Vous êtes sur la page 1sur 207

Renewable Energies series

Solar detoxification

Edited by
Julián Blanco Gálvez
Sixto Malato Rodríguez
In the SAME series:

Geothermal energy
Solar photovoltaic project development
Solar photovoltaic systems: technical training manual

The designations employed and the presentation of material throughout

this publication do not imply the expression of any opinion whatsoever
on the part of UNESCO concerning the legal status of any country,
territory, city or area or of its authorities, or concerning the delimitation
of its frontiers or boundaries.

The authors are responsible for the choice and the presentation of the
facts contained in this book and for the opinions expressed therein,
which are not necessarily those of UNESCO and do not commit the

Published in 2003 by the United Nations Educational,

Scientific and Cultural Organization
7, place de Fontenoy F-75352 Paris 07 SP
Typeset by S R Nova Pvt. Ltd., Bangalore,
India Printed by Jouve

ISBN 92-3-103916-4
© UNESCO 2003
All rights reserved
Printed in FRANCE

Everyone is aware of the strong capacity of sunlight to 1990, in particular in projects researching photocatalytic
degrade all kinds of objects. It is also well known that, to decontamination of used water at both national and
prevent degradation and deterioration, objects should be interna- tional levels. The experience they acquired with
sheltered from sunlight or light irradiation in general. solar detoxifi- cation systems at the engineering level led
This characteris- tic of sunlight is based on the property to the development and the installation at PSA of
of solar photons to directly degrade and break down Europe’s largest solar detoxi- fication facility at pilot-
organic molecules, an effect that it is known as plant scale, and this has been suc- cessfully used by
‘photolysis’. many European research institutions. The authors hope
In the 1970s, it was discovered that this effect could be that this book will illustrate the cross-linked synergistic
sub- stantially enhanced with the help of a catalyst, relationships that have been developed among different
opening up the field of solar photocatalysis. But it was not European research groups involved in the PSA
until the mid-1980s (coinciding with a growing photocatalysis research program over the last twelve years.
environmental consciousness across the world) that this This book summarizes most of the scientific and
process was given environmental application, and its use techno- logical research performed at PSA in this area
in the treatment and degradation of hazardous substances since 1990, through close collaboration with many
in water, air and soil began to be appre- ciated as a very scientific colleagues from a large number of universities,
attractive proposition. Of these three potential applications research institutions and private companies and through
(water, air and soil), the treatment of water con- taminants many collaborative efforts, networking activities and
using sunlight is considered of special relevance because joint projects.
of the magnitude of water-contamination problems and The book is divided into two parts of five chapters
its potential for worldwide application. each. Part A addresses the theory and fundamentals
Solar photocatalytic degradation of water of water decontamination by means of solar energy. The
contaminants (a process also known as ‘solar detoxification’) objective of this part is to provide the reader with a
is an outstand- ing example of how well solar applications background in readi- ness for Part B, which addresses
and environmen- tal problems fit together. This remarkable the practical applications and systems engineering of
application is currently the focus of many research the process. Our main motiva- tion in writing this book
institutions, and since 1990 it has also been the main was to collect, in a comprehensive and extensive way,
research and development object of the Solar Chemistry all the work performed at PSA and make it accessible not
group at Plataforma Solar de Almería (PSA, the largest only to people interested in solar and photo- catalytic
European facility for solar energy applications) and the applications, but also to all those who are interested in
solar centre of CIEMAT, a public insti- tution in Spain learning how environmental technology could help
devoted to research and technological devel- opment in the to solve environmental problems in general.
areas of energy and the environment.
The Solar Chemistry group at PSA have been Julián Blanco Gálvez
participating in the chemical application of water solar Sixto Malato Rodríguez
detoxification since

Our deep gratitude is owed to our colleagues Dr Manuel Mike Prairie and Jim Pacheco (Sandia National Labs,
Romero, Dr Benigno Sánchez, Dr Ana Isabel Cardona United States), Dr Detleft Bahneman and Dr Roland
and Alfonso Vidal from our mother institution CIEMAT, Goslich (ISFH, Germany), Professor Ezzio Pelizzetti and
and to Dr Christoph Richter (DLR-PSA), who has been Professor Claudio Minero (University of Torino, Italy),
deeply involved in most of our projects. This gratitude Professor Jean-Marie Herrmann (CNRS, France), Dr
is also extended to all CIEMAT-PSA staff for their great César Pulgarín (Ecole Politechnique Federale de
helpfulness and high skill, which made possible the Lausanne, Switzerland), Dr Manuel Collares-Pereira and
design, construction and operation of the photocatalysis Joao Farinha Mendes (INETI, Portugal), Dr Karl-Heinz
pilot plants that have been the bases of the success of all Funken and Dr Christian Sattler (DLR, Germany),
the projects we have been involved with. The Professor Marc Anderson (University of Wisconsin–
management and execution of these projects is directly Madison, United States), Dr Yogi Goswami (Solar
related to the knowledge we have gathered from them, Energy and Energy Conversion Laboratory, University of
most of it included in this book. We would also like to Florida, United States), Professor Leonardo Palmisano
give special recognition to Juan Antonio Camacho, Ginés (University of Palermo, Italy), Dr Charles Giannotti
García, Jose Manuel Molina, Jaime Aranda, Bernabé (CNRS, France), Dr Mirella Musci and Maria Cristina
Calatrava and Angel Soler, together with the Operation Casalle (CESI, Italy), Professor Rupert Bauer (Technical
and Maintenance team of the PSA. University of Vienna, Austria) and Professor Octav Enea
We are grateful to the directors of the Plataforma (University of Poitiers, France).
Solar de Almería and Centro de Investigaciones Our acknowledgement and gratitude are also due to
Energéticas, Medioambientales y Tecnológicas Martin Vincent (Ecosystem S.A., Spain), Joao Oliveira
(CIEMAT) for the support they gave us in producing this (Ao Sol, Portugal), Alex Ryer (International Light,
book. Inc.), Dr Dieter Doënitz (Schott-Rohrglas GmbH,
In addition, this book would not have been possible Germany) and Dr V.H. Kuester (ALANOD Aluminium –
with- out the help, contributions and assistance of Veredlung GmbH and Co, Germany).
many people around the world who either provided For reasons of space we cannot mention all the
text and graphic material, reviewed the manuscript people who have made a contribution to the content of
or simply provided valuable suggestions. this book but, finally, we cannot forget the crucial help
For these reasons, we want to acknowledge the of Carmen Montesinos and Deborah Fuldauer, who
contribu- tions of Professor Jaime Giménez and Dr devoted many long hours to the revision and
David Curcó (University of Barcelona), Professor correction of the various manuscripts.
Xavier Domènech and Dr José Peral (University To our families, for their patience and moral support.
Autónoma de Barcelona), Dr Daniel Blake and Alan
Lewandoswky (NREL, United States),

List of Figures xii The solar spectrum 19

List of Tables xvii Solar ultraviolet irradiation 21
Notes on the Editors xix Atmospheric attenuation of solar radiation 22
1. Introduction Solar radiation measurement 24
Detectors 25
Aims and objectives 1 Filters 26
Input optics 26
Solar chemistry 2
Water contaminants 3 Summary 28
Photodegradation principles 6 Bibliography 28
Definitions 6 Self-assessment questions 29

Heterogeneous PHOTOCATALYSIS 6
Application to water treatment 9 Experimental Systems
Gas-phase detoxification 10

Summary 13 Aims and objectives 31

Bibliography 14
Self-assessment questions 15 Laboratory systems 32
Solar detoxification pilot plants 34
Operation of pilot plants 36
2. Solar Irradiation Once-through OPERATION 36
Aims and objectives 17 Modelling once-through AND BATCH OPERATION 37
Evaluation of solar UV radiation inside
photoreactors 39
The power of light 18
Visible light 19 DATA 40
INFRARED light 19
x Contents

Collector efficiency 41 Composite semiconductors 73

Actinometric experiments 42 SURFACE SENSITIZATION 73
Simplified method for the evaluation of solar UV Recommended analytical methods 73
radiation inside photoreactors 43 ORIGINAL CONTAMINANTS 73
Bibliography 44 EXTRACTION methods 76

Self-assessment questions 46 Toxicity ANALYSIS 76

Summary 77
4. Fundamental Parameters Bibliography 77
in Photocatalysis Self-assessment questions 78

Aims and objectives 49

6. Solar Detoxification
Direct photolysis 50 Technology
Influence of oxygen 51
pH influence 51
Aims and objectives 81
Influence of catalyst concentration 52
Initial contaminant concentration influence 54 Solar collector technology generalities 82
Radiant flux influence 56 Collectors for solar water detoxification: features 85
Temperature effect 57 Specific FEATURES of SOLAR UV light UTILIZATION 85
Quantum yield 57 PARABOLIC trough collectors 86
One-sun (NON-CONCENTRATING) collectors 87

Bibliography 59 HOLOGRAPHIC collectors 91
Self-assessment questions 60 Concentrated versus non-concentrated sunlight 91
Technical issues 93
Reflective SURFACES 93
5. Water Decontamination
Catalyst issues 98
by Solar Detoxification
Slurry versus supported CATALYST 98
Aims and objectives 63
Summary 100

Detoxification of pollutants 64 Bibliography 101
TOTAL MINERALIZATION 65 Self-assessment questions 102
Toxicity reduction 68
Quantum yield improvement through 7. Solar Detoxification
additional Applications
oxidants 69
Hydrogen peroxide 70 Aim and objectives 105
Other OXIDANTS 71 Introduction 106
Catalyst modification 71 Industrial wastewater treatment 108
METAL semiconductor MODIFICATION 72
Contents xi

Phenols 108 9. Project Engineering

AGROCHEMICAL compounds 109
HALOGENATED HYDROCARBONS 110 Aims and objectives 151
Antibiotics, ANTINEOPLASTICS AND other
PHARMACEUTICAL biocide compounds 111 Feasibility study 152
REMOVAL of HAZARDOUS METAL ions from WATER 112 HAZARDOUS compounds 153
Seaport tank terminals IDENTIFICATION of most SUITABLE PHOTOCATALYTIC process 154
113 DETERMINATION of the optimum process PARAMETERS 154
Groundwater decontamination POST-TREATMENT process IDENTIFICATION 155
114 Contaminated landfill DETERMINATION of TREATMENT FACTORS 155
cleaning 115 Water disinfection Feasibility study example 156
Gas-phase treatments 117 EXPERIMENTATION: TiO2-PERSULFATE tests 157
Photo-Fenton tests 158

Summary 121 Conclusions AND TREATMENT FACTORS 159
Bibliography 121 Preliminary design 160
Self-assessment questions 123 Preliminary design example 161
Final design and project implementation 163
Example of final design and project
implementation 165
8. Economic Assessment

Summary 171
Aims and objectives 125 Bibliography 171
Self-assessment questions 172
Photochemical and biological

reactors 10. International

coupling 127
Cost calculations 129
Solar or electric photons? 133 Agency:
Solar resources assessment the SolarPACES programme 176
139 The European Union 178
Comparison with other technologies 142 The CYTED programme 180
THERMAL OXIDATION 142 Main research activities 181
Air stripping 143 SPAIN 182
Recommendations for successful water
Adsorption 143 treatment projects in developing countries
MEMBRANE technology 144 183
Ozone OXIDATION 145 Summary 184
ADVANCED OXIDATION processes 145 Bibliography 184
Self-assessment questions 186
Summary 146
Bibliography 146
Self-assessment questions 148
List of Figures

1.1 Schematic view of solar chemical applications 3 2.7 Normalized solar UV spectra shown
1.2 Furfural photo-oxidation and in Figure 2.6 24
pentachlorophenol mineralization 2.8 TBDUV at different periods of the year at
(photochemical processes) 3 Plataforma Solar de Almería (37 °N) 24
1.3 Behaviour of electrons and holes within a 2.9 Responsivities of three detectors. In the inset
particle of illuminated semiconductor in is shown a schematic of the effect of a filter
contact with an electrolyte 7 on detector responsivity 26
1.4 UV spectra of acrinathrin and of sunlight 2.10 Relative spatial response of an ideal cosine
between 200 and 400 nm 9 diffuser (up) and a radiance lens barrel (down)
1.5 Effect of UV radiation on a TiO2 particle 27
dispersed in water 9 2.11 A solar global UV detector (tilted 37° and facing
1.6 Diagram of an experimental system with south) with a cosine diffuser 27
a cross-section view of a gas-phase 2.12 A solar direct UV detector installed on a
monolithic photoreactor 11 solar tracking system 27
1.7 Schematics and photo of a tubular gas-
phase photoreactor 12 3.1 Typical stirred-tank laboratory reactor 33
1.8 Diagram of the experimental system
3.2 Typical recirculating laboratory system 34
and photo of a tubular gas-phase
photoreactor (dismounted) 13 3.3 a) Photocatalytic detoxification pilot-plant scheme.
b) CPCs at Plataforma Solar de Almería.
2.1 The optical portion of the electromagnetic c) Double-skin sheet collector at Plataforma Solar
spectrum a), and a light wave front modelled de Almería. d) Two types of non-concentrating
as a straight line b) 19 solar collectors at the solar detoxification test
2.2 Visible-light colour distribution 19 stand at the University of Florida 35
2.3 a 3.4 Schematic of two pilot plant operation concepts:
World solar irradiance, MWh m–2 year–1 20 a once-through operation and a batch operation 37
2.3b Spectral solar radiation plotted from 0.2 3.5 Procedure used to calculate the photon flux
to 4.5 µm 20 inside a solar reactor 40
2.4 Air mass and solar components 21 3.6 A diagram of the various loss factors affecting
2.5 Ultraviolet spectra at the Earth’s the photon flux inside a photoreactor
surface (standard ASTM) 22 42
2.6 Solar spectra at the Earth’s surface 3.7 Typical photocatalytic degradation (rDCA)
between 300 and 1,100 nm 23 in a solar pilot plant under different UV
solar light intensities 44
List of Figures xiii

3.8 Photocatalytic degradation in a solar pilot plant. 5.2 HPLC-UV chromatograms of

Concentration is plotted as a function of photodegradation of oxamyl before
experiment time (top) and accumulated energy photocatalytic treatment a) and when the
(bottom). oxamyl has completely disappeared b) 66
Solar UV power throughout the experiment is 5.3 Evolution of H+ and Cl– during pentachlorophenol
also shown 44 degradation. To more clearly demonstrate that
reaction 5.4 is completed, the concentration of
4.1 Effect of the concentration of dissolved oxygen TOC in mM is calculated assuming 1 mMol TOC
on photocatalytic mineralization. [O2]0 = 8.5 mg =
L–1. In the inset, the usual effect of partial 6 mMol of C = 72 mg of C 67
pressure of oxygen on the photocatalytic reaction 5.4 Chemical structures of pyrimethanil and
rate is its degradation products obtained during
shown 51 a photocatalytic treatment with TiO2
4.2 Mean particle size of TiO 2 (P-25) suspended in 68
water versus pH [TiO2] = 0.2 g/L 52 5.5 Electron capture by a metal in contact
4.3 Different laboratory photoreactor designs and with a semiconductor surface 72
zones of radiation penetration when illuminated 5.6 Concentrating solar reactor with a
in different ways 53 platinum/titanium dioxide catalyst on ceramic
4.4 Influence of catalyst concentration on the saddles. Tested on air contaminated by spray
rate of photocatalysis (normalized rates paint at Fort Carson Army Base in Colorado,
have been USA 72
used to make it more easily understood) in 5.7 The excitation process in a
different reactors. In the inset, relative reaction semiconductor- semiconductor
rates of degradation (two different photoreactors) photocatalyst 73
and mineralization (TOC) of the same 5.8 Steps of excitation with a sensitizer in
contaminant the presence of an adsorbed organic
are shown 53 electron acceptor 73
4.5 Typical photocatalytic degradation. The insert 5.9 Degradation pathway proposed for
shows data adjusted to Equation 4.6 55 pirimiphos-methyl dissolved in water
4.6 Graphics related to the adjustment of data when illuminated in the presence of TiO2
to an L–H type kinetic model 55 75
4.7 Initial degradation rate as function of the initial
substrate concentration. The insert shows the 6.1 Non-concentrating solar collectors for
linear transformation of Equation 4.4, from domestic water-heating applications 83
which the rate constant and the adsorption 6.2 Medium concentrating solar collector:
coefficient can be estimated from the intercept recirculating parabolic trough reactor for water
and the slope, respectively 56 purification using titanium dioxide slurry at
4.8 The relation between the photocatalytic NREL 83
reaction rate and the intensity of the 6.3 A high-concentration solar collector: a fixed-
radiation received 57 focus solar reactor 83
4.9 Plots of normalized concentration as a function 6.4 Yearly efficiency of solar collectors in an
of accumulated energy for the ideal cloudless year: a PTC-one axis with
photodegradation and mineralization of different orientations 84
phenol. C0 = 20 mg/L, 6.5 Yearly efficiency of solar collectors in ideal
cloudless year: flat plate with different
TiO2 = 200 mg/L, pH0 = 5 59 inclinations 84
6.6 The first engineering-scale outdoor solar
5.1 General processes for the photo-oxidative or detoxification reactor using a one-axis parabolic
photo-reduction degradation of organic trough collector. Part of the 465 m 2 parabolic
compounds in aqueous solution sensitized by trough system at Sandia National Laboratory 85
semiconductor particles. Examples of photo- 6.7 A CIEMAT 384 m2 solar detoxification facility
oxidation (PCP) and photoreduction (CCl4) are using two-axis parabolic trough collectors, at
shown 65 Plataforma Solar de Almería 86
xiv List of Figures

6.8 Solar ray reflection on a one-axis parabolic the absorbance of the solution at 800 nm;
trough collector 87 no absorbance means absence of TiO2 100
6.9 Experimental set-up of a thin-film fixed-bed
reactor tested by ISFH at PSA installations. The 7.1 Complete mineralization of a complex mixture
housing is made of Plexiglas and the catalyst is of organic contaminants containing phenols
fixed on a flat glass plate 88 using TiO2 (200 mg/L) and persulfate as an
6.10 One-sun water treatment reactors with PTFE electron scavenger 108
tubes at NREL: early reactor a) and reactor 7.2 Formation of phenolic resins 109
with titanium dioxide immobilized on 7.3 Reactor condensation wastewater from
fibreglass bundles b) 88 the manufacture of phenolic resins
6.11 Solar reflection on a CPC collector 89 109
6.12 Obtention of CPC involute 90 7.4 Photocatalytic degradation of
6.13 View of CPC shape a) and CPC dichloromethane, chloroform,
photoreactor array b) 90 trichloroethylene and tetrachloroethylene
6.14 Holographic concentration of solar light 91 using a titania catalyst manufactured by
6.15 Solar detoxification pilot plant in Cologne 92 ENEL (Italy) 110
6.16 Flux distribution on an absorber in the course of 7.5 Degradation of PCP at a PSA solar
the solar day (6:00 to 18:00). Simulation of CPC detoxification facility (Helioman’s collectors’
behaviour with the following data: collector loop) 112
7.6 Cr+6 to Cr+3 solar photocatalytic reduction at a
orientation east–west; semi-acceptance angle 60°;
PSA solar detoxification facility (CPC’s collectors’
truncation angle 80°; absorber radius 13.6 mm; loop). See also Figure 7.7 112
optical gap 2 mm; concentration ratio 1.17 92 7.7 Simultaneous oxidation of phenol and
6.17 Specular, diffuse and spread reflection reduction of Cr+6 to Cr+3 using solar
from a surface 93 detoxification technology 113
6.18 Reflectivity of fresh metal coatings for mirrors 94 7.8 Degradation of metham sodium wastewater
6.19 Reflectivity of different aluminium and plastic from a tank-cleaning process. Catalyst
film surfaces 95 deactivation is observed after partial
6.20 Transmittance of different materials suitable for degradation of the initial TOC content 114
the manufacture of photoreactor tubes 96 7.9 Generic concept of contaminated
6.21 Glass-tube manufacture. Different compositions groundwater treatment 114
mean that the glass can be used for a wide variety 7.10 Partial view of the 156 m 2 parabolic-trough
of applications 97 water treatment system tested on
6.22 Influence of iron on borosilicate glass light contaminated groundwater at a site at LLNL
transmission (oxidative conditions). (United States) in 1992 115
Samples: flat glass 3-mm in thickness 7.11 One-sun reactor built by American Energy
97 Technology, Inc. for treating contaminated
6.23 Zone of tubular reactor where light penetrates if groundwater in Florida (United States) in 1992 115
the catalyst concentration is 1 g L–1 (TiO2 7.12 Non-concentrating solar detoxification
heterogeneous photocatalysis) 98 system for BTEX-contaminated
6.24 a) Experimental concentrating solar groundwater at a commercial site in
reactor using titanium dioxide Gainesville, Florida (United States), 1996
immobilized on glass wool for treating 116
contaminated air streams. 7.13 Simulation of contaminated landfill treatment
b) Parabolic trough reactor for water using solar detoxification: mineralization of
purification with immobilized lindane 117
titanium dioxide 99 7.14 Common ultraviolet band designations based
6.25 Sedimentation experiments at different pH. on biological effects 117
[TiO2] = 0,2 g /L; [NaCl] = 0 M. The Y axis 7.15 Scheme of TCE gas-phase mineralization
shows with PCO and a monolithic catalyst based
on sepiolite/TiO2/Pt 118
List of Figures xv

