e
H2O + 2h 2H+ + O2

2H+ + 2e H2

V
FIGURE 6.14 Circuit of the hybrid cell of Morisaki et al (1976) and related electrochemical chain

The performance of such a hybrid cell is determined by the following factors:


1 the transparency of the TiO2-x overlayer; and
2 the effectiveness of the charge transfer across the silicon/TiO2-x interface.
Although the conceptual approach of this cell has great potential, the ECE achieved by
Morisaki et al in 1976 was only ~0.1 per cent.
Rahman et al (1996) have proposed a heterojunction solar cell formed of n-type TiO2
as an overlayer and p-type CuInSe2 as a sublayer. Due to the large band gap of undoped
TiO2, a conduction band discontinuity exists across the interface which results in energy
losses due to poor charge transfer. Rahman et al (1996) have shown that incorporation of
Pb into TiO2, resulting in the reduction of the band gap, leads to a substantial reduction of
the conduction band discontinuity and increases the theoretical conversion efficiency
from 1.5 per cent to 18.8 per cent when undoped TiO2 is replaced by Pb-doped TiO2. The
concept proposed by Rahman et al may be used for processing the photo-anode for PEC;
however, the junction design should address the charge separation resulting in different
polarity.

6.12 SOLAR CELL EQUIPPED WITH SPACE-BASED


SOLAR ENERGY COLLECTOR
The race to develop hydrogen technologies has expanded beyond the global realm and
has resulted in the initiation of a substantial research project by the National Space
Development Agency (NASDA) and the Institute for Laser Technology (ILT) in Japan
(Shimbun, 2001). The programmes aim is the development of a hydrogen generation
system using a space-based solar unit to harvest solar energy and to transfer it to a
TiO2-x-based electrochemical device located on Earth. This technology, shown
schematically in Figure 6.15, will include the following three devices:

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

206

Page 206

J. NOWOTNY, T. BAK, L. R. SHEPPARD AND C. C. SORRELL

Centre for Materials Research in Energy Conversion, UNSW

LASER
GENERATOR

SUN

SPACE-BASED
EQUIPMENT

SOLAR
CONCENTRATOR

EARTH-BASED
EQUIPMENT

O2

H2

H
TiO2

H2 2H + 2e

TiO2-BASED
ELECTROCHEMICAL CELL

2H + 2e + O2 H2O
HYDROGEN
FUEL CELL

Source: (Shimbun, 2001)

FIGURE 6.15 Schematic illustration of the solar-hydrogen generation system using space solar energy
collector

1 a space-based solar condenser collecting solar energy;


2 a laser generator transforming the solar energy into a laser beam, which is
transmitted to Earth-based TiO2-x acting as a photo-electrode; and
3 a ground-based photocatalytic device aimed at the collection of the laser beam and
the production of hydrogen.
This proposed technology has the following advantages over competing technologies
based on solar energy collected on the Earths surface:
solar irradiance in space is free of energy losses related to atmospheric effects and

so is substantially higher than that on Earth;


solar energy in space is available continuously and independently of the diurnal

cycle; and
solar energy in space is available independent of weather conditions.

It has been claimed that the cost of hydrogen manufactured using this technology will
be ~20 Japanese Yen for an amount of hydrogen equivalent to one litre of gasoline. It is
expected that the first experimental satellite will be launched by 2010 and the entire
system will be completed by 2020. A vision of this technology, using the solar energy
collector outside the atmosphere, was described by Bockris (1999).

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

Page 207

Solar-Hydrogen: A Solid-State Chemistry Perspective

207

Centre for Materials Research in Energy Conversion, UNSW

O2

H2

H
+

H2 2H + 2e

TiO2

TiO2-BASED
ELECTROCHEMICAL CELL

2H + 2e + O2 H2O
HYDROGEN
FUEL CELL

FIGURE 6.16 Schematic illustration of the use of both solar-hydrogen and solar-oxygen as input gases for
the production of electricity using a fuel cell

6.13 SOLAR-OXYGEN
Although the principal objective of the photo-electrochemical cells technology is the
production of hydrogen, the by-product of the photolysis of water is oxygen, whose wide
range of applications includes medical applications and metallurgy. Such solar-oxygen
also has the potential to be used in high-performance fuel cells instead of air. This
application would be advantageous because the combustion of any fuel, even hydrogen,
has the potential to result in NOx pollution owing to the fact that combustion usually takes
place in the presence of air, which contains ~80 per cent nitrogen. The NOx pollution at
the oxidation electrode could, therefore, be eliminated if the oxidation electrode is
exposed to solar-oxygen as the oxidizing agent instead of air as shown in Figure 6.16.

6.14 ECONOMIC AND ENVIRONMENTAL CONSIDERATIONS


OF SOLAR-HYDROGEN
6.14.1 SIGNIFICANCE
Hydrogen production will skyrocket within the next eight to ten years (Thomas et al,

1998).
Solar-hydrogen is the leading candidate for a renewable and clean energy carrier of

the future.
Hydrogen generated from water using solar energy constitutes a renewable form of

energy.
Hydrogen generated from water using solar energy constitutes a clean energy carrier

as neither its production nor its combustion produces greenhouse or pollutant gases.
Solar-hydrogen ultimately will reduce our total reliance on coal, gasoline and natural

gas, so it will provide energy security.


Solar-hydrogen technology based on TiO2-x as a PE will increase substantially the

range of practical applications for TiO2-x as a photo-sensitive material.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

208

Page 208

J. NOWOTNY, T. BAK, L. R. SHEPPARD AND C. C. SORRELL

6.14.2 BENEFITS
Many areas are subject to abundant sunlight and so are well placed to commercialize

solar-hydrogen technologies. For example, the northern parts of Australia have some
of the highest levels of sunlight in the world. When this energy is rationally utilized
for the production of fuel Australia could become an exporter of solar energy: a
singular concept.
Solar-hydrogen will allow reduction in emissions of greenhouse, pollutant and toxic
gases.
Solar-hydrogen encompasses both the production and utilization of a fuel that is
100 per cent environmentally clean.
Countries that possess the infrastructure for solar-hydrogen will be able to reduce
their reliance on fossil fuels.
Countries that produce excess solar-hydrogen can export solar energy, which is a
singular concept.

The technologies created for solar-hydrogen production will have far-reaching applications
in both domestic and industrial markets. For example, a solar-hydrogen panel of
approximately 40 40km, located in the Pilbara region of Western Australia, will be
sufficient to meet all of Australias current energy requirements.

6.14.3 ADVANTAGES
The fuel may be generated anywhere.
The process requires only water and solar energy;
The hydrogen-generating device does not have any moving parts, so maintenance on

this component is minimal.


The production process does not cause pollution.
The device can be marketed internationally.

6.14.4 COST ESTIMATES


So far, little is known of the actual costs of solar-hydrogen production using water
photolysis. Realistic price figures may become available when solar-hydrogen
technologies reach the stage of commercial maturity. However, it is inevitable that some
estimates of such costs must be made at this very pre-commercial stage of the
development of these technologies. While some estimates of these costs are available,
these figures must be considered to be premature.
The Japanese project discussed in section 6.12 (Shimbun, 2001) estimates that the
resultant technology will be able to produce a gasoline litre-equivalent of hydrogen at
a cost of ~20 Japanese Yen, which is ~17 US cents at present. This cost is equivalent
to US$6/GJ of fuel. This figure is similar to the current price of hydrogen from
methane/natural gas.
A perhaps more realistic price estimate is that by Bockris (2002), who is one of the
pioneers of solar-hydrogen. His estimate, which is based on the present cost of the use

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

Page 209

Solar-Hydrogen: A Solid-State Chemistry Perspective

209

of photovoltaic electricity to generate solar-hydrogen, leads to a cost estimate US$28/GJ.


This figure includes the costs of equipment manufacture and maintenance.
While the higher cost of US$28/GJ, compared to the present market price of
hydrogen, appears to be discouraging, when the cost of the pollution associated with the
combustion of gasoline or methane/natural gas is factored in, the conclusion is entirely
different. According to Bockris (2002) the cost of pollution is equivalent to US$33/GJ,
which gives the actual cost of the combustion of these fossil fuels of US$39/GJ.
Therefore, when the pollution costs are taken into account, the solar-hydrogen is a clear
winner. Further, when the high cost of the silicon solar cell is replaced by the low cost of
the TiO2-x PE, the cost estimate of US$28/GJ will fall substantially.
Momirlan and Veziroglu (2002) have recently reported an extensive overview of the
current status of hydrogen economy including the hydrogen production.

6.14.5 SIMPLE COMPARISONS


Assuming that the ECE is 20 per cent, which is estimated to be close to the achievable
level for a solar-hydrogen cell, a 10 x 10m solar panel, which could be accommodated on
the roof of a typical individual household, during 10 hours of unimpeded sunlight will
produce 4.9kg of hydrogen, which is equivalent to:
the amount of energy corresponding to ~20L of gasoline;
the amount of electricity sufficient to operate a 60W lightbulb for approximately

94 days (assuming 70 per cent efficiency of hydrogen fuel cell); or


44L of water as a raw material.

Assuming that there are 300 sunny days a year, under the preceding conditions, the
annual production of solar-hydrogen would be 1470kg. At the present hydrogen price of
US$1.3/kg, the value of this annual hydrogen production would be US$1911. Taking into
account the impact of pollution when using the hydrogen generated from natural gas and
the associated costs (see section 6.14.4), and the lack of pollution when using solarhydrogen, the real value of the hydrogen generated using solar energy should be elevated
to approximately US$8800 p.a.
At the above assumptions it is calculated (using 2001 data) that a 40 x 40km panel
would produce sufficient hydrogen to meet all of the energy needs of Australia. The above
surface area is equivalent to the surface of the individual household units if all homes in
Australia are equipped with such units.

6.16 SOLAR-METHANOL
Several Australian industrial organizations, including DUT, CC Energy Pty. Ltd and Isentropic
Systems Ltd., have proposed the integration of the UNSW solar-hydrogen technology with
CC Energys proposed coal-solar system for the production of zero-CO2-emission methanol
(Cummings, 2003). This initiative aims at the use of coal and solar-hydrogen for the
production of solar-methanol, with the surplus of CO2 entering into reaction with solarhydrogen, thus forming another production line of methanol. Further, the oxygen produced
along with the solar-hydrogen can be used for coal gasification, as shown in Figure 6.17.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

210

Page 210

J. NOWOTNY, T. BAK, L. R. SHEPPARD AND C. C. SORRELL

Centre for Materials Research in Energy Conversion, UNSW

COAL

GASIFIER
2 C + O2 2 CO

H2O

CO + H2O CO2 + H2

REACTOR-1
2 H2 + CO CH3OH

O2

CO2

WATER PHOTOLYSIS
H2O H2 + O2

H2O

REACTOR-2
CO2 + 3 H2 H2O + CH3OH

METHANOL

Source: Cummings, 2003

FIGURE 6.17 The concept of the Australian initiative of the solar-methanol technology

The system, in effect, sequesters surplus CO2, generated as a result of coal gasification,
and uses it to produce a low-emission fuel in the form of solar-methanol. Methanol has the
advantage that it is substantially easier to transport than hydrogen. Although the
combustion of methanol still results in the emission of greenhouse gases, the fuels
principal component is the solar energy. Therefore, this fuel should be considered to be a
substantial step forward, compared to fossil fuels, in the reduction of global warming.
According to Cummings (2003), solar-methanol is expected to be the most feasible
carrier of solar-hydrogen to be used for hydrogen transportation over large distances.

6.17 CONCLUSIONS
There has been growing awareness that the so-called hydrogen economy will soon be
part of our life. Introduction of this technology means that hydrogen will replace fossil
fuels. The move towards the hydrogen economy is rapid and it is being driven by a general
perception that hydrogen is an efficient and clean fuel. The perception of the latter must
be viewed with caution. Specifically, it must be appreciated that only hydrogen generated
using renewable energy leads to zero emissions of greenhouse gases and thus is
environmentally clean.
It is expected that the commencement of the hydrogen era will unfold a market
dominated by hydrogen generated from natural gas for economic reasons. However, the

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

Page 211

Solar-Hydrogen: A Solid-State Chemistry Perspective

211

issue of global warming will not go away and its increasingly damaging effects will
impose the need to generate hydrogen from renewable energy. In the present work, it has
been argued that generation from water decomposition using solar energy is one of the
most promising means to produce hydrogen.
The principle of hydrogen generation through water photolysis using TiO2-x as the
photo-electrode has been overviewed. The present work brought together the concepts
of photo-electrochemistry and the concepts of defect chemistry using TiO2 as an
example of photo-electrode. It was shown that photo-sensitivity of oxide materials is
closely related to their defect disorder. Therefore defect chemistry may be used as
a framework for engineering novel photo-sensitive materials. The defect chemistry of
TiO2-x and its impact on the properties that are critical for its performance as a
photo-electrode, including the chemical potential of electrons and the band gap, have
been emphasized.
It may be expected that, during the initial stage of commercialization, the price of
solar-hydrogen will be higher than that of hydrogen from methane. However, this cost
disadvantage will be compensated by reduction of pollution and global warming.
According to Bockris (2002), the price of pollution is so high that we cannot afford it. Solarhydrogen may be an ultimate remedy for pollution and global warming. This is the reason
why the discovery of solar-hydrogen by Fujishima and Honda (1972) may be considered
as one of the major discoveries of the 20th century.

NOTES
1 While in theory the combustion of hydrogen in oxygen results solely in the formation of
water, in practice an internal combustion engine uses air rather than oxygen. Since air
contains 80 per cent nitrogen then it is likely that the combustion in air may result in the
formation of a certain amount of NOx, irrespectively of the fuel used. However, this kind
of NOx pollution does not apply to the combustion of hydrogen in low temperature
hydrogen fuel cells.
2 This comment was made by the reviewer of the present paper.

ACKNOWLEDGEMENTS
The authors gratefully acknowledge financial support from The University of New South
Wales, the Australian Research Council, Rio Tinto Ltd and Sialon Ceramics Pty. Ltd. The
authors also wish to acknowledge the extensive comments on hydrogen economy-related
data kindly provided by Professor John OM. Bockris.

AUTHOR CONTACT DETAILS


Dr. J. Nowotny (corresponding author), School of Materials Science and Engineering, The University of
New South Wales, Sydney, NSW 2052, Australia
Tel: 61 2 9385 6465; Fax: 61 2 9385 6467; e-mail: J.Nowotny@unsw.edu.au

REFERENCES
Adamczyk, Z. and Nowotny, J. (1986) Effect of segregation on near-surface and bulk transport phenomena in ionic crystals,
J. Phys. Chem. Solids, no 47, pp1124

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

212

Page 212

J. NOWOTNY, T. BAK, L. R. SHEPPARD AND C. C. SORRELL

AGO (1999) Australian Greenhouse Office, Renewable Energy Commercialisation Program, 9 July 1999,
www.greenhouse.gov.au/renewable/renew3.html
Armstrong, P. (1996) Hydrogen power science facts or science fiction, Borderland Science Research Foundation,
www.borderlands.com/journal/h2.htm
Bach, U., Lupe, D and Gratzel, M. (1998) Solid-state dye-sensitized mesoporas TiO2 solar cells with high photon-to-electron
conversion efficiencies, Nature, no 395, pp583585
Bak, T., Nowotny, J., Rekas, M. and Sorrell, C. C. (2002) Photo-electrochemical hydrogen generation from water using solar
energy. Materials-related aspects, Int. J. Hydrogen. Energy, no 27, pp9911022
Bak, T., Nowotny, J., Rekas, M. and Sorrell, C. C. (2003) Defect chemistry and semiconducting properties of titanium dioxide.
I. Intrisic electronic equilibrium, J. Phys. Chem. Solids, no 64, pp10431056
Bak, T., Nowotny, J., Rekas, M. and Sorrell, C. C. (2003a) Defect chemistry and semiconducting properties of titanium
dioxide. II. Defect diagrams, J. Phys. Chem. Solids, no 64, pp10571067
Bak, T., Nowotny, J., Rekas, M. and Sorrell, C. C. (2003b) Defect chemistry and semiconducting properties of titanium
dioxide. III. Mobility of electronic charge carriers, J. Phys. Chem. Solids, no 64, pp10691087
Bak, T., Burg, T., Kang, S-J. L., Nowotny, J., Rekas, M., Sorrell, C. C., Vance, E. R., Yoshida, Y. and Yamaguchi, K. (2003c)
Charge transfer in polycrystalline titanium dioxide, J. Phys. Chem. Solids, no 64, pp10891095
Bak, T., Nowotny, J., Rekas, M. and Sorrell, C. C. (2003d) Application of the high temperature Fermi probe in studies of the
charge transfer at the gas/solid interface of electrochemical systems. Example of the oxygen/zirconia system, in
K. Uematsu (ed) Ceramic Interfaces, Trans Tech Publications, Zurich, Switzerland, pp151178
Balachandran, U. and Eror, N. G. (1988) Electrical conductivity in non-stoichiometric titanium dioxide at elevated
temperatures, J. Mater. Sci., no 23, pp26762682
Baumard, J. F. and Tani, E. (1977) Electrical conductivity and charge compensation in Nb doped rutile, J. Chem. Phys., no 67,
pp857860
Becker, J. H. and Frederikse, H. P. R. (1962) Electrical properties of nano-stoichiometric semiconductors, J. Appl. Phys.,
no 33, pp447453
Blumenthal, R. N., Kirk, Jr. J. C. and Hirthe, W. M. (1967) Electronic mobility in rutile (TiO2) at high temperatures, J. Phys.
Chem. Solids, no 28, pp10771079
Bockris, J. OM. (1980) Energy Options. Real Economics and the Solar-Hydrogen System, Australia & New Zealand Book
Company, Sydney, Australia, pp325364
Bockris, J. OM. (1999) Hydrogen economy in the future, Int. J. Hydrogen Energy, no 24, pp115
Bockris, J. OM. (2002), private communication to J. Nowotny, 6 June
Bockris, J. OM. (2003) On hydrogen futures: Towards a sustainable energy system, Int. J. Hydrogen Energy, no 28,pp131133
Bockris, J. OM. and Uosaki, K. (1976) The theory of hydrogen production in a photoelectrochemical cell, Heliotechnique and
Development, no 1, pp435470
Bockris, J. OM. and Uosaki, K. (1976a) Photoproduction of hydrogen: Potential dependence of the quantum efficiency as a
function of wavelength, Energy, no 1, pp143145
Bockris J. OM., Uosaki, K. and Kita, H. (1981) Interfacial electron transfer as a significant step in photo-electrochemical
reactions on some semiconductors, J. Appl. Phys., no 52, pp808810
Bockris, J. OM., Veziroglu, T. N. and Smith, D. (1991) Solar Hydrogen Energy. The Power to Save the Earth, Optima Series,
The Guernsey Press Co Ltd., UK, pp7789
Chandra, S. (1985) Photoelectrochemical Solar Cells, Gordon & Breach, New York, pp1619 and 63203
Cummings, R. (2003) The solar-methanol production system, communication to J. Nowotny, 8 May
El Zayat, M. Y., Saed, A. O. and El-Dessouki, M. S. (1998) Photoelectrochemical properties of dye sensitized Zr-doped SrTiO3
electrodes, Int. J. Hydrogen Energy, no 23, pp259266

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

Page 213

Solar-Hydrogen: A Solid-State Chemistry Perspective

213

Fievet, F., Germi, P., Bergevin, F. and Figlarz, M. (1979) Lattice parameter, microstrains and nonstoichiometry in nickel (II) oxide.
Comparison between mosaic microcrystals and quasi-perfect single microcrystals, J. Appl. Crystal., no 12, pp387394
Friedli, H., Lotscher, H., Oescher, H. and Siegenthaler, U. (1986) Ice core record of the 13C/12C ratio of atmospheric CO2 in the
past two centuries, Nature, no 324, pp237239
Fujishima, A. and Honda, K. (1972) Electrochemical photolysis of water at a semiconductor electrode, Nature, no 238, pp3738
Fujishima, A., Kohayakawa, K. and Honda, K. (1975) Hydrogen production under sunlight with an electrochemical photocell,
J. Electrochem. Soc., no 122, pp14871489
Fujishima, A., Hashimoto, K. and Watanabe, T. (1999) TiO2 Photocatalysis. Fundamentals and Applications, BKC, Inc., Tokyo,
pp14176
Gerischer, H. (1972) Electrochemical techniques for the study of photosensitization, Photochemistry & Photobiology, no 16,
pp243260
Gerischer, H. (1977) The electrochemical cell, in A. Heller (ed) Semiconductor Liquid-Junction Solar Cells, The
electrochemical Society, Inc., Princeton, NJ, US, pp119
Gerischer, H. (1997) Principles of electrochemistry, in P. J. Gellings and H. J. M. Bouwmeester (eds) The CRC Handbook of
Solid State Electrochemistry, CRC Press, Boca Raton, FL, US, pp1072
Green, M. (1986) Solar Cells. Operating Principles, Technology and System Applications, University of New South Wales,
Sydney, Australia, pp1731
Gulino, A., Condorelli, G. G., Fragalh, I. and Egdell, R. G. (1995) Surface segregation of Sb in doped TiO2 rutile, Appl. Surf.
Sci., no 90, pp289295
Hartley Management Group Pty Ltd (2003) The hydrogen economy. Challenges and strategies for Australia, including the tidal
energy link, www.hartleymgt.com.au/hydrogenbroome
Hepel, T., Hepel, M. and Osteryoung, R. A. (1982) Thermodynamic and photo-electrochemical behavior of n-TiO2 electrode in
fluorite-containing solutions, J. Electrochem. Soc., vol 129, no 9, pp12311241
Hirano, S. I., Kang, S-J. L., Nowotny, J., Smar, R. St. C., Sorrell, C. C., Sugihara, S., Taniguchi, T., Yamawaki, M. and Yoo, H-I.
(1999) Development of novel materials for reduction of greenhouse gases and environmental monitoring through interface
engineering, Korean Journal of Materials Science, no 9, pp635653
Hoffman, M. R., Martin, S. T., Choi, W. and Bahneman, D. W. (1995) Environmental applications of semiconductor
photocatalysis, Chem. Rev., no 95, pp6996
Ikeda, J. A. and Chian, Y. M. (1991) Space charge segregation at grain boundaries in titanium dioxide: I. Relationship
between lattice defect chemistry and space charge potential, J. Am. Ceram. Soc., no 76, pp24372446
Jonker, G. H. (1968) The application of combined conductivity and Seebeck effect plots for the analysis of semiconductor
properties, Philips Res. Rep., no 23, pp131138
Khan, S. U. M., Al-Shahry, M. and Ingler, W. B. (2002) Efficient photochemical water splitting by a chemically modified
n-TiO2, Science, no 297, pp22432245
Kofstad, P. (1972) Nonstoichiometry, Electrical Conductivity and Diffusion of Binary Metal Oxides, Wiley, New York
Kowalski, J. M., Johnson, K. H. and Tuller, H. L. (1980) Models for the photo-electrocatalytic decomposition of water at
semiconducting oxide anodes, J. Electrochem. Soc., no 127, pp19691973
Kowalski, K., Ijjaali, M., Bak, T., Dupre, B., Gleitzer, Ch., Nowotny, J., Rekas, M. and Sorrell, C. C. (2001) Semiconducting
properties of CoO thin films, Intern. J. Ionics, nos 46, pp394399
Kroger, F. A. (1974) The Chemistry of Imperfect Crystals, vol 3, North Holland, Amsterdam, p275
Licht, S., Wang, B., Mukerji, S., Soga, T., Umeno, M. and Tributsch, H. (2000) Efficient solar water splitting, exemplified by
ReO2-catalyzed AlGaAs/Si photoelectrolysis, J. Chem. Phys. B, no 104, pp89208924
Lo, W. J., Chung, Y. W. and Samorjai, G. A. (1978) Electron spectroscopy studies of the chemisorption of O2, H2 and H2O on
the TiO2 (100) surface with varied stoichiometry: Evidence for the photogeneration of Ti+3 and its importance in
chemisorption, Surf. Sci., no 71, p199