7.16 Destruction efficiencies of 400 ppm of TCE 9.1 Experimental mobile pilot plant for solar
using a MgSiO4/TiO2/Pt monolithic catalyst: water detoxification 153
thermocatalytic and photocatalytic processes at 9.2 Treating wastewater from the painting section at a
different temperatures 118 car assembly factory (ultrafiltration from
7.17 Flat-plate reactor for treating contaminated air cataphoresis process) 154
exiting from air-stripping units at McClelland 9.3 Catalyst mixing system for the LLNL water
Air Force Base, California, 1997 120 treatment system 155
7.18 Self-cleaning windows. Evolution of 9.4 Tonnes of pesticides used in the Almeria
photocatalytic treatment processes on glasses region (1995 data) 156
coated with titanium dioxide 121 9.5 a) Solar mineralization of TOC from the
insecticide abamectin; test performed in a
8.1 Conceptual scheme of photocatalytic + parabolic trough system; average direct UV light:
biological technologies coupling 127 38.1 watts m–2.
8.2 Elimination of tensioactive compounds b) Samples of photocatalytic degradation in
from agrochemical industrial wastewater CPC system; average global UV solar
after a few minutes of photocatalytic radiation was:
treatment 128 26.2 watts m–2 (acrinatrin test); 33.6 watts m–2
8.3 Photodegradation process: usual relationship (methamidophos) and 33.7 watts m–2
between the average oxidation state and time. (lufenuron test). TiO2 (Degussa P25): 200 mg
Points A and B are also related to Figure 8.4 129 L–1 157
8.4 Photodegradation process: normal relationships 9.6 TOC mineralization of a mixture of ten
between toxicity and time. Points A and B are selected pesticides (parabolic troughs); direct
also related with Figure 8.3 129 UV sunlight:
8.5 Ultraviolet spectra: solar (standard and PSA 36.3 watts/m2; TiO2 (Degussa P25) concentration:
measured spectrum) and two low-pressure 40 200 mg L–1; persulfate addition: 0.01 molar; treated
W mercury lamps (QUV fluorescent lamps: volume: 250 l. EC50 toxicity measured by
UVB-313 and UVA-340) 134 Microtox® 158
8.6 Comparative cost of UV photon 9.7 TOC mineralization of a mixture of ten
collection/generation with solar technology selected pesticides (CPCs). TiO 2 (Degussa
(CPCs) and electric lamps (electricity cost = P25) concentration: 200 mg L–1, slurry 158
€0.15/kWh), respectively. Data from tables 8.6 to 9.8 Pesticide degradation by Photo-Fenton
8.11. (Costs are indicated in process: comparison of different iron
euros, 1999) 139 concentrations for 100 ppm of pesticides in
8.7 Comparative cost of UV photon wastewater 159
collection/generation with solar technology 9.9 TOC degradation of a mixture of ten selected
(CPCs) and electric lamps (electricity cost = pesticides by TiO2 (Degussa P25) – persulfate
€0.05/kWh), respectively. Data from tables 8.6 to process. Degradation in function of UV
8.11. (Costs are indicated in collected energy (300 to 400 nm) 161
euros, 1999) 139 9.10 TOC degradation of a mixture of ten
8.8 Average direct and global UV irradiance (sunrise selected pesticides by Photo-Fenton
to sunset) at PSA (Almería, Spain). process and with different iron
Meteorological data records from 1991 to 1995 concentrations. Degradation in function of
139 UV collected energy
8.9 Average ‘cloud factor’ for global UV irradiance (300 to 400 nm) 161
(sunrise to sunset) at PSA (Almería, Spain). 9.11 Conceptual design of a solar
Meteorological data records from 1991 to 1995 140 detoxification plant for treatment and
8.10 Technology fit map: range of application of recycling of pesticide bottles 162
different water treatment technologies 142 9.12 Layout design of solar detoxification plant for
8.11 Optimization of solar detoxification and treatment and recycling of pesticide bottles 163
GAC technologies to PCP degradation 9.13 Laser device for manufacturing TiO2
146 catalyst powder 165
9.14 Modular CPC solar collector 165
xvi List of Figures

9.15 Manufacturing CPC reflectors 166 9.21 a) Installation of the main tanks in a solar
9.16 Solar detoxification plant layout 167 detoxification facility. From left to right:
9.17 Solar detoxification plant design, storage tank, catalyst separation tank, buffer
isometric drawing 167 tank.
9.18 Solar detoxification plant construction, b) Installation of tank fluid-level sensors.
showing lateral wall and pit to collect and c) Installation of catalyst separation system 169
contain any potential leaks 168 9.22 a) Installation and testing of the UV-A sensor
9.19 Installation of supporting structure device (front left). b) Testing the PLC and
and CPC units 168 electronic equipment 170
9.20 Photoreactor array input water manifold system 168 9.23 a) and b) Two views of the completed
solar detoxification treatment plant
List of Tables

1.1 Representative chemical reactions that can store 7.3 Examples of processes that are potentially
solar energy (thermochemical processes) 3 producers of antibiotic and antineoplastic
1.2 Some of the organic compounds included in wastewater 111
various lists of hazardous substances 7.4 Qualitative degradation rates of different
identified by the US EPA 4 compounds screened for photocatalytic activity:
1.3 Oxidation potentials of common substances slow (s), medium
and agents for pollution abatement. The more (m) and fast (f ) 119
positive the potential, the better the species is
as an oxidizing agent 5 8.1 Estimated cost of typical sizes of solar
1.4 Photocatalytic properties of detoxification facility (euros, 1999) 130
selected semiconductors 7 8.2 Estimated operating cost of a TiO 2-
1.5 VOCs amenable to treatment via persulfate solar detoxification plant (euros,
photocatalytic oxidation 12 1999) 132
3.1 Radiometric and spectroradiometric UV 8.3 Estimated annual treatment cost of a TiO2-
measurements at different times of the persulfate solar detoxification plant (euros,
1999) 132
day (n = 400 nm) 40
8.4 Estimated operating cost of a Photo-Fenton
3.2 Radiometric and spectroradiometric UV
solar detoxification plant (euros, 1999) 133
measurements at different times of the
8.5 Estimated annual treatment cost of a Photo-
day (n > 400 nm) 40
Fenton solar detoxification plant (euros, 1999)
4.1 Relative photonic efficiencies for the two 133
pesticides treated, with phenol as the standard 8.6 Estimated yearly cost of collecting 1.E+28
reference solar UV-photons at different yearly average
(C0 = 20 mg/L) 59 UV global
irradiation (Costs are indicated in euros, 1999.)135
5.1 Some examples of TiO2-sensitized 8.7 Estimated yearly cost of collecting 5.E+28
photodegradation of organic substrates 65 solar UV photons at different yearly average
7.1 Example of processes producing phenol residues UV global
in wastewater 109 irradiation (Costs are indicated in euros, 1999.)136
7.2 Examples of processes that are potentially 8.8 Estimated yearly cost of collecting 1.E+29
producers of wastewater containing solar UV photons at different yearly average
agrochemical residues 110 UV global irradiation (euros, 1999) 136
8.9 Estimated yearly cost collecting 5.E+29 solar
UV photons at different yearly average UV
global irradiation (euros, 1999) 136
8.10 Estimated yearly cost of collecting 1.E+30
solar UV photons at different yearly average
UV global irradiation (euros, 1999) 137
xviii List of Tables

8.11 Estimated yearly cost of UV photon generation 9.1 Selected pesticides for the feasibility
with electric lamps. Electricity cost = 0.15 study assessment, CIEMAT (Spain),
Euros/kWh; FCR = 17 per cent. In Table 8.11A 1996 156
the cost per lamp is €100; in Table 8.11B the cost 9.2 Degradation of pesticide mixture: mass
per lamp is €200. (euros, 1999) 138 treatment factor (Tfm) and volumetric treatment
8.12 Extraterrestrial solar irradiation 141 factor (Tfv) obtained for TiO2-persulfate and
Photo-Fenton processes, CIEMAT (Spain),
1998 159
Notes on the Editors

Julián Blanco Gálvez Sixto Malato Rodríguez

Julián Blanco Gálvez obtained a Diploma in Industrial Sixto Malato Rodríguez obtained a Diploma in
Engineering from Seville University and a Masters in Chemical Engineering from the Facultad de Ciencias of
Environmental Sciences from the International Open University of Granada, a Masters in Environmental
University. He has sixteen years of experience Sciences from the Instituto de Investigaciones
working in different industrial sectors and during 1989 Ecológicas, and a PhD in Chemical Engineering from the
also worked as a consultant for the Spanish University of Almería.
Normalization Institution AENOR. His professional career began in 1987 as Junior
Since 1990 he has worked at the Plataforma Solar de Researcher in the Chemical Engineering Department of
Almería (PSA) on R&D projects linked to solar and the University of Almería. In 1988, he joined the
environ- mental technology. From 1990 to 1993, he was Production Department in an oil refinery (REPSOL S.A.)
the project leader for the design and construction of the in Puertollano. Since 1990 he has worked at the
first European pilot plant for experimentation in solar Plataforma Solar de Almería. He participated in the
detoxification of industrial wastewaters. In 1993 he development of all solar photochemistry activities and
became the head of PSA’s Solar Chemistry department, all projects linked to the solar detoxification of water
which is devoted to the development of technology that took place in the early 1990s at the Plataforma
required for the utilization of solar radiation in Solar de Almería. At present, he has a permanent
photochemical processes. Since 1995 he has also headed position as Senior Researcher of the Spanish Ministry of
the CIEMAT Solar Chemistry department of PSA, Science and Technology.
directing staff and scientific installations in Almería and He has taken part in nine European Union Framework
Madrid on many national and international projects. In Programmes, nine National R&D Projects, and five
the same year, he was appointed Spanish National R&D Contracts related to the development of solar
Representative for the Task II group of International technologies for wastewater treatment. He has been
Energy Agency – SolarPACES (Solar Power and involved in the design and construction of all the
Chemical Energy Systems), with the objective of European pilot plants for experi- mentation in solar
worldwide coordination and knowledge sharing of detoxification of industrial wastewaters.
international solar research activities. He is the author and co-author of several books as well
He has given multiple lectures in courses, conferences as contributing chapters and articles to more than 50
and seminars, including EURESCO conferences and is publica- tions. He has co-authored 14 articles in
the co-author of 4 books, more than 30 international technical journals and 85 contributions to different
publica- tions and about 70 contributions to International International Congresses and Symposiums. He has
Congress and Symposiums. assisted in 21 Workshops and Conferences relating
to water treatment and has taught courses related to
advanced wastewater treatment.
Introduction 1 CHAPTER 1

Aims Objectives
This chapter describes an alternative By the end of this chapter you will
source of energy that combines sunlight understand the main factors that cause
and chemistry to produce chemical photochemical reactions and you will be
reactions. able to do five things:
It outlines the basic chemical and
physical phenomena involved in solar 1. Distinguish perfectly between
chemistry. thermochemical and
It reviews approaches that have been taken photochemical processes.
and progress that has been made, and gives
some projections for the short- and long- 2. Understand the impact of pollutants
term prospects for the commercialization
on the environment.
of solar photochemistry. It also introduces
the focus of this book: solar detoxification.
3. Calculate the energy flux of a light
source and its relationship with
semiconductor excitation.

4. Understand the basic principles

of advanced oxidation processes.

5. Describe the most important

features of heterogeneous
photocatalysis and their
application in the treatment of
contaminated aqueous effluents.
2 Solar Detoxification


Symbol Description Units
Aλ Absorbance at wavelength λ
AOPs Advanced Oxidation Processes
c Speed of light nm/s, m/s
ci Concentration of component i mol
EG Semiconductor band-gap energy eV, J
Eλ Spectral irradiance W m–2 nm–1
Eλo Spectral irradiances incident into the medium W m–2 nm–1
Eλ l Spectral irradiances at a distance l W m–2 nm–1
EC50 Concentration that produce an effect in 50% of a population mg/L,
GAC Granulated activated carbon
h Planck’s constant Js
LC50 Concentration that produces death in 50% of a population mg/L,
NOEL No observed effect level mg/kg/day
pi Partial pressure of component i atm
U Energy of a photon eV, J
αλ Absorption coefficient cm–1 atm–1
ελ Extinction coefficient mol–1 cm–1
8 Quantum yield
λ Wavelength nm, µm

1.1. Solar chemistry from the thermal energy obtained from the Sun for the
general purpose of substituting for fossil fuels.
• PHOTOCHEMICAL processes: solar photons are directly
The dramatic increases in the cost of oil beginning in 1974 absorbed by reactants and/or a catalyst, causing a
focused attention on the need to develop alternative sources reaction. Photo- chemical processes produce a chemical
of energy. It has long been recognized that the sunlight reaction from the energy of the Sun’s photons for the
falling on the Earth’s surface is more than adequate to general purpose of carrying out new processes.
supply all the energy that human activity requires. The
challenge is to collect this dilute and intermittent energy It should be emphasized, as a general principle, that the first
and to convert it to forms that are convenient and case is associated with processes that are feasible with con-
economical, and to use solar photons in place of those from ventional sources of energy. The second is related only to
lamps. It must be kept in mind that today there is clear completely new processes or reactions that are presently
worldwide consensus regarding the need to find a long- car- ried out with electric arc lamps, fluorescent lamps or
term replacement for fossil fuels (which were produced lasers.
millions of years ago and today are merely consumed) by From the outset, it was recognized that direct
identifying other inexhaustible or renewable energy conversion of light into chemical energy held promise for
sources. Under these circumstances, the growth and the production of fuels and chemical feedstock and the
development of solar chemical applications can be of storage of solar energy. Production of chemicals by
special relevance. These technologies can be divided in two reactions that are thermodynamically ‘uphill’ can
main groups (Figure 1.1): transform solar energy and store it in forms that can be
used in a variety of ways. Wide ranges of such chemical
• THERMOCHEMICAL processes: solar radiation is transformations have been pro- posed. A few
converted into thermal energy, which causes a chemical representative examples are given in Table 1.1 to
reaction. Thermochemical processes produce a chemical illustrate the concept.
Introduction 3

h < 700 nm
Methylene blue/O2

h < 390 nm
Heat Photons C6Cl5OH + 9/2O2 + H2 6CO2 + 5HCl
O TiO2

Increase of Modification of FIGURE 1.2. Furfural photo-oxidation and pentachloro- phenol

temperature chemical bonds mineralization (photochemical processes).
Thermochemical process Photochemical process
Steam reforming of methane Excitation of a semiconductor that give useful products. Photochemical reactions can be
CH4 + H2O  CO+ 3H2 – 206 kJ/mol h + SC  e– + p+ used to carry out a wide variety of chemical syntheses,
600º–850 ºC h  EG of SC
rang- ing from the simple to the complex. Processes of this
FIGURE 1.1. Schematic view of solar chemical applications. type may start with more complex compounds than fuel-
producing or energy-storage reactions and convert them
TABLE 1.1. Representative chemical reactions that can
to substances for which the photochemical step provides
store solar energy (thermochemical processes).
additional value or destroys harmful by-products. The
AH (kJ/mol) principles of photo- chemistry are well understood and
there is a wide range of known types of synthetic
CO2(g) → CO(g) + 1/2O2 286 transformations (Figure 1.2). Given this, the problem
CO2(g) + 2H2O(g) → CH3OH(l) + 3/2O2 727 becomes one of identifying applications in which the use
2 2 2 of solar photons is possible and economically feasible.
H O(l) → H (g) + 1/2O 286 The processes of interest here are photochemical;
CO2(g) + 2H2O(l) → 1/6C6H12O6(s) + O2 467 processes in which some component of the reacting
These processes generally start with substances in system is capable of absorbing photons in the solar
low- energy, highly-oxidized forms. The essential feature spectrum. As pho- tons should be treated like any
is that these reactions increase the energy content of the other chemical reagent in the process, their number is
chemicals using solar energy. For such processes to be a critical element in solar photochemistry (see Chapter
viable, they must fulfil the following requirements, as 2).
outlined by the National Renewable Energy Laboratory Because they are very attractive technologically and
(NREL) (1995) and slightly modified by the authors: envi- ronmentally, solar chemical processes have seen
• The thermochemical reaction must be endothermic. spectacular development in recent years. In the
beginning, research in solar chemistry was centred only
• The process must be cyclic and have no side reactions
on converting the solar energy into chemical energy,
that could degrade the photochemical reactants.
which could then be stored and transferred over long
• The reaction should use as much of the solar spectrum distances. But together with this impor- tant application,
as possible. other environmental uses have since been developed, so
• The back reaction should be very slow, to allow storage that today the entire range of known solar chemical
of the products, but rapid when triggered to recover applications has a promising future. In principle, any
the energy content. reaction or process requiring an energy source can utilize
• The products of the photochemical reaction should be solar energy.
easy to store and transport.
The other way that sunlight is used in photochemistry 1.2. Water contaminants
is to use solar photons as replacements for those from
artificial sources. The goal in this case is to provide a cost-
effective and energy-saving source of light to drive Environmental pollution is a pervasive problem with wide-
photochemical reactions spread ecological consequences. Recent decades have
witnessed increased contamination of the Earth’s drinking-
water reserves. The inventory of priority pollutants
4 Solar Detoxification

by the United States Environmental Protection Agency In any case, a consensus exists that the environmental
(US EPA) provides a convenient frame of reference for impact of a given contaminant depends on the degree of
under- standing the importance of removing such exposure (its dispersion and the resulting concentration
contamination from the Earth (Table 1.2). in the environment) and on its toxicological properties.

TABLE 1.2. Some of the organic compounds included in various lists of hazardous substances identified by the US EPA.
Acetaldehyde o-,m-,p- Maneb Poly-chlorinated
Cresols Mechlorethamine biphenyls (PCBs)
Acetone Cumene Melamine Pyrene
Acetonitrile Cyclohexane Methanol Quinone
2,4-Diaminoanisole Methoxychlor Quintozene
4,4’-Diaminodiphenyl ether Methyl acrylate Safrole
e Acrolein
Diazomethane Methyl Set-Butyl alcohol
Dibenzofuran isocyanate Sevin (carbaryl)
Acrylic acid
1,2-Dibromoethane Methyl tert-butyl Styrene
1,2- ether Methylene Terephthalic acid
Dichlorobenzene bromide 4,4’- Tert-Butyl
1,1-Dichloroethane Methylenedianiline alcohol
2,4-Dichlorophenol Methylhydrazine Tetrachlorvinpho
1,2- 4- s
o-Anisidine hydrochloride
Dichloropropane Methylphenol 1,1,2,2-Tetrachloroethane
Dichlorvos Mirex Tetrahydrofuran
Dicofol Mustard gas Thioacetamide
Diepoxybutane Nitrilotriacetic Thioure
Diethanolamine acid o- a
Dimethyl Nitroaniline Toluene
phthalate 2,4- Nitrobenzene 2,4-Toluene
Benzyl chloride
Dinitrophenol Nitrofen diamine Toluene
1,2- Nitrogen mustard diisocyanate
de Beta-Propoxur
Dinitrotoluene Nitroglycerin o-Toluidine
Biphenyl hydrochloride Total
Bis(2- 2,4- 5-Nitro-o-anisidine
Dinitrotoluene Nitrophenol xylenes
Chloroethoxy)methane Toxaphene
Bromoethane 1,4-Dioxane 2-
1,2-Diphenylhydrazine Nitropropane Triaziquon
Captan e
Carbary Disulfoton n-Butyl
Endosulfan alcohol Trichlorfon
l 2,4,6-Trichlorophenol
Carbon disulfide Epichlorohydrin n-Dioctyl phthalate
Ethylbenzene N- Trifluralin
Carbon 1,2,4-Trimethylbenzene
tetrachloride Ethylene glycol Nitrosodiethylamine
Ethylene thiourea N-Nitrosopiperidine 2,2,4-Trimethylpentane
Catechol 2,4,6-Trinitrotoluene
Chlordane Fluometuron N-Nitroso-N-
Formaldehyde ethylurea Urethane (ethyl
Chloroacetic acid carbamate) Vinyl
Chlorobenzene Hexachlorobenzen Octachloronaphthalen
e e Octane bromide
Chlorodibenzodioxin Vinyl chloride
s, Hexachloroethane Oxirane
Hexane Parathion Vinylidene chloride
various Xylene (mixed
Chlorodibenzofurans Hydroquinone (DNTP)
Isophorone Pentachlorobenzene isomers) Zineb
2-Chloroethyl vinyl
ether 2-Chlorophenol Isopropyl alcohol Pentachlorophenol
p-Chloro-m-cresol Lindane Phenanthrene
Malathion Phosgene
Phthalic anhydride
Introduction 5

The assessment of exposure involves an understanding of To solve this problem, apart from reducing emissions,
the dispersion of a chemical in the environment and estima- two main water-treatment strategies are followed: first,
tion of the predicted concentration to which organisms chemical treatment of drinking water and contaminated
will be exposed. For example, the pesticide fenaminphos surface and groundwater, and second, chemical treatment
oxidizes very quickly (a ten-day half-life) into sulfoxide and of waste- waters containing non-biodegradable
sulfone, while its pesticidal properties remain unaffected. A compounds.
half-life of seventy days has been found for degradation of Chemical treatment of polluted surface and
fenaminphos and its two metabolites. Furthermore, the two groundwater or wastewater is part of a long-term strategy
metabolites are more mobile (soluble) than fenaminphos to improve the quality of water by eliminating toxic
(Hayo and Van Der Werf, 1996). Assessment of the compounds of human origin before returning the water to
contam- inant’s effect involves summarizing data on the its natural cycles. This type of treatment is suitable when
effects the chemical has on selected representative a biological processing plant cannot be adapted to certain
organisms and using these data to predict a no-effect types of pollutants that did not exist when it was
concentration in a specific niche. designed. In such cases, a poten- tially useful approach is
Organisms may consume chemicals through ingestion to partially pre-treat the toxic waste using oxidation
of food and water, respiration or skin contact. When a technologies to produce intermediates that are more
chemical crosses the various barriers of the body, it readily biodegradable. Light can be used, under certain
reaches the meta- bolic tissue or a storage depot. The conditions, to encourage chemicals to break down the
toxicity of a chemical is usually expressed as the pollutants into harmless by-products. Light can have a
effective concentration or dose of the material that would dramatic effect on a molecule or solid because when it
produce a specific effect in 50 per cent of a large absorbs light its ability to lose or gain electrons is often
population of test species (EC 50 or ED50). If the effect altered. This electronically excited state is both a better
recorded is lethal, the term ‘LC50’ (or LD50) is used. The ‘no oxidizing and a better reducing agent than its counterpart
observed effect level’ (NOEL or NOEC) is the dose in the ground. Electron transfer processes involving
immediately below the lowest level eliciting any type of excited- state electrons and the contact medium (for
tox- icological response in the study. For example, the example water) can therefore generate highly reactive
pesticide methamidophos, which has been classified as a species like hydroxide ( •OH) and superoxide (O •–)
‘Restricted- Use Pesticide’ (RUP) by US EPA, is highly radicals (see Table 1.3). These can then be used to
toxic for mammals (acute oral LC50 = 16 mg/kg in rats decompose a pollutant chemically into harmless end-
and 30 to 50 mg/kg in guinea pigs), birds (bobwhite products. Alternatively, light can be used directly to
quail, CD50 = 8 to 11 mg/kg) and bees. The ninety-six- break up pollutant molecule bonds photolytically.
hour LC50 is 25 to 51 mg/L in rainbow trout, but
concentrations as low as 0.22 ng/L are lethal to larval
crustaceans in ninety-six-hour toxicity tests. A fifty-six- TABLE 1.3. Oxidation potentials of common substances and
day rat-feeding study resulted in a NOEL of agents for pollution abatement. The more positive the
0.03 mg/kg/day (Tomlin, 1997). potential, the better the species is as an oxidizing agent.
Decontamination of drinking water is mainly by
pro- cedures that combine flocculation, filtration, Oxidizing reagent Oxidation potential, V
sterilization and conservation, to which a limited number
of chemicals are added. Normal human sewage water can Fluorine 3.06
be efficiently treated in conventional biological •
Hydroxide radical ( OH) 2.80
processing plants. But very often these methods are
Ozone 2.07
unable to reduce the power of the contaminant. In these
cases, some form of advanced biolog- ical processing is Hydrogen peroxide 1.77
usually preferred in the treatment of effluents Chlorine dioxide 1.57
containing organic substances. Biological treatment Chlorine gas 1.36
techniques are well established and relatively cheap.
However, these methods are susceptible to toxic com- Oxygen 1.23
pounds that inactivate the waste-degrading Hypochlorite 2 0.94
microorganisms. Iodine 0.54
Superoxide radical (O•–) –0.33
6 Solar Detoxification

These processes are called ‘Advanced Oxidation Processes’ processes employ semiconductor slurries for catalysis,
(AOPs). Many oxidation processes, such as TiO 2/UV, whereas homogeneous photochemistry is used in a
H2O2/UV, Photo-Fenton and ozone processes (O 3, single- phase system. Any mechanistic description of a
O3/UV, O3/H2O2) are currently employed for this photo- reaction begins with the absorption of a photon,
purpose. sunlight being the source of photons in solar
photocatalysis. In the case of homogeneous
1.3. Photodegradation principles photocatalytic processes, the interac-
tion of a photon-absorbing species (transition metal com-
1.3.1. Definitions plexes, organic dyes or metalloporphyrines), a substrate
the contaminant) and light can lead to chemical modification
of the substrate. The photon-absorbing species (C) is acti-
For the benefit of those who may have a limited vated and accelerates the process by interacting through a
background state of excitation (C*). In the case of heterogeneous
in photochemistry, a brief outline of some basic concepts photo- catalysis, the interaction of a photon produces the
of photochemistry is presented here. In order for photo- appear- ance of electron/hole (e– and h+) pairs, the
chemistry to take place, photons of light must be catalyst being a semiconductor (e.g. TiO 2 or ZnO). In
absorbed. The energy of a photon is given by this case, the excited electrons are transferred to the
reducible specimen (Ox 1) while concurrently the catalyst
U (1.1) accepts electrons from the oxidizable specimen (Red 2)
λ that occupies the holes. In both directions, the net flow of
electrons is null and the catalyst remains unaltered.
where h is Planck’s constant (6.626 × 10–34 J.s), c is the
speed of light and λ is the wavelength. For a molecule’s
bond to be broken, U must be greater than the energy
that bond. Chν C* (1.4)
When a given wavelength λ of light enters a medium, C*  R  R*  (1.5)
its spectral irradiance Eλ (W m nm ) is attenuated
–2 –1
C* R*  P (1.6)
according to the Lambert-Beer law, which is expressed
in two ways, one for the gas phase and the other for the
liquid phase: Chν C(e   h  ) (1.7)
ln(Eλo / Eλl )  αλ pil gas phase (1.2) h  Red  Ox (1.8)
2 2
log(Eλo / Eλl )  ελ cil liquid phase e Ox1  (1.9)
Eo and El are the incident spectral irradiance and at a
λ λ
distance l into the medium, αλ is the absorption 1.3.2. Heterogeneous photocatalysis
coefficient (cm–1 atm–1), pi is the partial pressure (atm)
of component i, ελ is the extinction coefficient (M–1 cm– The concept of heterogeneous photocatalytic degradation
1), and c is the concentration (M) of component i. The
i is simple: a stable solid semiconductor is irradiated to
absorbance Aλ at wavelength λ is the product ελcil. The stimulate a reaction at the solid/solution interface. By
photochemical quantum yield (φ) is defined as the definition, the solid can be recovered unchanged after
number of molecules of the target compound that react many turnovers of the redox system. When a
divided by the number of photons of light absorbed by semiconductor is in contact with a liq- uid electrolyte
the compound in a fixed period of time. Normally, this solution containing a redox couple a charge transfer
unit is the maximum quantum yield attainable. occurs across the interface to balance the potentials of the
The term ‘photocatalysis’ implies the combination of two phases. An electric field is formed at the surface of the
photochemistry and catalysis. Both light and catalyst are semiconductor. The bands bend as the field forms from
necessary to achieve or to accelerate a chemical reaction. the mass of the semiconductor towards the interface.
Photocatalysis may be defined as the ‘acceleration of a pho- During pho- toexcitation (a photon with appropriate
toreaction by the presence of a catalyst’. Heterogeneous energy is absorbed), band bending provides the
conditions for carrier separation. In the case of
semiconductor particles, there is no ohmic
Introduction 7

h TABLE 1.4. Photocatalytic properties of selected semi-


Band gap energy Corresponding
Material required to wavelength
activate required (nm)
catalyst (eV)
BaTiO3 3.3 375
Red1 Oxid
CdO 2.1 590
CdS 2.5 497
Oxid1 Red2
CdSe 1.7 730
Fe2O3 2.2 565
GaAs 1.4 887
GaP 2.3 540
FIGURE 1.3. Behaviour of electrons and holes within a SnO2 3.9 318
particle of illuminated semiconductor in contact with an
electrolyte. SrTiO3 3.4 365
TiO2 3.0 390
contact to extract the majority carriers and to transfer them WO3 2.8 443
by an external conductor to a second electrode. This ZnO 3.2 390
means that the two charge carriers should react at the
ZnS 3.7 336
semiconductor/ electrolyte interface with the species in
solution. Under steady-state conditions, the amount of Summarizing, a semiconductor particle is an ideal
charge transferred to the electrolyte must be equal and photo- catalyst for a specific reaction if:
opposite for the two types of carriers. The • The products formed are highly specific.
semiconductor-mediated redox processes involve
electron transfer across the interface. When electron/ hole • The catalyst remains unaltered during the process.
pairs are generated in a semiconductor particle, the elec- • The formation of electron/hole pairs is required
tron moves away from the surface to the mass of the (gener- ated by the absorption of photons with energy
semi- conductor as the hole migrates towards the surface greater than that necessary to move an electron from
(see Figure 1.3). If these charge carriers are separated the valence band to the conduction band).
quickly enough they can be used for chemical reactions at • Photon energy is not stored in the final products,
the surface of the photo- catalyst: that is, for the oxidation being an exothermic reaction that is only retarded
or reduction of pollutants. kinetically.
Metal oxides and sulfides represent a large class of
semi- conductor materials suitable for photocatalytic 1.3.3. Homogeneous photodegradation
Table 1.4 lists a selection of semiconductor materials that
have been used for photocatalytic reactions, together
with the band-gap energy required to activate the The use of homogeneous photodegradation (a single-
catalyst. The final column in the table indicates the phase system) to treat contaminated water dates back
wavelength of radia- tion required to activate the to the early 1970s. The first applications concerned
catalysts. According to Planck’s the use of UV/ozone and UV/H O . The use of UV
equation, the radiation required to produce this gap must be light for the 2 2

of a wavelength (λ) equal to or less than that calculated photodegradation of pollutants can be classified into two
by equation 1.10. principal areas:
hc • PHOTO-OXIDATION: light-driven oxidative processes
λ (1.10) princi- pally initiated by hydroxyl radicals.
EG • Direct PHOTODEGRADATION: light-driven processes in
where EG is the semiconductor band-gap energy, h is which degradation proceeds following direct excitation
Planck’s constant and c is the speed of light. of the pollutant by UV light.
8 Solar Detoxification

Photo-oxidation involves the use of UV light plus an Since the net result of ozone photolysis is the
oxidant to generate radicals. The hydroxyl radicals conversion of ozone into hydrogen peroxide, UV-ozone
then attack the organic pollutants to initiate oxidation. would appear to be only a rather expensive method of
Three major oxidants are used: hydrogen peroxide making hydrogen per- oxide. However, there are other
(H2O2), ozone and Photo-Fenton reaction. H2 O2 oxidation-related processes that occur in solution, such
absorbs photons fairly weakly in the UV region, with as the direct reaction of ozone with a pollutant (see Table
absorption increasing as the wavelength decreases. At 1.3). Ozone may have advantages in water, with high
254 nm, ελ is 18 M–1 cm–1, whereas at 200 nm it is 190 inherent UV absorbance, but it faces the same problem as
M–1 cm–1. The primary process for the absorption of light hydrogen peroxide as far as its use in solar- energy
below 365 nm is dissociation, to yield two hydroxyl processes is concerned.
radicals: The essential process of the Fenton reaction
(Safarzadeh- Amiri et al., 1996) is the same as for all
AOPs. Highly
2 2
H 2O hν (1.11) reactive radicals (such as HO • and HO2•) oxidize nearly all

OH organic substances to yield CO 2, water and inorganic salts.
The use of hydrogen peroxide is now very common In the case of Photo-Fenton, Fe2+ ions are oxidized by
for H2O2 producing one •OH ion (1.18), and the Fe 3+ or
the treatment of contaminated water due to several complexes obtained then act as the light-absorbing
practical advantages: species producing
• H2O2 is available as an easily handled solution that another radical while the initial Fe 2+ is recovered (1.19 and
can be diluted in water to give a wide range of 1.20) 
concentrations. . Fe2  H2O2  Fe3  OH  (1.18)
• There are no air emissions. OH
• A high-quantum yield of hydroxyl radicals is (1.19)
generated (0.5). Fe3  H2O hν  Fe2  H  OH
[Fe(OOCR)]2  hν  Fe2  CO2  R
The major drawback is the low molar-extinction Note that in Equation 1.20, the ligand R-COO– can
coefficient, which means that in water, when UV be replaced by other organic groups (e.g. ROH or
absorption is high, the fraction of light absorbed by H 2O2 RNH2). Compared with other homogeneous photo-
may be low unless very large concentrations are used. oxidation processes, the advantage of Photo-Fenton is
Furthermore, especially as concerns the focus of this the improved light sensitivity (up to a wavelength of
text, H2O2 photon absorption is very low in the Solar UV 600 nm, correspon- ding to 35 per cent of the solar
range (up to 300 nm). radiation). On the other hand, disadvantages, such as
Ozone is generated as a gas in air or oxygen in the low pH values required (usually below pH 4) and
concen- trations generally ranging from 1 to 8 per cent the necessity of removing iron after the reaction, remain.
(v/v). It has a strong absorption band centred at 260 Some pollutants are able to dissociate only in the
nm with ελ = 3,000 M–1 cm–1. Absorption of light at pres- ence of UV light. For this to happen, the
this wavelength leads to the formation of H2O2: pollutant must
absorb light emitted by a lamp (or the Sun) and have a
O3 
hν O( 1 D) 
(1.12) reasonable quantum yield of photodissociation.
Organic pollutants absorb light over a wide range of
O O(1 D)  H2O  H2O2 (1.13) wavelengths, but generally absorb light more strongly at
lower wavelengths,
Hydroxyl radicals are then formed by the reaction of especially below 250 nm (Figure 1.4). In addition, the
ozone quantum yield of photodissociation tends to increase
with the hydrogen peroxide conjugate base: lower wavelengths, since the photon energy is increasing
H2O2  H2O  HO2  (1.14 (Eq. 1.1). The net chemical result of photodissociation is
H3O  ) usually oxidation, since the free radicals generated can
react with dissolved oxygen in the water. In practice, the
HO2  O3 O3  HO2 (1.15
range of wastewaters that can be successfully treated by
O  H O  HO  UV alone is very limited. This defect is more relevant
3 2 3 (1.16 when solar energy
OH )
HO3  OH  O2 (1.17
Introduction 9

h  3.0 eV
Arbritary units

O2 e–

TiO2 Particle
UV solar spectrum
h+ H2O
UV acrinathrin spectrum

200 220 240 260 280 300 320 340 360 380 400
Wavelength, nm OH+H +

FIGURE 1.4. UV spectra of acrinathrin and of sunlight Figure 1.5. Effect of UV radiation on a TiO 2
between 200 and 400 nm. particle dispersed in water.

is used (see Figure 1.4) because only photons up to 300 Whenever different semiconductor materials have been
nm are available. tested under comparable conditions for degradation of the
same compounds, TiO2 has generally been demonstrated to
1.4. Application to water be the most active, equalled only by ZnO. TiO 2’s
treatment strong resistance to chemical and photocorrosion, as well
as its safety and low cost, limits the choice of convenient
alterna- tives (Herrmann, 1999). Furthermore, TiO 2 is of
As mentioned above, UV light can be used in several special interest as it can use natural (solar) UV light. This is
ways. However, direct photolysis can occur only when because it has an appropriate energetic separation between
the contam- inant to be destroyed absorbs incident light its valence and conduction bands, which can be surpassed
efficiently. In the case of UV/ozone and UV/hydrogen by the energy content of a solar photon (see Table 1.4).
peroxide this does not happen. Absorption by some Other semi- conductor particles, for example, CdS or GaP,
sensitizer must initiate the reaction, and limited absorb larger fractions of the solar spectrum and can form
absorption by the solute or the additive restricts chemically activated surface-bond intermediates, but
efficiency. Furthermore, these mixtures often still require unfortunately these photocatalysts are degraded during
large quantities of added oxidant. By contrast, in het- the repeated catalytic cycles involved in heterogeneous
erogeneous photocatalysis dispersed solid particles photocatalysis. Therefore, degradation of the organic
absorb larger fractions of the UV spectrum efficiently pollutants present in wastewater using irradiated TiO 2
and generate chemical oxidants in situ from dissolved suspensions is the most promising process, and R&D in this
oxygen or water (see Figure 1.5). These advantages make field has grown very quickly during the last few years.
heterogeneous photo- catalysis a particularly attractive
method for environmental detoxification. The most
important features of this process that make it applicable
to the treatment of contaminated
aqueous effluents are: TiO2 hν e   h   TiO
2 (1.21)
• The process takes place at ambient temperature. e  h  TiO2  TiO2  heat and/or
 
• Oxidation of the substances into CO 2 is hν’
complete. (TiOIV  O2  TiIV )  OH2  h+
• The oxygen necessary for the reaction is obtained  (TiOIV  O2  TiI )  OH  H
from V

the atmosphere.  