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

214

Page 214

J. NOWOTNY, T. BAK, L. R. SHEPPARD AND C. C. SORRELL

Marucco, J. F., Gautron, J. and Lemasson, P. (1981) Thermodynamic properties of titanium dioxide, niobium dioxide and their
solid solutions at high temperature, J. Phys. Chem. Solids, no 46, pp709717
Matzke, Hj. (1981) Diffusion in nonstoichiometric oxides, in O. T. Sorensen (ed) Nonstoichiometric Oxides, Academic Press,
New York, pp155230
Mavroides, J. G., Tchernev, D. I., Kafalas, J. A. and Kolesar, D. E. (1975) Photoelectrolysis of water in cells with TiO2 anodes,
Mater. Res. Bull., no 10, pp10231030
Memming, R. (1980) Solar energy conversion by photoelectrochemical process, Electrochem. Acta, no 25, pp7788
Momirlan, M. and Veziroglu, T. N. (2002) Current status of hydrogen energy, Renewable and Sustainable Energy Reviewes,
no 6, pp141179
Morisaki, H., Watanabe, T., Iwase, M. and Yazawa, K. (1976) Photoelectrolysis of water with TiO2-covered Solar-cell
electrodes, Appl. Phys. Lett., no 28, pp338340
Nakato, Y., Shioji, M., Tsubomura, H. (1982) Photoeffects on the potentials of thin metal films on a n-TiO2 crystal wafer. The
mechanism of semiconductor photocatalysis, Chem. Phys. Lett., no 99, pp453456
Neftel, A., Moore, E., Oeschger, H. and Stauffer, B. (1985) Evidence from polar ice cores for the increase in atmospheric CO2
in the past two centuries, Nature, no 315, pp4547
Nowotny, J. (1991) Interface defect chemistry and its impact on properties of oxide ceramic materials, in J. Nowotny (ed)
Science of Ceramic Interfaces, Elsevier, Amsterdam, pp79204
Nowotny, J. (1997) Interface electrical phenomena in ionic solids, in P. J. Gellings and H. J. M. Bouwmeester (eds) The CRC
Handbook of Solid-State Electrochemistry, CRC Press, Boca Raton, FL, US, pp121159
Nowotny, J. and Rekas, M. (1994) Defect structure, electrical properties and transport in barium titanate, VII. Chemical
diffusion, Ceramics International, vol 20, no 4, pp265275
Nowotny, J., Radecka, M. and Rekas, M. (1997) Semiconducting properties of undoped TiO2 J. Phys. Chem. Solids, no 33,
pp927937
Nowotny, J., Radecka, M., Rekas, M., Sugihara, S., Vance, E. R. and Weppner, W. (1998) Electronic and ionic conductivity of
TiO2 single crystals within the n-p transition range, Ceram. Intern., no 24, pp571577
Nowotny, J., Sorrell, C. C., Sheppard, L. R. and Bak, T. (2005) Solar-hydrogen: Environmentally safe fuel for the future, Int.
J. Hydrogen Energy, vol 30, no 5, pp521544
Onishi, T., Nakato, I. and Tsobomura, H. (1975) The quantum yield of photolysis of water on TiO2 electrodes, Ber. Bunsen-Gess.
Phys. Chem., no 79, pp523525
Ooki, H. (1998) Global warming and issues for Japan in Global Environment Section (ed), The Environment Agency of Japan
Booklet #24, Tsuyosji Akiyama, Tokyo, Tokyo Shimbun Shuppankyoku (Newspaper Publishing), pp1312
ORegan, B. and Gratzel, M. (1991) A low-cost, high efficiency solar cell based on dye-sensitized colloidal TiO2 films, Nature,
no 353, pp737739
Prather, M. J. (2003) An environmental experiment with H2?, Science, no 302, pp581582
Rahman, M. M., Soga, T., Jimbo, T. and Umeno, M. (1996) Novel low-cost solid-state heterojunction solar cell based on TiO2
and its modification for improved efficiency, Jpn. J. Appl. Phys., no 35, pp33343342
Rahman, M. M., Krishna, K. M., Soga, T., Jimbo, T. and Umeno, M. (1999) Optical properties and X-ray photoelectron
spectroscopic study of pure and Pb-doped TiO2 thin films, J. Phys. Chem. Solids, no 60, pp201210
Rand, D. A. J. and Badwal, S. P. S. (2005) Australian Hydrogen Activity, Department of Industry, Tourism and Resources,
Australian Government, Canberra
Reisfeld, R., Eyal, M. and Brusilovsky, D. (1988) Luminescence enhancement of rhodamine 6a in sol-gel films containing silver
aggregates, Chem. Phys. Lett., no 153, pp210214
Ruiz, A. M., Sakai, G., Cornet, A., Shimanoe, K., Morante, J. R. and Yamazoe, N. (2003) Cr-doped TiO2 gas sensor for
exhaust NO2 monitoring, Sensors and Actuators B, no 93, pp509518

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

Ch-06

2/6/07

7:06 PM

Page 215

Solar-Hydrogen: A Solid-State Chemistry Perspective

215

Ruiz, A. M., Dezanneau, G., Arbiol, J., Cornet, A. and Morante, J. R. (2004) Insight into the structural and chemical
modification of Nb additive on TiO2 nanopartcles, Chem. Mater., no 16, pp862871
Russ, M. (1994) Cost-Effective Strategies for an Optimised Allocation of Carbon Dioxide Emission Reduction Measures,
Umwelttechnik Series (Technique of the Environment), Verlag Shaker, Aachen, Germany, pp1180
Sawatari, H., Iguchi, E., Tiley, R. J. D. (1982) Formation energies of point defects in rutile (TiO2), J. Phys. Chem. Solids, no
43, pp11471155
Seraphin, B. O. (1979) Spectrally selective surfaces and their impact on photothermal solar energy conversion, in B. O.
Seraphin (ed) Solar Energy Conversion, Springer-Verlag, Berlin, pp556
Service, R. F. (2002) Catalyst boosts hopes for hydrogen bonanza, Science, no 297, pp21892190
Sharma, R. K., Bhaynagar, M. C. and Sharma, G. L. (1997) Mechanism of highly sensitive and fast response Cr-doped TiO2
Oxygen gas sensors, Sensors and Actuators B, no 45, pp209215
Sheppard, L. R. (2004) Defect disorder and semiconducting properties of Nn-doped titanium dioxide, Ph.D. thesis (in
progress), School of Materials Science and Engineering The University of New South Wales, Sydney, Australia
Shimbun, Y. (2001) Satellite system would generate clean fuel, The Daily Yomiuri, Tokyo, 18 August
Soliman, A. A. and Seguin, H. J. J. (1981) Reactively sputtered TiO2 electrodes from metallic targets for water electrolysis
using solar energy, Solar Energy Mater., no 5, pp95102
Son, J. and Yu, I. (1996) A study on the defect structure of TiO2 (rutile) by electrical conductivity Measurements, Korean J.
Ceram., no 2, pp131136
Tang, H., Prasad, K., Sanjines, R., Schmidt, P. E. and Levy, F. (1994) Electrical and optical properties of TiO2 anatase thin
films, J. Appl. Phys., vol 75, no 4, pp20422047
Thevuthasan, S., Shivaparan, N. R., Smith, R. J., Gao, Y. and Chambers, S. A. (1997) Rutherford backscattering and
chanelling studies of a TiO2, (100) substrate, epitaxially grown pure and Nb-doped TiO2 films, Appl. Surf. Sci., no 115,
pp381385
Thomas, C. E., James, B. D. Jr and Lomax, F. D. (1998) Market penetration scenarios for fuel cell vehicles, Int. J. Hydrogen
Energy, no 23, pp949966
Tromp, T. K., Shia, R-L., Allen, M., Eiler, J. M. and Young, Y. L. (2003) Potential environmental impact of a hydrogen economy
on the stratosphere, Science, no 300, pp17401741
United Nations Environmental Programme (2002) in Full List of Vital Climate Graphics, GRID, Arendal, Norway,
www.grida.no/db/maps/collection/climate6/austral.htm
Veziroglu, T. N. (1998) Dawn of the hydrogen age, Int. J. Hydrogen Energy, no 23, pp10771978
Veziroglu, T. N. (2000) Quarter century of hydrogen movement 19742000, Int. J. Hydrogen Energy, no 25,
pp11431150
Wagner, C. (1972) The thermoelectric power of cells with ionic compounds involving ionic and electronic conduction, Progr.
Solid State Chem., no 7, pp137
Wang, Q., Lian, G. and Dickey, E. C. (2004) Grain boundary segregation in yttrium-doped polycrystalline TiO2, Acta Materialia,
no 52, pp809820
Wilke, K. and Brauer, H. D. (1999) The influence of transition metal doping on the physical and photocatalytic properties,
J. Photochem. and Photobiol. A: Chemistry, no 121, pp4953
Yahia, J. (1963) Dependence of the electrical conductivity and thermoelectric power of pure and aluminium-doped rutile on
equilibrium oxygen pressure and temperature, Phys. Rev., no 130, pp17111719
Zhao, G., Kozuka, H. and Yoko, T. (1996) Sol-gel preparation and photo-electrochemical properties of TiO2 films containing Au
and Ag metal paricles, Thin Solid Films, no 277, pp147154

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 169215

ch-07

2/7/07

2:58 PM

Page 216

Solar Heat for Industrial


Processes
H. Schweiger, J. Farinha Mendes, Ma. J. Carvalho,
K. Hennecke and D. Krger

Abstract
The idea of using solar heat in industry has been extensively discussed since the 1980s,
and several pilot plants have been built. In recent years the costs have been
substantially reduced and the technology much improved with highly efficient solar
collectors and improved system technology (controls, pumps and so on).
Several studies on industrial heat demand have confirmed that more than 50 per cent
of industrial heat demand is at temperatures in the low (< 60oC), medium (60oC150oC)
and medium-high (150oC250oC) temperature ranges. The potential is especially high in
the food industry, pulp and paper industry, and textile industry. The technical potential
for solar process heat in just the Iberian Peninsula is estimated to be 5804GWh (20.9PJ).
This corresponds to 3.6 per cent of the industrial heat demand and 0.7 per cent of the
total final energy demand of this region.
This paper is a summary of the results of the European research project POSHIP, a study
on the potential for solar heat in industrial processes, funded by the European Commission
within the 5th Framework Programme. Many industries in Spain and Portugal have been
analysed in this project. Case studies have been carried out for solar systems in industries
with favourable conditions resulting in possible solar industrial plants larger than 25,000m2.

Keywords solar thermal; process heat; application potential; industry; medium temperature

7.1 INTRODUCTION
The idea of using solar heat in industry was extensively discussed in the 1980s and
several pilot plants were built. Since the 1980s the costs have come down substantially
while technologies for high efficiency solar collectors and other system components such
as controls and pumps have improved.
These developments have improved the potential for solar energy for industrial
process heat even more. This paper is a summary of the results of the European research
project POSHIP, a study on the potential of solar heat for industrial processes, funded by
the European Commission within the 5th Framework Programme.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216260

ch-07

2/7/07

2:58 PM

Page 217

Solar Heat for Industrial Processes

217

Within this project a large number of industries in Spain and Portugal were analysed.
Case studies were carried out for solar systems in industries with favourable conditions
resulting in possible solar industrial plants larger than 25,000m2.

7.2 APPLICATION POTENTIAL


7.2.1 GENERAL REMARKS
Today, the thermal utilization of solar energy is usually confined to domestic hot water
systems and space heating at temperatures up to 60C. However, industrial process heat
has a considerable potential for solar energy as well. Industries use up to 50 per cent of the
national final energy consumption in developing countries while in industrial countries the
amount is between 25 and 40 per cent (Garg, 1987). About 30 per cent of the industrial
energy consumption is for process heat (Kreider, 1979; Winter, 1997). Considering preheating, about 50 per cent of the process heat is used at temperatures below 300C (Kreider,
1979) and up to 25 per cent is used below 180C. In total 57 per cent of the final energy is
used as process heat below 300oC (Nitsch and Luther, 1990). The conditions for solar
thermal technologies are favourable in this temperature range (Huerdes and Lachal, 1986).
The industrial uses of solar energy may become an important contribution towards
meeting the goal of supplying 12 per cent of the energy demand in the European Union
(EU) with renewable energy sources by 2010. The total potential for industrial process heat
at medium temperature (below 150oC) for the 12 countries that formed the EU in 1994 was
estimated to be 202.8TWh (million MWh). The present energy demand in the EU for
medium and medium-high temperature processes (below 250C) can be estimated to
about 300TWh, which is 7 per cent of the total energy demand (Laue and Reichert, 1994).
Favourable conditions for the application of solar energy are found in processes with
a continuous heat demand during sunlight hours and throughout the year.
These processes may, for instance, be the heating of liquid baths for washing, drying
and chemical treatments, air heating for drying and low-pressure steam generation for
several uses.
Another industrial application of solar systems is in refrigeration and cooling by means
of absorption machines or other thermal chillers. One advantage of this method is the
match between the peak demand for cooling and maximum sunlight rate.
Studies on industrial heat demand at different temperature levels have been carried
out in the US, Switzerland (Huerdes and Lachal, 1986), Sweden (Laue and Reichert, 1994),
Germany (Laue and Reichert, 1994; Schreitmller, 1987), the UK (Lewis, 1980), The
Netherlands (TNO, 1995), Japan (Laue and Reichert, 1994), Spain (Schreitmller, 1986) and
Portugal, yielding a representative overview of the typical process heat demand up to
250oC. In spite of some differences among these countries some general conclusions
were drawn from the analysis:
In all recent studies a general observation is confirmed: about 50 per cent of the

industrial heat demand is found at temperatures in the low (60oC), medium


(60150oC) and medium-high (150250oC) temperature range.
A very high percentage of the heat demand in the medium and medium-high range is
found in the food, paper, textile and chemical industries. These industries require

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216260

ch-07

2/7/07

218

2:58 PM

Page 218

H. SCHWEIGER et al

more than 50 per cent of their total process heat in the temperature range up to
200oC.
The biggest heat demand is located in the paper and food industries. A considerable
heat demand is also found in the textile and chemical industries.
Most of the process heat in the 100200oC range is used in the food, textile and
chemical industries for such diverse applications as drying, cooking, cleaning and
extraction.

7.2.2 RESULTS OF THE POSHIP STUDY


The POSHIP project analysed the heat demand for a large number of industries in Spain
and Portugal. The focus of the analysis was on the evaluation of the heat demand by the
temperature levels of the processes. Based on the results of this analysis, the potential
for implementation of solar thermal systems was studied.
Table 7.1 gives an overview of the industries analysed in the POSHIP project. Even if
the number of companies for which the available data is too little to draw exact
quantitative conclusions by extrapolating them to entire industrial sectors, the obtained
figures can be considered as a rough estimate of the potential of solar thermal energy in
industry.
Figure 7.1 shows the heat demand in the analysed industries grouped by industrial
sectors and the fractions of the heat demand at low and medium temperatures. In all
the studied sectors (except for paper industry), more than 60 per cent of the heat demand
is at temperatures below 160oC, and in several sectors nearly all heat is consumed at
temperatures below 60oC.
For all the industries the application potential of solar thermal systems has been
analysed. In Figure 7.2 the technical potential (given by the maximum available roof area
and for a maximum solar contribution of 60 per cent) and the economic potential
(economic competitiveness with fossil fuels, considering a cost reduction potential of 50
per cent for solar systems on a short term and 25 per cent public funding) for solar
thermal energy are shown. In many industries, the limiting factor is the available roof or
ground area, so that there is no difference between technical and economical potential.
Using data on the overall process heat demand in the Spanish and Portuguese
industry from Figures 7.1 and 7.2, the order of magnitude of the overall solar potential can
be estimated by simple proportional extrapolation to the whole industry (Table 7.2). The
TABLE 7.1 Overview of the industries studied in the POSHIP project (number of companies)

Source: POSHIP (2001)

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216 260

ch-07

2/7/07

2:58 PM

Page 219

Solar Heat for Industrial Processes

Total

Low and medium temperature (<160C)

219

Low temperature (<60C)

100

80

60

%
40

20

0
Food

Wine and
Beverage

Paper

Textile

Malt

Automobile

Industrial sector
Note: Industries studied in POSHIP grouped by industrial sectors

FIGURE 7.1 Distribution of the heat demand by temperatures

100% Heat load

Technical potential (%)

Economical potential (%)

100
90
80
70
60

50
40
30
20
10
0

Food

Wine and Beverage

Paper

Textile

Malt

Automobile

Industrial
Note: Industries studied in POSHIP grouped by industrial sectors
Source: POSHIP (2001)

FIGURE 7.2 Technical and economical potential of solar industrial process heat (percentage of total heat
demand)

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216260

ch-07

2/7/07

220

2:58 PM

Page 220

H. SCHWEIGER et al

TABLE 7.2 Potential for solar industrial process heat in the Spanish industry (data in GWh)

Sources: IDAE (2001) and POSHIP (2001)

Heat demand
Low and medium temp.

Total

10%

13%

5%
10%

0%

14%

41%

3%

57%
15%

18%

14%

Notes: Data for Spain, 1999 ; asterisked data estimated as no case studies were available
Source: IDAE (2001) for total heat demand in the different sectors; own data for share of low and medium temperature heat
demand

FIGURE 7.3 Distribution of the heat demand in the industry according to different industrial sectors

total technical application potential for solar industrial process heat in Spain and Portugal
can be roughly estimated to be 5.8TWh or 3.6 per cent of the total industrial heat demand.
The relative importance of each of the sectors can be seen in Figure 7.3. The category
other sectors contains all the industries that consume heat mainly at high temperatures
(for example metallurgy, ceramics and the extractive industry).

7.3 AVAILABLE SOLAR COLLECTOR TECHNOLOGY


7.3.1 INTRODUCTION
Industrial processes usually require energy within a range of temperatures, from ambient
temperature to 250C. Active solar liquid or air heating systems can be used to deliver

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216 260

ch-07

2/7/07

2:58 PM

Page 221

Solar Heat for Industrial Processes

221

FIGURE 7.4 Instantaneous efficiency for different solar collector types


Notes: Collector efficiency for a solar radiation of 1000W/m2 (800W/m2 direct normal radiation and 200W/m2 diffuse radiation).
Data based on collector gross area.
Source: POSHIP (2001)

energy to these processes. Depending on the temperature range needed, different solar
collector technologies and different concepts for their integration into the industrial
heating system can be considered. The selection of the appropriate collector type
depends mainly on the desired working temperature and the climatic conditions.
A graphical representation of the instantaneous efficiency for different collector
technologies is given in Figure 7.4.
Flat plate collectors may be used for application with temperatures up to 80C;
however, for temperatures over 80C flat plate collectors may not be appropriate because
of low efficiency. Concentrating collector technologies may be used at higher
temperatures (Rabl, 1985).
Two major types of solar collectors can be considered for industrial heat production:
1 Stationary collectors. These collectors do not use any mechanisms to track the
sun. They can produce heat at low and medium temperatures up to 150C. Flat plate
collectors, evacuated tube collectors and compound parabolic concentrator (CPC)
type concentrators belong to this group of collectors.
2 Collectors with tracking systems. Mainly one-axis tracking collectors are used
both in solar process heat plants and in large power plants for solar thermal electricity
generation. Temperatures up to 400oC can be obtained with good efficiency.

7.3.2 SOLAR COLLECTOR TYPES


7.3.2.1 Flat plate collectors
A flat plate collector (FPC) is the simplest way to transform the energy from the sun into
heat. Various methods and technologies are used in these collectors to control heat

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216260

ch-07

2/7/07

222

2:58 PM

Page 222

H. SCHWEIGER et al

losses. Radiation heat losses are reduced in most standard flat plate collectors using
selective coatings (Wackelgard et al, 2001). Heat losses can be further reduced by using
a double transparent cover, which may consist of a glass sheet and a transparent film
behind it, or two glass sheets with anti-reflective coating. The use of transparent
insulation materials is another possibility to build high efficiency stationary flat plate
collectors (Carvalho et al, 1995).

7.3.2.2 Evacuated collectors


Evacuated collectors eliminate the conduction and convection heat losses by evacuating
the space surrounding the absorber (see Figure 7.5).

7.3.2.3 CPC-type collectors (stationary concentrators)


The concentration of solar radiation can be obtained by so-called non-imaging optics,
where maximum concentration of radiation within the acceptance angle allowed for a two
dimensional geometry is given (Welford and Winston, 1978) by:

where is half the acceptance angle.


Collectors using non-imaging concentrators are also called CPC (compound parabolic
concentrator) type collectors (see Figure 7.6) since a combination of parabolas was the

Source: S. L.Viessman

FIGURE 7.5 Evacuated tube collector (ETC)

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216 260

ch-07

2/7/07

2:58 PM

Page 223

Solar Heat for Industrial Processes

223

Source: POSHIP (2001)

FIGURE 7.6 Layout of a CPC type collector with tubular absorbers

first configuration discovered to operate on the above-mentioned limit (Welford and


Winston, 1978). Mirrors are produced with the proper shape and reflect the radiation onto
the absorber.
The wide acceptance angle of these devices allows them to collect some diffuse
radiation in addition to the beam radiation. This is an advantage for this type of
concentrator in comparison with point or line focus concentrators.
For a fully stationary collector (six hours operation centered at solar noon without
adjustments), must be large, in other words the concentration must be low. For an ideal
concentrator the acceptance angle should be at least = 30, for which the concentration
would be equal to two, with eastwest orientation of the CPC-axis. If northsouth
orientation is required then the CPC collector has to be designed with a higher acceptance
angle. For the same six hours stationary operation, the acceptance angle should be =
45 with a maximum concentration of 1.4.