O2(ads)   (1.24)
• The catalyst is cheap, innocuous and 2(ads)
 eB O
reusable. C
• The catalyst can be attached to different types of inert
To date, evidence supports the idea that the hydroxyl
radical (•OH) is the main oxidizing specimen responsible
For all these reasons, from now on only this method will for photo- oxidation of the majority of the organic
be dealt with in this text. compounds studied. The first effect, after absorption of
near ultraviolet radiation,
10 Solar Detoxification

λ < 390 nm, is the generation of electron/hole pairs, • Pesticides: contaminate waters with a high level of
which are separated between the conduction and valence agricul- tural runoff. Among the recently investigated
bands (Eq. 1.21). In order to avoid recombination of the compounds are triazines, organophosphorous,
pairs generated (Eq. 1.22), if the solvent is carbamates, phenoxy- acids and organochlorines.
oxidoreductively active (i.e. water) it also acts as a • SURFACTANTS: surface active agents are entering
donor and acceptor of electrons. Thus, on a hydrated domestic and industrial wastewater in increasing
and hydroxylated TiO 2 surface, the holes trap •OH amounts. Because their biodegradability may be one of
radicals linked to the surface (Eq. 1.23). In any case, it the more important constraints to their use,
should be emphasized that even trapped electrons and photocatalytic degradation has received increasing
holes can rapidly recombine on the surface of a particle attention.
(Eq. 1.22). This can be partially avoided through the • Dyes: strongly coloured compounds can be
capture of the electron by pre-adsorbed molecu- lar removed by adsorption but it is always better to
oxygen, forming a superoxide radical (Eq. 1.24). destroy them by oxidation.
Whatever process the photo-oxidation takes, O2 and
water are essential for photo-oxidation with TiO2. There Four exhaustive reviews by Blake (1994, 1995, 1997,
is no degradation in the absence of either. Furthermore, 1999) describe almost 3,000 studies in this field carried
the oxidative species formed (in particular the hydroxyl out before 1999.
radicals) react with the majority of organic substances. Despite encouraging laboratory-scale data and some
For example, in aromatic compounds, the aromatic part industrial-scale tests, chemical oxidation detoxification is
is hydroxylated, then successive steps in still restricted to a few experimental plants (Dillert et al.,
oxidation/addition lead to breaking of the rings. The 1999). The broader application of those technologies
resulting aldehydes and carboxylic acids are requires:
decarboxylated and finally produce CO2. However, the • reactor optimization and modelling, and
important issue governing the efficiency of • assessment of the efficiency of oxidation technology
photocatalytic oxidative degradation is to minimize to reduce the toxicity of effluents.
electron-hole recombi- nation by maximizing the rate of
interfacial electron transfer to capture the The following chapters of this book will attempt to
photogenerated electron and/or hole. This issue is highlight these matters.
discussed in more detail later.
Photocatalytic degradation has been
commonly investigated with monoaromatics, and
1.5. Gas-phase detoxification
consequently these pollutants appear as model
compounds in dozens of scien- tific papers. Some Airborne pollutants (such as volatile compounds) can be
monoaromatics investigated include: benzene, treated during the gas phase with the UV/TiO 2 process.
dimethoxybenzenes, halobenzenes, nitrobenzene, Gas- phase treatment offers several advantages. In
chlorophenols, nitrophenols, benzamide and aniline, general, substrate mass-transport is an order of
most of which are recognized as priority pollutants (see magnitude faster in the gas phase than in the liquid
Table 1.2). In addition to these, several other types of phase. This in turn leads to much faster reaction rates.
molecules have also been investigated as substrates for Oxidant starvation may be less of a problem in the gas-
photocatalytic degradation: phase. There is also no interference on the photocatalytic
surface from other species that are invari- ably present in
• HALOALIPHATICS (e.g. trichloroethylene or TETRACHLOROMETHANE): aqueous treatment media (for example, anions). In
important because so many of these compounds have addition, photocatalysis separation after use is not a
been released into the environment and contaminate problem, unlike aqueous slurry suspensions. As solar
waters. Some also originate during water treatment by energy is used to drive the process, no fuel is required,
chlorination. gaseous affluent volume is reduced, no NO x is generated,
• WATER-MISCIBLE solvents (e.g. ETHANOL or ALKOXYETHANOL): no products of incomplete combustion are produced,
these compounds are very difficult to detoxify, as they CO2 emanating from fuel burning is avoided, and
are resistant to treatment and are poorly adsorbed substantial fuel saving may be achieved. Since no
on granulated activated carbon (GAC). burning takes place,
Introduction 11










FIGURE 1.6. Diagram of an experimental system with a cross-section view of a gas-phase monolithic photoreactor.
oxygen is only necessary at a stoichiometric ratio. Solar hydroxylation on the oxide surface. There are also
con- centrators enable small-size solar furnaces and even indications that product (or intermediate) adsorption on
mobile solar parabolic dishes to be sufficient for on-site the TiO2 surface may be problematic during the course of
destruction of low productions of highly toxic the reaction.
compounds. Pollutant substrates such as trichloroethylene,
On the other hand, there are indications that acetone, formaldehyde, m-xylene and NOx have been
mineraliza- tion may not be complete with some treated with TiO2/UV in the gas-phase in bench-scale
organic substrates in the gas-phase. The TiO2 tests. Field tests have also been conducted to treat effluent
photocatalyst becomes less effective after prolonged use air emissions using this technique at different
and must be reactivated with moist air, which manufacturing plants in the United States (Rajeshwar,
presumably restores the original degree of 1996).
12 Solar Detoxification

In figures 1.6, 1.7 and 1.8, different gas-phase photo- from the front part of the reactor. A 4,000 W Xenon
catalytic reactors are shown. Figure 1.6 presents a reactor lamp, which supplies the photonic flux, has been
designed for use with monolithic catalysts (Avila et adapted for this reactor. A Pyrex window is used to filter
al., 1998). The reactor (40 mm inner diameter, 200 mm wavelengths below 300 nm (see Figure 1.4), a filter is
length) permits a change in the direction of gas flow used to remove IR radia- tion, and an electric heating
from the rear to the front and vice versa. The monolythic jacket used to control tempera- ture. A series of
catalysts are placed inside the reactor so that light thermocouples are installed to monitor the temperature
illuminates the internal sur- faces of the channels from of the reactor inlet and outlet and, at various points
the front. Gases are usually fed inside different channels of the monolith, to obtain
axial and radial temperature profiles. A reactant gas
consist- ing of air with a flow of several litres per
TABLE 1.5. VOCs amenable to treatment via photocatalytic minute is mixed with the VOC (hundreds of ppm)
oxidation ( Jacoby et al., 1996). before entering the reac- tor. Contaminants are
pumped through the monolith and outlet gases, heat
Class of compound Chemicals tested traced to avoid condensation, are sub- sequently
analysed by direct sampling (on-line) using a gas
Aromatics Benzene, toluene
chromatograph (GC) with a flame ionization detector
Nitrogen-containing ring Pyridine, picoline, nicotine (FID). Manual (off-line) samples that can be removed
Aldehydes Acetaldehyde, through a sampling valve are Tenax adsorbent to trap
formaldehyde volatile compounds from the air. In addition, CO2
Ketones Acetone concentration is measured with an infrared gas
Alcohols Methanol, ethanol, propanol analyser, to calculate the carbon-mass balance.
Ethylene, propene, tetramethyl In figures 1.7 and 1.8 a different type of gas-
Alkane ethylene phase photoreactor is shown. The tubular photoreactor
sTerpenes α-pinene consists of a Pyrex tube (Figure 1.7) containing a
Sulfur-containing organics Methyl thiophene tubular matrix of TiO2-based catalyst surrounding a
Dichloroethylene, UV lamp. The stream
Chlorinated tetrachloroetylene of gas flows between the inner wall of the catalyst and the
ethylenes Dichloroacetyl chloride, surface of the UV lamp. The UV lamp irradiance is
tetrachloroacetyl similar to the solar spectrum. A heat-trace box
Acetyl chlorides chloride (Figure 1.8) that controls process temperature
encloses the photoreactor.

Catalytic monolith

Outlet gases Inlet gases

Thermocouple Fluorescent lamp

FIGURE 1.7 Schematics and photo of a tubular gas-phase photoreactor.

Introduction 13

Secondary pollutant
H2 O

Peak H2 Air He


Vent TCD
GC (HP6890)
C Analizador

Relief valve Tenax

Vent Venteo

FIGURE 1.8 Diagram of the experimental system and photo of a tubular gas-phase photoreactor (dismounted).

Two thermocouples, located at the inlet and the outlet,

give the reaction temperature of gases in contact with the Summary
catalyst surface. In this case a stream of dry CO 2-free air
is delivered from an air generator at flow rates of several A description is given of how solar chemistry could
litres per minute. The VOC is bubbled into the airstream become a significant segment of the chemical
and through the catalyst in the reactor. Outlet gases are industry and how it can be used, under certain
analysed by on-line GC and manual samples are Tenax conditions, to provoke the chemical breakdown of
adsorbent. pollutants into harmless by- products. The behaviour of
contaminants in environmental
14 Solar Detoxification

water is summarized. The basic concepts of HAYO, M.G. AND VAN DER WERF, H. 1996. Assessing the
photochemistry relating to photolysis of chemical bonds, impact of pesticides on the environment. AGRICULTURAL
homogeneous photodegradation and heterogeneous Ecosystems AND Environment, No. 60, pp. 81–96.
photocatalysis are reviewed. The use of semiconductors
for wastewater treat- ment, with particular reference to HERRMANN, J.M. 1999. Heterogeneous photocatalysis:
TiO2, is discussed. Examples of the waste materials that fundamentals and applications to the removal of
have been treated suc- cessfully using TiO2 are various types of aqueous pollutants. CATALYSIS TODAY,
presented. Gas-phase photocatalysis is also introduced. No. 53, pp. 115–29.


Bibliography S.K. 1996.
Heterogeneous photocatalysis for control of volatile
organic compounds in indoor air. JOURNAL of the Air WASTE
B.; CARDONA, A.I. AND ROMERO, M. 1998. Gas-
photo-assisted mineralization of volatile compounds by NATIONAL RENEWABLE ENERGY LABORATORY. 1995. SOLAR
monolithic titania catalysts. In: Applied CATALYSIS B: Photochemistry – Twenty YEARS of Progress: WHAT’S Been
ENVIRONMENTAL, No. 17, pp. 75–88. Accomplished, AND Where Does It LEAD? Golden, Colo.,
NREL Report NREL/TP-433-7209.
BLAKE, D.M. 1994. BIBLIOGRAPHY of Work on the
RAJESHWAR, K. 1996. Photochemical strategies for abating
PHOTOCATALYTIC REMOVAL of HAZARDOUS Compounds from environmental pollution. Chemistry AND Industry, No.
WATER AND Air, May. Springfield, Va., 22161, 17, pp. 454–8.
National Technical Information Service, US Depart. of
Commerce. UPDATE Number 1: To June 1995, October 1995. SAFARZADEH-AMIRI, A.; BOLTON, J.R. AND CATER, S.R. 1996.
UPDATE Number 2: To October 1996, January 1997. UPDATE The use of iron in advanced oxidation processes. JOURNAL of
Number 3: To JANUARY 1999, May 1999. ADVANCED OXIDATION Technology, Vol. 1, No. 1, pp. 18–26.

TOMLIN, C.D.S. 1997. The Pesticide MANUAL. 11th ed.

DILLERT, R.; CASSANO, A.E.; GOSLICH, R. AND BAHNEMANN, D. Farnham, Surrey, United Kingdom, British Crop
1999. Large (1) scale studies in solar catalytic wastewater Protection Council.
treatment. CATALYSIS TODAY, No. 54, pp. 267–82.
Introduction 15

Self-assessment questions Part B

Part A: True or false? 1. What is the most important difference

between thermochemical and photochemical
solar processes?
1. Solar energy is useful only to substitute for fossil
fuels; it is converted into thermal energy, thus provoking 2. What is the usual way of expressing the toxicity
chemical reactions. of a chemical in the environment?
2. The toxicity of a chemical is the same for 3. Why is biodegradation (a major mechanism in
all species. wastewater treatment) quite inefficient in treating
certain types of wastewater?
3. Biological treatment techniques are the
cheapest wastewater treatment methods. 4. What is the percentage of absorbed photons in a
solution with the following characteristics:
4. The energy of a photon depends on the extinction coefficient = 1327 cm–1 M–1,
ambient temperature. concentration of substrate
0.01 M, illuminated pathlength = 5.6 cm, with an extinction
5. Heterogeneous photocatalysis employs coefficient of 0.3?
liquid catalysts.
5. What wavelength is required to excite a
6. Light-driven oxidative processes are initiated by semiconductor with a band-gap of 4.0 eV?
excited electrons on the catalyst surface.
6. Name three important characteristics of
7. Ozone can be produced from air. heterogeneous photocatalysis when used in a water-
treatment process.
8. The most important characteristics of photocatalysts
are: resiliance against chemical and photocorrosion,
7. Why is TiO2 the most suitable photocatalyst
safety, cost and band-gap.
for wastewater treatment?
9. The electron/hole recombination can be avoided
8. What is the most important electron acceptor in water?
by increasing reaction temperature.
9. What is the most important product of
10. Heterogeneous photocatalysis can only be applied photocatalytic degradation of organic contaminants?
to monoaromatics.
10. Why do hydroxyl radicals react with
organic substances?
16 Solar Detoxification

Answers 3. Because in the case of very toxic compounds,

microorganisms need an extended period of
adaptation, during which they are not completely
Part A non-viable.

4. 100 per cent and 3.8 per cent.

1) False. 2) False. 3) True. 4) False. 5) False. 6)
False. 5. λ Š 310 nm.
7) True. 8) True. 9) False. 10) False.
6. The process takes place at ambient temperature, the
Part B
oxygen necessary for the reaction is obtained from the
atmosphere, and the catalyst is cheap, innocuous and
1. In thermochemical processes, solar radiation is can be reused.
converted into thermal energy; in photochemical
processes the solar photons are absorbed directly by the 7. It has exhibited the highest activity. It is highly
reactants, giving rise to the reaction. stable to chemical and photocorrosion. It can use natural
2. The toxicity of a chemical is usually expressed as the
effective concentration or dose of the material that 8. Dissolved oxygen.
would produce a specific effect in 50 per cent of a large
population of test species (EC 50 or ED50). 9. Carbon dioxide.

10. Because of its very high oxidation potential.

Introduction 17


Solar Irradiation

Aims Objectives
This chapter describes the power of light At the end of this chapter you will
as a source of energy. It outlines the understand the main factors affecting
basic solar radiation behaviour and you will
principles that are related to the light be
spectrum and specifically to the solar
spectrum. able to do six things:
It discusses solar UV radiation and its
photon flux in more detail, because this part 1. Discriminate between the
of the solar spectrum is the most important different components of solar
for driving chemical processes. The major radiation and their principal
atmospheric variables determining the characteristics.
amount of UV solar radiation on the Earth’s
surface are discussed. A method for 2. Recognize typical solar spectra and
calculating UV attenuation at a given site is understand the effect of Sun position on the
presented. Finally, solar radiation solar power reaching the Earth’s surface.
measurement systems are described.
3. Find the photon flux of a
polychromatic source of energy with
simple calculations.

4. Describe the most important

components of the Earth’s atmosphere and
their effect on power and spectral
distribution of the solar radiation.

5. Understand the procedures that

permit solar power to be calculated from
available radiation at any given site.

6. Comprehend the basic principles on

which solar radiation measurement is
18 Solar Detoxification


Symbol Description Units

AM Air mass ratio

fn Cloud factor
fλ Fraction of power associated with a wavelength nm–1

H Radiance exposure monthly average kJ m–2
HTBD TBDUV radiance exposure kJ m–2
I Photon flux density Einstein s–1 m–2
NA Quantity of photons absorbed by the system Photons s–1
N0 Avogadro’s number, 6.023 × 10 23
Photons mol–1
Nλ Number of photons supplied by a source of light of wavelength λ Photons s–1
Qλ Energy of a monochromatic source of light of wavelength λ W m–2 µm–1
T Transmittance
Tλ Transmittance of direct-bean solar radiation under cloudless skies at a specific wavelength
TA,λ Transmittance related to absorption and dispersion by aerosols
Tg,λ Transmittance resulting from absorption of atmospheric gases
To,λ Transmittance related to the effect of the ozone layer
TR,λ Transmittance related to the molecules of air
Tv,λ Transmittance resulting from absorption by steam
TBDUV Typical ‘best day’. Completely clear sky during all the hours of sunlight
U Energy of one photon eV, J
UVD Direct ultraviolet light W m–2
UVG Global ultraviolet light W m–2
UVλ Ultraviolet irradiance associated with a wavelength W m–2 nm–1
8 Quantum yield
λ Wavelength nm, µm

2.1. The power of light Like all electromagnetic waves, light waves can interfere
with each other, become directionally polarized, and bend
slightly when passing through an edge. These properties
Light is just one of various electromagnetic waves allow light to be filtered by wavelength or amplified coher-
present in space. The electromagnetic spectrum covers an ently as in a laser. In radiometry, light’s propagating wave
extremely broad range, from radio wavelengths of a front is modelled as a ray travelling in a straight line (See
metre or more, down to x-rays with wavelengths of less Figure 2.1b). Lenses and mirrors redirect these rays along
than one billionth of a metre. Optical radiation lies predictable paths. Wave effects are insignificant in a large-
between radio waves and x-rays on that spectrum and scale optical system, because the light waves are randomly
has a unique combination of ray, wave and quantum distributed and there are plenty of photons.
properties. At x-ray and shorter wavelengths,
electromagnetic radiation tends to be quite particle-like
in its behaviour, whereas toward the long wave- length 2.1.1. Ultraviolet light
end of the spectrum behaviour is mostly wavelike. The
UV-visible portion occupies an intermediate position, Short wavelength UV-light exhibits more quantum
having both wave and particle properties in varying proper- ties than its visible or infrared counterparts.
degrees (See Figure 2.1a). Ultraviolet light
Solar Irradiation 19

Blue Green Yellow Red

400 500 600 700 nm

FIGURE 2.2. Visible-light colour distribution.

FIGURE 2.1. The optical portion of the electromagnetic about 10 photons per second at 555 nm; this corresponds to
spec- trum a), and a light wave front modelled as a straight a radiant power of 3.58 × 10–18 W (or J s–1). Similarly, the
line b). eye can detect a minimum flux of 214 and 126 photons per
second at 450 nm and 650 nm, respectively.
is arbitrarily broken down into three bands, according to 2.1.3. Infrared light
its effects. UV-A (315 to 400 nm), which is the least
harmful type of UV light, because it has the least
energy (recall Eq. 1.1), is often called ‘black light’, and Infrared light contains the least amount of energy per
is used for its relative harmlessness and its ability to photon of any band and is unique in that it has primarily
cause fluorescent materials to emit visible light – thus wave prop- erties. This can make it much more difficult
appearing to glow in the dark. UV-B (280 to 315 nm) is to manipulate than ultraviolet and visible light. Infrared
typically the most destructive form of UV light, because is more difficult to focus or refract with lenses, diffracts
it has enough energy to damage biological tissues, yet more, and is difficult to diffuse. Since infrared light is a
not quite enough to be completely absorbed by the form of heat, infrared detec- tors are sensitive to
atmosphere. UV-B is known to cause skin cancer. Since environmental changes such as a person moving in the
the atmosphere blocks most of the extraterrestrial UV-B field of view. Night vision equipment takes advantage of
light, a small change in the ozone layer could this effect, amplifying infrared to distinguish people and
dramatically increase the danger of skin cancer. UV-C machinery that are concealed in the darkness.
(100 to 280 nm) is almost completely absorbed in air
within a few hundred metres. When UV-C photons
collide with oxygen atoms, the energy exchange causes
2.2. The solar spectrum
the formation of ozone. UV-C is never observed in
nature, however, since it is absorbed so quickly. The total energy coming from that huge reactor, the Sun,
Germicidal UV-C lamps are often used to purify water from which the Earth receives 1.7 × 1014 kW, amounts to
because of their capability to kill bacteria. 1.5 × 1018 kWh per year, or approximately 28,000 times
world consumption (Figure 2.3a). Radiation beyond the
2.1.2. Visible light atmosphere has a wavelength of between 0.2 µm and 50
µm, which is reduced to between 0.3 µm and 3.0 µm when
reaching the Earth’s surface, because of the absorption of
Visible light is concerned with the radiation perceived by
part of it by different atmospheric components (ozone, oxy-
the human eye. The lumen (lm) is the photometric
gen, carbon dioxide, aerosols, steam and clouds). The solar
equivalent of the watt, weighted to match the eye response
radiation that reaches the ground without being absorbed or
of the ‘standard observer’. Yellowish-green light receives
scattered is called ‘direct radiation’; radiation that reaches the
the greatest weight because it stimulates the eye more
ground but has been dispersed is called ‘diffuse radiation’; and
than blue or red light of equal radiometric power (1 W at
555 nm = 683.0 lumens). To put this into perspective: the
human eye can detect a flux of
20 Solar Detoxification

FIGURE 2.3a. World solar irradiance, MWh m–2 year–1.

2,000 2,000
Irradiance, Wm2 µm1

1,500 Extraterrestrial 1,500

Direct air mass 1.5
Global 37° air mass 1.5
1,000 1,000

500 500

0 0
0.6 1.2 1.8 2.4 3.0 3.6 4.2
Wavelength, µm

FIGURE 2.3b. Spectral solar radiation plotted from 0.2 to 4.5 µm.

the sum of both is called ‘global radiation’. In other words, it Figure 2.3b shows the standard solar radiation spectra
is the direct radiation that produces shadow when an opaque (Hulstrom et al., 1985) at ground level on a clear day.
object blocks it; diffuse radiation does not. In general, the The dotted line corresponds to the extraterrestrial
direct component of global radiation on cloudy days is min- radiation in the same wavelength interval. When this
imum and the diffuse component is maximum, and the radiation enters the atmosphere, it is absorbed and
opposite is true on clear days. scattered by atmospheric components such as air
molecules, aerosols, water vapour,
Solar Irradiation 21

UV radiation (direct + diffuse) generally increases with

regard to total global solar radiation when atmospheric
transmissivity decreases, mainly because of clouds, but
also because of aerosols and dust. In fact, the average
ratio of UV to total radiation on cloudy days is up to two
percentage points more than on clear days.
The efficiency of a chemical reaction is calculated
from the ratio of the products to the departing reactants.
In photo- chemistry, it is very common to use the
quantum yield con- cept, which is calculated from a
known amount of photons absorbed in the reaction.
Quantum yield (8) is defined as the ratio of the
number of reacting molecules (An) to the quantity of
photons absorbed by the system (NA):
FIGURE 2.4. Air mass and solar components.
 (2.1)

liquid water droplets and clouds. The spectral irradiance Experimentally, quantum yield is expressed as the number
data are for the Sun at a solar zenith angle of 48.19°. of moles of reactant in an interval of time t, divided into
This zenith angle corresponds to an air mass of 1.5, the number of moles of photons absorbed during the
which is the ratio of the direct-beam solar-irradiance path same period. Knowledge of the quantum yield is rather
length through the atmosphere at a solar zenith angle of important for an understanding of the mechanism of a
48.19° to the path length when the Sun is in a vertical photochemical reaction. If every absorbed photon
position. AM = 1 when the Sun is directly overhead produces a molecular transformation, 8 = 1. If it is less
(zenith). As air mass increases, the direct beam traverses than 1, it means that deactivation processes have
longer path lengths in the atmos- phere, which results in occurred or other reactions have competed with the one
more scattering and absorption of the direct beam and a studied. 8 > 1 indicates a series of reactions has taken
lower percentage of direct-to-total radiation (for the place, the promoter of which has been excited by a
same atmospheric conditions). photon. In the case of photocatalysis by UV radi- ation,
The AM 1.5 global irradiance is shown for a flat the number of photons that reach the reacting mixture
surface facing the Sun and tilted 37° from the horizontal. and are thereby susceptible to being absorbed will be in
The 37° tilt angle is used because it is the latitude of the rela- tion to the UV solar spectrum. For reactors using
Plataforma Solar de Almería in Spain, where most of the solar radi- ation, knowledge of the solar UV spectrum is
research pre- sented here was done. The limited part of important for the following reasons:
the solar spectrum that can be used in photocatalysis with • The radiation (sunlight) that reaches them is not con-
TiO2 may be clearly seen (see Table 1.4), but as the stant. This prevents direct comparison between
energy source is so cheap and abundant, even under experi- ments carried out at different times of the day
these limitations its use is of interest. or seasons of the year or under different atmospheric
• The extensive literature on photocatalytic
2.3. Solar ultraviolet irradiation decomposition of organic compounds indicates that
the majority of the experiments in which photon flux
Solar ultraviolet radiation is, as explained above, only a is known are carried out in laboratory reactors
very small part of the solar spectrum, amounting to illuminated by lamps. In order to compare these
between 3.5 per cent and 8 per cent of the total solar results with solar radiation or to use the information
spectrum, as demon- strated by measurement (this figure contained in these reports, it is necessary to know the
may vary for a given location on cloudy and clear days). photon flux inside the solar reactor.
The percentage of global • The quantum yield of the reaction tested under a
given experimental condition provides information
on the
22 Solar Detoxification

optimum conditions for decomposition of the Equation 2.3 gives the ratio between photonic and radio-
contami- nant. Knowledge of the photon flux in metric quantities, defining the photon flux density I
[Einstein s–1 m–2] as the number of incident photons per
this situation is basic to the determination of the
unit of surface and time:
efficiency of the solar reactor components
(reflective surface, absorber tube, control system,
concentration factor and so on) and to development
of any possible modifications to these I 
components to improve photodegradation N0dtdA (2.4)
• Any economic comparison between solar radiation
and electric lamps as the UV photon source requires where N0 is Avogadro’s number (6.023 × 1023). 1 Einstein =
knowl- edge of the photon flux incident on the solar 1 mol of photons 6.02 × 1023 photons.
reactor. Using the spectrum data and the above equations
The two spectra shown in Figure 2.5 correspond to the in congruent units [S.I.], it is possible to determine the
same spectra shown in Figure 2.3 for the solar UV photon flux density I (I D = 3.6 × 1019 photons m–2 s–1
spectrum range at ground level. The shorter spectrum = 6 × 10–5 Einstein m–2 s–1, IG = 8.4 × 1019 photons m–2
(direct UV) reaches 22 W m–2 between 300 and 400 s–1 = 14 × 10–5 Einstein m–2 s–1). These two values give
nm, the longer spectrum (global UV) reaches 46 W m– an idea of the energy coming from the Sun and
2. available for photocatalytic reactions with TiO2, which
The number of photons, Nλ, supplied by a monochro- only uses the part of the UV spectrum up to 390 nm,
matic source of light with wavelength λ and energy Qλ as explained below. In any case, the UV radiation
is related to the energy of one photon, Uλ, by Planck’s values described vary from one location to another and,
equation (Eq. 1.1): obviously, vary at differ-
ent hours of the day and in different seasons, making it
Q λ
N λQ (2.2) necessary to know these data for the particular location
λ λ
Wλ hc and in real time. These equations will be very useful in
When a source of light is polychromatic, as is solar those cases where this is not possible.
radiation, the number of photons is given by an integral
covering the whole range of wavelengths of that source:
2.4. Atmospheric attenuation
N   N(λ )dλ 

Q(λ (2.3) of solar radiation
λ1 hc λ1

A general expression for transmittance (T) of direct-

beam solar radiation under cloudless skies at a specific
wavelength (λ) is (Iqbal, 1983):

Tλ  TR,λTA,λTo,λTg,λTv,λ (2.5)

TR,λ is the spectral transmittance resulting from the

 disper- sion produced by molecules of air (dimensions
of many of which are =1 Å, Raleigh dispersion). TA,λ is the
spectral trans- mittance related to absorption and
dispersion by aerosols (solid or liquid particles
suspended in the air). To,λ corre- sponds to the effect of
 the ozone layer. Tg,λ is the transmit- tance resulting from
absorption of atmospheric gases (such as carbon dioxide
FIGURE 2.5. Ultraviolet spectra at the Earth’s surface and oxygen). Tν,λ corresponds to absorp- tion by water
vapour. The effect of each of these parameters
(standard ASTM). within the range in question (300 nm–400 nm) would be
Solar Irradiation 23

the following (Riordan et al., 1990): the UV range, but in the near-infrared region (see Figure
2.6) they cause increased absorption, due to the water
• TR,λ = exp(–0.008735 λ–4.08 M'), where M' is air vapour and liquid water they contain (Tv,λ).
mass corrected according to its density, which Clouds modify the total UV energy reaching the Earth’s
depends on pressure and, therefore, altitude. surface, but wavelength distribution is not affected. This
Accordingly, this factor would be practically constant can- not be guaranteed, however, if the data for the spectra
for a given site. shown in Figure 2.6 are not represented in a standardized
• TA,λ = exp(–λ–α β M) where M is air mass, β is the manner as in Figure 2.7. This can be done for any
turbid- ity coefficient, which usually varies between 0 wavelength inter- val by the following operation.
and 0.5 and is a reflection of the amount of aerosols in Summations have been used to treat the discrete values nm
the air, and α is an index of the size of the aerosol to nm:
molecules. β will vary depending on the atmospheric
contamination at the
site. Where there is no contamination it varies only
slightly. The same reasoning is valid for α. fλ 
UVλ nm
• T , therefor fλ
o,λ is constant for a specific site since the ozone layer
λ400 nm
e λ300
a practically constant thickness (for now).  UVλ
λ300 nm
• Tg,λ only influences wavelengths over UV.
• Tv,λ does not affect the UV spectrum. where fλ, is the fraction of power associated with
wavelength λ and UVλ is the irradiance, W m–2 nm–1
Keeping in mind, then, the different transmittances, it
corresponding to each wavelength and measured with a
may be assumed that the solar UV spectrum does not
spectroradiometer. In Figure 2.7, the homogeneity of the
vary sub- stantially at a specific site throughout the year,
spectra recorded may be observed. If the spectrum of
unless atmos- pheric conditions (other than clouds) do
UV radiation is assumed to have a fixed form, then a
so. The dominating attenuator of solar radiation is cloud.
standardized spectrum can be considered as standard for
Under overcast skies there is no direct-beam radiation,
each site. Therefore, the number of photons
and under partly cloudy skies there is intermittent direct-
corresponding to this range of wavelengths is only a
beam radiation when clouds are not obscuring the Sun’s
function of the intensity (measurable in real time with
disk. Clouds are often assumed to have a wavelength-
radiometers, see the following section in this chapter).
independent attenuation function in


Clear sky Clear sky



0.2 0.1
0.0 0.0
400 600 800 1,000 1,200 300 325 350 375 400
Wavelength, nm Wavelength, nm

FIGURE 2.6. Solar spectra at the Earth’s surface (Plataforma Solar de Almería) between 300 and 1,100 nm.
Solar Irradiation 25


λ400 nm
U λ
fλ  λ400Vnm , therefor  fλ
 U λ
e λ300 nm =1

λ300 nmV
f, nm-1


Area below each curve = 1

300 325 350 375 400
Wavelength, nm
FIGURE 2.7. Normalized solar UV spectra shown in Figure 2.6.

2.4.1. Annual available ultraviolet

A general index of atmospheric transmittance
incorporating all the processes described above is the so- FIGURE 2.8. TBDUV at different periods of the year at
called ‘cloud fac- tor’ (fn). For the calculation of fn, the Plataforma Solar de Almería (37 °N).
annual average ultra- violet radiation level must be found.
Knowledge of this factor enables the amount of energy that
reaches the Earth’s surface to be predicted for a given throughout the day. In this case, being discrete values, it
place at any time of the year. It has been calculated by summing hourly irradiance valves
should be noted that the cloud factor for global radiation – –2
is sunrise to H (kJ ) is the average of the month.
always lower than for direct radiation, as the diffuse sunset. m
nent of solar radiation is maximum when direct radiation This is calculated by multiplying the monthly average irradi-
is absent (global = direct + diffuse). This factor is ance by the monthly average of hours of sunlight. The
calculated (Eq. 2.7) from the ratio of average radiation cloud factor for UV radiation is calculated from data
(affected by all atmospheric phenomena) to the highest collected with radiometers that are UV-specific. The UV-
attainable radiation at all times of the year. This is radiation database must be large enough to be considered
usually calculated for each month separately. To find out statistically correct (at least four to five years).
the highest UV radiation avail- able each month, a typical
‘best day’ (completely clear sky during all daylight hours)
is selected for each month (TBDUV) from among all the
2.5. Solar radiation measurement
days for which average radiation is calcu- lated. The
parameters taken into account in selecting the TBDUV Detectors translate light energy into an electrical
are absence of clouds and proximity to the fifteenth day current (Ryer, 1997). Light striking a silicon photodiode
of each month (See Figure 2.8). Several days per month causes a charge to build up between internal ‘P’ and ‘N’
(corresponding to each year for the period to be layers. When an external circuit is connected to the cell,
analysed) must be chosen and compared to find the day an electrical current is produced. This current is
with the maximum HTBD. linear with regard to the incident light over a
 H  dynamic ten-decade range. A wide dynamic range is a
prerequisite for most applica-
fn  1    tions. The radiometer should be able to cover the entire
 HTBD 
dynamic range of any detector that will be plugged into it.
where HTBD (kJ m–2) is the TBDUV radiance exposure This usually means that the instrument should be able
(direct or global). This is obtained by the sum of the UV to cover at least seven decades of dynamic range with
irradiance minimal
24 Solar Detoxification

linearity errors. The current or voltage measurement stability and durability are considerations. Some detector
device should be the least significant source of error in types must be cooled or modulated to remain stable.
the system. The second thing to consider when choosing High voltages are required for other types. In addition,
a radiometer is the type of features offered. Ambient some can be burned out by excessive light, or have their
zeroing, integration ability and a ‘hold’ button should be windows permanently ruined by a fingerprint.
standard. The ability to multiplex several detectors to a
single radiometer or to con- trol the instrument 2.5.1. Detectors
remotely may also be desired. Last, portability and
battery life may be an issue for measure- ments made
Planar-diffusion-type silicon photodiodes are perhaps the
in the field.
most versatile and reliable sensors available. The P-
Light is all around us every day, yet it remains the most layer material at the light-sensitive surface and the N
elusive form of energy to measure accurately. A single pho- material at the substrate from a P–N junction, which
ton of light travels in a straight line in one direction at a operates as a photoelectric converter, generate a current
given wavelength. A light bundle consists of a jumbled that is propor- tional to the incident light. Silicon cells
mixture of billions and billions of photons at different operate linearly over a ten-decade dynamic range, and
wavelengths, going in different directions, at different remain true to their original calibration longer than any
moments in time. The watt (W), the fundamental unit of other type of sensor. For this reason, they are used as
optical power, is defined as a rate of energy of one joule ( transfer standards at the United States National
J) per second. Optical power is a function of both the Institute of Standards and Technology (NIST).
number of photons and the wavelength. Each photon
The phototube is a light sensor that is based on the
carries an energy that is described by Planck’s equation
photo- emissive effect. It is a bipolar tube consisting of a
(Eq. 1.1). All light measure- ment units are spectral, spatial
photoemis- sive cathode surface that emits electrons in
or temporal distributions of optical energy. The biggest
proportion to incident light, and an anode that collects
hurdle in light measurement is the spatial nature of light.
the electrons emit- ted. The anode must be biased at high
Irradiance is a measure of the energy density received
voltage (50 to 90 V) in order to attract electrons strongly
from a light source. Since light expands outward from a
enough to cause them to jump through the vacuum of the
point source, irradiance decreases with distance. Irradiance
tube. Some phototubes use a forward bias of less than 15
also decreases with incident angle. Carefully designed input
volts, however. The cathode material determines the
optics cannot prevent measurement errors caused by lax
spectral sensitivity of the tube. Solar- blind vacuum
attention to the measurement geometry upon which the
photodiodes use Cs-Ta cathodes to ensure sensitivity
unit’s systems are based. Spectral responsiv- ity and
only to ultraviolet light, providing as much as a million
detectivity present a very different problem. Many
rejecting long wavelengths for every UV wavelength
properties of light are dependent on wavelength, and the
accepted. A UV glass window is required for sensitivity in
energy of one photon is inversely proportional to its wave-
the UV down to 185 nm, with fused silica windows
length. Since a detector measures only absorbed light, it
offering transmission down to 160 nm.
cannot differentiate between one photon (200 nm) and ten
photons (2,000 nm). The light must be filtered by The thermopile is a heat-sensitive device that
wavelength before it reaches the detector. measures radiated heat. Infrared light contains the least
amount of energy per photon of any band. Because of
Sensitivity to the band of interest is a primary this, an infrared photon often lacks the energy required to
considera- tion when choosing a detector. You can pass the detection threshold of a quantum detector.
control peak respon- sivity and bandwidth through the Infrared light is usually measured using a thermal
use of filters, but you must have an adequate signal to detector such as a thermopile, which measures
start with. Filters can suppress out-of-band light but temperature change due to absorbed energy. While these
cannot boost the signal. Another con- sideration is thermal detectors have a very flat spectral respon- sibility,
blindness to out-of-band radiation. If you were they suffer from temperature sensitivity, and usually
measuring solar ultraviolet in the presence of must be artificially cooled. The sensor is usually sealed
massive amounts of visible and infrared light, for in a vacuum to prevent heat transfer except by radiation.
example, you would select a detector that was A ther- mopile consists of a number of thermocouple
insensitive to the long wave- length light that you junctions in
intended to filter out. Last, linearity,
26 Solar Detoxification

the filter. Given a desired overall sensitivity and a known

detector responsivity, you can then draw a transmission
curve for the ideal filter. Filter bandwidth decreases with
thickness according to Lambert-Beer’s law (see
1.2 and 1.3), so by varying filter thickness you can selec-
tively modify the spectral responsivity of a sensor to
match a particular function. Multiple filters cemented in
layers give a net transmission equal to the product of the
individual transmissions. Filters operate by absorption or
interference. Coloured glass filters are doped with
materials that selec- tively absorb light by wavelength,
and obey Lambert-Beer’s law. The peak transmission is
, , inherent to the additives, while bandwidth is dependent
on thickness. Sharp-cut fil- ters act as long-pass filters,
and are often used to subtract out long-wavelength
FIGURE 2.9. Responsivities of three detectors. In the inset radiation in a secondary measurement. Interference
is shown a schematic of the effect of a filter on detector filters rely on thin layers of dielectric to cause
respon- sivity. interference between wave fronts, providing very
narrow bandwidths. Any of these filter types can be
series, which convert energy into a voltage via the Peltier combined to form a composite filter that matches a
effect. Thermopiles are convenient sensors for measuring particular photo- chemical process.
infrared light, because they offer adequate sensitivity and
a flat spectral response in a small package. More 2.5.3. Input optics
sophisticated bolometers and pyroelectric detectors need
to be chopped and are generally used only in calibration
laboratories. When selecting input optics for a measurement
The best method of operating a thermal detector is application, consider both the size of the source and the
by chopping incident radiation, so that drift is zeroed out viewing angle of the intended real-world receiver.
by the modulated reading. The quartz window in most Suppose, for example, that you were measuring the
thermopiles is adequate for transmitting from 200 to erythemal (sunburn) effect of the Sun on human skin.
4,200 nm, but for long wavelength sensitivity, up to While the Sun may be considered very much a point
40 microns, Potassium bromide windows are used. source, skylight, refracted and reflected by the
Another strategy employed by thermal detectors is to atmosphere, contributes significantly to the overall
modulate incident light with a chop- per. This allows amount of light reaching the Earth’s surface. Sunlight is a
the detector to measure differentially between the combina- tion of a point source and a 2π steradian area
dark (zero) and light states. Quantum-type detectors source. The skin, since it is relatively flat and diffuse, is an
are often used in the near infrared range, especially effective cosine receiver. It absorbs radiation in
below 1,100 nm. Specialized detectors such as InGaAs proportion to the incident angle of the light. An
offer excellent responsivity from 850 to 1,700 nm. appropriate measurement system should also have a
Typical silicon photodiodes are not sensitive above cosine response. If you aimed the detector directly at the
1,100 nm. These types of detectors are typically Sun and tracked the Sun’s path, you would be measuring
employed to measure a known arti- ficial near-IR the maximum irradiance. If, however, you wanted to
source without including long wavelength measure the effect on a person lying on the beach, you
background ambient radiation. Most radiometric IR might want the detector to face straight up, regardless of
meas- urements are made without lenses, filters or the Sun’s position. This example can be extended to solar
diffusers, relying on just the bare detector to measure collectors (see Chapter 6).
incident irradiance. Different measurement geometries necessitate
specialized input optics. Radiance and luminance
2.5.2. Filters measurements require a narrow viewing angle (< 4°) in
order to satisfy the conditions underlying the
measurement units. Power measurements,
A detector’s overall spectral sensitivity is equal to the
prod- uct of the responsivity of the sensor and the
transmission of
Solar Irradiation 27

FIGURE 2.11. A solar global UV detector (tilted 37° and

fac- ing south) with a cosine diffuser.