7.3.2.4 Tracking solar collectors


Solar tracking concentrators are classified depending on the way they track the
movement of the sun:
One-axis solar tracking and line focusing systems follow the suns movement in one

dimension, and accept varying incident angles in the other. Usually mirror and
receiver are moved as a unit. Exceptions are the fixed focus collector with moving
mirrors and a fixed receiver (Krger et al, 2000) and a roof integrated collector with a
stationary curved mirror and a moving absorber pipe, currently being demonstrated in
a project in Raleigh, US (Gee et al, 2003).
Two-axis tracking and point focusing systems are parabolic dishes, central receiver
tower with heliostats and solar furnaces.
A typical one-axis tracking collector is the so-called parabolic trough collector (PTC, see
Figure 7.7). Parabolic trough collectors are the most mature concentrating solar

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216260

ch-07

2/7/07

224

2:58 PM

Page 224

H. SCHWEIGER et al

Source: POSHIP (2001)

FIGURE 7.7 Parabolic trough collector

technology to generate heat. Reflectors with a parabolic shape concentrate the direct
solar radiation onto the receiver located in the focal line of the parabola. The receiver
consists of an absorber tube of an area usually 1525 times smaller than the aperture
area. The fluid to be heated is circulated through the absorber piping. Water and thermal
oil are typically used as working fluids.
Parabolic trough collectors have a very low thermal loss coefficient and are therefore
well suited also for applications at higher temperatures. They do not use the diffuse part of
the solar radiation; however, they make better use of the direct (beam) radiation than
stationary collectors due to the sun-tracking mechanism optimizing the position towards
the sun.

7.3.2.5 Solar air collectors


In industrial processes where large volume flows of heated air are needed (for example
drying of products), solar air collectors are recommended. The air type of solar flat plate
collectors present some advantages in comparison with the liquid types, because they
have no freezing, overheating or leakage problems; system technology and installation
are also simpler. Their disadvantages include lower efficiency, more difficult energy
storage and large pressure losses. The construction of solar air collectors is similar to
solar water heating collectors. The solar collectors which are available on the market can
be categorized as:
glazed flat plate collectors; and
unglazed flat plate collectors.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216 260

ch-07

2/7/07

2:58 PM

Page 225

Solar Heat for Industrial Processes

225

In solar air collectors pressure drop in the collector is a very important issue and because
of that sometimes the collector is designed to take into account a compromise between
thermal performance and electric energy consumption for the air flow, as can be seen
from those available on the market.

7.3.3 SOLAR SYSTEM CONCEPTS


7.3.3.1 Introduction
An industrial solar installation consists of a solar collector field through which water or
water and glycol is circulated (primary circuit). A control system controls this circulation
depending on the available solar radiation intensity. Heat exchangers are used to transfer
heat to the liquids or air.
A solar system may be coupled to a conventional heat supply system in several ways,
such as direct coupling to a specific process, preheating of water and steam generation
in the central system (see Figure 7.8).

Steam generation
Direct coupling
to the process
Central steam supply

Process

Process

Process

Return water
Feed-up water
Pre--heating of
feed up water
Source: POSHIP (2001)

FIGURE 7.8 Possibilities for the coupling of the solar system with the conventional heat supply

7.3.3.2 Industrial solar systems without storage


In many industries the heat demand is so high that there is no need for storage of solar energy.
This allows the building of very low cost solar systems by eliminating storage-related costs.
The simplest case is an industrial solar system supplying heat for a process with
continuous operation and a load always (or at least during most operating hours) higher
than the solar gains (process operating at least 12 hours per day during daytime).

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216260

ch-07

2/7/07

2:58 PM

226

Page 226

H. SCHWEIGER et al

Solar
irradiation

Heat to
process

Solar
collectors

Return
Heat exchanger
Source: POSHIP (2001)

FIGURE 7.9 Solar system without storage

In these cases, the solar system can be designed without storage. The solar heat
produced is fed directly to the process or to the heat supply system. A schematic diagram
of such a system is given in Figure 7.9.
This is an indirect system using a heat exchanger to separate the collector circuit
(primary circuit) from the load circuit, which allows the use of special fluids to protect the
collector and its materials from freezing and corrosion.

7.3.3.3 Industrial solar systems with heat storage


If, as is most common, the industrial process operates only five or six days a week and is
idle during the weekend, the system can be designed with storage of energy collected
during the weekend breaks. The stored energy is then used during the rest of the week.
Storage may also be necessary if there are strong fluctuations of the process heat
demand during the operational periods (for example demand peaks or short breaks in
operation).
The principle of a solar thermal system with storage is shown in Figure 7.10.

Solar
irradiation

Storage

Heat to
process

Solar
collectors

Return
Heat exchanger

Source: POSHIP (2001)

FIGURE 7.10 Solar system with heat storage

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216 260

Heat exchanger

ch-07

2/7/07

2:58 PM

Page 227

Solar Heat for Industrial Processes

227

7.3.4 THERMAL AND ECONOMICAL SYSTEM EVALUATION


Solar systems are used to save conventional energy in industrial processes. The
evaluation of the energy savings can be performed using various simulation methods,
such as TRNSYS.
As an illustration, a system without storage such as the one described in section
7.3.3.2 was considered. The system consists of a collector field (collector gross area of
1000m2) and the necessary control elements for the circulation of transfer fluid in the
collectors. The yearly energy delivered was determined by dynamical system simulation
using the TRNSYS program (TRNSYS, 2000).
The non-tracking collectors considered are oriented towards the south with an
inclination of about 30. For other locations, a tilt equal to latitude is close to the optimum
for annual energy gain.
Figure 7.11 shows the yearly energy gains for three locations and for the five collector
types already mentioned. The energy gains are represented as a function of the process
temperature, in other words the temperature of the fluid in the collector outlet.
It can be seen that the energy delivered decreases when temperature increases.
Standard selective flat plate collectors (FPCs) show a poor performance in all locations for
temperatures above 60C. These collectors, however, are the most economical choice for
industrial applications at low working temperatures.
For temperatures above 100C the most appropriate collectors are evacuated tube
and parabolic trough collectors. For medium temperatures below 100C non-evacuated
CPCs and evacvated flat plate collectors (EFPCs) can be used.
It must be noted that the results obtained are valid for solar systems with 100 per cent
utilization. This means that all of the produced heat can be used directly in the process
without storage. The performance of real systems will be somewhat below these values,
depending on the profile of the process heat demand.
Based on these results an economic evaluation has been performed, considering the
estimated system costs given in Table 7.3. The system costs have been estimated based
on the prices of solar collectors quoted by the manufacturers. It is assumed that the
collector field costs (including mounting, supports, foundations and field piping) amount
to 80 per cent of the total costs. The additional 20 per cent are for piping, heat
exchangers, pumps, control devices and planning. The given costs are valid for large
collector fields (1000m2).
TABLE 7.3 Total investment costs for the solar collector field related to the gross collector area
COLLECTOR TYPE
Evacuated tube (ETC)
Evac. tube with CPC
Evacuated flat-plate (EFPC)
Parabolic trough (PTC)
Flat plate (FPC)

INVESTMENT COSTS FOR COLLECTORS IN /m2


GROSS AREA
ABSORBER AREA
350
500
350
370
320
370
240

220
250

Source: POSHIP (2001)

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216260

2/7/07

228

2:58 PM

Page 228

H. SCHWEIGER et al

1000

FPC

Ener
ergy y
yiel
eld [[kWh
Wh/m2]
2]

900

Barcelona
elona

800

EFPC
ETC

700

CPC

600

PTC

500
400
300
200
100
0
60

80

100

120

140

160

180

200

Temperat
perature
ure [[C]
1000

Energy yield
y
[kWh/m
[k
/m2]

900

FPC

Lisbon
Lis
bon

EFPC

800

ETC

700

CPC

600

PTC

500
400
300
200
100
0
60

80

100

120

140

160

180

200

Temper
perat
atur
ure
e [C]
[
1000

FPC

900

Huel
uelva

EFPC

800

Ener
ergy y
yiel
eld [kWh
Wh/m2]
2]

ch-07

ETC

700

CPC

600

PTC

500
400
300
200
100
0
60

80

100

120

140

160

180

200

Temper
peratur
ure [[C]

Note: All values are related to gross collector area


Source: Own data/POSHIP (2001)

FIGURE 7.11 Yearly energy yield delivered to a process for three sites, depending on the process temperature

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216 260

ch-07

2/7/07

2:58 PM

Page 229

Solar Heat for Industrial Processes

229

Note: Only heat costs below 0.5/kWh are considered


Source: Own data/POSHIP (2001)

FIGURE 7.12 Heat costs for the 3 systems at different process temperatures

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216260

ch-07

2/7/07

230

2:58 PM

Page 230

H. SCHWEIGER et al

Figure 7.12 shows the heat costs in /kWh at process temperatures of 60C, 100C,
150C and 200C for the same locations as in Figure 7.11. The heat costs are calculated
including the investment costs for the collector field as given in Table 7.3. The annuity
(yearly amortized payment plus interest) is assumed to be 10.5 per cent based on a
15 years lifetime and 6 per cent interest. Maintenance costs are assumed to be
2.5/m2/year for stationary collectors and up to 5/m2/year for parabolic trough
collectors.
The solar heat costs for the most economic solution for each temperature range vary
from 0.04/kWh to 0.22/kWh depending mainly on the climate and the working
(process) temperature. The climatic conditions should therefore be carefully regarded
during planning.
Applications requiring temperatures below 150C can be supplied with solar heat at
significantly lower costs then those operating at higher temperatures. In the central and
northern Mediterranean climates of the Iberian Peninsula (for example Barcelona, Lisbon
or Madrid), heat costs can be below 0.08/kWh for a supply temperature of 100C. In the
southern Portuguese and Spanish climates solar heat at this temperature can even be
delivered at costs down to 0.04/m2.

7.3.5 EXISTING PROJECTS AND PROJECTS IN PROGRESS


Since the 1980s, several solar thermal systems for industrial applications have
been developed and are currently operating. A list of existing projects is given in
Table 7.4.
In the US several parabolic trough installations with collector areas between 500m
and 2500m were erected in the 1990s by Industrial Solar Technology (IST). They are
operating reliably with little maintenance. Several companies in Europe and the
Mediterranean offer parabolic trough collectors for the temperature range between 50C
and 300C.
A solar cooling system involving a double-effect absorption chiller driven by parabolic
trough collectors was installed in 2004 at a hotel in Sarigerme, Turkey (the project
planning is discussed in Krger et al, 2002). The chiller has a coefficient of performance
(COP) of 1.2 (according to the producer), which is significantly higher than the COP of 0.7
of the more common single-effect absorption chillers. In this case, only 58 per cent
thermal energy is necessary to produce the same amount of cooling, resulting in a
considerable reduction of the solar field area. The chiller is driven by saturated steam at a
pressure of 4 bars from a steam generator which receives pressurized hot water from the
solar field.
Paraboloid mirrors for heat production are mainly in use as small solar cookers. More
than 200 mirrors of a larger version, the so-called Scheffler Mirror, are in use as solar
cookers, including an installation for steam generation. (For more information see
www.charity-india.de.)
Two existing solar thermal systems for industrial applications are shown in Figure 7.13
(page 238).

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216 260

Mapag, Austria

Gillich Galavanik

First Class Holz

Peter Lisec GmbH MEA

Bilderland GmbH

Mapag

Gillich

Hamminger

Lisec

Bilderland

Xsolar

ISL

Solution Solartechnik
GmbH, Austria

Sunwash, Austria

Graz

Hausmening

Lohnsburg

Graz

Gumpoldskirchen

Kflach

1979

1981

1994

1999

2001

2002

Actual state

Temperature

60

60

of use

Drying of high quality


wood boards
Hot water & heating
for workshop

dismantled Hot water for photo


(operating
processing
19791999)

40

Air 50C

Air 60C

Degreasing and
75
removal of laquer from
metal parts in baths

Cleaning of samples
for analytical lab

Self service car wash


stations

Process

shut down Drying of laquered


machine parts
(operating
19811998)

operating

operating

operating

operating

2003

(October )

Type of

Collector

FPC

FPC

FPC

ETC

FPC

FPC

(1)

1284

916

88

1x10
1x32

45

(2)

collectors area

2:58 PM

Sunwash
Viessmann

operation

construction)

Year of
start of

(design and

Site

Companies

Company (user

of the plant)

2/7/07

AUSTRIA

Short name

TABLE 7.4 List of solar process heat plants

ch-07
Page 231

Solar Heat for Industrial Processes


231

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216260

FPC
FPC
FPC

Freudenstadt
Walz
Barnau

Clean Park

Filling Station

JVA Barnau

FREUDENSTADT

WALZ

BARNAU

FPC

FPC

PTC

ETC

170C

Mlheim

Refuse Collection
Department

Thermal

process heat for


pharmaceutical
company

Sun Car Wash


Mlheim

in
commissio
ning (first
steam was
produced
on 20 Oct
2003)

(1)

MLHEIM

1998

2003

Type of

Collector

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216 260

89

24

24

25

151

770

1200

(2)

collectors area

Dresden

Targassonne

Cairo

Temperature
of use

Stadtreinigung
Dresden GmbH

Basic design,
specification and site
supervision by Fichtner
Solar GmbH, detailed
design, construction and
erection by Lotus

2003

Process

DRESDEN

GERMANY

TARGASSONNE

FRANCE

user of the
steam: El Nasr
Pharmaceuticals
and Chemicals

Actual state

232

(October )

2:58 PM

NREA

operation

construction)

Year of
start of

(design and

Site

Companies

Company (user

of the plant)

2/7/07

EGYPT

Short name

TABLE 7.4 List of solar process heat plants (Contd)

ch-07
Page 232

H. SCHWEIGER et al

Tannery; preheating of
water for the steam
boiler

1993

Athens

Tripou-Katsouri
S.A.

TRIPOUKATSOURI

Dairy; yoghurt
production

1993

Korinth

Mandrekas S.A.

Textile Dyeing and


Finishing; preheating
of water for steam
boiler

1993

MANDREKAS

Wine producer: Bottle


washer

1993

Patras

Achaia Clauss,
S.A.

ACHAIA CLAUSS

Heraklion

Greenhouse; Space
heating

1994

Kozani

Kozani
Greenhouses S.A.

KOZANI

KASTRINOGIAN Kastrinogiannis
NIS
S.A.

Cosmetics stock
warehouse; adsorption
cooling

1998

Dairy; preheating of
water for steam
boiler; washing
machine

2000

Oinofita

Thessaloniki

Dairy; Preheating of
water for steam boiler

2000

Sarantis S.A.

Mevgal S.A.

MEVGAL

Thessaloniki

Pulling

40

70-85

40-45

40

40

FPC

308

170

180

308

80

2700

727

576

138

2:58 PM

SARANTIS

Alpino S.A.

Krimmer Pulling

2/7/07

ALPINO

GREECE

PULLING

ch-07
Page 233

Solar Heat for Industrial Processes


233

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216260

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216 260

Quinta do Mendanha, 1987


Carregado

Pretec

Knorr Best Foods


S.A.

KNORR

operating

out of
service

1985

Vialonga

out of
service

operating

2003

1985

Sociedade Central TECSO, Tcnica Solar


de cervejas
Lda. (Valado de Frades
Nazar, Pt.)

guas de Moura

2001

VIALONGA

TECNIVEST, SA

Univ. Mannheim Modulo Cuernavaca


Solar / AIGUASOL

UCAL

Guetermann
Mexico

Pisticci

UCAL, AGUAS
DE MOURA

PORTUGAL

GUETERMANN

MEXICO

PISTICCI

Actual state
(October )

Dairy; hot water for


washing of work tools

Brewery

Dairy

Adsorption cooling of
production hall

Chemical

Process

Temperature

40-45

63

188

of use

Type of

Collector

1120
*1280

438

1728

(2)

area

None
440
selective
flat plate

Non
192.8
evacuated
1.5 x CPC

PTC

PTC

PTC
92 axis)

(1)

collectors

2:58 PM

ITALY

operation

construction)

Year of
start of

(design and

Site

Companies

Company (user

234

of the plant)

2/7/07

Short name

TABLE 7.4 List of solar process heat plants (Contd)

ch-07
Page 234

H. SCHWEIGER et al

food industry:
farinaceous products
factory (Drying)
food industry
(Pasteurization)

1988

1983

Hallau

HALLAU

Desalination of water

Hattwill

shut down

80-120

1981

Alcorcn (Madrid)

Gran Canaria

207 (1)

1982
dismantled Sterilisation of meat
definitively (operating products (steam for
1985
19841988) back-up)

Merida

dismantled Processing of steam


(since 1987 for pre-sterilisation
out of use)

20-300

dismantled Cold generation with


(operating
absorption machine
19831988)

1995

Washing of cisterns,
cars, etc

1994

23-26

Juzbado
(Salamanca)

Huelva

Fish growing

1996

HUTTWILL

SWITZERLAND

Vinejard

Lactaria Espaola

LACTARIA
CASTELLANA

ARINAGA

Carnes y
Conservas
Espaolas S.A.

CARCESA

Autoclavados
Carte S.A.

CARTE

Huelva

Olives; preheating of
water

1997

ETC

ETC

500

400

360

600

1024

PTC

PTC
(CasaAuxini)

1080

PTC

138

1316

260
2:58 PM

Empresa Nacional IER-CIEMAT ENDESA /


de Uranio S.A.
GESA

Acuinova
Andalucia S.A.

ACUINOVA

Sevilla

2/7/07

ENUSA

TE-PE S.A.

TE-PE

SPAIN

ch-07
Page 235

Solar Heat for Industrial Processes


235

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216260

Hotel Grand
Kaptan

ALANYA

Austin Energy

Solargenix Energy Solargenix Energy

Phoenix Federal
Prison

Austin Energy

Parker Lincoln
Power Roof

PHOENIX

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216 260

IST

Solargenix Energy

Ford

SOLEL

Solitem

Solitem

FORD Visitor
Center

U.S.A.

Iberotel
Sarigerme Park

operating

operating

operating

2002

1999

Phoenix, Arizona

Raleigh, NC

Austin, TX

2003

hot water

50-ton 2E absorption
chiller

30-ton 1E absorption
chiller

In
Absorption
construction

2003

operating

in
Absorption Cooling
construction and Laundry

Process

2003

2003

Dearborn (Michigan) 2003

Alanya

Sarigerme

Actual state

Temperature

60

170

95

144
(steam)

144
(steam)

of use

Type of

Collector

240

180

180

(2)

area

PTC

1673

Tracking 530
RoofIntegrate
d Power
Roof

Fixed nonevacuated
CPC

PTC

PTC

PTC

(1)

collectors

236

(October )

2:58 PM

SARIGERME

operation

construction)

Year of
start of

(design and

Site

Companies

Company (user

of the plant)

2/7/07

TURKEY

Short name

TABLE 7.4 List of solar process heat plants (Contd)

ch-07
Page 236

H. SCHWEIGER et al

IST

ACDF

Tehachapi, California

TEHACHAPI

320-ton 1E absorption
chiller plus space
heating

Shut down
in 1987

1980

Textile
Malt

operating
operating

Druomg

Laundry

operating

1980

Hot water pre-heating

20-ton 2E absorption
chiller and space
heating

process-water

ICPC VTC
Collectors
Operating

1983

Notes: (1) PTC: parabolic trough collectors; ETC: vacuum tube collector; FPC: flat plate collector
(2) Useful collector area
Source: POSHIP (2001)

Texas

SAN CH07
IST

Georgia

MACON

California
Correctional
Institution

Alabama

FAIRFAX

Gotham Construction Co. New York, NY

California

PASADENA

Wagner College, Wagner College


NYC

Arizona

CHANDLER

1986

1997

105/55

110

95

85

83 to 165

1000

5620

725

1150
2676

PTC

1000

900

PTC

PTC

PTC

Sunmaste 1100
r ETC
with
external
CPC
reflectors

PTC

PTC

PTC

Non107
tracking
Integrated
Compound
Parabolic
Concentra
ting ETC

2:58 PM

Brighton, Colorado

Sacramento, CA

2/7/07

Adams County
Detention Facility

Solar Enterprises
International and BES

Sacramento CEC Bergquam Energy


Project
Systems (BES)

ch-07
Page 237

Solar Heat for Industrial Processes


237

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216260

ch-07

2/7/07

2:58 PM

238

Page 238

H. SCHWEIGER et al

Top: Production of hot water with parabolic trough collectors, California, US; Bottom: Knorr factory, Portugal
Source: POSHIP (2001)

FIGURE 7.13 Solar industrial process heat plants in operation

7.4 GUIDELINES FOR EVALUATION AND SYSTEM DESIGN


In the following sections some general guidelines for the feasibility evaluation and design
of solar industrial process systems are outlined. A summary of the evaluation criteria is
given in Table 7.6.

7.4.1 FEASIBILITY ANALYSIS EVALUATION CRITERIA


In this section the most important aspects to be taken into account for the feasibility
evaluation of solar industrial process heat plants are discussed.

7.4.1.1 Selection of appropriate interfaces for the coupling of a solar system


As a first step in the feasibility study, the most appropriate interfaces (processes) of
coupling a solar system to the existing heat supply have to be selected. The selection
criteria are the following:
Low temperature level. Solar heat at temperatures above 150C is technically feasible

but economically reasonable only at favourable locations. Applications at low


temperature (60C) offer the best economics;

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216 260

ch-07

2/7/07

2:58 PM

Page 239

Solar Heat for Industrial Processes

239

Continuous demand (otherwise storage is needed); and


The technical possibility of introducing a heat exchanger for the solar system in the

existing equipment or heat supply circuit.


Whenever possible, a direct coupling of the solar systems to one or several processes is
preferred, as the working temperatures in this case are lowest (see section 7.3.3). Direct
coupling to a process can mainly be done in the following two ways:
1 Preheating of a circulating fluid (e.g. feed water, return of closed circuits, air preheating): this possibility exists if fluid circulation is either continuous or periodic (e.g.
periodic replacement of water in baths). If circulation is discontinuous, a storage tank
has to be introduced.
2 Heating of liquid baths or hot (drying) chambers: the energy demand in this case can be
divided into the demand for heating at the operational start-up in the early morning hours
or after each replacement of bathwater and the demand to maintain the temperature
throughout the operation, which is generally constant during the operating hours.
The existing heat exchangers for bath heating generally require steam at temperatures
that are too high for a simple solar system. The introduction of additional heat exchangers
with a larger exchange area into existing baths is not always possible, due to lack of space
or other technical restrictions. In some cases, an external heat exchanger in combination
with a circulation pump can be used.
If the process baths are well insulated, they can be used for solar heat storage, in
other words maintaining the temperature of the solar system during a weekend without
operation can reduce the heat demand for start-up on Monday morning.
In almost all industries, the coupling of the solar system to the central heat supply system
is possible. This can be done either by the preheating of the feed water for the steam
boilers (the temperature level in this case increases with increasing condensate recovery)
or by a solar steam generator. The latter is only recommended in climates with a high level
of solar radiation. Concentrating collectors are the best choice for this kind of application.

7.4.1.2 Influence of the working temperature


The influence of the working temperature on the useful solar energy gains has been shown
in section 7.3. The upper limit for the working temperature depends on the climate.
As a rule of thumb, it can be stated that the solar systems for temperatures above
100C are only recommended in high insolation regions (southern and central regions of
the Iberian Peninsula). In the northern regions only low temperature systems should be
considered.
It has to be taken into account that the working temperature in the solar system must
always be somewhat higher than the required process temperatures, due to losses in the
piping and the temperature drop in the heat exchangers. In well-designed systems this
temperature difference can, nevertheless, be less than 10K.
In the case of liquid bath preheating, the mean working temperature of the solar
system is lower than the required final process temperature. The smaller the solar fraction

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216260

ch-07

2/7/07

240

2:58 PM

Page 240

H. SCHWEIGER et al

(percentage of the heat demand covered by the solar system), the lower the mean
working temperature. In very low solar fractions, the mean working temperature may be
close to the fluid inlet (or return) temperature.

7.4.1.3 Continuity of the load and storage


In order to obtain a reasonable economic performance, solar systems should be designed
for close to 100 per cent utilization. This means that the heat demand should always be
higher than the maximum possible output of the solar system. Otherwise, and if no
storage is used, the useful heat drawn from the solar system is reduced.
Short-term mismatches of consumption and demand (daily demand profile) can be
buffered by a small short-term storage (up to 25 litres per m2 of solar collector) that does
not significantly increase the overall system costs (Figure 7.14).
Heat storage for breaks of one or several days (for example weekends or public
holidays) is more cost intensive. For a two-day weekend storage, about 250 litres per m2 of
collector are necessary. Weekend storage is generally not recommended for small systems.
As a rough estimation, the annual solar gains of a real system without long-term
storage can be calculated as:
Qreal = Qideal ndays / 365
where Qideal is the solar energy gain of an ideal system with 100 per cent utilization and
ndays are the days of operation per year.
Nevertheless, if holidays are concentrated during the summer (the period with
maximum potential solar production) the reduction of solar gains during this period is
underestimated by this simple equation. Systems with only seasonal utilization (less than
six months of operation a year) are generally not economical.

Ttank (C)
P(W)
DISCHARGE
CHARGE
USEFUL SOLAR
HEAT

DISCONTINUOUS

PROCESS

OF THE STORAGE

STORAGE
TEMPERATURE

HEAT

00:00 h

Source: POSHIP (2001)

FIGURE 7.14 Short-term mismatch between heat demand and available solar heat

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216 260

24:00 h

ch-07

2/7/07

2:58 PM

Page 241

Solar Heat for Industrial Processes

241

7.4.1.4 Climatic zones on the Iberian Peninsula


The distribution of the global solar radiation on a horizontal surface throughout the Iberian
Peninsula is shown in Figure 7.15.
The geographic regions with very good meteorological conditions (H > 1600kWh/m2a)
are the south of Portugal, Andaluca, Murcia, Valencia and Central Spain. Medium levels
of radiation (14001600kWh/m2a) are obtained in Catalonia, Galicia and northern Portugal.
A relatively low level of radiation (H < 1400kWh/m2a), comparable to Central European
climates, dominates in the northern coast of Spain.
Regarding the Spanish and Portuguese islands, a very high level of radiation exists on
the Canary Islands, a medium level (about 1600kWh/m2a) on Madeira and a low level on
the Azores.
A strong influence of the climatic conditions on the annual solar gains can be seen
from the results presented in section 7.3.4.

7.4.1.5 Summary of evaluation criteria


The influence of a combination of determining factors working temperature level,
demand continuity and solar radiation level are shown in Figure 7.16. The costs of useful
solar heat for several industries are plotted as a function of these factors.
In low temperature systems (60C), reasonable energy costs (2060/MWh, without
subsidies) can be obtained even in non-optimum conditions (medium level of solar
radiation or discontinuous demand). Medium temperature systems with working

Bilbao

Barcelona
Madrid
Lisboa

Huelva
1200-1300
1300-1400
1400-1500
1500-1600
1600-1700
1700-1750
Source: Meteotest

FIGURE 7.15 Global solar radiation on a horizontal surface (H) on the Iberian Peninsula

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216260

2/7/07

2:58 PM

242

Page 242

H. SCHWEIGER et al

temperatures above 100C at present are only recommended if all other conditions are
favourable (high solar radiation and continuous demand zones).
Another important factor to be considered is the systems size. System costs for small
systems (100m2) can be more than 50 per cent higher than those of large solar systems
(> 1000m2), with a corresponding increase in heat costs.

7.4.2 GUIDELINES FOR SYSTEM DESIGN


7.4.2.1 Solar collector field
The selection of an appropriate solar collector basically depends on the operating
temperatures (Table 7.5), though other aspects, such as the possibility of roof integration
or system size, also have to be considered.
The necessary collector field size can be estimated from the results presented in
section 7.3. The maximum annual energy gain that can be obtained from a solar thermal
system varies from 350 to 1100kWh/m2, depending on the site and the working
temperature. For an accurate design and sizing of the systems, dynamical system
simulations should be carried out taking into account the specific demand profile. The
peak thermal power of solar collectors is about 500W/m2 for medium temperature and as
high as 800W/m2 for low temperature systems.

160

140

4
11

120

Energy cost [/MWh]

ch-07

100

17
3

20

80

22

23

24
60

18
19

15

10

12
1

16
40

13
20

21
2
14

0
0

20

40

60

80

100

120

140

Working temperature

Notes: Colours: very high (black), high (grey), medium (white) solar radiation.
demand;

: systems with continuous

: Systems with continuous demond during the week and weekend breaks;

: systems with

continuous demand during the week and weekend breaks, but with long-term (weeked) storage; : systems
with only seasonal operation.
Source: POSHIP (2001)

FIGURE 7.16 Costs of useful heat for the different systems as a function of the mean annual
working temperature

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216 260

160

ch-07

2/7/07

2:58 PM

Page 243

Solar Heat for Industrial Processes

243

TABLE 7.5 Selection of solar collector types according to working temperatures


TEMPERATURE RANGE
40C
4070C
70100C
> 100C

PROCESS
Unglazed selective collectors or low cost standard flat plate collectors
Highly selective flat plate collectors or CPC collectors
CPC collectors, evac. tubes or other high efficient stationary
collectors; concentrating collectors for medium and large systems
Concentrating collectors; evac. tubes with CPC

Source: POSHIP (2001)

Optimum annual heat gains per unit collector area in the Iberian Peninsula (latitude from
37 to 42) are obtained for non-tracking systems at collector inclinations of 30 and a row
separation of at least 1.5 times the gross collector (row-) height. The fraction of gross
collector area to roof area in this case is less than two thirds. If a limited roof area is available
and maximum total solar heat gains are desired, the collector inclination can be set down to
10 approximately. Nevertheless, for glazed collectors a lower limit of 20 for the collectors
inclination is recommended in order to avoid rain water infiltration and enhance self-cleaning
of the glass sheet by rain water. A deviation from south orientation of up to 45 can be
tolerated, leading to about a 10 per cent reduction in the systems output.
Parabolic trough collectors should be tracked around an axis oriented NS, to maximize
the annual heat production. The row spacing should be about three times the aperture width
to avoid excessive shadowing in the morning and evening. By an appropriate design of flow
rates, pipe diameters and pipe insulation, the electricity consumption for fluid circulation can
be below 1 per cent of the overall heat gains. Thermal losses in the piping and storage
should not be above 5 per cent of the overall heat gains for medium and large size systems.
In earlier studies, depending on the load characteristics and profile, different solutions
have been studied, like single-pass versus multi-pass strategies (Gordon and Rabl, 1982;
Collares Pereira et al, 1984) or constant versus variable flow rate (Gordon and Zarmi, 1985)
in order to assure good temperature stratification in the storage.
Solar gains are highest if the load and solar supply flow rates are matched with the
help of flow regulators. A bypass for the preheating of the primary circuit during the
morning hours and the extraction of the residual heat after sundown by appropriate
regulation can improve solar gains by up to 20 per cent.

7.4.2.2 Storage
An introduction to solar systems with heat storage has been given in section 7.3.3.
Short-term heat storage is recommended whenever a mismatch between available
solar radiation and heat demand occurs. For short-term storage (several hours), volumes
of about 25 litres/m2 are recommended. Short-term storage may even be recommended
for continuously operating processes, in order to lower the mean working temperature of
the solar system and thereby improving its efficiency, especially if low cost solar
collectors with high thermal loss coefficients are used.
The larger the systems size, the more effective the heat storage over longer periods
(for example weekends). Weekend storage becomes economic from about 500m2 and
above. Weekend storage volumes should be about 250 litres/m2. In this case, storage

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216260

2/7/07

244

2:58 PM

Page 244

H. SCHWEIGER et al

costs can range from 10 to 20 per cent of the total system costs. Storage for longer
periods (seasonal storage) can only be considered for very large systems (> 5000m2).

7.4.2.3 Coupling to the existing heat supply and control


Coupling the solar system to the existing heat supply should always be attempted at the
lowest possible temperature. Nevertheless, for liquid and air preheating, solar heat should
Solar
irradiation

Hot water
to process

Solar
collectors

Waste
water
Heat
exchanger
Heat
recovery

Source: POSHIP (2001)

FIGURE 7.17 Combination of solar thermal system and waste heat recovery

3000
2500
2000
MWh

ch-07

1500
1000
500
0
solar heat

solar heat &

cogeneration

cogeneration

electricity

high eff.

low eff.

Note: Conversion efficiency (electrical) for: (a) highly efficient (combined cycle) cogeneration plants (50%); (b) low efficiency
cogeneration plants (25%)
Source: POSHIP (2001)

FIGURE 7.18 Primary energy saving with respect to a conventional steam boiler, for 1000MWh of
industrial process heat produced either by solar thermal systems or conventional cogeneration system

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216 260

ch-07

2/7/07

2:58 PM

Page 245

Solar Heat for Industrial Processes

245

TABLE 7.6 Evaluation criteria for solar industrial process heat systems for the Iberian Penninsula
CRITERION
Working temperature

INFLUENCE ON SYSTEM PERFORMANCE


Working temperatures not higher than 150C, best performance
below 100C
Very good conditions in the southern and central regions of the
Iberian Peninsula.
In regions with medium or low levels of solar radiation
(H < 1600kWh/m2) systems should only be considered if all other
conditions are favourable (low temperature, continuous demand).

Climate

Continuity of the demand:


Annual variation
Daily variation

System size

Annual solar gains


Solar fraction
Available roof or ground area

Static load aspects of the roof

Waste heat recovery and cogeneration

Breaks in summer reduce system performance. Losses in solar gains


are more than proportional to the time interval of the break.
Continuous demand or demand with peaks at daytime are favourable.
Short interruptions (several hours) can be buffered by low volume
storage with little increase in system cost.
The economic performance of solar thermal systems depends
strongly on the system size. Resulting solar energy costs are up to
50% lower for large systems (> 1000m2, thermal peak power >
0.5MW) than for small systems (100m2, thermal peak power 50kW).
Annual solar gains of a solar system should be at least 500 kWh/m2
for economic profitability.
Systems should be designed for solar fractions not higher than about
60% (for continuous demand).
Sufficient roof or ground area should be available in order to obtain
solar fractions from 5 to 60%.
Orientation to the south with inclination of about 30 is the optimum.
Small deviations from these values are tolerable (45 from south
orientation, 15 from optimum inclination).
Long piping should be avoided.
The need for reinforcement of roof structures increases system cost
and therefore reduces economic performance. The additional static
load of solar collectors is 2530kg/m2 for standard collectors.
First all means of improving energy efficiency by waste heat recovery
and by co-generation should be exploited. Solar systems should be
designed to cover (part of) the remaining heat demand.

Source: POSHIP(2001)

be introduced only after first preheating by waste heat recovery systems (Figure 7.17), and
not as an alternative to these systems. Even if the waste heat recovery raises the working
temperature in the solar system, the combination of both systems yields better results
than a solar system at lower temperature without heat recovery.
If solar heat is being fed to several processes, a control strategy should be chosen
in order to reach an optimum overall energy saving. In most cases feeding solar energy

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216260

ch-07

2/7/07

246

2:58 PM

Page 246

H. SCHWEIGER et al

to the process with the lowest temperature might be the best choice, but in some cases
giving preference to heat production at higher temperatures during the hours around
noon with high insolation results in a better overall performance. The control strategy
should be optimized for the specific cases using dynamic system simulation
techniques.

7.4.2.4 Solar thermal systems and cogeneration


As a general rule, it is possible to state that solar thermal systems should be designed
as a complementary system to cogeneration plants covering the remaining heat
demand.
Only in some cases can solar thermal energy be considered as an alternative heat
supply system. Shutting down an existing cogeneration plant during periods when solar
heat is available in excess would imply a higher importation of electricity from the grid
during this period. In Figure 7.18 the primary energy saved by producing 1000MWh of
process heat by solar thermal systems or by cogeneration, instead of a conventional
steam boiler, is compared. The primary energy savings for the produced electricity are
taken into account, using as a reference the mean efficiency of the present electricity
supply (about 35 per cent). In addition, the possibility of simultaneous heat and electricity
production from solar thermal systems is considered, assuming a conversion efficiency of
30 per cent from heat to electricity.
It can be seen that solar heat saves primary energy compared to cogeneration plants
only if either the electrical efficiency of the installed cogeneration plant is lower than the
mean efficiency of the grid, or if the solar system itself produces both heat and
electricity.

7.4.2.5 Summary
The most important evaluation criteria for solar process heat plants are summarized in
Table 7.6.

7.5 CASE STUDIES


Within the POSHIP project, case studies have been carried out for more than 20
companies in Spain and Portugal that showed some interest in applying solar thermal
energy in their factories. Based on a detailed analysis of the heat (and cold) demand of
the industries, appropriate solar systems have been proposed. A prediction of the annual
energy gains of these systems based on dynamic system simulations (TRNSYS) and
economic analyses have been performed.
A description of some selected systems is given in the following sections. For an
overview see Table 7.7. The geographical locations of the proposed demonstration plants
are shown in Figure 7.19.

7.5.1 MALTING FACTORY, ANDALUCA (SPAIN)


The malting factory is located in the province of Seville, in southern Spain. The malt
produced is used for beer production in the same factory. About 80 per cent of the heat
and cold consumption in the malting factory is used for air heating for drying the malt (hot

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216 260

ch-07

2/7/07

2:58 PM

Page 247

Solar Heat for Industrial Processes

247

TABLE 7.7 List of the most relevant parameters of the demonstration projects proposals
COMPANY
SECTOR
LOCATION

PROCESS

WORKING
TEMPERATURE
RANGE

COLLECTOR
TYPE**

PROJECTED
COLLECTOR
AREA (m2)

ESTIMATED
COST ()

Malte Ibrica
Malting factory
Poceiro, Portugal

Air for malt


drying
preheating

Low temperature
(2060C)

FPC or CPC

3500

624,000

Heineken Espaa S.A.


Malting factory
Sevilla, Spain

Air preheating
for malt drying

Low temperature
(1080C)

SFPC

3046

930,000

Beirala Lanificios
Textile factory
Seia, Portugal

Water preheating Low temperature


in dyeing process (2080C)

FPC or CPC

1750

340,000

CPC

60

55,000*

Bodegas Mas Martinet Cooling of wine Medium


Wine cellar
stores and space temperature
Tarragona, Spain
heating
(5085C)
Notes: * cost for solar system and absorption cooling system.

**FPC: flat plate collector; CPC: compound parabolic concentrator; SFPC: selective flat plate collector.
Source: POSHIP (2001)

Source: POSHIP (2001)

FIGURE 7.19 Geographical location of the proposed demonstration plants

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216260

ch-07

2/7/07

248

2:58 PM

Page 248

H. SCHWEIGER et al

Cold generation
19%

Malt drying
81%
Source: POSHIP (2001)

FIGURE 7.20 Heat demand by processes in the malting factory in Seville

air at about 60C). Most of the remaining 20 per cent is required for air cooling in the
germination process (Figure 7.20).
The profile of the heat demand in this factory is ideal for the implementation of a solar
system:
The heat demand is high and continuous during 7 days a week and 24 hours a day,

so nearly all the collected solar energy can be used directly in the process without
need of large storage volumes.
The required temperatures are low. Air is preheated by the solar system only a few
degrees above ambient temperature.
The factory is located in a region with a high solar radiation (1770kWh/m2a).
Energy costs are a significant cost factor in the malt production process, so a
reduction of energy consumption can help improve competitiveness.
A solar system is proposed for the preheating of air for the malt drying process. An
additional hot water to air heat exchanger will be installed in series before the existing
heat exchanger using steam from the conventional heat supply system. The solar system
consists of a solar collector field of a total surface of 5000m2 and a storage tank of 1000m3
(Figure 7.21 and Table 7.8). The solar collectors will be integrated in the saw-tooth shaped
roof structure forming a rain-tight solar roof.

7.5.2 MALTIBRICA SOCIEDADE PRODUTORA DE MALTE S.A. (PORTUGAL)


Maltibrica is a PortugueseSpanish factory producing malt for breweries pertaining to
the UNICER Group in Portugal and also exporting it to Spain. Although it is located not
far from a very industrialized zone of Portugal, the Setbal Peninsula, around 50km from
Lisbon, no other factories are located nearby, increasing the difficulties to extend the
natural gas network to that area. As a result, it is very expensive to change from the
actual usage of heavy fuel oil in great quantities to a cleaner conventional energy
source.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216 260

ch-07

2/7/07

2:58 PM

Page 249

Solar Heat for Industrial Processes

249

Source: POSHIP (2001)

FIGURE 7.21 Scheme of the proposed solar system for the malting factory in Seville

TABLE 7.8 Characteristic data of the proposed system for solar air heating in a maltting factory
in Seville, Spain
Company
Location
Ind. sector
Process
Working temperature
Solar collector area
Solar collector type

Heineken Espaa S.A.


Seville, Spain
Malting factory
Air preheating for malt drying
1080C
5000m2
Selective collectors without glass cover integrated into
a saw-tooth roof structure
1000m3
4770MWh
7%
1,320,000
410,000
105,000
4 years

Solar storage
Annual solar gains (useful heat)
Solar fraction
Total investment
Total own contribution
Yearly net savings
Payback
Source: POSHIP (2001)

The factory produces malt, which undergoes three steps during the process:
1 wetting of barley;
2 germination of green malt (cold); and
3 drying of malt (heat).
The largest part of the energy consumption is for the production of superheated water
(180C, 14 bars), which is used with the waterair heat exchangers for drying the malt

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216260

ch-07

2/7/07

250

2:58 PM

Page 250

H. SCHWEIGER et al

Source: POSHIP (2001)

FIGURE 7.22 Malting factory at Poceiro, Portugal: Ground layout

(6080C). A small part is consumed for keeping thick fuel at its appropriate storage
temperature. The factory is quite new (less than 10 years old) and its design incorporates
heat recovery measures that improve the potential use of solar energy when considering
new energy-saving actions.

preheated air
2550C

Collector field

Source: POSHIP (2001)

FIGURE 7.23 Layout of the proposed solar system for the malting factory at Poceiro, Portugal

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216 260

ch-07

2/7/07

2:58 PM

Page 251

Solar Heat for Industrial Processes

251

There are several favourable factors for the implementation of a solar system in the
factory:

the factory is in operation all day throughout the year round;


energy can be used at the time of its collection, which avoids the need solar storage;
solar energy is used at a low temperature to preheat the outside air for drying;
the factory is located in a site with very good annual sunny conditions: more than
2800hours of sunshine and 1720kWh/m2a;
the solar energy system would remain totally independent from the existing
conventional system.
The ground layout of the factory is given in Figure 7.22 and the layout of the proposed
solar system in Figure 7.23. The most relevant data of the proposed plant are summarized
in Table 7.9.
TABLE 7.9 Characteristic data of the proposed system for air preheating for malt drying in Portugal
Company
Location
Ind. sector
Process
Working temperature
Solar collector area
Collector type
Storage volume
Annual solar gains (useful heat)
Solar Fraction
Total investment
Total own contribution
Yearly net savings
Pay back

Grupo Unicer
Poceiro (Setbal)
Malting factory
Air preheating for malt drying
2080C
3500m2
Selective flat plate or CPC collectors
3762MWh/year
20%
624,121
405,679
118,320
4 years

Source: POSHIP (2001)

7.5.3 BEIRAL LANIFCIOS S.A. (PORTUGAL)


Beiral is a Portuguese textile factory that produces woollen garments and cloth. It is
located in the north-west of central Portugal, at the foot of the highest Portuguese
mountain, Serra da Estrela, with average annual solar radiation conditions: 1520kWh/m2a
and 2400 hours of sunshine (Figure 7.24).
In the dyeing process, only a small fraction of the process heat consumption (20 per
cent) is at 120C and most does not exceed 80C. A great quantity of hot water at 6070C
is consumed for washing the dyeing vessels.
There is a variety of solar collector technologies available on the market for this level
of temperature appropriate for this application.
Favourable conditions for the implementation of a solar system apply due to the
following:

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216260

ch-07

2/7/07

2:58 PM

252

Page 252

H. SCHWEIGER et al

Source: POSHIP (2001)

FIGURE 7.24 Solar radiation in Portugal and location of the Beiral plant
access to the national natural gas network is not expected in the short term;
due to the high transportation costs, it is very expensive to change from heavy fuel

to a cleaner fossil fuel like propane gas;


solar energy can complement the natural gas, therefore lowering the final energy

costs and is ideal to meet the strategic ISO 14000 certification target;
a large area, with good orientation, is available on the roof of the factory for placing

solar collectors;
the solar system can be designed in a very common way, storing the collected

energy in an already existing concrete storage tank; and


the dyeing process and/or the washing process will be supplied by water from the

solar tank after increasing its temperature by the back-up conventional energy
system as necessary.