FIGURE 2.10. Relative spatial response of an ideal cosine

diffuser (up) and a radiance lens barrel (down).

on the other hand, require a uniform response to

radiation, regardless of input angle, to capture all light.
There may also be occasions when the need for additional
signal or the desire to exclude off-angle light affects the
choice of input optics. A high-gain lens, for example, is
often used to amplify a distant point source. A detector can
be calibrated to use any input optics as long as they
reflect the overall goal of the measurement.
FIGURE 2.12. A solar direct UV detector installed on a solar
Cosine diffusers tracking system.
A bare silicon cell has a near-perfect cosine response, as
do all diffuse planar surfaces. As soon as you place a signal. Radiance optics merely limits the viewing angle to
filter in front of the detector, however, you change the less than the extent of a uniform area source. This input
optic is used to measured direct sunlight, and is
spatial re- sponsivity of the cell by restricting off-angle
light. Fused silica or optical quartz with a ground mounted on a mobile sun-tracking platform (one loop
(rough) internal hemisphere makes an excellent per day) to follow the Sun from sunrise to sunset.
diffuser, with adequate transmission in the ultraviolet.
Teflon is an excellent alterna- tive for UV and visible Fibre optics
applications, but is not an effective diffuser for infrared Fibre optics allows measurements in tight places or
light. where irradiance levels and heat are very high. Fibre optics
consists of a core fibre and a jacket with an refraction
Radiance lens barrels index chosen to maximize total internal reflection. Glass
Radiance and luminance optics frequently employ a dual fibres are suitable for use in the visible, but quartz or fused
lens system that provides an effective viewing angle of less silica is required for trans- mission in the ultraviolet. Fibres
than 4°. The trade-off of a restricted viewing angle is a are often used to continuously
reduction in
28 Solar Detoxification

monitor UV curing ovens, due to the attenuation and heat fluxes (Einstein) from radiometric measurements (W)
pro- tection they provide. Typical fibre optics restricts the and spectral data (nm–1) has been introduced. The
field of view to about ±20° in the visible and ±10° in the attenuating components of the atmosphere and their
ultraviolet. effect on UV radia- tion have been discussed to achieve a
final conclusion: the UV spectrum is constant at a
Integrating spheres definite emplacement under defined circumstances. This
An integrating sphere is a hollow sphere coated inside characteristic permits standard- ization of the solar-UV
with barium sulfate, a diffuse white reflectance coating spectrum, which is very helpful for finding a standard
that offers greater than 97 per cent reflectance between photon flux. The cloud factor index has been described
450 and 900 nm. The sphere is baffled internally to and its calculation from a UV-radiation data- base has
block direct and first- bounce light. Integrating spheres been explained. Finally, solar radiation measure- ment
are used as sources of uni- form radiance and as input systems have been described, with special emphasis on
optics for measuring total power. Often, a lamp is placed their main components. Their correct combination will
inside the sphere to capture light that is emitted in any per- mit accurate analysis of solar radiation and correct
direction. evaluation of the quantum yield of photochemical
High-gain lenses
In situations with low irradiance from a point source,
high- gain input optics can be used to amplify the light by Bibliography
as much as fifty times while ignoring off-angle ambient
light. Flash sources such as tower beacons often employ
HULSTROM, R.; BIRD, R. AND RIORDAN, C. 1985. Spectral
Fresnel lenses, making near field measurements difficult.
With a high-gain lens you can measure a flash source solar irradiance data sets for selected terrestrial conditions.
from a distance without compromising signal strength. SOLAR Cells, No. 15, pp. 365–91.
High-gain lenses restrict the field of view to ±8°, and so
cannot be used in full immersion applications where a IQBAL, M. 1983. An Introduction to SOLAR RADIATION. Canada,
cosine response is required. Academic Press.


Summary Influences of Atmospheric Conditions AND Air MASS on the
The three principal components of light (ultraviolet, Colo., Solar Energy Research Institute (SERI)/TP-215-3895.
visible and infrared) and their wavelength distribution
have been described. Typical solar spectra and air mass RYER, A. 1997. Light MEASUREMENT HANDBOOK. Newburyport,
effects (Sun position) have been shown. The calculation Mass., International Light Inc.
of photonic
Solar Irradiation 29

Self-assessment questions Part B

Part A: True or false? 1. What is the most important difference

between ultraviolet, visible and infrared light?
1. The visible portion of light is more powerful than 2. What are the usual units used to express solar
the UV portion. spectrum power?
2. Most solar UV light at the Earth’s surface is between 3. Why is the air mass at Sun zenith called AM 1?
100 and 280 nm.
4. Cite at least two reasons to justify the importance
3. The solar radiation that reaches ground level without of understanding the solar spectrum in
being absorbed or scattered is called ‘direct radiation’. photochemistry.
4. Any economic comparison between using solar 5. What is the basic measurement unit in
radiation and electric lamps as the UV photon source, photochemistry, and why?
requires knowledge of the photon flux incident on the
solar reactor.
6. Why do clouds not absorb UV light?
5. The UV photon flux density at the Earth’s surface is
7. What is the usefulness of equation 2.6?
in the range of thousands of photons per square metre.
8. Convert the following fraction of power associated
6. Clouds absorb UV light. with each wavelength (corresponding to a standardized
spectrum from 300 to 400 nm) into solar spectrum power
7. Clouds scatter UV light. given that the total measured power is 40 W/m2.
8. The cloud factor for UV light is determined TABLE 2.1
using specific UV radiometers.
λ, nm fλ
9. The best detector to measure infrared light
is a phototube. 300 0.0002
302 0.0004
10. A detector’s overall spectral sensitivity is equal to 304 0.0007
the product of the transmission of the sensor and the
responsivity of the filter. 306 0.0011
308 0.0017
310 0.0023
312 0.0032

9. Why is the measurement of infrared light

with a quantum radiometer very difficult?

10. How is it possible to vary the spectral responsivity

of a sensor without changing its type of filter?
30 Solar Detoxification

Answers 5. Einstein. Because it is 1 mol of photons and

therefore, the quantum yield of a photochemical
reaction (rate
Part A 1 M/min) that absorbs 1 Einstein L –1 min–1 is 1.

1) False. 2) False. 3) True. 4) True. 5) False. 6) False. 6. Because pure water does not absorb UV light.
7) True. 8) True. 9) False. 10) False.
7. To represent spectra in a standardized manner. This
Part B will permit comparison between spectra recorded at
different sites, hours of the day and/or different seasons.

1. Wavelength: ultraviolet light is from 100 to 400 8. TABLE 2.2

nm, visible light is from 400 to 770 nm, and infrared
light is from 770 to 106 nm. λ, nm UVλ, W m–2 nm–1

2. W m–2 nm–1 or W m–2 µm–1. 300 0.0080

302 0.0160
3. Because zenith corresponds to the minimum path 304 0.0280
length through the atmosphere of direct-beam solar
irradiance. 306 0.0440
308 0.0680
4. a) The radiation that reaches the solar reactors is 310 0.0920
not constant. 312 0.1280
b) In order to compare laboratory results with
solar results or to use the information obtained
with lamps. 9. Infrared light contains the least amount of energy
c) To determine the efficiency of the per photon of any band. Because of this, an infrared
components of the solar reactor and the possible photon often lacks the energy required to pass the
modifications to improve the conditions of detection threshold of a quantum detector.
d) To perform economic comparisons between 10. By varying filter thickness.
solar radiation and electric lamps.
Solar Irradiation 31


Experimental Systems

Aims Objectives
This chapter describes the experimental By the end of this chapter, you will
systems necessary for performing understand the most important features
pilot-plant-scale solar photocatalytic of large-scale outdoor photocatalytic
experiments. Various laboratory set-ups experiments and you will be able to do
are also shown in order to emphasize six things:
differences between these small 1. Design a simple, versatile solar
photoreactors and pre-industrial-scale pilot detoxification pilot system to fit the
plants. present and future needs of the research
The chapter outlines the basic components to be performed with it.
of these pilot plants and the different
possibilities for operating them. Modelling
2. Understand pilot-plant operation
of the experiments is reviewed from a
and decide among different options.
chemical engineering point of view.
Finally, an extensive overview of the
3. Calculate the kinetic constant using
relationships between radiometer
measurements and the photons that the appropriate method depending on
pilot-plant characteristics and operation.
actually reach the photoreactor is
4. Perform on-site calibration of solar-
UV radiometers.

5. Find the relationship between solar

radiometer measurements and the
photons that actually reach the

6. Employ chemical actinometers to

validate all the calculations performed to
measure the photon flux inside
32 Solar Detoxification


Symbol Description Units
C or Ci Reactant concentration M or mg L–1
EUV Accumulated energy (per unit of volume) incident on the reactor kJ L–1
I Photon flux density Einstein s–1 m–2
IE Effective photon flux corresponding to the UV inside the absorber of a photoreactor Einstein s–1 m–2
I*E Volumetric effective photon flux corresponding to the UV inside the absorber
of a photoreactor Einstein L–1s–1
fλ Fraction of power associated with a wavelength nm–1
k First order kinetic constant min–1, h–1
N Photon flux density Photons s–1 m–2
Q Flow rate L min–1, m3 h–1
r Reaction rate M min–1, mg L–1 min–1
Sp Surface area of the reflector of a photoreactor capturing the radiation m2
ST Surface area of the tube of a photoreactor m2
texp Experimental time min, s, h
tR Residence time min, s, h
UV Measurements provided by a radiometer W m –2
UVλ Ultraviolet irradiance associated with a wavelength W m–2 nm–1
–2 –1
UV*λ Spectral data calculated with the standard spectrum and radiometric data W m nm
UVΣ Summation of the measurements provided by a spectroradiometer W m –2
V Volume L, m3
VTOT Volume of the entire pilot plant L, m3
x Conversion
8 Estimated quantum yield using moles of incident photons instead of moles of
photons absorbed by the catalyst
η Loss factors affecting the photon flux
λ Wavelength µm, nm
µλ Coefficient of absorptivity cm–1
τ One residence time min, s, h

3.1. Laboratory systems continuous stirred tank reactor (Figure 3.1) and the other
a recirculating system (Figure 3.2).
Figure 3.1 shows the basic set up of a
The treatment of contaminated water necessarily includes microphotoreactor used in laboratory experiments. The
the design of an efficient photoreactor. Basic laboratory photoreactor consists of a flask made of Pyrex with a flat
research on the process has mostly been performed with bottom, generally made of quartz, which constitutes an
experimental devices in which efficiency is not as
important as obtaining appropriate conditions that will optical filter permitting UV- visible radiation with λ Š
permit reproducibility of the results and provide exhaustive 220 nm to enter. For an isothermal reaction, or when one
data on the effects of all impor- tant parameters (Cassano, wants to change the temperature of the medium to
2000). This is correct when the goal is fundamental determine the activation energy of the reac- tion (see
knowledge of the process, but is not always sufficient to Section 4.7), the system is equipped with a jacketed
attempt a change of scale. Two different types of envelope, through which flows a temperature-constant
laboratory photoreactors are presented here. One is a fluid (for example, water) delivered by a thermostat.
There are
Experimental Systems 33

or Valve


Pollutant Magnetic
+ TiO2


Optical H2O

UV lamp

FIGURE 3.1. Typical stirred-tank laboratory reactor (courtesy of Laboratoire de Photocatalyse, Catalyse et
Environnement, (E.R. au CNRS), Ecole Centrale de Lyon).

several upper apertures to introduce the reacting mixture, The radiant flux (in W/m 2) is measured for each
the catalyst, or a controlled gas atmosphere, and also to experiment by a radiometer positioned in the place of the
with- draw samples, either from the slurry or from the photoreactor. The photograph represents a simple
gas phase through the ground top and the valve for photoreactor, without a cooled jacket and open to the air,
analysis by GC (gas chromatography) or GC/MS (gas used for a routine experi- ment of water detoxification
chromatography-mass spectrometry, see Chapter 5). A using an illuminated slurry of titania. Mixing is sufficient
good stirring is obtained with a magnetic stirrer, to supply gaseous oxygen to the medium (see Section
curiously working perpendicularly to the rotating magnet 4.2).
stirrer to ensure it is at the right level. The determination Figure 3.2 shows the set up and a photograph of the
of gaseous CO2 is made in an oxygen- confined recirculating system. Water, contaminants and titanium
atmosphere by sampling amounts of supernatant gas dioxide are continuously recirculated through a 250 mL
phase through the ground top and the valve, linked to a Pyrex reactor. During photocatalytic degradation
GC by stainless-steel tube. experi- ments, parameters such as pH, CO 2 and O2 can
The UV light is provided by a lamp introduced in a be moni- tored in situ via specific electrodes (probes). Air
water- cooled envelope that is closed by a silica optical is introduced into the reactor flask through a flow cell
disk – used to prevent the formation of ozone in the and distributed via a sintered-glass tip to maintain a
laboratory atmos- phere. All the IR beams, which could constant oxygen concentra- tion in the system. A solar
heat the slurry sus- pensions, are removed by a water simulator equipped with a high- pressure xenon arc lamp
cell, which contains an optical disk transparent to the and a parabolic reflector carries out irradiation. As solar
wavelength domain desired. simulators contain a significant thermal
34 Solar Detoxification

Probes above, a simple modification of the parabolic-trough

solar thermal collector has been successfully developed
and is used for experiments in which the catalyst is
Magnetic deployed in suspension. Parabolic concentrating-type
mixer reactors have been used for applications such as
groundwater remediation and removal of metals from
Photoreactor Xenon lamp water. A disadvantage of concentrat- ing reactors is that
they cannot concentrate diffuse solar radiation. This is
not a major problem for solar thermal applications,
Pump Neutral filter because diffuse radiation forms a small fraction of total
solar radiation. However, solar photocatalytic detox-
Sampling ification with TiO2 as a catalyst uses only the ultraviolet
valve por- tion of solar radiation, and as much as 50 per cent or
more of this may be in diffuse form, especially at humid
locations and during cloudy or partly cloudy periods.
Therefore, concen- trating reactors would be more useful
at dry, high, direct- insolation locations. Another
disadvantage of concentrating reactors is that quantum

efficiency is low, because their rate has an inverse
square, rather than a linear, dependence on light flux (see
Chapter 4). These disadvantages tend to favour the use of
non-concentrating reactors. In any case, these are aspects
which will be discussed in detail in Chapter 6.
Whatever the type of collector, solar detoxification
pilot plants have several components and design
characteristics in common, which can be discussed here.
The design procedure for a pilot solar detoxification
FIGURE 3.2. Typical recirculating laboratory system. system requires the selection of reactor, catalyst
operating mode (suspension or fixed matrix), reactor-
field configura- tion (series or parallel), treatment-
component, it is necessary to decrease the temperature system mode (once- through or batch), flow rate,
by using a small heat exchanger. The entire reactor is pressure drop, pre-treatment, catalyst and oxidant
con- structed with chemically resistant materials (Pyrex, loading method, and pH control method. A pilot plant has
Teflon and Viton). to be as versatile as possible to allow for these variables
and, at the same time, provide sufficient confidence in
3.2. Solar detoxification the experiments carried out in it. A pilot plant must fulfil
pilot plants all the present and future requirements of the research to
be performed in it. In Figure 3.3a, a detailed drawing of
a plant is given. Usually, a detoxification pilot plant is
The first outdoor engineering-scale photoreactor constructed with several solar collectors (Figures 3.3b,
developed was a transformed solar thermal parabolic- c and d show different solar collector designs). All the
trough collector in which the absorber/glazing-tube modules are connected in series, but with valves that
combination had been replaced by a simple Pyrex glass permit bypassing any number of them. Sampling valves
tube through which contaminated water could flow are in the outlet of each of the modules. All the tubes and
(Goswami, 1995). Since that time, research all over the valves are made of black HDPE (chosen because it is
world has advanced a number of reactor concepts and strongly resistant to chemicals and is weatherproof and
designs, including concentrating and non-concentrating opaque) in order to avoid any photochemical effect
reactors (Malato et al., 2002). The catalyst can be outside the collectors. The test mix- tures are prepared in
deployed in several ways, including using a fixed cat- storage-feeder tanks, also made of HDPE, of various
alyst, a slurry or large neutral-density particles. As capacities. Four different operating modes
Experimental Systems 35

Solar collector
T n Closed valve
in series
Opened valve
Refrigeration system (optional)

Pr O2

GAC + TiO2 Holding
Flow control
filter tank
Pr Batch tank
Clean water

b) c)


FIGURE 3.3. a) Photocatalytic detoxification pilot-plant scheme (batch mode is shown). b) CPCs at Plataforma Solar de
Almería (Spain). c) Double-skin sheet collector (ISFH-Hannover design) at Plataforma Solar de Almería (Spain). d) Two
types of non-concentrating solar collectors at the solar detoxification test stand at the University of Florida. (Courtesy
of the Solar Energy and Energy Conversion laboratory, University of Florida.)
36 Solar Detoxification

are possible: recirculation, once-through, partial

recirculation and system cleaning.
3.3. Operation of pilot plants
When concentrating solar collectors are used, the
temperature of the water flowing through them rises con- 3.3.1. Once-through operation
siderably. Obviously, the slower the flow rate used in
once- through experiments, and the longer the The experimental procedure begins when the pump is
recirculation experiment, the greater the increase of turned on and the system is filled with clean water. The
temperature is. Therefore, to avoid evaporation and required modules are selected and the corresponding
damage to plastics, cool- ing is necessary, and a closed- valves are set to bypass the rest. Then the water is
circuit water-cooling system has to be installed. pumped through the circuit and the modules are covered.
A centrifugal pump with an electric motor Obviously, the max- imum pump flow rate is necessary
(calculated to provide sufficient flow when the for this procedure. The amounts required to obtain the
maximum length of the sys- tem is used) has to be initial concentrations of catalyst, contaminant and any
installed to pump the water to be treated through the other ingredient in the exper- iment are added to the
reactor. The flow rate (in batch mode) has to be such holding tank. When the time needed for the mixture to
that only a small amount of reactant is con- verted combine has expired, this is verified by taking samples at
each time it passes through the reactor, and the two different points in the reactor at the same time for
concentration throughout the system remains analysis. A few minutes later two more samples are taken
relatively constant (this reasoning will become clear and, if the four are the same, the concentration of the
below). Either a flow-rate control loop made up of a reactives may be considered to be the same throughout the
flow meter connected to a controller (which in turn reactor. Simultaneously, the automatic control sets the
governs an automatic electric valve) or an electric flow rate (Q), which will then be kept constant during
pump with a speed controller must be installed to the experiment, oxygen injection is activated, and valves
regulate the flow to the rate desired. The most are adjusted so that the fluid flows to the discharge tank.
important sensors required for the system are After that the modules to be used are put into operation.
temperature, pressure and dissolved oxygen sensors (at This marks the beginning of the experiment.
least in the reactor outlet). Other sensors, such as pH The modules are kept illuminated a little longer than
and selective electrode sensors, could be useful nec- essary to allow the water in the holding tank to go
depending on the type of experi- ments to be carried through the reactor and approach the discharging tank.
out. As constant pure oxygen is required for the This experi- mental time (texp) is:
oxidation of organic substances, an injection system at
the reactor inlet allows oxygen to be added to the reactor –
either at planned intervals by opening and closing in a Vtube  nVmod
pre- defined cycle, or continuously. Atmospheric oxygen texp  (3.1)
can also be stirred into the reaction medium in the Q
reservoir tank. In
this case, the dissolved O2 is kept at a concentration of where Q is the flow rate, Vtube is the volume in the pipes
around 6 to 8 mg/L. A UV-radiation sensor must be between the modules and the tank and Vmod is the volume in
placed in a position where the solar UV light reaching the each module, with n the number of modules in the series.
photoreactor can be measured, permitting evaluation of At this time, samples are taken at all the valves in the
incident radiation as a function of the hour of the day, outlets of each of the modules in the experiment. This
cloud cover, and atmos- pheric or other environmental provides ‘n’ number of samples with different residence or
variations. All these data have to be sent by an illumination times (tR,i) to enable determination of kinetics.
appropriate transmitter from the sensors to a computer, Under these conditions, the reactor behaves according to the
which stores the results for later evaluation. To clean the ideal plug- flow model, as explained later. The residence
system, a drainage tube, with an active carbon filter to time corre- sponding to each sample collected at the end
retain any organic compound not decomposed during the of the experiment is calculated with the following equation:
experiments, must be hooked up to the sewage pipelines.
Use of demineralized water (conductivity less than 10
organic carbon content less than 0.5 mg/L) is niVillu
recommended for cleaning as well as in the tests tR,i  (3.2)
Experimental Systems 37

where i is the number of modules through which the

sam- ples have passed before being collected and Villu
is the volume in the illuminated section of each module.
When the test is over, n samples have been obtained
with a reactor residence time that is a function of the
flow rate. Thus, if the procedure is repeated at a
different flow rate, that group of samples has a
different tR and the number of points (tR,i,
concentration) necessary to evaluate any experiment
can be obtained.