Back-up

Source: POSHIP (2001)

FIGURE 7.25 Scheme of the proposed solar system for the Beiral plant

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216 260

2/7/07

2:58 PM

Page 253

Solar Heat for Industrial Processes

253

The scheme of the proposed solar system is given in Figure 7.25. The most relevant data
for the proposed plant are summarized in Table 7.10.
TABLE 7.10 Characteristic data of the proposed system for a textile drying process
Company
Location
Ind. sector
Process
Working temperature
Solar collector area
Collector type
Storage volume
Annual solar gains (useful heat)
Solar fraction
Total investment
Total own contribution
Yearly net savings
Payback

Beiral
Seia
Textile
Preheating of water for dyeing process
2080C
1750m2
Stationary collectors, selective flat plate or CPC type
75m3
1331MWh/year
33%
340,430
221,280
84,496
3 years

Source: POSHIP (2001)

7.5.4 BODEGAS MAS MARTINET, TARRAGONA (SPAIN)


The Bodegas Mas Martinet wine cellar is located in the Priorat region, in the south of
Catalonia. The company produces a high quality wine for national consumption and for
export. The cellar has no connection to the electricity grid, so it tries to cover most of the
energy demand by solar energy (thermal and photovoltaic). Most of the heat and cold
consumption in the wine cellar arises for the cooling of the cellar during the summer
months and for the heating of an office building and a residential building during the
winter season. A small fraction of the heat demand is used for domestic hot water
production. The yearly distribution of the heating and cooling demands are shown in
Figure 7.26. Electricity generation using the installed diesel generator is expensive, so
Cooling

Heating

DHW

10000

8000

6000

[kWh]
4000

2000

be
r

be
r
De
ce
m

No
ve
m

O
ct
ob
er

be
r
Se
pt
em

Au
gu
s

Ju
ly

Ju
ne

ay
M

Ap
ril

ar
ch
M

Fe
br
ua
ry

ry

Ja
nu
a

ch-07

Source: POSHIP (2001)

FIGURE 7.26 Heat, cold and domestic hot water demand of the installations of Bodegas Mas Martinet

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216260

2/7/07

2:58 PM

254

Page 254

H. SCHWEIGER et al

Solar contribution

Auxiliary Heat Demand

10000

8000

6000

4000

2000

be
r

be
r
De
ce
m

No
ve
m

O
ct
ob
er

be
r
Se
pt
em

Au
gu
s

Ju
ly

Ju
ne

ay
M

Ap
ril

ar
ch
M

Fe
br
ua
ry

ry

Ja
nu
a

ch-07

Source: POSHIP (2001)

FIGURE 7.27 Total heat demand of Bodegas Mas Martinet and solar contribution throughout the year

favourable conditions are encountered for an alternative thermal cooling system that
substitutes the electricity consumption of the compression chillers.
A solar system is proposed for the air conditioning of the wine cellars and the office
building using a lithium bromide absorption cooling machine (Yazaki Company). The
excess heat from the solar collectors during the winter months is used for space heating
and domestic hot water production. In the proposed system, fan-coils are used for the
distribution of heat and cold to the wine cellar and the connected office building, and a
floor heating system for a nearby residential building.
The solar system consists of a solar collector field of 60m2 of CPC collectors
producing hot water at 8090C and 1500L of hot water storage. The absorption chiller
has a nominal cooling power of 35kW and is buffered by a 3000L cold storage. The total
heat demand of the factory (including the heat demand of the absorption chiller) and the
solar contribution is given in Figure 7.27. Figure 7.28 shows the hydraulic scheme of the
plant and Table 7.11 gives the characteristics of the solar system.

7.5.5 PRE-FEASIBILITY STUDY ON SOLAR PROCESS STEAM GENERATION FOR


THE PRODUCTION OF POROUS CONCRETE
A pre-feasibility study on solar process steam generation for the production of porous
concrete was carried out in cooperation with the YTONG Company (Hennecke et al, 2002).
Porous concrete is a building material with low heat conductivity, supporting energy
conservation in the building sector. The manufacture of 1m3 of this material consumes
some 155kg steam at 10 bars, equivalent to about 87kWhth. The motivation for this
investigation is to partly substitute this energy by steam produced in solar parabolic
trough collectors, reducing the fossil energy consumption and improving further the
overall energy balance of this material.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216 260

ch-07

2/7/07

2:58 PM

Page 255

Solar Heat for Industrial Processes

Zona casa

floor
heating
system

cooling
from pool

Parte ya instalada
ACS casa

Bacs-casa
Suelo radiante
Ttemp-SR

Vcasa Vtemp-SR

255

Retorno a balsa

Agua de
la balsa

T
BSR

T
VSR
Caldera
calefaccin

Bcasa

Tbalsa

Filtro

TSR-ret

Bbalsa
Campo de colectores

Acumulador de calor

Thcol-o

Vdep1500

Ftorre

Mquina de absorcin

Caldera auxiliar

Ths-top

Thx-hi

Tcaldera

Thot-ma
Vhot-ma

Bhot-ma

Vbalsa

1 500 l

B1ario
Thx-ho

B2ario

Ths-bot

Tret-2ario

Benfriamiento

Agua de
renovacin

Tevap-o
ACS oficina

Vinter-dep

Tacs-of-top

solar
thermal
system
Sala de calderas

Bacs-of

Vcal-ida
T cal-ida

Vcr-of-ret

Tcs-top

TFC-ret

Vfro-aux2

absorption
cooling
system

VD-evap2

Vcr-of

3 000 l

Bof

VD-evap1
Tcs-bot
Acumulador de
fro / calor

Vfro-ida

Vfro-aux1

Balm

Tfro-ida
Bbod
F.C.- B2

F.C.- B1

F.C.- B3

F.C.- B4

F.C.- Alm.

F.C-Of.1
F.C-Of.2
F.C-Of.3

Source: POSHIP (2001)

FIGURE 7.28 Scheme of the proposed solar system for Bodegas Mas Martinet

TABLE 7.11 Characteristic data of the proposed system for space heating and cooling of a wine cellar
Company
Location
Ind. sector
Process
Working temperature
Solar collector area
Solar collector type
Solar storage
Annual solar gains (useful heat)
Solar fraction
Total investment
Total own contribution
Yearly net savings

Bodegas Mas Martinet


Tarragona, Spain
Wine cellar
Space cooling and heating
5085C
60m2
CPC collectors
1500L (hot) + 3000L (cold/hot)
21MWh
58% (cooling)
35% (heating)
55,000
35,000
3000

Payback

11 years

Source: POSHIP (2001)

Reference data for this study were obtained from a typical factory in Hamm, Germany.
The main steam-consuming process is the curing of the material under precisely
controlled conditions in large autoclaves (Figure 7.29). Several autoclaves are operated in
a staggered mode, and Ruth-type (saturated steam) storages are commonly used to

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216260

ch-07

2/7/07

256

2:58 PM

Page 256

H. SCHWEIGER et al

pressure [bar]
steam flow [t/h]
15
10
5
0
-5

time [hours]
10

10
15
Source: Hennecke et al (2002)

FIGURE 7.29 Autoclave process (simplified)

manage the steam flows efficiently during charging and discharging. It should be pointed
out that the flexibility of the system, designed to cope with the severe load changes
caused by the batch processes, also provides distinct advantages for the integration of
solar process steam. A solar field was designed to fit the roof of the main factory hall,
consisting of small collectors like the parabolic trough collector from Industrial Solar
Technology (IST), with a total aperture area of about 1400m2 (Figure 7.30). The nominal
power output (600kWth) represents about 5 per cent of the total fossil-fired boiler capacity
installed. The area of the surrounding storage yards is large enough to provide up to 100

Source: Hennecke et al (2002)

FIGURE 7.30 Solar field integration via Ruth-type (saturated steam) storage

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216 260

ch-07

2/7/07

2:58 PM

Page 257

Solar Heat for Industrial Processes

257

per cent of the factorys total steam demand with collector systems mounted on elevated
structures.
Care must be taken in the long, horizontal and asymmetrically heated boiler tubes to
avoid potentially dangerous flow patterns of the two-phase flow. Theoretical
investigations indicate that the problem of flow stratification will be reduced at lower
pressures. However, below about 5 bars the low steam density leads to high velocities
and pressure drops, which might prevent the achievement of high steam qualities at the
outlet of the solar steam generator. The recirculation process provides a convenient
solution, particularly if the existing ruths steam storages can be used as the steam drum
to separate steam and water. During times when no steam is extracted for the production,
the solar energy input will lead to rising pressure in the system, increasing the exergetic
value of the stored energy and consequently reducing live steam consumption
requirements by the same amount. At the same time, any direct interference between the
solar steam generator and the delicate production process can be avoided.
The annual system performance has been simulated for different sites as a basis for
cost estimations, indicating that in the Mediterranean area the solar steam may be
competitive with conventional boilers (Table 7.12). At favourable sites, 10 per cent of the
annual steam demand can be satisfied from the solar field sized to provide 5 per cent of
the fossil boiler capacity.
TABLE 7.12 Annual system performance for different sites
LOCATION
Wrzburg, Germany
Portoroz, Slovenia
Mugla, Spain

DNI
kWh/m2a
894
1330
1900

ANNUAL HEAT PRODUCTION


kWh/m2a
294
472
761

LEVELIZED HEAT COST


/kWh
0.11
0.07
0.045

Note: DNI: direct normal irradiance


Source: Hennecke et al (2002)

7.6 CONCLUSIONS
Industrial solar thermal systems as a renewable energy source can cover a significant fraction
of the industrial heat and electricity demand. This industrial heat demand constitutes about a
third of the total final energy demand in southern European countries. About 7 per cent of the
total final energy is consumed in the form of industrial process heat at temperatures below
250C. Fulfilling the goal of installing 2,000,000m2 of solar thermal collectors for industrial
process heat and solar cooling, as outlined in the Campaign for Take-Off of the European
Commission would represent a primary energy saving of about 2,000,000MWh/year.
Therefore, solar thermal energy in industry can be an important contribution to a reliable,
clean, safe and cost-effective energy supply based on renewable energy sources.
Solar collectors for hot water production at low temperatures are a well-known and
well-extended standard technology. With recently developed high performance solar
collectors, heat at temperatures up to 250oC can be produced with excellent efficiency.
Heat at these temperatures is required in many industrial processes, such as steam
generation, washing, drying, distillation and pasteurization.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216260

ch-07

2/7/07

2:58 PM

258

Page 258

H. SCHWEIGER et al

At present, industrial solar thermal systems are operating in Europe with a total
installed collector surface of about 10,000m2. Solar collector arrays can be integrated
either into industrial roofs or installed on an available ground area.
Large-scale industrial installations may lead to very low cost systems, meaning solar
systems for industrial process heat production can become economically competitive
with fossil fuels on a very short term. Present investment costs for solar thermal systems
range from 250 to 500/m2 (corresponding to 2501000/kW of thermal power), leading
to average energy costs in southern Europe from 2 to 5 cents/kWh for low temperature
applications and from 5 to 15 cents/kWh for medium temperature systems.
The present cost per kWh of saved primary energy in industrial applications is less
than that of small- to medium - scale solar domestic hot water applications and can be
further reduced. Cost reductions can be obtained by large-scale production, reducing
operation and maintenance costs and improving collector efficiency and system design,
especially for medium temperature solar collectors. Cost reductions of up to 50 per cent
on a medium term are expected by 2010.

AUTHOR CONTACT DETAILS


H. Schweiger (corresponding author),Creu dels Molers, 15, 2-1, 08004 Barcelona, Spain
Tel: (+34) 93 441 53 93; Fax: (+34) 93 441 53 95; e-mail: hans.schweiger@energyxperts.net
Dr. Schweiger is director of the engineering company energyXperts.BCN, working in the fields of energy
efficiency and renewable energies (www.energyxperts.net).
J. Farinha Mendes, INETI (Instituto Nacional de Engenhara e Tecnologa Industrial), Lisbon, Portugal,
www.ineti.pt
Ma. J. Carvalho, INETI, Lisbon, Portugal, www.ineti.pt
K. Hennecke, DLR (Deutsches Zentrum fr Luft- und Raumfahrt e.V.), www.dlr.de
D. Krger, DLR, www.dlr.de

REFERENCES
Carvalho, M. J., Collares Pereira, M., Oliveira, J. C., Mendes, J. F., Haberle, A. and Wittwer, V. (1995) Optical and thermal
testing of a new 1.12X CPC solar collector, Solar Energy Materials and Solar Cells, no 37, pp175190
Collares Pereira, M., Gordon, J. M., Rabl, A. and Zarmi, Y. (1984) Design and optimisation of solar industrial hot water
systems with storage, Solar Energy, no 32, p121
Garg, H. P. (1987) Advances in Solar Energy Technology, vol 2, Reidel Publishing Company, Dordrecht, The Netherlands
Gee, R., Cohen, G., Winston, R., Greenwood, K. and McGuffey, B. (2003) Design, installation and early operation of a roofintegrated solar cooling and heating system, proceedings of the ASES Annual Conference, Dordrecht, The Netherlands,
June 2126
Gordon, J. M. and Rabl, A. (1982) Design, analysis and optimisation of solar industrial process heat plants without storage,
Solar Energy, vol 28, no 6, pp519530
Gordon, J. M. and Zarmi, Y. (1985) Single-pass open-loop solar thermal energy systems: On maximum energy delivery by
variable flow rate, Journal of Solar Energy Engineering, no 107, p273
Hennecke, K., Meinecke, W. and Krger, D. (2000) Integration of solar energy into industrial process heat and cogeneration
systems, proceedings of 9th International Symposium on Solar Thermal Concentrating Technologies, Font-Romeu, France, June

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216 260

ch-07

2/7/07

2:58 PM

Page 259

Solar Heat for Industrial Processes

259

Hennecke, K., Ktter, J., Michel, O. and Peric, D. (2002) Solar process steam generation for the production of porous
concrete, in A. Steinfeld (ed) 11th SolarPACES International Symposium on Concentrated Solar Power and Chemical
Energy Technologies, Zurich, Switzerland, 24 September, Paul Scherrer Institut, Villigen, Switzerland
Huerdes, J. V. and Lachal, B. (1986) Calor Solar Industrial, Spanish edition, Delta Energy Ltd, Rafz, Switzerland
IDAE Instituto para la Diversificacin y el Ahorro de la Energa (2001) Private communication: official data from the Spanish
Ministry for Industry and Energy (MINER) for 1999
Kreider, J. F. (1979) Medium and High Temperature Solar Processes, Academic Press, New York
Krger, D., Hoffschmidt, B., Hennecke, K., Pitz-Paal, R., Rietbrock, P. and Fend, T. (2000) Results from parabolic trough collectors for
process heat at the DLR Cologne, in H. Kreetz, K. Lovegrove and W. Meike (eds) Proceedings of the 10th International
Symposium on Solar Thermal Concentrating Technologies, Sydney, 810 March, Australia National University, Canberra, Australia
Krger, D., Pitz-Paal, R., Lokurlu, A. and Richarts, F. (2002) Solar cooling and heating with parabolic trough collectors in the
Mediterranean, in A. Steinfeld (ed) Proceedings of the 11th International Symposium on Concentrated Solar Power and
Chemical Energy Technologies, Zurich, Switzerland, 24 September, Paul Scherrer Institut, Villigen, Switzerland
Laue, H. J. and Reichert, J. (1994) Potential for medium and large sized industrial heat pumps in Europe, Directorate General
for Energy (DGXII), European Commission, contract no XVII/7001/90-8, final report
Lewis, C. W. (1980) The prospects for solar energy use in industry within the United Kingdom, Solar Energy, no 24, pp4753
Nitsch, J. and Luther, J. (1990) Energieversorgung der Zukunft, Springer-Verlag, Berlin
POSHIP (2001) The potential of solar heat for industrial processes, Project NNE5-1999-0308, final report, funded by the
European Commission
Rabl, A. (1985) Active Solar Collectors and Their Applications, Oxford University Press, Oxford, UK
Rabl, A. (1976) Comparison of solar concentrators, Solar Energy, no 18, p93
Schreitmller, K. (1986) Erfahrungen mit solaren Prozewrmeanlagen Solar Thermische Prozewrme, Klaus Scharmer,
Gesellschaft fr Entwicklungstechnologie mbH, Aldenhoven
Schreitmller, K. (1987) Stand und Perspektiven der solaren Prozewrmebereitung unter besondrere Bercksichtigung der
Aspekte von Entwicklungslndern, DFVLR, Institut fr technische Thermodynamik
TNO Bouw (1995) Potentieelstudie Hoogrendement-zonnecollectoren en Nederland, Report 94-BBI-R 1368, TNO, Delft, The
Netherlands
TRNSYS (2000) Software for Dynamical System Simulation, Solar Energy Laboratory, Madison, Wisconsin, US, Version 14.2
(1997) and 15.0
Wackelgard, E., Niklasson, G. A. and Granqvist, C. G. (2001) Selectively solar-absorbing coatings, J. Gordon (ed) Solar
Energy, the State of the Art, James & James, London
Welford, W. T. and Winston, R. (1978) The Optics of Non-Imaging Concentrators Light and Solar Energy, Academic Press
Winter, C.-J. (1997) Sonnenenergie nutzen. Technik, Wirtschaft, Umwelt, Klima, VDE-Verlag GmbH, Berlin

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 216260

ch-07

2/7/07

2:58 PM

Page 260

ch-08

2/7/07

2:38 PM

Page 261

Solar Energy Technology in the


Middle East and North Africa
(MENA) for Sustainable Energy,
Water and Environment
W. E. Alnaser, F. Trieb and G. Knies

Abstract
Our study reveals that concentrating solar power (CSP) can be used to fulfil the total
electricity demand in Middle East (ME) and North African (NA) countries of 1700TWh/y
in 2025, rising to 3600TWh/y in 2050, especially as an expanding technology with a
growth of 25 to 35 per cent a year. A strong competition of several CSP technologies
could lead to higher overall CSP growth rates than those assumed in the present study
(around 30 per cent per year). In 2025 14 per cent of the electricity demand of MENA
may be covered by CSP, which will become the dominating power source with a share
of 57 per cent by 2050.
According to our scenarios the mix of renewable energies will cost around 5
cents/kWh in 2050. The cost of CSP will be reduced from the current 8 cents/kWh to
6 cents/kWh in 2010 to 5 cents/kWh in 2020 (other scenarios quote 15 cents/kWh). From
2020 until 2050, the MENA region will save US$250 billion compared to the business as
usual policy scenario, where it was assumed that the fuel prices start at US$25/bbl for
oil and US$49/ton for coal and escalate by only 1 per cent per year (currently the prices
are at a level of US$55/bbl and US$65/ton respectively and the escalation rates have
amounted to 40 per cent per year since 2003).
The study shows that solar thermal power plants are suitable for seawater
desalination. A concentrating solar thermal collector array required for desalinating
1 billion m3/y would cover a total land area of approximately 10 x 10km, corresponding
to about 10m3 of desalinated water per m2 of collector area. About 10 per cent of the
desalinated water would suffice to irrigate the desert land beneath the collectors with
a water column of 1m/y.
The study shows that the total carbon emissions of electricity generation of all
MENA countries can be reduced from 770 million tons per year to 475 million tons in
2050, instead of increasing to 2000 million tons.
Furthermore, the greenhouse emissions from constructing the plants for most
renewable energy technologies are between 10 and 25g/kWh of useful energy. In
the case of photovoltaic systems, reductions are possible in the medium term to

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

ch-08

2/7/07

2:38 PM

262

Page 262

W. E. ALNASER, F. TRIEB AND G. KNIES

about 50g/kWh. The scenario reaches a per capita emission of 0.58 tons/cap/y in the
power sector in 2050, which is acceptable in terms of the recommended total emission
of 11.5 tons/cap.

Keywords installed power capacity; global solar radiation; water deficiency; MENA; wave and tidal
power

8.1 INTRODUCTION
The MED-CSP study Concentrating solar power for the Mediterranean region (DLR,
2005), which aims to catalyse the cooperation between Europe and Africa in solar energy
application, focused on the electricity and water supply of the regions and countries
illustrated in Figure 8.1, including southern Europe (Portugal, Spain, Italy, Greece, Cyprus
and Malta), North Africa (Morocco, Algeria, Tunisia, Libya and Egypt), western Asia (Turkey,
Iran, Iraq, Jordan, Palestine & Israel, Lebanon and Syria) and the Arabian Peninsula (Saudi
Arabia, Yemen, Oman, United Arab Emirates, Kuwait, Qatar and Bahrain).
The Arab countries in North Africa (NA) and in West Asia (WA), in general, lack energy
and water. The average annual rainfall in Arab countries is from 5 to 45mm/y, while in
European countries (mainly those in the Mediterranean EU) it ranges from 200 to
400mm/y. The total internal water reserve is only 100km3 for Arab countries, while in
Europe it may exceed 400km3 (bear in mind that the world internal water reserve is
43,764km3). Nearly 50 per cent of the Arab countries have water availability per capita less
than the absolute water scarcity level (200m3/capita/year) while the rest, except Iraq, are
in water scarcity threshold level (1000m3/capita/year) (ESCWA, 2003). In Europe, the per

Southern Europe
Western Asia
Arabian Peninsula
North Africa

FIGURE 8.1 Countries of the EU-MENA region analysed in this study

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

ch-08

2/7/07

2:38 PM

Page 263

Solar Energy Technology in the Middle East and North Africa

263

capita water resource ranges as high as 85,478m3/year (Norway). In fact among the
countries with least water resource (UNDP, 2002) we have 14 Arab countries (Kuwait, UAE,
Qatar, Libya, Saudi Arabia, Jordan, Bahrain, Yemen, Oman, Algeria, Tunisia, Egypt,
Morocco and Palestine) where the per capita water supply ranges from 10m3/year (Kuwait)
to 971m3/year (Egypt).
In 1999, electricity consumption per capita in Western Europe was never less than
4000kWh, while it was as low as 46kWh in Sudan. In the majority of the Arab countries,
except Arabian Gulf countries (GCC), the average person consumes electricity at the rate
of 1200kWh annually. Probably more than 1 million Arab citizens (of some 300 million)
have no access to electricity (UNDP, 2002).
Furthermore, the European Organisation for Economic Co-operation and Development
(OECD) countries emitted 3800 million tons of CO2, which is 15.2 per cent of the global
emission (25,000 million tons), in 2001. The CO2 emissions from the Middle East are only
4.8 per cent compared to North America (27.7 per cent), Eastern Europe (12.6 per cent),
Western Europe (15.6 per cent), Africa (8.8 per cent), Central and South America (4.1 per
cent), and the Far East and Oceania (31.5 per cent) (World Almanac, 2004). According to
the latest reports, Germany had managed to reduce emission of CO2 by only 20 per cent
(compared to the base year 1990), while others had actually increased their emissions or
stayed the same for example the UK 12.5 per cent, Italy 6.5 per cent, France 0 per cent,
Russia 6 per cent, the US 7 per cent and Japan 6 per cent (Global Emission Newsletter,
2004). These industrialized nations are required to reduce their CO2 emissions according
to the Kyoto Protocol, which took effect on 16 February 2006 by 5.2 per cent of their
emissions in 1990. The increased temperature of the globe due to global warming due
to extensive use of fossil fuels has led to more extreme weather and hence
catastrophes. According to Swiss Re-insurance Company (World Almanac, 2004) in its
report on natural catastrophes and man-made disasters in 2003, nearly 36 natural
accidents had occurred in Europe with 424 victims wasting US$2173 million (11.8 per
cent of the total wasting cost budget) while in the Middle East and Asia 178 accidents
had occurred with 51,894 victims wasting US$1447 million (7.8 per cent). Therefore,
mutual cooperation between Arab and European countries in the field of energy
production using solar radiation is strongly advisable.
The Arab countries are characterized and blessed with abundant direct solar radiation,
ranging from 4.1kWh/m2/day in Mosul, Iraq to 6.7kWh/m2/day in Nouakchott, Mauritania
(Alnaser et al, 2004; ESCWA, 2001). Of course, the total solar radiation will be larger than
these figures by nearly 30 per cent. Moreover, the maximum recorded annual mean
sunshine duration ranges from 7.5 hrs in Tunis to 10.7 hrs in Egypt. These figures are
much larger, at least three times, compared to European countries (Alnaser et al, 2004).
The electricity demand and temporal behaviour among MENA and European countries
is found to complement each other. Therefore cooperation will enhance renewable energy
utilization worldwide and it will increase to more than the current level of 13.8 per cent of
the total primary energy supply, where 2.3 per cent is from hydro, 11.0 per cent from
combustible renewable and waste and only 0.5 per cent covers solar (0.039 per cent), wind
(0.026 per cent), tidal (0.004 per cent) and geothermal (0.440 per cent) power. The solar
power market is already growing it was 1000MW in 2000 and is expected to be

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

ch-08

2/7/07

264

2:38 PM

Page 264

W. E. ALNASER, F. TRIEB AND G. KNIES

14,000MW by 2010 and 70,000MW in 2020 (UNDP, 2002). This, of course, will minimize the
cost per watt for each renewable energy source, especially solar thermal and photovoltaic.
The present study is the result of an analysis of the Middle East and North Africa
(MENA) countries with respect to their electricity demand and the possibility to satisfy this
demand by solar thermal power plants in the coming decades. Most of the findings of this
paper are derived from the final report Concentrating solar power for the Mediterranean
Region (DLR, 2005).
This paper describes the importance of clean and sustainable energy exports among
MENA countries and Europe in the form of high voltage potential produced in Arab
countries, using solar thermal (the most suitable technology for the region) or other
renewable energy resources; this is, of course, after using part of this energy for water
desalination in Arab countries, where 65 per cent of the water resources are politically in
debate with non-Arab countries, which may ignite conflicts. It will also show how the
environment will be protected against CO2 emissions if solar thermal technology is utilized.

8.2 DETERMINATION OF THE ELECTRICITY DEMAND IN MENA


8.2.1 DRIVING FORCES
The scenarios for electricity demand without electricity used for desalination were
calculated using GDP per capita and population as driving forces. The GDP per capita and
population scenarios are the same as in the scenarios for the demand for desalination.

8.2.1.1 Population
All scenarios share the same population prospect the World Population Prospect by the
UN, middle path, and revision 2002, including the age structure of the population. The last
is not taken into account explicitly but will be used to discuss the feasibility of economic
growth and what might happen if the work force and not the population would be
considered as a driving force.
In addition, it should be noted that according to the scenario in the next 50 years most
of the states under investigation are in a demographic transformation due to decreasing
birth rates, which will raise the share of the workforce in the population to a high level. At
the end of the period the share of old people will increase significantly, however. Even with
given wage levels, such a demographic transformation offers the opportunity to reap a socalled demographic dividend within the next 30 years when the share of working people
will increase. Therefore rising income per capita and rising savings can be expected.

8.2.1.1 GDP per capita


When talking about economic development, you must decide which indicator should be
used. In the following Gross National Income (GNI) in Purchasing Power Parity (PPP),
US$-2001 per capita will be employed. While GDP is a better measure for economic
activities within a country the GNI is a better measure of the populations income. The PPP
used is calculated on the basis of a US-basket of goods and services. A regression suggests
that a 1.45 per cent increase in GNI in Atlas-$ is necessary to reach a 1 per cent increase in
GNI (PPP). Different functions suggest that at the lower end of the income scale a lower
increase in GNI in Atlas-$ is necessary while at the upper end a higher increase is required.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

ch-08

2/7/07

2:38 PM

Page 265

Solar Energy Technology in the Middle East and North Africa

265

8.2.2 THE LINK BETWEEN GDP PER CAPITA AND ELECTRICITY CONSUMPTION
PER CAPITA
The essential part for deriving the electricity demand is a link between GDP per capita and
electricity consumption per capita. With the population growth, the overall electricity
demand can obviously be calculated.
The link between the per capita values used the following steps:
1. For every year in the period 19602001, regressions were calculated between GDP
per capita and total final consumption (TFC) for electricity per capita (IEA, 2002;
Heston et al, 2002).
2 For the two parameters of the regression equations, time trends were estimated
(using power and alternatively linear functions). A power function gave a
significantly better fit for the absolute term. For the power it was difficult to
distinguish a linear trend from a power trend. So both were used, resulting in two
alternative links between GDP per capita and electricity consumption per capita.
The linear trend gives a scenario with high efficiency increase while the power
trend results generally in increasing electricity intensities. It should be noted that
with these parameter trends, the link between GDP per capita and electricity per
capita no longer depends solely on growth rates. Also the absolute value shows an
influence.
3 From the TFC for electricity, the domestic demand was derived using International
Energy Agency (IEA) data for 2001 (IEA, 2002) on distribution losses, consumption in
the energy sector and the so-called own use. These consumptions were split into a
proportional and a fixed term. The fixed term is meant to accommodate the use for
oil production. The proportional term was linearly reduced to a level which is now
common in industrial countries (8 per cent). It should be noted that the data on these
terms are not of a high quality and are sometimes missing. Fortunately, the impact of
these terms is generally small.
4 The resulting general functions were calibrated to individual countries
assuming a linear mix of the current values and the estimated value. The
weight of the estimated value is assumed to increase linearly from a current
0 to 1 in 2050.
5 The two scenarios are obtained by combining high economic growth with high
efficiency growth and low economic growth with low efficiency growth, as the
increase of efficiency is coupled to investment and the higher growth rates result in
higher investment rates and a higher share of new machineries.

8.3 RENEWABLE ENERGY RESOURCES IN EU-MENA


The renewable energy resources in the Euro-Mediterranean region were assessed on the
basis of spatial information available from different sources described later in this chapter.
The direct normal irradiance (DNI) used by concentrating solar power systems was
assessed by the German Aerospace Centers (DLRs) high resolution satellite remote
sensing system (SOLEMI, 2004), while the data for the other renewable energies were
taken from materials provided by the renewable energy scientific community.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

ch-08

2/7/07

266

2:38 PM

Page 266

W. E. ALNASER, F. TRIEB AND G. KNIES

Both the technical and economic potentials were defined for each renewable energy
resource and for each country. For each resource and for each country, a performance
indicator was defined that represents the average renewable energy yield with which the
national potential could be exploited (Table 8.1). The economic potentials are those with a
sufficiently high performance indicator that will allow new plants in the medium and long
term to become competitive with other renewable and conventional power sources,
considering their potential technical development and economies of scale (Table 8.2).
The renewable energy potentials for power generation differ widely in the countries
analysed within this study. Altogether they can cope with the growing demand of the
developing economies in MENA. The economic wind, biomass, geothermal and hydropower
resources combined amount to about 400TWh/y. Those resources are more or less locally
concentrated and not available everywhere, but can be distributed through the electricity grid,
which will be reinforced in the future in line with the growing electricity demand of this region.
The greatest resource in MENA is solar irradiance, with a potential that is by several orders
of magnitude larger than the total world electricity demand. The solar energy irradiated on the
ground equals one to two barrels of fuel oil per square metre per year of primary energy. This
magnificent resource can be used both in distributed photovoltaic systems and in large
central solar thermal power stations. Thus, both distributed rural and centralized urban
demand can be covered by renewable energy technologies.
The accuracy of a global resource assessment of this kind cannot be better than 30
per cent for individual sites as it depends on many assumptions and simplifications.
However, it gives a first estimate of the order of magnitude of the renewable energy
treasures available in Europe and MENA.

8.3.1 RESOURCES FOR CONCENTRATING SOLAR POWER


Weather satellites such as the Meteosat-7 of the European Organization for the
Exploitation of Meteorological Satellites (EUMETSAT) are geo-stationary satellites at a
distance of 36,000km at a fixed point over the globe that send half-hourly images for
weather forecasting and other purposes. From those images, the optical thickness of
clouds can be derived obtaining half-hourly cloud values for every site. Of all atmospheric
components, clouds have the strongest impact on the direct irradiation intensity on the
ground. Therefore the very high spatial (5 x 5km) and temporal (0.5 hour) resolution
provided by Meteosat is required for this atmospheric component.
Aerosols, water vapour, ozone and so forth have less impact on solar irradiation. Their
atmospheric content can be derived from several orbiting satellite missions like NOAA and
from re-analysis projects like GACP or NCEP/NCAR and transformed into corresponding
maps/layers of their optical thickness.
The spatial and temporal resolution of these data sets can be lower than that of
clouds. The elevation above sea level also plays an important role as it defines the
thickness of the atmosphere. It is considered by a digital elevation model with 1 x 1km
spatial resolution. All layers are combined to yield the overall optical transparency of the
atmosphere for every hour of the year. Knowing the extraterrestrial solar radiation intensity
and the varying angle of incidence, the direct normal irradiation can be calculated for every
site and for every hour of the year. Electronic maps and GIS data of the annual sum of

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

ch-08

2/7/07

2:38 PM

Page 267

Solar Energy Technology in the Middle East and North Africa

267

direct normal irradiation can now be generated, as can an hourly time series for every
single site.
The mean bias error of the annual sum of direct normal irradiation, which is decisive
for economic assessment, is usually in the order of 5 per cent. More information
TABLE 8.1 Renewable electricity performance indicators representing the average renewable electricity
yield of a typical facility in each country

5000m

TABLE 8.2 Economic potentials of renewable energy sources in the southern EU and MENA region in
TWh/y

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

ch-08

2/7/07

268

2:38 PM

Page 268

W. E. ALNASER, F. TRIEB AND G. KNIES

FIGURE 8.2 Annual direct normal irradiance for 2002

can be found at the websites www.dlr.de/steps, www.solemi.com and http://


swera.unep.net.
The analysis was performed for the countries is shown in Figure 8.2 for 2002. A oneyear basis is not sufficient for the development of large CSP projects, as the annual
climatic fluctuations can be in the range of 15 per cent. For project development
purposes, at least 515 years of data should be processed. However, for the assessment
of national solar electricity potentials and their geographic distribution, this basis is
sufficient, especially because in most MENA countries the total solar energy potential is
some orders of magnitude higher than the demand.

8.3.2 OTHER RENEWABLE ENERGY RESOURCES


8.3.2.1 Hydropower
The national technical and economic hydropower potentials were taken from the literature
(WEC, 2004; Horlacher, 2003). The annual full load hours are used as the performance
indicator. They were calculated from the installed capacity and the annual electricity
generation of the plants installed at present in each country (ENERDATA, 2004). The map
of gross hydropower potentials (Figure 8.3) illustrates the geographic distribution of the
hydropower potentials.
The total economic hydropower potential of all countries analysed within the study is
432TWh/y. In 2000 approximately 70GW of hydropower were installed, producing
155TWh/y of electricity.
There is certain evidence that climate change is possibly having an increasing impact
on hydropower generation with the possibility of reductions of up to 25 per cent in the
long term in the southern Mediterranean countries (Bennouna, 2004; Lehner et al, 2005).
Although we have not quantified such impacts in the study we believe that this is a

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

ch-08

2/7/07

2:38 PM

Page 269

Solar Energy Technology in the Middle East and North Africa

269

GWh/y on 30 x 50 km pixel

Source: Lehner et al (2005)

FIGURE 8.3 Gross hydropower potentials in EU-MENA

serious concern that should be taken into account in energy planning. Efficiency of
hydropower use should be enhanced systematically in order to, at least, partially
counteract such effects.

8.3.2.2 Geothermal power


Considerable conventional geothermal resources are available in Italy (already used to a
great extent), Turkey and Yemen. Conventional geothermal resources were taken from the
literature (Gawell et al, 2004). For Europe, medium-term geothermal power potentials
from the literature were taken for cross-checking (EU, 2004).
A map of subsoil temperatures at 5000m depth was taken to assess the total areas
with temperatures higher than 180C as the economic potential for hot dry rock (HDR)
technology. It was assumed that a layer with 1km thickness at 5000m depth was used as
a heat reservoir (Bruchmann, 2004; Hurter et al, 2000). The technical HDR potential for
temperatures below 180C was not assessed.
The temperature at 5000m depth was used as the performance indicator (Figure 8.4).
With that information, the efficiency and the specific investment cost (Inv) of a HDR
plant was compared with that of a reference plant and calculated according to the
following equation, using a scaling exponent of 0.7:
Inv = InvReference(Reference/)0.7

(1)

The data of the reference plant were taken from Chopra (2003).

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

ch-08

2/7/07

270

2:38 PM

Page 270

W. E. ALNASER, F. TRIEB AND G. KNIES

Source: Wittig (2004)

FIGURE 8.4 Temperature at 5000m depth for hot dry rock geothermal power technology

The annual amount of electricity that can be generated from hot dry rock depends on
the heat in place and the time of extraction. That time was assumed to be 1000 years in
order to ensure that the geothermal potentials can be renewed within this time span. At
such a slow rate, the geothermal power potentials can be considered as renewable
energies that could be used continuously without limitations in time like other renewable
energy sources.
In 2000 approximately 600MW of conventional geothermal power capacity was
installed in the analysed countries producing 4.6TWh/y of electricity. The total economic
potential was estimated to be around 400TWh/y; this is, however, a quite rough and
conservative estimate.

8.3.2.3 Electricity from biomass


The electricity potential of municipal waste, solid biomass (wood) and agricultural
residues was calculated. From the literature, agricultural residues such as biogases, which
at present are mainly unused for power purposes, were taken as reference (WEC, 2004).
An electricity conversion factor of 0.5MWh/ton of biomass was assumed for the
calculation of the potential electricity yield from agricultural waste biomass. It was
assumed that 80 per cent of this potential will be used in 2050. The amount of potentially
available municipal waste was calculated in proportion to the growing urban population in
each country. The growth of population was taken from the UN medium growth model.
Due to growing urban population, the biomass potential from municipal waste grows

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

ch-08

2/7/07

2:38 PM

Page 271

Solar Energy Technology in the Middle East and North Africa

271

steadily with each year. We have assumed a constant municipal waste productivity of
0.35tons/cap/year and a waste-to-electricity conversion factor of 0.5MWh/ton. It was
estimated that 80 per cent of this potential will be used until 2050.
Solid biomass (mainly wood) potentials were assessed from a global map of biomass
productivity in tons/ha/year and from the existing forest areas of each country (Figure 8.5).
A possible change in productivity or forest areas in the future has been neglected. Results
were cross-checked for plausibility with historical data from European countries (WEC,
2004). There will also be competition with traditional fuel wood use in most MENA
countries which must be taken into consideration. Therefore the rate of use of the fuel
wood potential was assumed to be 40 per cent until 2050. Annual full load hours are used
as performance indicator.

Source: Bazilevich (1994)

FIGURE 8.5 Map of biomass productivity

The total installed capacity of biomass power plants in the analysed countries in 2000
amounted to 1.8GW; plants were generating a total of 6.4TWh/year of electricity. For the
total region, a biomass electricity potential of 400TWh/y was identified, of which about 50
per cent might be used until 2050. Potential from residues dominate in MENA, while
power from solid and other biomass sources is also very important in Europe.

8.3.2.4 Wind energy


Wind power resources are given in the literature for European countries, including Malta
and Cyprus, and for Morocco, Tunisia, Egypt and Turkey (EU, 2004; Gardner et al, 2002;
OME, 2002). There is additional information on wind power potentials for Morocco,
Jordan, Egypt and Turkey in GTZ (2002, 2004).

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

ch-08

2/7/07

272

2:38 PM

Page 272

W. E. ALNASER, F. TRIEB AND G. KNIES

Source: Gral (2003a)

FIGURE 8.6 Annual average wind speed at 80m above ground level in m/s

For the other countries, electricity potentials were estimated from the wind map in
Figure 8.6 taking into account wind speed and area restrictions and site exclusion similar to
that used for CSP but adapted to wind power. The original wind speed was taken from
ECMWF (2002) for 33 and 144m height and was interpolated to 80m height. This map gives
a very rough estimate of the distribution of wind speed as an average for an area of 50 x
50km. The original data has a geographic resolution of 1.12 degrees.
Wind electricity potentials were calculated as functions of the average wind speed
according to well-known equations. We have assumed a maximum installed capacity of
10MW per square kilometre of land area. Areas with annual full load hours over 1400h/y
equivalent to a capacity factor of 16 per cent were considered as long-term economic
potential. Results were cross-checked and eventually corrected for those countries that
have made a national resource assessment (WEC, 2004; OME, 2002; GTZ, 2004; Alnaser
and Al-Karaghouli, 2000; Gral et al, 2003a). Annual full load hours (capacity factor) define
the performance indicator. These have been derived from the literature, the World Wind
Atlas (WWA, 2004) for a selection of sites in each country, and from the wind speed map.
Potentials include both onshore and offshore.
In 2000 a total of 3.3GW of wind capacity was installed in the analysed region producing
7.2TWh/y of wind electricity (ENERDATA, 2004). The total economic wind power potential in
the region amounts to 440TWh/y, of which 285TWh/y could be exploited by 2050.

8.3.2.5 Photovoltaic energy


Photovoltaic (PV) applications are in principle unlimited. There are no criteria for site
exclusion for PV systems, as they can be installed almost anywhere. However, their

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

ch-08

2/7/07

2:38 PM

Page 273

Solar Energy Technology in the Middle East and North Africa

273

expansion is still limited by their high investment cost. Using present growth rates and
scenarios for very large PV systems and distributed applications, PV potentials were
assessed in a relatively intuitive way. For EU states, the literature gives mid-term
potentials for PV (EU, 2004). The global irradiance on a surface tilted according to the
latitude was used as a performance indicator (Figure 8.7) (Meteonorm, 2004). Although
we have not introduced any economic threshold, the learning curves of PV suggest that
this technology will become competitive by the middle of this century under the irradiance
conditions of the MENA region.

Source: ECMWF (2002), Gral (2003a)

FIGURE 8.7 Annual global irradiation on surfaces tilted south with latitude angle in kWh/m /year

PV systems are especially suited to decentralized small scale applications in remote


regions where they often are already competitive with conventional diesel motor
generator power supply schemes. This special market segment has been assessed by
several studies (OME, 2002; GTZ, 2002, 2004). In our study we have only included global
PV potentials without quantifying grid-connected and remote systems separately. In 2000
about 0.02TWh/y of solar electricity was produced by PV, mainly in Italy and Spain.
Including very large scale PV systems (VLS-PV) of up to 1.5GW of capacity in the desert
regions, by 2050, as suggested by IEA (2003), we estimate the PV potential in the region
to be about 218TWh/y with a total installed capacity of 125GW.

8.3.2.6 Wave and tidal power


Wave and tidal power potentials were taken from the literature (EU, 2004). Performance
indicators are the annual full load hours which have been set for all locations at 4000h/y.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

ch-08

2/7/07

274

2:38 PM

Page 274

W. E. ALNASER, F. TRIEB AND G. KNIES

More information on the equations used to estimate the potential of each type of
renewable energy resource can be obtained from Alnaser (1995).

8.4 CSP DEMAND-SIDE POTENTIAL IN MENA


According to the estimates of the reference scenario Closing the Gap for each MENA
country, assuming high economic growth rates and at the same time high efficiency gains
in the power sector, the total MENA electricity demand will increase from 800TWh/y today
to 1710TWh/y in 2025 and to 3605 TWh/y in 2050 (Figure 8.8). This reflects an increase by
a factor of 4.5.
In the coming decades, the necessity of re-investing in the power sector in MENA
countries by replacing older conventional power plants with solar energy plants is the
most suitable and economic decision to take. The concentrating solar power (CSP) plant
is the candidate. These new plants will coincide with the increasing growth rates of the
power demand by that time the CSP capacities will still be very limited by their possible
maximum speed of expansion.
Although CSP will expand rapidly with growth rates of 25 to 35 per cent per year, major
CSP shares of the domestic power supply will not be seen before 2025 (Figure 8.9). Very
high growth rates will be achieved in the first five years. There will probably be competition
between various collector designs, which will accelerate the production capacity

Export

Desalination

MENA

FIGURE 8.8 Domestic electricity demand in MENA and electricity supplied by existing power stations, by
new power plants and by CSP for domestic consumption, export and seawater desalination for the scenario
Closing the Gap

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

2/7/07

2:38 PM

Page 275

Solar Energy Technology in the Middle East and North Africa

275

100
90

CSP Growth Rate [%/y]

80
70
60
50
40
30
20
10

20
50

20
45

20
40

20
35

20
30

20
25

20
20

20
15

20
10

20
05

20
00

ch-08

Year
FIGURE 8.9 Growth rate of CSP production in MENA in the scenario Closing the Gap including domestic
power supply, export electricity and seawater desalination

expansion. A long-term growth rate of 30 per cent, as achieved in the scenario between
2015 and 2025, is quite realistic and has already been observed in other technologies.
Nevertheless, until 2025 considerable production of conventional and other renewable
electricity of 1440TWh/year will have to be built in the form of new conventional power
plants or other renewable power technologies such as wind, hydro or geothermal power.
Once installed, this new power plant capacity will be there for a time span of about 3040
years, until it will again be replaced.
The scenario assumes that once CSP has reached appropriate production volumes,
the expansion of new fossil-fuel-fired power plants will stagnate. After 2025 most new
plants will be hybrid, using CSP and fossil fuels as input energies (Table 8.3). Part of the
existing fossil-fuel-powered capacity could then be replaced subsequently with
concentrating solar power, if the land area required for the collector fields is considered
in the planning of those power plants. This additional CSP potential has been neglected
except for the islands of Cyprus and Malta.
In our scenario the CSP capacity of each country is proportional to each countrys
share of the total MENA power deficit. Countries showing larger deficits will also require
larger CSP capacities. The total supply of electricity by CSP in MENA will reach 235TWh/y
by 2025 and 2045TWh/y by 2050. It is assumed that the installed CSP capacities would be
distributed evenly among the MENA countries.
In the real world, there might be competition among the MENA countries to achieve
higher growth rates within each individual country. A strong competition among several
CSP technologies could lead to higher overall CSP growth rates than those assumed here
(around 30 per cent per year). In 2025 14 per cent of the domestic electricity demand of

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

ch-08

2/7/07

276

2:38 PM

Page 276

W. E. ALNASER, F. TRIEB AND G. KNIES

TABLE 8.3 Electricity demand in MENA and generation by conventional old and new
plants and by CSP technologies in MENA until 2050 (all numbers rounded to 5)
YEAR
Old Plants
New Plants
CSP Supply
MENA Electricity Demand
CSP Desalination
CSP Export
Total CSP in MENA

TWh/y
TWh/y
TWh/y
TWh/y
TWh/y
TWh/y
TWh/y

2000
785

785

2025
40
1440
235
1710
115
45
395

2050
0
1560
2045
3605
525
645
3215

MENA will be covered by CSP, which will become the dominating power source with a
share of 57 per cent by 2050.
The necessary solar capacities for seawater desalination are quantified in section 8.7.
A CSP production of 115TWh/y in 2025 and 525TWh/y in 2050 would be used for
desalination purposes. After 2030, the CSP desalination capacities would be large enough
to cope with the freshwater demand and desalination will grow much more slowly. While
in 2025 about 29 per cent of the total CSP production would be used for desalination, in
2050 only 16 per cent would be used for that purpose.
Export of solar electricity will be limited to minor volumes before the scheduled
finalization of the Mediterranean Ring Interconnection in 2015. Also after that, additional
transmission capacities will be required to export considerable quantities of electricity
from MENA to Europe, as the Mediterranean Ring transmission capacity will be limited to
about 500 to 1000MW. The scenario assumes that export through additional high voltage
direct current (HVDC) interconnection of North Africa and Europe will start in 2015 with
2TWh/y. An extended HVDC grid including other MENA regions will transmit 45TWh/y by
2025 and 645TWh/y by 2050, which would be equivalent to about 15 per cent of the
European electricity demand of that time. In the scenario, export capacities are distributed
to all MENA countries in proportion to their share of the total MENA domestic power
demand. The idea behind this is that if a country installs high capacities for domestic
supply, it will also have a good industrial base for export. However, there are different
priorities for electricity export and different starting times for interconnections, depending
on the distance to Europe, the solar energy resources and the priority of the domestic
demand for power and water. By 2025, 11 per cent of the total CSP production of MENA
will be exported, increasing to about 20 per cent by 2050. The islands are not considered
as potential CSP exporters in this scenario. On the other hand, the electricity
interconnection between the Gulf Cooperation Council countries (southwest Asia Saudi
Arabia (SA), Kuwait (KU), Qatar (QR), United Arab Emirates (UAE), Bahrain (BN) and Oman
(ON)) has already been agreed on, with a budget of US$1.2 billion; it started in early 2005
and will be complete in 2008.
The scenario is a rough estimate of the CSP potential in MENA. There will be three
types of plants distributed among all three CSP categories (domestic, export and
desalination) used in different combinations:

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

ch-08

2/7/07

2:38 PM

Page 277

Solar Energy Technology in the Middle East and North Africa

277

1 CSP plants for cogeneration with coupled seawater desalination must be at the
coast, as the cogenerated heat cannot be transferred over long distances. Their
electricity can be used for additional reverse osmosis (RO) desalination, domestic
use or export. As the coastal regions in MENA are strongly used by other human
activities, this plant type will be limited to regions with appropriate site conditions
and available land area.
2 CSP plants used exclusively for power generation can be anywhere on the grid. Their
electricity can be transmitted to any other place and used for domestic supply, export
or RO desalination. This type of plant will be placed where good irradiation coincides
with good infrastructure conditions.
3 CSP plants for industrial cogeneration will be limited to appropriate industrial sites.
While their heat will be used on site, their electricity might be used on site too or be
sold to the grid for domestic use, export or RO desalination. Cogeneration plants are
considered as part of the domestic CSP production potential.
In the real world, there will be a mix of these three plant types, which will vary according
to the regional demand of each country and the local supply-side conditions. The scenario
can only give a rough estimate of the overall potentials of the region, showing the
amounts of energy potentially used for domestic supply, export or desalination. However,
it cannot distinguish and quantify the different plant types that will be erected in each
country, which will be subject to national strategic power expansion planning.