3.3.2. Batch operation

FIGURE 3.4. Schematic of two pilot plant operation
Solar detoxification pilot plants are frequently operated concepts: a once-through operation (top) and a batch
in a recirculating batch mode, as depicted in Figure 3.3. operation (bottom).
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. The systems
are operated in a discontinuous manner by recirculating 3.3.3. Modelling once-through
the slurry solution using an intermediate reservoir tank and batch operations
and a centrifugal pump. This type of operation differs
little from the previous one. In general, once-through operations can be modelled as
When concentration of the reactives is the same plug-flow. The variation in concentration (dC) over time
throughout the reactor, oxygen injection (if necessary) is (dt) for once-through experiments and for first-order
activated and the position of the valves is maintained so reactions (typical in photocatalysis) would be:
that the fluid begins and ends in the holding tank (now
called the ‘batch tank’). The automatic control sets the max- Ci
ln  ktR,i (3.4)
imum flow rate, which has to be such that it guarantees that C0
only a small amount of reactant is converted each time it
goes through the reactor. Then the modules that are going where tR,i is calculated by Equation 3.2. For most
to be used are put into operation. This begins the experi- laboratory- scale batch systems, the amount of
ment. Recirculation is continued and the test lasts as long as conversion that occurs each time the mixture passes
required, even up to several days. Samples may now be through the reactor is small, and the concentration
taken at any of the sampling ports – as the system is in recir- throughout the system is relatively constant. The entire
culation mode, tR is the same for samples taken at any point system can easily be modelled as a sim- ple batch system
in the system. The (tR,i, concentration) pairs are thus in which only a fraction (i.e., the part in the reactor) of
obtained (Eq. 3.3). the total mixture is undergoing reaction. In this case,
Equation 3.4 above can also be applied, using Equation
3.3 to calculate tR,i.
Villu In a large field system, the amount of conversion each
tR,i  t exp, (3.3)
VTOT i time the mixture passes through the reactor is noticeable.
As the relatively clean water in the reactor is mixed with
where VTOT is the volume of the entire pilot plant. Villu the ‘dirty’ water in the batch tank, the water sent from it to
and VTOT are defined at the beginning of the experiment the reactor has a lower and lower concentration of
by the number of modules used and the level of water in reactant. Because the conversion rate usually decreases
the batch tank. The experimental time (texp) is the with concen- tration, the overall rate in the reactor
difference in time between the initial sample (initial responds likewise. Thus, unless properly accounted for,
concentration of the pollutant, t = 0) and samples the presence of the tank will alter the perceived
collected during the experiment (t > 0). performance of the photoreactor. Two solutions are
available to solve this problem. The first,
38 Solar Detoxification

and also the easier of the two, is to use a very high flow • Regime 2: 0 < tR < τ. Under these conditions, the fluid
rate to achieve low conversion each time through. This exiting the reactor has been illuminated for less than one
high flow rate (Eq. 3.5) must ensure a rate of less than 1 complete reactor residence time. In Equation 3.6, the
per cent con- version per pass (C1(t) = C2(t)). This 1 time spent in the reactor is equal to tR, and the starting
per cent has been selected because is very far from the concentration is C0, thus:
error associated with any chemical method applied for Ci
analysis. C1 tR
    dτ (3.7)
V r C0 r 0
Q  TOT (3.5)
C0 x
• Regime 3: tR Š τ. Now the fluid exiting the reactor has
where r is the reaction rate, C0 the initial concentration of been illuminated for one complete residence time and
the reactant and x the conversion (0.01). For example, if the τ* = τ. However, to solve for the reactor outlet concen-
initial reactant concentration is 50 mg L–1, the reaction rate tration at time t, what went into the reactor at time t – τ
is R
0.1 mg L–1 min–1, VTO is 250 L, the flow rate must must be known. This defines the lowest limit on the rate
and T be
– 1
over 50 L min . When this is not possible because the integral
reaction rate is very high, initial concentration very low : τ
and/or the pump is not strong enough, the mathematical C1 (tR )
   dτ
 (3.8)
formulation is the following: C2 (tR τ) r 0
In steady state, the concentration in the reactor
outlet at time t, C1(t) is determined from the inlet Having defined these three regimes, the balance of
concentration (which does not change with time) and material in the well-mixed tank may now be written as:
the reaction kinetics. However, because the system is a
transient process, the nor-
mal steady-state plug flow reactor equation cannot be dC2 (tR )
used to model the photoreactor. Because the inlet  Vbatch  Q[C1(tR )  C2 (tR (3.9)
concentration changes with time, C2(t) is defined by dt
the concentration of
what went into the reactor one residence time prior to t,
where Vbatch is the volume in the batch tank. Now, it only
C2(t–τ), and the kinetics. The batch tank is still modelled remains to link Equation 3.9 with the different forms of
as a well-mixed tank. Solving these equations is more Equation 3.6. To do this, we must define an expression
difficult than the low-conversion-per-pass case, and the for the rate. A simple first-order rate expression (typical
numerical routine is required to fit the data from the in photocatalysis) can be used as an example (see Eq.
batch test. This routine also takes into account the volume 3.4). The algorithm that solves Equation 3.6 for C1(t) as
of piping between the reactor outlet and the batch tank. a function of C2, at the different time regimes using first-
The defining equations for the system components order kinetics is the solution:
shown in Figure 3.4 are given by basic chemical
engineering principles. For the plug flow reactor: • Regime 1: C1 = C2 = C0
• Regime 2: C1(t) = C0 exp(–kt)
C1 dC τ*
C    dτ (3.6) • Regime 3: C1(t) = C2(t – τ) exp(–kτ)
2 r 0
This expression for C1(t) is substituted into Equation 3.9
In the batch process, both concentrations vary with time, and the value for C2(t) predicted. This prediction is
that is, C1 = C1(tR) and C2 = C2(tR). A solution for C1(tR) then com- pared to the experimental values for the
and C2(tR) for all tR can be obtained if three different time concentration in the batch tank, with iteration until the k
regimes are defined as follows: value providing the best fit to the data is obtained. This
methodology can be used to solve numerically for the
• Regime 1: tR = 0 (Reactor exposed to sunlight at t = 0). concentration profile in any batch process. Although it
By definition we have C1 = C2 = C0 is only really necessary when the recir- culation flow is
not high enough (See Eq. 3.5).
Experimental Systems 39

3.4. Evaluation of solar where N0 is Avogadro’s number (6.023 × 1023). Using

UV radiation inside photoreactors solar spectrum data (See Figures 2.6 and 2.7) and the
former equations in congruent units [S.I.], it is possible to
determine the photon flux density. In any case, the UV
Solar ultraviolet radiation is an essential parameter for the radiation values vary from one location to another and,
cor- rect evaluation of data obtained during photocatalytic obviously, during the course of a day and from season to
experi- ments in a solar water-decontamination pilot season, so that it is neces- sary to know these data for a
plant. The kinetic constants of photocatalytic processes given location and in real time. The calculation of the
can be obtained by plotting substrate concentration as a photon flux in any photochemical reactor could be
function of three dif- ferent variables: time, incident undertaken following the flow diagram shown in Figure
radiation power inside the reac- tor, and photonic flux 3.5.
absorbed by the catalyst. Depending on the procedure, the
complexity of obtaining these constants, as well as their
applicability, varies. When the photonic flux absorbed by 3.4.1. Calibration of radiometers
the catalyst is used as an independent variable,
extrapolation of the results is better. However, many Solar UV radiometers (See Chapter 2.5) provide data in
parame- ters (incident photons passing through the terms of the incident W m –2 reaching their surface, which
reactor without interacting with the catalyst, the directions gives an idea of the energy reaching any surface in the
in which light scat- ters, distribution of the sizes of TiO2 same position as they are with regard to the Sun.
particle suspended in the liquid and so on) must be known According to their manufacturers, UV instruments are
for this, making it impracti- cal in a reactor used for pilot- sensitive to ultravi- olet radiation (see Figure 2.9). But in
plant experiments. their technical descrip- tions, the interval of wavelengths
Use of the experimental time as the calculation unit covered in their calibration is not usually reported.
could give rise to misinterpretation of the results, Therefore, that specification, so important in the
because the reac- tor consists of illuminated and non- calculation of photonic flux, is not certifi- able. In order
illuminated elements. Large experimental reactors to resolve this uncertainty, on-site calibration is
require much instrumentation and the reactor must also necessary. This is performed by comparing the data pro-
be as versatile as possible, substan- tially increasing the vided by the UV radiometers with those from a
non-illuminated volume. With use of residence time – spectrora- diometer (which gives data in terms of W m –2
that is, the time the water has been exposed to the
nm–1) in the same position as each of the radiometers.
radiation – the conclusions would also be erroneous.
This is because, when time is the independent variable, This gives:
dif- λn
ferences in the incident radiation in the reactor
an experiment are not taken into account. Furthermore, UV   UVλ (3.11)
the more different the environmental conditions of the λ300
ments to be compared (different days, different periods
of the year or atmospheric variations), the more critical where UVΣ (W m–2) is the summation of the radiation
this becomes. The only way to avoid this problem is to meas- urements (from 300 nm up to n, in 1 nm intervals)
use a relationship between experimental time, plant provided by the spectroradiometer. An example (with
volume, collector surface and radiant power density, only a few measured data) is shown in Table 3.1. The
measured by radiometers. This permits extrapolation of two ways of measuring UV at different times of the day
known data from one scale to another (using the same are compared, with n = 400 nm, which is the ultraviolet-
type of solar collector), as well as avoiding the problem visible threshold, above which the radiometers should
arising from a variable source of radiation (sunlight). not measure.
Recalling Equation 2.4, the photon flux density I In view of these results, the conclusion is that the UV
[Einstein s–1 m–2] is the number of incident photons per radiometer measures beyond 400 nm. In order to find the
unit of surface and time: real interval, the same procedure is carried out for n
values over 400 nm, at 1 nm intervals. The results are
d2N given in Table 3.2, together with the percentage of error
at each interval.
I= (3.10)
It seems that, taking as the upper limit a wavelength of
N0dt dA 403–404 nm, the radiation measured by both instruments is
40 Solar Detoxification

Spectrophotometric measurement of UV
It is impossible to
evaluate the photon Radiometer calibration
flux without on-line
spectrum NO
measurements Is the uv spectrum constant?
It is possible to obtain an equation that correlates the
radiometric and spectrophotometric data

Determination of spectral collector

efficiencies (reflectivity, trasmissivity, etc.)

Equation to obtain photon flux inside the photoreactor

as function of on-line uv measurements

Actinometric experiments

Efficiencies NO
Experimental actinometric quantum yield =
calculation calculated (by previous equation) quantum yield?
not correct

FIGURE 3.5. Procedure used to calculate the photon flux inside a solar reactor.

TABLE 3.1. Radiometric and spectroradiometric UV measurements at different times of the day (n = 400 nm).
UV, W m–2 Local time UVΣ, W m–2 UV–UVΣ, W m–2 100(UV–UVΣ)/UVΣ,%
13.69 10:31 11.68 2.01 +17.2
17.43 11:00 16.01 1.42 +8.9
26.96 12:59 25.42 1.54 +6.1
17.68 15:34 16.00 1.68 +10.5

TABLE 3.2. Radiometric and spectroradiometric UV measurements at different times of the day (n > 400 nm).
UV (1)
, W m–
UV, W m–2 2
n = 401 n = 402 n = 403 n = 404 n = 405
13.69 11.99(+14.2) 12.33(+11.0) 13.30(+2.9) 13.53(+1.2) 13.10(+4.5)
17.43 16.42(+6.2) 16.87(+3.3) 17.33(0.0) 17.79(–2.1) 18.25(–4.7)
26.96 26.01(+3.6) 25.4(+6.1) 27.32(–1.3) 27.0(0.0) 28.65(–6.3)
17.68 16.39(+7.6) 16.81(+5.2) 17.24(+2.5) 17.67(0.0) 18.51(–4.4)
1. Under each n in the table is shown the UV obtained with the data from the spectoradiometer when Equation 3.11 is applied. To the right of
each value (in parentheses), the percentage of error between these measurements and the UV that appears in the column to t he left is
indicated:100(UV – UVΣ)/UVΣ.

in agreement. The radiometer measurement interval is 3.4.2. Correlation between radiometric

funda- mental for calculation of the photon flux that and spectroradiometric data
reaches the interior of the reactor. The alternative might
be on-line meas- urement of the solar spectrum with a
solar spectroradiometer, but this is very often not possible Now, recalling Equation 2.6, it is possible to standardize
because of the price and/or the labour force required to the UV spectrum up to the radiometer cut-off wavelength.
operate a spectroradiometer. In this case, 404 nm is used. It is also used as an
example in the
Experimental Systems 41

above calculations to clarify the explanation, but similar the number of photons corresponding to the average
reasoning could be applied to any wavelength interval. radia- tion at any given instant for the radiometer is:

λ404  N300404  1.856  1018 (3.16)
fλ  λ404 nmλ , therefor nm
fλ (3.12)
e  =1

 UV λ300
where N is the nmber of incident photons,
λ300 λ 300–404
nm nm
between 300 and 404 nm, per m2 per second, with UV
If the UV radiation spectrum is assumed to have a measured in W m–2 and λ in nm, speed of light c = 2.988 ×
fixed shape similar to that shown in Figure 2.6, a 1017 nm s–1 and Planck’s constant h = 6.626 × 10–34 J s.
standardized spectrum is available in the wavelengths If Einstein (moles of photons) is used as the unit, the
result is:
measured by the radiometer. Therefore, using the
standardized spectrum and the irradiance data (W m–2)
measured by the radiometer, the
N300404 nm
spectral distribution can be calculated for all of these I300404   3.083  106 (3.17)
nm UV
UV*λ  fλ UV (3.13)

where UV*λ are the spectral data calculated with the 3.4.3. Collector efficiency
spectrum (fλ) and the radiometer data. Therefore, the number
of photons is only a function of intensity (measurable in real
time with the radiometers). Once the spectral distribution All the factors that contribute to collector efficiency are
of the radiometer measurements is known, the number of shown in Figure 3.6. Using the combination of these
inci- dent photons per unit of time on surface (N) factors as defined below and Equation 3.17, photon flux
corresponding to those measurements can be found. may be calculated.
Recalling Equation 2.3, which relates the number of
photons from a given polychro- matic source of light to • ηC includes all the errors produced while the collector
the energy corresponding to each wavelength, this can be was built up.
transformed for this case into the fol- lowing (using • ηR,λ is the reflectivity of the parabolic mirror surface
summations of discrete values): (usually aluminium, see Chapter 6). Since the reflection
spectrum does not vary over a wide range, and variation
1 λ404 nm
N300404 nm   UV*λ λ
hc λ300
(3.14 is uniformly distributed over the whole interval
) (300–400 nm), an average spectral distribution may
be considered. However, as the surface is
outdoors and
where N300–404 nm are incident photons. But, as the spectral could become dirty or damaged, it has to be measured
distribution is assumed to be constant, Equations. 3.13 periodically, in which case the value is ηR,i. With
and frequent cleaning (before every experiment), it may
3.14 yield: be considered constant and is then ηR.
UV λ404 nm
N300404 nm 
 fλ λ (3.15 • ηT,λ is the spectral transmissivity of the absorber tube.
λ300 )
nm The solar radiation that reaches ground level without being
Therefore, the wavelength considered equivalent to the absorbed or scattered is called ‘direct radiation’, the radia-
summation would be: tion that has been dispersed but reaches the ground is
λ404 nm ‘global‘diffuse
called radiation’,
radiation’ and 2).
(see Chapter theGlobal
sum ofradiation
both is called
is col-
 fλ λ  368.79 lected directly by the transparent absorber tube without
λ300 nm intervention of the collector. In Figure 3.6 the path
42 Solar Detoxification

To calculate the values corresponding to other intervals, followed by I until it arrives inside the absorber tube is
the same procedure is followed. Returning to Equation shown. It must arrive at the surface and be reflected (part
3.15, is
Experimental Systems 43

calculated for the reactor volume. Therefore, Sp/ST

in Equation 3.20 is substituted by this ratio (collector
area/col- lector volume). In this way, photon flux is
obtained in units congruent with the reaction rate (M s–1)
so that an estimated quantum yield (8 ) similar to that
in Equation 2.1 Ecan be obtained, but using moles of
incident photons inside the tube instead of moles of
photons absorbed by the catalyst.

reaction rate
E  
I E

FIGURE 3.6. A diagram of the various loss factors (η) E

where I* is photon flux in Einstein L –1 s–1 of UV
affecting the photon flux (I) inside a photoreactor.
irradiance incident inside the tube. To calculate the
values correspon- ding to other wavelength intervals, the
lost due to ηR,i) in the right direction by the real mirror same procedure is followed.
surface (ηC), before penetrating (ηT,λ) in the tube.
Furthermore, the parabolic concentration factor must 3.4.4. Actinometric experiments
be considered (the ratio of the surface area of the parabola
capturing the radiation to the surface area of the tube,
Sp/ST) if it is a concentrating solar collector (see Chapter In a chemical actinometer, the photochemical conversion
6). Therefore, the effective photon flux of the UV inside is directly related to the amount of photons absorbed.
the absorber (IE) is: This method has been used since the 1930s, but because
of recent progress in radiation sensors and semiconductor
and elec- tronic equipment development, physical
measuring devices
 SP  have become more popular for photochemistry. In the
I  f I, , η , η ,η (3.18)
E   S C R,i T,λ   case of reactors with simple geometries, they are
preferable because they are very quick, simple and
precise. In the case
in hand, the chemical actinometer is needed for
On cloudy days, only the transmissivity of the glass, ηT,λ,
validation (see Figure 3.5) of all the calculations
affects the radiation reaching the absorber without being
performed to obtain the equations detailed above. A good
reflected by the mirror (1' ). In this case Equation 3.18 is chemical actinometer meets the following specifications:
• The photochemical system should be simple and the
IE  f (I,ηT,λ ) (3.19) reaction should be reproducible, under well-defined and
easily controlled conditions. The quantum yields should
Since ηT,λ depends on the wavelength, in order to evaluate be well known for a wide range of wavelengths if
IE, Equation 3.15 must be recalculated as follows: poly- chromatic wavelength radiation has to be
S  UV λ404 nm • The quantum yield should be independent of the
intensity of radiation, actinometer concentration and
IE   P  ηC  ηR,i   fλ ηT,λ λ (3.20 temperature (large pilot plants cannot be thermo-
 ST  hc λ300
• The reagents and products should be reasonably stable,
where, if UV is W m–2 and the units of I Eare Einstein so errors do not arise between the time the sample is
m–2 s–1. S /S , as accurately as possible, is used to taken and the time it is analysed.
determinep T
IE in each case. For this it is necessary to make the corre- • The analytic methods should be simple and the
sponding trigonometric calculations based on the reagents should be easily synthesized and, even
collector characteristics. Once this is known, the same better, commer- cially available. This is especially
ratio can be important in the case of
44 Solar Detoxification