8.4.1 CSP DEMAND-SIDE POTENTIAL IN NORTH AFRICA


According to the reference scenario Closing the Gap, the domestic North African
electricity demand will increase from 160TWh/y today to 490TWh/y in 2025 and to
1230TWh/y in 2050 (Figure 8.10). This reflects an increase by a factor of 7.7, which is
almost twice the MENA average. The export share of the Northern African countries is
assumed to become relatively large, due to their proximity to Europe. Export will start in
2015, achieve 20TWh/y, equivalent to 14 per cent of the total CSP production, by 2025 and
385TWh/y, equivalent to 25 per cent of the total CSP production, by 2050.
Due to the considerable growth of demand, domestic electricity production will be the
dominant potential for CSP in North Africa, with equally important short-term potentials
for desalination and considerable long-term potentials for export electricity. In 2025 15 per
cent of the domestic electricity demand will be covered by CSP, which will grow to
become the dominating electricity source with a share of 65 per cent in 2050 (Table 8.4).
On the basis of country statistics from FAO (2004), a considerable demand for solar
powered desalination can only be detected for Libya and Egypt. However, the desalination
potential of those two countries alone accounts for 38 per cent of the total CSP
production in North Africa in 2025 and 20 per cent in 2050.
CSP export of the western Asian countries is of relative low priority because the
distance to Europe is quite large for most countries. However, land and solar radiation
resources are available as in Northern Africa. Export may start in 2025 with approximately
10TWh/y, equivalent to 8 per cent of the total CSP production, and achieve 120TWh/y,
equivalent to 19 per cent of the total CSP production in 2050.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

ch-08

2/7/07

278

2:38 PM

Page 278

W. E. ALNASER, F. TRIEB AND G. KNIES

TABLE 8.4 Domestic electricity demand in North Africa and generation of domestic
power, export electricity and power for desalination by CSP technologies until 2050
YEAR
Old Plants
New Plants
CSP Supply
North Africa Demand
CSP Desalination
CSP Export
Total CSP in North Africa

TWh/y
TWh/y
TWh/y
TWh/y
TWh/y
TWh/y
TWh/y

2000
160

160

2025
10
415
70
490
55
20
145

2050
0
430
800
1230
285
385
1470

FIGURE 8.10 Domestic electricity demand in North Africa and electricity supplied by old power stations,
by new power plants and by CSP for domestic consumption, export and seawater desalination for the scenario
Closing the Gap

The interconnection of electricity between European and MENA countries will


resemble that interconnecting the seven Mediterranean states (Spain, Italy, Morocco,
Algeria, Libya and Egypt) through the UCTE (Union of Coordination for Transmission of
Electricity) which is supervising projects in more than 20 European countries providing
more than 40 million consumers with 2100TWh. There is also an interconnection of
electricity between African states which link Egypt to Sudan and to the Congo. This link
will be illustrated later.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

ch-08

2/7/07

2:38 PM

Page 279

Solar Energy Technology in the Middle East and North Africa

279

8.4.2 CSP DEMAND-SIDE POTENTIAL IN WESTERN ASIA


The domestic electricity demand in western Asia will increase from 375TWh/y today to
810TWh/y in 2025 and 1590TWh/y in 2050 (Figure 8.11 and Table 8.5). This reflects an
increase by a factor of 4.2, which is about the MENA average.
CSP export of the western Asian countries is of relative high priority. However, land and
solar radiation resources are more limited than in Northern Africa and the distance to Europe
is greater. Export may start in 2020, achieve 10TWh/y, which is equivalent to 8 per cent of total
CSP production, in 2025 and 145TWh/y, equivalent to 13 per cent of the total CSP production,
in 2050. Export volumes of CSP from western Asia are smaller than those of North Africa, as

FIGURE 8.11 Electricity demand in western Asia and electricity supplied by old power stations, by new
power plants and by CSP for domestic consumption, export and seawater desalination for the scenario Closing
the Gap

TABLE 8.5 Domestic electricity demand in western Asia and generation of domestic
power, export electricity and power for desalination by CSP technologies until 2050
YEAR
Old Plants
New Plants
CSP Supply
Western Asia Demand
CSP Desalination
CSP Export
Total CSP in Western Asia

TWh/y
TWh/y
TWh/y
TWh/y
TWh/y
TWh/y
TWh/y

2000
375

375

2025
20
680
110
810
10
10
130

2050
0
685
905
1590
60
145
1110

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

ch-08

2/7/07

280

2:38 PM

Page 280

W. E. ALNASER, F. TRIEB AND G. KNIES

the site conditions in western Asia are not as conducive as North Africa and the domestic
demand in western Asia is very large and will probably be binding the capacities.
On the basis of country statistics from the FAO (2004), a considerable demand for solar
powered desalination can only be detected for Jordan, Israel and Syria. The potential for
desalination will achieve 8 per cent of the total CSP production in 2025 and 5 per cent in 2050.
Due to the considerable growth of demand, domestic electricity production will be the
dominant potential for CSP in western Asia, with comparatively small potentials for
desalination and for export electricity. In 2025 14 per cent of the domestic electricity
demand will be covered by CSP, which will grow to become the dominate electricity
source with a share of 57 per cent in 2050.

8.4.3 CSP DEMAND-SIDE POTENTIAL OF THE ARABIAN PENINSULA


The domestic electricity demand of the Arabian Peninsula will increase from 235TWh/y
today to 410TWh/y in 2025 and 790TWh/y in 2050 (Figure 8.12 and Table 8.6). This reflects
an increase by a factor of 3.4, which is well below the MENA average. The low value is
due to an already rather high consumption of electricity per capita in most countries of the
peninsula except Yemen and Oman.
On the basis of country statistics from the FAO (2004), all countries of the Arabian
Peninsula have a large freshwater deficit and hence a large demand for solar powered
desalination. The potential for desalination will achieve 43 per cent of the total CSP
production in 2025 and will still make up 28 per cent in 2050.

FIGURE 8.12 Electricity demand of the Arabian Peninsula and electricity supplied by old power stations, by
new power plants and by CSP for domestic consumption, export and seawater desalination for the scenario
Closing the Gap

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

ch-08

2/7/07

2:38 PM

Page 281

Solar Energy Technology in the Middle East and North Africa

281

TABLE 8.6 Domestic electricity demand of the Arabian Peninsula and generation of
domestic power, export electricity and power for desalination by CSP technologies until
2050 (rounded to 5)
YEAR
Old Plants
New Plants
CSP Supply
Demand Arabian Peninsula
CSP Desalination
CSP Export
Total CSP Arabian Peninsula

TWh/y
TWh/y
TWh/y
TWh/y
TWh/y
TWh/y
TWh/y

2000
235

235

2025
10
345
55
410
50
10
115

2050
0
445
345
790
180
120
645

Due to the considerable growth of demand, especially in Saudi Arabia and Yemen,
domestic electricity production will be the dominant potential for CSP on the Arabian
Peninsula, followed by a large potential for desalination and comparatively small
potentials for export. In 2025 14 per cent of the domestic electricity demand will be
covered by CSP, which will grow to become a very important electricity source with a
share of 44 per cent in 2050.

8.4.4 SUMMARY
From this analysis we can conclude that there are several consequences related to the
scenario Closing the Gap:
Until 2025, the CSP market in MENA will be limited neither by the available solar

energy resources nor by the expected energy demand, but only by the intrinsic
achievable speed of CSP capacity expansion.
Under the assumptions of the scenario Closing the Gap, CSP would cover about 60
per cent of the expected electricity demand for power and water in MENA in 2050,
the rest would have to be covered by conventional power technologies using fossil
fuels. In spite of the considerable expansion of CSP and other renewable energies,
CO2 emissions in MENA would still double by 2050. However, without CSP CO2
emissions in 2050 would become roughly five times those of 2000.
Economic growth and a reasonable growth of the population do not necessarily lead
to an excessive consumption of natural resources if sustainable pathways are taken
immediately. This includes the intensification of energy and water efficiency and the
use of all available renewable energy technologies.
Due to the abandoning of CSP in the beginning of the 1990s in California, about 15
years of development, learning and expansion have been lost. These missing 15
years will lead to a considerable gap in the energy and water supply in MENA during
the 2020s, which must be bridged by non-sustainable resources like fossil fuels and
groundwater reserves, hoping that those resources will be able to carry on long
enough. An expansion of CSP in time could have avoided those gaps. Now, every
year of delay of the introduction and expansion of CSP will lead to more critical

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

ch-08

2/7/07

282

2:38 PM

Page 282

W. E. ALNASER, F. TRIEB AND G. KNIES

situations in the MENA countries. The cost of CSP varies widely. It depends on the
producer and consumer ability. Fully exporting this technology, relying fully on
expatriates and foreign manufacturing and services of CSP will make the cost twice
that found in our scenario
Export of solar electricity and of CSP technology can become an important economic
factor in those MENA countries that will be able and willing to use this new, clean
and everlasting natural resource to increase their national income, while their income
from fossil reserves will inevitably decrease.

8.5 THE CRITICAL ISSUE OF WATER AND ENERGY


IN THE GCC COUNTRIES
The majority of the population in the Arab world is concentrated in urban areas. In the
Arabian Gulf countries, which are the poorest in water resources, approximately 15.7 per
cent of its people inhabit rural areas with no access, or very limited access, to suitable
water and energy resources (figures range from 22.3 per cent in Oman to 16.5 per cent in
Saudi Arabia, 8.4 per cent in Qatar, 2.5 per cent in Kuwait and probably 1.5 per cent
in Bahrain (ESCWA, 2001). These countries are the poorest in natural water resources. In
2000 Oman had only 320m3/capita/year, Bahrain 170m3, Saudi Arabia 100m3, Kuwait 20m3
and Qatar 10m3. Given that the poverty level is 500m3/capita/year and the critical level is
100m3, one can see how critical the situation is. Freshwater scarcity has been a serious
challenge and freshwater supplies from conventional sources are unable to meet basic
demands. The desalination option continues to be favoured. With desalination, which
consumes power derived from oil and natural gas, the annual per capita water
consumption in the GCC in 2000 had become reasonable (Table 8.7, including the
projected demands for 2010).
The GCC countries are, generally, the wealthiest in thermal heat, which can be used
for desalination, either directly using trough solar concentrating or solar tower or solar
dish technology, or possibly photovoltaic technology, to run the pumping of the reverse
osmosis system. Table 8.8 shows the solar potential in these countries.

TABLE 8.7 Annual per capita water consumption in 2000 and projected demand in 2010 for the GCC
countries The poorest in natural water resources
COUNTRY

Bahrain
Kuwait
Oman
Qatar
Saudi Arabia
United Arab Emirates
Total

DOMESTIC
DEMAND (m3)
2000
237
308
54
258
140
340
1337

2010
279
302
76
296
137
352
1442

AGRICULTURAL
DEMAND (m3)

TOTAL
DEMAND (m3)

2000
197
64
467
341
978
573
2622

2000
436
372
521
599
1118
913
3959

2010
215
60
417
349
880
638
2559

Source: Dabbagh (1995)

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

2010
494
362
493
645
1017
990
4001

ch-08

2/7/07

2:38 PM

Page 283

Solar Energy Technology in the Middle East and North Africa

283

TABLE 8.8 Solar resources in the GCC countries


COUNTRY
Bahrain
Kuwait
Oman
Qatar
Saudi Arabia
United Arab Emirates

GLOBAL SOLAR RADIATION


(kWh/m3/day)
6.4
6.2
5.1
5.5
7.0
6.5

DIRECT NORMAL SOLAR RADIATION


(kWh/m3/day)
6.5
6.5
6.2
5.6
6.5
6.0

Source: Global Emission Newsletter (2004)

With relatively large land area available in these countries, it is possible to utilize the solar
power to produce enough water in the GCC (and other Arab states), while the electricity can
be transmitted to Europe, which suffers from heavy air pollution due to burning conventional
fuel, especially in those countries with heavy industry and advanced technology.
Cooperation between MENA and Europe is indeed needed. The investment in solar
water desalination may be very successful. According to certain newspaper reports, the
GCC is planning to spend more than US$40 billion.
Table 8.9 shows the useful electrical equivalent per m3 of distillate (UEPCMD) in kWh,
and the specific electrical energy input per m3 of distillate (SEEPCMD) in kWh/m3 for typical
multi-stage flash (MSF), multiple effect distillation (MED), thermal vapour compression
(TVC), mechanical vapour compression (MVC), reverse osmosis (RO) and electrolysis (ED).
TABLE 8.9 Comparison of energy consumption for six types of water desalination
DESALINATION TYPE
UEPCMD(kWh)
SEEPCMD(kWh/m3)

MSF
16
4

MED
11.5
2.0

TVC
14.56
1.7

MVC

11.5

RO

1.5

ED

Source: Jawad (2001); ESCWA (2003)

The future of the electrical power transmission between different countries is


affordable. If this happens, then there will be no problem at all in producing electricity
and water using solar energy systems, such as the concentrating solar power CSP for
desalination or by photovoltaic. We have to bear in mind that, today, the cost of water
desalination using fossil fuel ranges from US$0.5/m3 (reverse osmosis) to US$1/m3
(thermal desalination methods).
According to the ESCWA (2003) report, MSF desalination plants and RO systems are
the most predominant installed desalination plants in the GCC countries. These systems
achieve 7.859 and 1.811 million m3 per day respectively. VC systems achieve 19,469m3
per day, ED 64,424m3 per day. The minimum installed capacity for ME is only 21,204m3
per day (Table 8.10). In conclusion:
the GCC countries have the largest installed MSF desalination capacities in the world,

accounting for 80 per cent of the world total of 9.8 million m3 per day;

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

ch-08

2/7/07

2:38 PM

284

Page 284

W. E. ALNASER, F. TRIEB AND G. KNIES

TABLE 8.10 The available desalination capacity in the GCC countries, in 2000, using different desalination
technologies, in cubic metres per day
DESALINATION
TYPE
FEED
WATER
RO

MSF

VC

MED

ED

Brackish
Sea
Waste
and others
Total
Brackish
Sea
Total
Brackish
Sea
Waste
Total
Brackish
Brine
Sea
Total
Brackish
Total

Total

BAHRAIN KUWAIT

GCC COUNTRIES
OMAN QATAR

SAUDI
ARABIA

65,504
48,217

44,103
3000

5151
13,840

4154
5000

974,201
444,363

1130

70,133

113,721
47,103

297,202 1,468,036
297,202 1,468,036

1635

1635

1135

1272

1200
1135
2470
3350
2418
3350
2418
416,861 1,520,029

UNITED
ARAB
EMIRATES
73,969
58,344

TOTAL BY
TECHNOLOGY
(m3/day)
1,167,082
572,764

71,263

18,991 10,284 1,488,697 132,313

1300

161,015 499,954 3,491,385 1,940,596


161,015 499,954 3,492,685 1,940,596
1515

6362

1600 19,590
59,012 103,735

1242

3115 19,590
66,616 103,735

500

3000
2542
6900
4655
3000
2542
6900
5155

56,156
2500

56,156
2500
186,121 532,370 5,111,054 2,184,299

1,811,109
1300
7,858,006
7,859,306
7877
185,572
1242
194,691
1635
1262
18,297
21,204
64,424
64,424

Source: ESCWA (2001)

the GCC countries have 9.35 million m3 per day or some 19.4 per cent of the total RO

world capacity; and


the world total of ED, ME and VC technologies in 2000 was 2.95 million m3 per day

(280,319m3 per day for GCC countries, representing only 9.5 per cent of the world
total) (ESCWA, 2003).

8.6 WATER RESOURCES AND WATER DEMAND IN MENA


The analysis leads to a scenario for the prediction of the demand and the resources of
freshwater at the country level. Inside a country, there might be regions with deficits that
cannot be identified on the basis of countrywide statistical data. The analysis of Spain or
Italy at that level would not yield any deficits. Excessive withdrawal of groundwater is also
a common problem in many regions. This study concentrates on those cases that can be
identified as problematic on the basis of national statistics. Sub-national demand for nonconventional freshwater resources is not considered.
Most of the actual data on water resources and use has been obtained from the
AQUASTAT database of the Food and Agriculture Organization of the United Nations (FAO)
(FAO, 2004; AQUASTAT, 2004). Extrapolations have been made on the basis of population
and GDP growth rate expectations as described in this report.
ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

ch-08

2/7/07

2:38 PM

Page 285

Solar Energy Technology in the Middle East and North Africa

285

The extrapolation of future water demand at country level is based on the


assumptions that:
Agricultural production and its water demand per capita will be maintained as today.

This means that the demand of the agricultural sector will be growing proportionally
to population.
The demand of the domestic and industrial sector will grow proportionally to the
GDP, which is calculated for every country adding the population growth rate to the
per capita GDP growth rate.
The efficiency of water use in the agricultural and municipal sector will be increased
from todays country-specific values to a maximum value which depends on the
selected scenario, the water demand growth rate thus becoming lower than the
population or GDP growth rates. Enhanced technologies will additionally de-couple
water demand from economic growth as experienced in Australia in recent decades
(AQUASTAT, 2004; Gleick, 1998).
In the analysis of the power sector, two different economic scenarios have been used to
determine baselines for water demand predictions:
1 The scenario Following Up assumes an average per capita GDP growth rate of only 1.2
per cent for every country from today until 2050. This implies that the relative distance
between the actual GDP/capita (US$-PPP) of the respective country and the US will
remain constant because the GDP of the US at the same time will also be growing by
1.2 per cent. Efficiencies of the agricultural and the municipal water supply system and
the reuse of wastewater increase gradually from the present national performance
values to a future better value of an enhanced system. However, the efficiency
enhancements are limited by the slow economic development. Population growth and
the agricultural sector dominate the water demand growth rates in this case. Decoupling of the water demand from the economic growth by using enhanced water
supply technologies is also limited in this scenario (Gleick, 1998; Heston et al, 2002).
2 The scenario Closing the Gap assumes that the relative distance between the actual
GDP/capita (US$-PPP) in the US and the respective country is reduced to 50 per cent
by 2050 while the GDP of the US at the same time is growing by 1.2 per cent. This
scenario assumes that the MENA countries will by 2050 achieve GDP per capita
values close to that of the European countries. In this case, the industrial and
domestic sectors will dominate the water demand growth. However, efficiencies will
also be increased and a significant de-coupling of water demand and economic
growth as experienced in Australia in the past decades will take place.
The water demand in the MENA region today is made up of 85 per cent agricultural use,
9 per cent domestic use and 6 per cent industrial use. The future demand is calculated
individually for each country and aggregated to the regions of North Africa, western Asia
and the Arabian Peninsula.
Under the assumptions of the scenario Following Up, the share of agricultural water
use will fall to about 80 per cent and the domestic and industrial share will increase to 12
ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

2/7/07

286

2:38 PM

Page 286

W. E. ALNASER, F. TRIEB AND G. KNIES

per cent and 8 per cent respectively. The total water demand will increase from 300 billion
m3/y today to about 510 billion m3/y in 2050. The scenario reflects the influence of
enhanced water management, policies and efficiencies that are of highest priority for a
sustainable water future in MENA, but that are limited by the slow economic growth
within this scenario.
Under the assumptions of the scenario Closing the Gap, the share of agricultural
water use will fall to about 66 per cent and the domestic and industrial share will increase
to 18 per cent and 16 per cent respectively, more and more dominating the water
demand. The total water demand will increase from 300 billion m3/y today to about 540
billion m3/y in 2050 (Figure 8.13). The scenario also reflects the pronounced influence of
enhanced water management, policies and efficiencies, giving them highest priority for a
sustainable water future in MENA, especially since this scenario is oriented towards a
high economic growth.
In terms of water demand, both scenarios are rather optimistic compared to other
scenarios that predict a doubling of demand for 2025. This is determined by extrapolating

600
500

Freshwater Demand
[billion m/y]

ch-08

Industrial
Domestic

400

Sus tainable
Water

300
200

Agricultural

100
0

00
20

05
20

10
20

15
20

20
20

30
25
20
20
Year

35
20

40
20

45
20

50
20

Note: The white line indicates the sustainable renewable freshwater resources.

FIGURE 8.13 Water demand structure in MENA and its evolution until 2050; scenario Closing the Gap

the water demand growth rates as experienced in the last decades. However, we believe
that a reduction of the agricultural sector and the successive de-coupling of economic
growth and water demand are more realistic approaches. At first glance, it is surprising
that both scenarios culminate in a rather similar water demand of 510 to 540 billion m3/y
by 2050, which obviously will be achieved with or without economic growth. It reflects the
positive impact of economic stability and development on the water supply. In the
scenario Following Up, consumption is limited by availability, while in the scenario
Closing the Gap, it is rather limited by the enhanced efficiency of the supply system.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

ch-08

2/7/07

2:38 PM

Page 287

Solar Energy Technology in the Middle East and North Africa

287

As the future deficits and the additional demand for non-conventional resources will
not change considerably assuming one scenario or the other, the scenario Closing the
Gap (which is more desirable from the point of view of the MENA countries) will be used
as reference in further analyses.
Western Asia still has large sustainable water resources that will be increasingly
exploited in the future. However, even in this region, non-sustainable use as from fossil
fuelled desalination and from unsustainable groundwater withdrawal is already
experienced on a local level and shows an increasing trend in the future. Unsustainable
water supply from fossil-fuelled desalination and from excessive groundwater withdrawal
is considered as potential future deficits (Figure 8.14).
The sustainable freshwater resources of North Africa are being used today almost to
their limits, and therefore no considerable increase of their exploitation can be expected
for the future. Unsustainable use from fossil desalination and from excessive
groundwater withdrawal is already taking place to a considerable extent, with a dramatic
increase of this situation ahead. On the Arabian Peninsula, the relation of sustainable and
unsustainable use of water is even more dramatic.
The total annual water deficits in MENA will increase from 35 billion m3/y today, at
present supplied by excessive groundwater withdrawals and fossil-fuelled desalination, to
about 155 billion m3/y by 2050. There is no sustainable resource in sight to supply such
deficits except renewable energies. The cost of fossil fuels is already today too high (US$50
in August 2004) for intensive seawater desalination and its volatility and the fact that fossil
fuels are limited in time eliminate fossil fuels as a resource for sustainable water security in
MENA. Nuclear power is a very limited and costly resource, and in addition faces unsolved
problems such as nuclear waste disposal, proliferation and other serious security issues.

FIGURE 8.14 Water supply from sustainable sources and deficits in MENA (Closing the Gap)

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

2/7/07

2:38 PM

288

Page 288

W. E. ALNASER, F. TRIEB AND G. KNIES

(a)

Consumption Growth Rate [%/y]

The water demand growth rates will decline in all three MENA regions from about 1.5
to less than 1 per cent per year. The per capita water demand and its future trend is
different in the three regions (Figures 8.15a and 8.15b). The MENA average per capita
demand is expected to stay almost constant at about 800m3/capita/y. Western Asia will
reduce its per capita demand from 1000 to about 900m3/capita/y, while the demand in
North Africa will grow from 700 to about 800m3/capita/y, due to a relatively moderate
growth of the population and an increasing importance of the domestic and industrial
sector, mainly in Egypt. The specific consumption on the Arabian Peninsula will fall from
600m3/capita/y (currently) to about 400m3/capita/y in 2050 due to a strong growth of the

3.0
2.5
2.0

North Africa
Western Asia
Arabian Peninsula
Total MENA

1.5
1.0
0.5
0.0

00 05 10 15 20 25 30 35 40 45 50
20 20 20 20 20 20 20 20 20 20 20
Year

1400

(b)

Consumption [m/capita/y]

ch-08

1200
1000
North Africa
Western Asia
Arabian Peninsula
Total MENA

800
600
400
200
0
00 05 10 15 20 25 30 35 40 45 50
20 20 20 20 20 20 20 20 20 20 20

Year
FIGURE 8.15 a) Water consumption growth rates in MENA (Closing the Gap); b) Water consumption per
capita in MENA (Closing the Gap)

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

2/7/07

2:38 PM

Page 289

Solar Energy Technology in the Middle East and North Africa

289

population and the persisting importance of the agricultural sector, coupled with very
limited natural water resources.

8.6.1 NORTH AFRICA


The scenario assumptions lead to a linear growth of the water demand in North Africa
from 100 billion m3/y today to 200 billion m3/y in 2050 (Figure 8.16). The reduction in the
agricultural sector is compensated by the growth of the domestic and industrial sectors.
Sustainable sources in North Africa cannot be exploited to a greater extent than today. All
countries will experience growing deficits, with Egypt being by far the dominating case,
due a very strong agricultural sector and large population, followed by Libya and the
Maghreb countries. The deficit of Egypt expected for 2050 might rise to the present water
capacity of the Nile River of about 70 billion m3/y. An official expectation of a deficit of 35
billion m3/y until 2025 was recently published.
All countries will experience a reduction of their water demand growth rates of about
0.5 per cent by 2050. The per capita demand is highest in Egypt and Libya (about

Consumption [billion m/y]

ch-08

250
200
Industrial
Domestic
Agricultural

150
100
50
0

0
20

05

20

10

20

15 020 025 030 035 040 045 050


20
2
2
2
2
2
2
2
Year

FIGURE 8.16 Water demand structure in North Africa and its evolution until 2050

1000m3/cap/y), and lowest in Algeria and Malta (200m3/cap/y), with a slightly increasing
trend in all countries.
The strong economic growth of the scenario Closing the Gap reveals the challenge
of this path, as the water demand of the industrial and domestic sector will grow very
quickly and overcompensate for possible reductions in the agricultural sector.

8.6.2 WESTERN ASIA


The water demand in western Asia will increase from 175 billion m3/y today to about 275
billion m3/y in 2050, showing a slight stabilization trend by that time (Figure 8.17).
There are vast sustainable water resources in that region which will be increasingly
exploited in the future. However, local deficits will occur in Syria, Jordan and Israel and
later also in Iraq.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

2/7/07

290

2:38 PM

Page 290

W. E. ALNASER, F. TRIEB AND G. KNIES

Consumption [billion m/y]

ch-08

300
250
200

Industrial
Domestic
Agricultural

150
100
50
0
20

00

20

05

20

10

20

15

20

20

20

25

20

30

20

35

20

40

20

45

20

50

Year
FIGURE 8.17 Water demand structure in western Asia and its evolution until 2050

The demand growth rates are high in Jordan but at a very low level of per capita
demand, as can be appreciated from Figure 8.20. Strong consumers are Iraq, Turkey and
Syria, with only Syria facing a short-term deficit. The average per capita demand of the
western Asian region will be slightly reduced from 950 to 850m3/cap/y, while in all
countries the consumption growth rates will be reduced.

8.6.3 THE ARABIAN PENINSULA


The Arabian Peninsula is characterized by a strongly growing population and a dominating
agricultural sector. The demand will increase from 30 to 65 billion m3/y (Figure 8.18). The
regions water demand is dominated by Saudi Arabia and Yemen, both relying to a great
extent on non-sustainable sources, including fossil-fuelled desalination and excessive
groundwater withdrawal. Due to the combination of high population and high dependency
on agriculture, both countries will be facing considerable deficits if their water supply is
persistently based on the limited resources of fossil fuels and non-renewable
groundwater. The sustainable natural resources of this region are very limited.
The average per capita consumption on the Arabian Peninsula will be reduced from
600 to 450 m3/cap/y. Saudi Arabia and UAE will have the highest consumption per capita
(about 700800m3/cap/y). The strongest decrease of per capita consumption will be
experienced in Yemen.
In terms of population growth and share of the agricultural sector, Yemen is a very
specific case among the MENA countries. The per capita consumption will decrease from
400 to 250m3/cap/year, but the consumption growth rates will not decrease until after
2030. The scenario Closing the Gap would require a continuous GDP growth rate of
Yemen of 11 per cent until 2050 (a necessary 7.8 per cent per capita growth rate to close
the GDP per capita gap with the US plus a 3.2 per cent population growth rate), which is
rather unrealistic.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

2/7/07

2:38 PM

Page 291

Solar Energy Technology in the Middle East and North Africa

Consumption [billion m/y]

ch-08

291

70
60
50
Industrial
Domestic
Agricultural

40
30
20
10
0

20

00

20

05

20

10

20

15

20

20

20

25

20

30

20

35

20

40

20

45

20

50

Year
FIGURE 8.18 Water demand structure for Arabian Peninsula and its evolution until 2050

8.7 THE POTENTIAL FOR DESALINATION BY CONCENTRATING


SOLAR POWER
The previous analysis of water deficits in MENA showed that there is a pressing need for
new, non-conventional, sustainable water sources in many countries of this region. The
hot spots can be found in North Africa (mainly Egypt and Libya) and the Arabian Peninsula
(mainly Yemen and Saudi Arabia), while the situation is by far less critical in the countries
of western Asia. However, Syria, Jordan and Israel also face considerable deficits.
Although the demand of the agricultural sector, which in MENA makes up 85 per cent of
the total, will not grow as fast as in the past decades. This will be compensated for by a
quickly growing demand of the urban centres and industry (FAO, 2003).
The use of water is today heavily subsidized in many MENA countries (Al-Zubari,
2002). This reflects the fact that the cost of supplying water is already too high today
considering the per capita income level, especially in the agricultural sector. Today, the
cost of desalinating water using fossil fuels ranges between US$0.5/m3 (reverse osmosis)
and US$1/m3 (thermal desalination methods), which would be higher than the prices paid
for water in most MENA countries.
Economies building their water supply to a great extent on desalination with fossil
fuels would suffer from additional subsidy loads, from the volatility of fossil fuel costs and
from the gradual depletion and cost escalation of fossil energy resources. A severe
stagnation of investments in the water sector is the consequence of this situation, the
total water sector becoming more and more dependent on national and international
subsidization.
Today, many countries try to avoid an increasing dependency on desalination and
fossil fuels by exploiting their groundwater resources. However, in many countries the
exploitation rate is much higher than the renewable groundwater resources (Figure 8.19),
making this solution not more sustainable than the dependency on fossil fuels. Therefore,

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

ch-08

2/7/07

292

2:38 PM

Page 292

W. E. ALNASER, F. TRIEB AND G. KNIES

Saudi Arabia

1456%

Kuwait

1275%

Qatar
Gaza
Bahrain
Jordan
Yemen
Syria
UAE
Oman
Algeria
Tunisia
Morocco
Cyprus
0%

50%

100%

150%

200%

250%

300%

350%

400%

Source: Saghir (2000)

FIGURE 8.19 Groundwater withdrawals as percentage of safe yield for selected countries

a renewable, sustainable freshwater source with much lower and more stable costs than
fossil fuels is required.
In the present study, we have assumed that the unsustainable water supplied by
groundwater depletion and fossil-fuelled desalination represents a potential future deficit
that could be covered by concentrating solar power (CSP) plants in cogeneration with
multi-effect desalination (MED) and additionally using the remaining electricity for reverse
osmosis (RO) desalination.
To estimate the minimum CSP capacity potential in the water supply sector, we have
assumed that all plants would be coupled to MED desalination plants, while the electricity
generated is completely used for RO desalination in order to produce larger amounts of
desalinated water. In view of the quick increase of future deficits of water in MENA, this
could become necessary in order to substitute for the unsustainable water supply more
quickly. This approach leads to the minimum installed capacity of CSP that is necessary
to cover the future water deficits in MENA.
The capacity potential for CSP would, in reality, be higher, as part of the plants would
be only used for cogeneration of city power and MED desalination, but without RO
desalination. The installation of such plants would be limited to the coast. Another part
would only be used for power generation for RO, but without making use of cogeneration

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

2/7/07

2:38 PM

Page 293

Solar Energy Technology in the Middle East and North Africa

293

with MED plants. Those CSP plants could be anywhere on the grid, while only the RO
desalination plant must be located at the coast.
To this capacity, the potential of CSP electricity generating and cogeneration plants
outside the water sector may be added. This is described in other chapters of this study.

8.7.1 POTENTIAL FOR CSP DESALINATION IN MENA


Today, 35 billion m3/y of the water consumption in MENA are covered by non-sustainable
water sources. According to the reference scenario Closing the Gap, this deficit will
increase to about 155 billion m3/y by 2050 (Figures 8. 20 and 8.21).
The figures show how these deficits could be subsequently covered by CSP
desalination plants, reducing the non-sustainable water supply and providing most of the
non-conventional water by 2030 and afterwards using solar energy. Deficits will have to
be bridged by fossil-fuelled desalination and groundwater withdrawals, hoping that those
resources will remain available and affordable until then. This may seem optimistic, but
there are no sustainable and affordable alternatives.
It is a reassuring fact that the potential of CSP is neither limited by the solar energy
resource nor by its cost, but only by the possible speed of CSP capacity expansion, and
that there is a solution for the freshwater deficits in MENA that can be realized by 2030.
However, a lot of time has been unnecessarily lost, leading to a considerable increase of
non-sustainable water until 2030. This calls for the intensive additional use of other
renewable sources like geothermal and wind power for non-conventional water
production, and also calls for an intensive freshwater management programme with
urban and rural applications. Only a decided employment and efficient combination of all
possible measures will lead to satisfactory and sustainable water supply security in
MENA. The capacity of CSP plants until 2050, if installed exclusively for seawater

600
Water Demand and Supply
billion m/y

500
400
CSP-Desalination
Non-Sustainable Supply
Sustainable Supply

300
200
100

45

40

35

30

25

20

15

10

05

50
20

20

20

20

20

20

20

20

20

20

00

0
20

ch-08

Year
FIGURE 8.20 Water demand, sustainable freshwater resources, non-sustainable supply and potential
future supply by CSP via cogeneration by MED plus direct generation via solar electricity in RO plants

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

ch-08

2/7/07

294

2:38 PM

Page 294

W. E. ALNASER, F. TRIEB AND G. KNIES

FIGURE 8.21 Covering the future freshwater deficits in MENA by transitory non-sustainable sources and
by CSP plants using MED in cogeneration plus solar electricity for RO

desalination, could amount to a total of 67GW. North Africa (35GW) has the largest
potential for CSP desalination plants, followed by the Arabian Peninsula (24GW) and the
western Asian countries (8GW) (Figure 8.22).

FIGURE 8.22 Capacity potential for CSP-Desalination plants with MED and RO in the three main MENA
regions

8.7.1.1 North Africa


The deficit in North Africa will grow from 15 billion m3/y in 2000 to 80 billion m3/y in 2050
with a major share taken by Egypt. The CSP capacity potential for desalination amounts
to 30GW for Egypt and 4GW for Libya, while the other countries have minor shares. On

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

ch-08

2/7/07

2:38 PM

Page 295

Solar Energy Technology in the Middle East and North Africa

295

the basis of country statistics, no potential can be detected for Morocco, Malta and
Tunisia, although there may be deficits at the local level.

8.7.1.2 Western Asia


The deficit in western Asia will grow from 0.5 billion m3/y in 2000 to 20 billion m3/y in 2050
with a major share taken by Syria. After 2040, Iraq will be part of the deficit mix. The CSP
capacity potential for desalination amounts to 10GW for Syria and 1GW each for Israel and
Jordan. The other countries have minor shares. On the basis of country statistics, no
potential can be detected for Cyprus, Lebanon, Turkey and Iran, although there may be
deficits at the local level.

8.7.1.3 The Arabian Peninsula


The deficit on the Arabian Peninsula will grow from 20 billion m3/y in 2000 to 55 billion m3/y
in 2050 with the major share taken by Saudi Arabia and Yemen. The CSP capacity potential
for desalination amounts to 32GW for Saudi Arabia and 17GW for Yemen, while the other
countries have minor shares.
It is evident that it is very feasible to use CSP technology in order to produce electricity
using solar energy in the Arab countries, which have ample solar radiation (Table 8.11). In
fact, if only 1 per cent of the Saudi Arabian desert was used for CSP, the electricity gained
would be enough to provide Europe and the Arab countries with electrical power and
water. European countries would provide the technology and training and, in return, could
get enough electrical power from the Arab states through the interconnection shown
below. By this, European countries could minimize their air pollution while Arab countries
would be receiving 100 billion m3 of water annually.

8.8 ENVIRONMENTAL IMPACT OF USING SOLAR TECHNOLOGY


IN MENA
The main environmental impact of our scenario is the reduction of carbon emissions by
about 40 per cent from electricity generation in spite of growing demand. In the following
we highlight two important issues in this regard.

8.8.1 LAND USE FOR CSP


The specific land requirement of hydropower ranges between 10km2/TWh/y for microhydropower to over 400km2/TWh/y for very large schemes like the Aswan dam (Table
8.12). The average value results in 165km2/TWh/y for the total analysed region.
Geothermal power requires little land (1 to 10km2/TWh/y), and the areas affected are in the
subsoil at thousands of metres depth. In our scenario, biomass is produced mainly by
agricultural and municipal residues (no extra land use) and from wood, resulting in an
average land use of only 2km2/TWh/y. Energy crops with a very high land use were not
considered in the MENA countries, as they would compete with food and water supply.
For wind power, the average land use was 46km2/TWh/y. The specific values differ
considerably according to the different performance indicators in each country.
Table 8.12 the two columns on the right show the total area of each country and the
percentage of this area used for power generation by renewable energy sources in 2050.

ADVANCES IN SOLAR ENERGY 2007 VOLUME 17 PAGES 261304

ch-08

2/7/07

296

2:38 PM

Page 296

W. E. ALNASER, F. TRIEB AND G. KNIES

TABLE 8.11 Global solar radiation, monthly and annual means (in kWh/m2/day)
CITY
Amman
Aquaba
Abu Dhabi
Mohraaq
Alain
Tunis
Cabes
Alger
Oran
Tamanrset
Riyadh
El-Medina
Taif
Khartoum
Juba
Damas
Almsleamea
Bagdad
Mosul
Mascat
Oman
Jerusalem
Doha
Kuwait
Tripoli
Sirte
Cairo
Aswan
Casablanca
Marrakech
Nouak Chatt
Sanaa
Mocha

JAN
2.7
3.0
4.3
3.6
3.9
2.4
2.9
2.2
2.8
5.2
3.5
4.5
4.4
5.5
5.5
3.1
2.0
3.0
2.0
4.0
4.6
2.7
3.7
3.1
2.9
3.4
5.7
4.6
2.7
3.4
5.7
4.0
4.9

FEB
3.7
4.6
5.0
4.8
4.3
3.1
3.7
3.0
3.7
6.1
4.6
5.4
5.2
6.4
5.5
3.5
3.0
3.8
2.8
4.7
4.7
3.4
4.4
4.1
4.0
4.2
4.0
5.6
3.3
4.2
6.2
4.4
5.4

MAR
5.0
5.9
5.7
4.6
5.3
4.4
4.9
4.1
4.9
6.9
5.1
6.2
5.6
6.8
5.4
4.6
4.1
4.8
3.6
5.5
5.5
5.0
4.9
5.5
5.0
5.0
5.2
6.5
4.5
5.2
7.2
4.8
6.1

APR
6.8
6.7
6.7
6.1
6.6
5.7
5.9
4.9
6.2
7.7
5.5
6.6
5.8
7.2
5.6
5.8
5.3
5.7
4.5
6.3
6.3
6.0
5.7
6.2
6.0
5.6
6.3
7.3
5.4
6.0
7.8
4.4
6.3

MAY
7.8
7.6
7.6
6.9
7.5
6.5
6.8
6.0
6.9
7.8
5.6
7.4
5.8
6.6
5.6
7.4
6.2
6.5
5.5
6.9
6.9
6.9
6.2
7.1
6.5
6.2
7.1
7.4
6.3
6.7
7.7
5.6
6.2

JUN
8.4
8.3
7.6
7.4
7.5
7.3
7.2
6.2
7.4
7.7
6.1
8.2
6.4
6.3
5.2
8.0
7.9
7.3
6.3
6.7
6.7
7.6
6.5
7.9
7.0
6.9
7.6
8.0
6.5
7.3
7.7
5.3
5.8

JULY
8.2
8.0
7.0
7.0
6.8
7.4
7.3
7.0
7.5
7.6
6.1
8.3
6.3
6.1
5.0
7.9
7.7
7.2
6.2
6.1
6.1
7.6
6.0
7.5
7.1
6.9
7.4
7.8
6.4
7.6
7.2
4.7
4.3

AUG
7.5
7.5
6.7
6.3
6.6
6.6
6.5
6.4
6.8
7.4
5.9
8.0
5.9
6.0
5.4
7.2
7.1
6.6
5.6
6.0
6.0
7.2
5.8
7.1
6.5
6.3
6.8
7.4
5.9
7.0
7.1
4.6
4.8

SEP
6.4
6.5
6.5
5.8
6.4
5.3
5.4
5.1
5.6
6.7
5.7
6.8
5.5
6.0
5.9
6.1
6.1
5.7
4.8
5.8
5.8
6.1
5.5
6.2
5.5
5.6
5.9
6.6
5.0
5.9
6.7
5.2
5.3

OCT
4.8
5.1
5.7
4.9
5.7
4.0
4.2
3.3
4.0
6.0
5.3
5.8
5.2
5.9
5.8
4.5
4.1
4.4
3.5
5.2
5.2
4.8
4.8
4.8
4.0
4.1
4.7
5.7
3.8
4.6
6.3
4.0
6.1

NOV
3.6
4.0
4.8
3.9
4.7
2.9
3.2
2.7
2.9
5.2
4.5
4.8
4.6
5.7
5.5
3.0
2.8
3.3
2.3
4.4
4.4
3.5
4.1
3.4
3.1
3.2
3.5
4.8
2.7
3.6
5.5
4.0
5.4

DEC
2.7
3.4
4.0
3.0
4.6
2.3
2.7
2.0
2.4
4.8
3.6
4.3
4.4
5.5
5.3
2.3
2.0
2.7
1.8
3.8
3.8
2.6
3.5
2.9
2.3
3.0
5.9
4.3
2.4
3.2
5.1
3.9
5.1

ANNUAL
5.6
5.9
6.0
5.3
5.8
4.8
5.1
4.4
5.1
6.6
5.1
6.4
5.4
6.2
5.5
5.2
4.9
5.1
4.1
5.4
5.4
5.3
5.1
5.5
5.0
5.0
5.4
6.3
4.6
5.4
6.7
4.6
5.5

Source: Global Emission Newsletter (2004)

Hydropower surface demand varies strongly between countries. Photovoltaic surface


demand considers only 50 per cent of the total because many plants will be installed on
roofs. Wind power and CSP surface demand is calculated as if exclusively used for power
generation. Biomass surface demand is only considered for fuel wood energy.
Concentrating solar thermal power schemes have a specific land use of
610km2/TWh/y. However, land could be g