pilot plants, because of the large volumes of

actinometer that have to be prepared.
3.5. Simplified method for the
• The system should be sufficiently sensitive for low evaluation of solar UV radiation
radia- inside photoreactors
tion intensities, and the evaluation of photons absorbed
should be simple.
The above procedure is usually the most appropriate, but
For this case, a common uranyl-oxalate system provides several parameters are very often not available: such as
an example. Very complete information on actinometric meas- urement of the on-site solar spectrum, collector
sys- tems has been summarized by Kuhn et al. (1989): reflectivity, and absorber transmissivity. In these cases, a
hν shorter proce- dure can be used. Although not the best
UO2  (UO2 )* solution, it is fre-
quently very useful. Nevertheless, a UV radiometer mounted
2 2 (3.22 at the same angle as the solar collector is always necessary
(UO22 )*  H2 C2 O4 UO22  CO  CO 2 H O2 ) for data evaluation. This radiometer sends a signal to a
computer in which the data are stored. As radiation data
Absorptivity of this solution may be estimated are collected continuously, it is very easy to calculate the
from: average incident radiation on the collector surface (UV),
for each period of t,
Absorbed Photons and apply Equation 3.24 to that average. Consequently, the
1 λ
2b (3.23
Incident Photons amount of energy collected by the reactor (per unit of
exp(µ ) ) volume) from the start of the experiment until each sample
where b is the inside reactor radius, 2b is the light collected may be found by:
path length and µλ is the coefficient of absorptivity.
General characteristics of the actinometer and EUV ,n = EUV,n1 + tn
experimental details may be found in Curcó et al. (1996) VTOT (3.24)
and references therein.
Photon flux inside the reactor during the actinometric tn = tn  tn1
experiments may be calculated from Equation 3.20. where tn is the experimental time at which each sample
However, in this case the range of wavelengths is was taken, SP is the collector surface, VTOT is the total
widened to 536 nm (actinometer activity cut-off) and plant volume and EUV,n is the accumulated energy (per
solar spectra up to 536 nm are used. If the photon flux unit of volume, kJ L–1) incident on the reactor for each
inside the reactor and the characteristics of the sample taken during the experiment. This procedure is
actinometer are known, the oxalic acid degradation rate correct only if the rate of the photoprocesses under
can be calculated from IE(300–536), examination has a linear relationship to the intensity of
µλ, the oxalic acid quantum yield in the mixture and the radiation (see Chapter 4). For this purpose, several
area/volume ratios in each case. The comparison different experiments must be carried out in the
between this calculated acid decomposition rate and photoreactor, to determine this effect under very
the rate measured in the actinometric experiments different solar-radiation conditions (morning, noon,
gives an idea of the validity of the equations in Section with and without clouds, summer, winter and so on)
3.4.3. The oxalic acid degradation rate calculated without changing the kinetic order of the reaction.
through the experiments or by Equation 3.21 (combined The results of a typical test are shown in Figure 3.7.
with Eq. 3.23) should be quite close. Therefore, all the The reaction rate is calculated at the beginning of the
equations developed are assumed to be valid. experiment (where reaction rate is constant)
However, actinometry would not be useful for finding corresponding exactly to the period used for the
the photon flux at any given time inside a photoreactor calculation of r. Therefore, the rate is linear with regard
illuminated by solar radiation. The variations in solar to the radiation intensity under the same experimental
intensity as a result of changes in weather, and the conditions and in the same reactor where Equation 3.24
impossibility of using an actinometer inside the reactor is going to be applied. Figure 3.8 shows the
while the photocatalytic experiments are being improvement obtained using this equation to calculate
performed, make it impossible. the reaction rate in a two-day photocatalytic
Experimental Systems 45

same substance using different solar collectors would

indi- cate that their collector efficiency is different.

A pilot plant has to be as versatile as possible in order for
any photocatalytic experiment to be performed with
suffi- cient confidence. The pilot plant must fulfil all the
actual and future necessities of the research to be
performed in it. Plants may be operated in once-through
or recirculation mode. The calculation of residence time
in the reactor is dif- ferent in each case, affecting the
conversion rate. In any case, the use of residence time
to calculate reaction rates (very common in chemical
FIGURE 3.7. Typical photocatalytic degradation (rDCA) in engineering) is not always advis- able because solar
a solar pilot plant under different UV solar light power is not constant. Knowledge of the wavelength
intensities. interval at which the UV radiometers used are active is
basic to data treatment. On-line measurement of
40 1st day 2nd day
UV power is essential and reactor efficiency must

30 60

always be calculated considering, at least, the amount

C, mg L

20 of energy
10 30 incident on it.
0 Therefore, it is necessary to find a relationship between
0 250 500 750 0 radiometer measurements and the photons that actually
t, min reach the reaction. The estimation of quantum yields
Equation 3.21 enables results obtained in the photoreactor
to be compared with others, and thereby makes use of
C, mg L

exist- ing literature on the compounds that are going to

2nd day
be tested.
1st day 30 This is very important when working with a large
reactor, where any test means a considerable outlay of
time and
0 expense. So, A priori, the more information available, the
0 10 20 –1 30 40
fewer experiments are necessary and the faster useful
conclu- sions may be arrived at. Actinometric
experiments have been
FIGURE 3.8. Photocatalytic degradation in a solar pilot shown to be useful to contrast the validity of equations
plant. Concentration (•) is plotted as a function of related to photon flux calculation.
experiment time (top) and accumulated energy (bottom).
Solar UV power throughout the experiment is also
shown. Bibliography
experiment. UV power varies during the day and clouds,
on the first day, make this variation still more noticeable, CASSANO, A.E. 2000. Reaction engineering of
but with Equation 3.24, the data for both days can still be suspended solids in heterogeneous photocatalytic
com- bined and compared with other photocatalytic reactors.
experiments. Consequently, with EUV, the reaction rate CATALYSIS TODAY, No. 58, pp. 167–97.
(r) is expressed in terms of mass of reactant degraded per
kJoule of UV incident on the collector surface. If r is CURCÓ, D.; MALATO, S.; BLANCO, J. AND GIMÉMEZ, J. 1996.
expressed in these units, col- lector efficiency is already Photocatalysis and radiation absorption in a solar plant.
included in it through the use of incident surface SOLAR Energy MATERIALS AND SOLAR Cells, No. 44,
radiation. Different values for r with the pp. 199–217.
46 Solar Detoxification

GOSWAMI, D.Y. 1995. Engineering of solar photocatalytic MALATO, S.; BLANCO, J.; VIDAL, A. AND RICHTER, C. 2002.
detoxification and disinfection processes. In: K.W. Photocatalysis with solar energy at a pilot-plant scale: an
Böer (ed.), ADVANCES in SOLAR Energy. Boulder, Colo., overview. Applied CATALYSIS B: ENVIRONMENTAL, Vol. 37,
American Solar Energy Society. pp. 1–15.
Chemical actinometry. Pure & Applied Chemistry., Vol.
61, No. 2, pp. 187–210.
Experimental Systems 47

Self-assessment questions 2. Why does a pilot plant have to be as versatile as possible?

3. What are the main characteristics of the material to

Part A: True or false? be used for connecting the pilot plant photoreactors?

1. The photocatalytic laboratory microreactors 4. Which are the most important sensors to install
must be open to the atmosphere. in a photocatalytic pilot plant?
2. The power of the pilot plant pump should be low 5. Why is the flow rate very important in
to avoid high cost. batch experiments?
3. The residence time corresponding to each
sample collected at the end of once-through 6. Why is residence time not the best variable to use
experiments is calculated with the following for obtaining kinetic constants in solar
equation: photocatalysis?
tR,i  i illu
VTOT 7. What are the essential data (necessary for obtaining
4. Generally, a once-through operation system can an equation similar to the following) for correlating
be modelled as a plug-flow reactor. radiometric measurements and photonic flux inside a
photocatalytic solar photoreactor?
5. The kinetic constants of photocatalytic processes
can only be obtained by plotting substrate Nλ1λ2
concentration as a
function of time. Iλ1λ2   f (UV)
6. The interval of wavelengths covered in their
is not usually reported in UV instruments’ technical 8. What wavelength is considered equivalent to
descriptions. the following normalized spectrum?
7. The spectral transmissivity of the photoreactor
absorber tube does not affect the photocatalytic reaction TABLE 3.3
λ, nm fλ
8. Quantum yield estimated by using moles of
incident photons is always lower than quantum yield 300 0.0002
calculated by using moles of photons absorbed by the 310 0.0023
catalyst. 320 0.0063
9. The quantum yield of a chemical actinometer 330 0.0109
should increase when the intensity of radiation does. 340 0.0118
10. The amount of energy collected by the 350 0.0124
reactor (per unit of volume) from the 360 0.0127
commencement of the
370 0.0159
experiment until each sample is collected as calculated
by the following equation is only a simplified 380 0.0156
procedure. 390 0.0160

EUV,n = + tn UVn P
tn = tn  tn1 9. If Einsteins (moles of photons) are used as the unit,
what is the number of photons corresponding to the
Part B average radiation at any given instant in the previous

1. Why is it necessary to filter IR in 10. What data are essential for radiation evaluation in a
laboratory photoreactors? solar photocatalytic reactor?
48 Solar Detoxification

Answers 4. Temperature, pressure, dissolved oxygen and

solar irradiation sensors.
Part A 5. Because it is directly related to the conversion rate
per pass and, therefore, to the modelling of the
1) False. 2) False. 3) False. 4) True. 5) False. 6) True. photoreactor.
7) False. 8) True. 9) False. 10) True.
6. Because when time is the independent variable,
differences in the incident radiation in the reactor
Part B during an experiment are not taken into account.

1. Because IR beams could heat the slurry suspensions 7. Solar spectra.

very quickly.
λ390 λ390
2. Because of the need to accomplish all the
fλ  1; nm
fλ λ  373.6 nm
experiments with confidence and fulfil all the actual 8.  
λ300 λ300
and future nm nm
requirements of the research to be performed in 9. 3.135 × 10–6 UV (Einstein s–1 m–2).
10. Radiation data from a radiometer mounted at the
3. All the tubes and valves must be strongly resistant to same angle as the solar collector.
chemicals, weatherproof and opaque, in order to avoid
any photochemical effect outside the photoreactors.
Fundamental Parameters
in Photocatalysis

Aims 3. Comprehend the influence of

catalyst concentration and know how
to find the optimum catalyst mass in an
This chapter describes the fundamental experimental reactor.
parameters related to heterogeneous
photocatalysis reactions. Examples are
4. Understand how a Langmuir–
shown to provide a better understanding of
Hinshelwood (L–H) model, although it
oxygen, pH, catalyst concentration, initial
does not explain the photocatalytic
substrate concentration, radiation intensity
mechanisms, may be used in
and the effect of temperature. The chapter
heterogeneous photocatalysis to obtain the
outlines the basic tests used to understand
kinetic constants for reactor optimization.
experimental system behaviour when these
parameters change, and why these changes
affect the photocatalytic reaction rate. 5. Arrange experiments to obtain L–
Finally, an overview of the different ways H parameters.
of evaluating quantum yields is presented.
6. Describe the interdependence of the
photocatalytic reaction rate and
Objectives illumination intensity and know when
this relationship is directly proportional.
By the end of this chapter, you will
understand the main aspects of 7. Determine useful parameters for
experimental photocatalysis variables and describing photocatalysis efficiency
you will be able to do seven things: without knowing the amount of photons
absorbed by the system, and understand
1. Understand why blank experiments why quantum yield is almost impossible to
are always necessary in photocatalysis. calculate in heterogeneous photocatalysis
experimental systems.
2. Design experiments that avoid the
effects of oxygen and pH and thereby
reduce the number of variables affecting
the tests.
50 Solar Detoxification


Symbol Description Units

C or Ci Reactant concentration M or mg L–1
C0 Initial reactant concentration M or mg L–1
CS Solvent concentration M or mg L–1
E Activation energy kJ mol–1
I Photon flux corresponding to the UV inside a photoreactor Einstein L–1 s–1
k' Apparent reaction rate constant min–1
Kr Reaction rate constant M min–1
K Reactant adsorption constant M–1
KS Solvent adsorption constant M–1
L–H Langmuir–Hinshelwood model
NA Number of photons absorbed Photons s–1
pKa –log (acid ionization constant)
pO2 Oxygen partial pressure atm
r Reaction rate M min–1
An Number of reacting molecules Molecules s–1
8 Quantum yield
θx Fraction of surface covered by the substrate
ζr Relative photonic efficiency

4.1. Direct photolysis in the photoreactors, it is necessary to carry out these tests
at the beginning of the experiment, before the catalyst
comes into contact with the photoreactors. In pilot-plant-
As mentioned in Chapter 1, some pollutants can only be scale experi- ments, removal of the thin coating of catalyst
dissociated in the presence of UV light. For this, the on the tubes after TiO2 suspensions have circulated through
pollutant must absorb the light of the lamp (or the Sun) them is a very diffi- cult, complex and expensive task. In
with a reason- able photodissociation quantum yield. laboratory tests, TiO2 can only be removed with an
Although organic pol- lutants absorb light over a wide ultrasonic cleaner or by abrasion. This type of experiment
range of wavelengths, absorption is generally stronger at will focus on demonstrating whether the following are true
the lower wavelengths. However, such natural or evaluating their importance:
photodegradation is usually very slow. For example, at
least ten days under perfectly sunny condi- tions is • The treatment is not feasible without a catalyst.
necessary to reduce 50 mg /L acrinathrin to half this • Increase in temperature (due to illumination) in
quantity. One-half of 100 mg/L pentachlorophenol at pH the reactor does not cause product loss.
7.3 decomposes in forty-eight hours. So the photolytic • There is no adsorption of pollutant or its metabolites
reaction rate is usually different from one substance to in the materials of the pilot plant.
another, even in the same experimental device. The
photolytic half-lives of hun- dreds of substances have been After these tests have been performed, the photocatalytic
summarized by Tomlin (1997). In any case, the focus here experiment results may be considered accurate and the
is on fundamental photo- catalytic parameters and kinetic parameters can be determined properly. Any side-
therefore the photolytic effect will be discussed from this effect of the photocatalytic reaction rate can be
point of view. Tests have to be performed in order to find quantified and subtracted from the global rate, resulting
out the decomposition rates without the influ- ence of the in the real photocatalytic reaction rate.
semiconductor. As TiO2 readily sticks to the glass
Fundamental Parameters 51

4.2. Influence of oxygen consumed, photodecomposition of TOC comes to a

halt. At the moment injected oxygen reaches the reactor,
photo- decomposition continues. Therefore, injection of
In semiconductor photocatalysis for water purification, pure O2 becomes necessary in once-through experiments
the pollutants are usually organic and, therefore, the (see Figures
overall process can be summarized by Equation 4.1. 3.3 and 3.4) at low flow rates. At high flow rates or with
Given the reaction stoichiometry of this equation, there recirculation, the addition of oxygen is not always
is no photo- mineralization unless O 2 is present. The necessary since the illumination time per pass is short.
literature provides a consensus regarding the influence of The water again recovers the oxygen consumed when it
oxygen. It is necessary for complete mineralization and reaches the tank (open to the atmosphere and stirred).
does not seem to be in com- petition with other reactives
during the adsorption on TiO 2 since the places where
oxidation takes place are different 4.3. pH influence
from those of reduction+
(see Figure 1.5). The O 2 avoids the
recombination of e–/h (Eq. 1.24) and photoactivated
oxygen (O2•–) also reacts directly (Table The pH of the aqueous solution significantly affects TiO2,
1.3). including the charge of the particle and the size of the
aggregates it forms. The pH at which the surface of an
organic pollutant     
ultra-bandgap (4.1 oxide is uncharged is defined as the ‘zero point charge’
O2 energy )
CO2  H2O  mineral acids (pHzpc), which for TiO2 is around 7. Above and below
this value, the catalyst is negatively or positively
The concentration of oxygen also affects the reaction charged according to:
rate, which is faster when the partial pressure of oxygen
TiOH  TiOH  H
(p ) in 2
the atmosphere in contact with the water increases. In 2 (4.2)
any TiOH  TiO  H
case, it seems that the difference between using air (pO2 (4.3)
0.21 atm) or pure oxygen (pO2 = 1 atm) is not drastic
Figure 4.1). In an industrial plant this would be The equilibrium constants of these reactions
purely a (Hoffmann et al., 1995) are pKTiOH+ = 2.4 and pKTiOH =
matter of economy of design. In Figure 4.1 it is clear that 8.0, the abun- dance of2 all the species as a function of pH
when all the oxygen contained in the water has is given by –TiOH

been Š 80 per cent when 3 < pH < 10; –TiO Š 20 per cent if pH
> 10; –TiOH2 Š 20 per cent when pH < 3. Under these con-
3.0 ditions, the photocatalytic degradation of the ionizable
organic compounds is affected by the pH. At first sight,
2.5 O2 and for pollutants for which pKa is outside the range of 1
to 13, a very acidic solution appears to be detrimental
Reaction rate

2.0 and a very basic solution to be favourable, since the

variations are mod- est or non-existent around

neutrality. Because even at extreme pHs the change in
the photocatalytic rate is generally less than one order of
magnitude, the TiO2 water treatment definitively
0 20 40 60 80 100
possesses an advantage over other processes. In many
cases, a very important feature of photocatalysis is not
0.5 taken into account when it is to be used for
TOC decontamination of water: namely, that during the
0.0 reaction a multitude of intermediate products are
0 10 20 30 40
produced that may behave differ- ently depending on the
Illumination time, min pH of the solution. To use only the
FIGURE 4.1. Effect of the concentration of dissolved oxygen
on photocatalytic mineralization. = 8.5 mg L–1. In rate of decomposition of the original substrate could yield an
[O2]0 the
inset, the usual effect of partial pressure of oxygen on the photocatalytic reaction rate is shown.
52 Solar Detoxification

erroneous pH as the best for contaminant degradation.

Therefore, a detailed analysis of the best pH conditions
Fundamental Parameters 53

3,000 4.4. Influence of catalyst

pHZPC = 6.9 concentration
Mean size, nm

2,000 Whether in static, slurry or dynamic flow photoreactors,

the initial reaction rates were found to be directly
1,500 proportional to catalyst mass. This indicates a truly
heterogeneous catalytic regime. However, above a
certain value, the reaction rate levels off and becomes
500 independent of catalyst mass. This limit depends on the
geometry and working conditions of the photoreactor and
0 is calculated for a definite amount of TiO2 in which all
2 4 6 8 10 12
pH the particles – that is, all the surface exposed – are
totally illuminated. When catalyst concentra- tion is very
high, after travelling a certain distance on an opti-
FIGURE 4.2. Mean particle size of TiO2 (P-25) suspended in cal path, turbidity impedes further penetration of light in
water versus pH [TiO2] = 0.2 g/L. the reactor. In any given application, this optimum
catalyst mass ([TiO2]OPT) has to be found in order to
avoid excess catalyst and ensure total absorption of
efficient photons.
There are a number of studies in the literature on the
should include not only the initial substrate but also the influence of catalyst concentration on process efficiency.
rest of the compounds produced during the process. The results vary widely, but from all of them it may be
The mean TiO2 particle size (presented in Figure deduced that radiation incident on the reactor and path
4.2) has been found to be constant at pH far from pH = length inside the reactor are fundamental in determining
7. Sizes increase to 2 to 4 µm when dispersion reaches the optimum catalyst concentration:
pHzpc. The zero surface charge yields zero electrostatic • If the lamp is inside the reactor and coaxial with it
surface potential, which cannot produce the interactive (see Figure 4.3a), [TiO2]OPT is very high (to the order
rejection necessary to separate the particles within the of several grams per litre) if the path length is short
liquid. This induces a phenomenon of aggregation and (several mm). On the other hand, [TiO 2]OPT is low
the TiO2 clusters become larger. The large mean size in (hundreds of mg per litre) if several centimetres are
suspension at pH = 7 becomes much smaller at pH far crossed (see Figure
from 7. This effect is clearly related to the capability of
4.4 solid circles).
the suspension of transmitting and/or absorbing light.
Furthermore, larger clusters sedi- ment more quickly • If the lamp is outside the reactor (see Figure 4.3c), but
than small particles; thus the agitation necessary to the path length is short (1 to 2 cm max.), the maximum
maintain perfect homogeneity must be more vigorous. In rate is obtained with 1 to 2 g L –1 of TiO (see Figure
4.4 solid squares). 2
contrast, these variations in particle size could be an
advantage when separating the catalyst from water (by • If the lamp is outside, but the path length is several
sedimentation and/or filtration) at the completion of the cen- timetres long, similar to a reactor illuminated by
photocatalytic treatments. solar radiation (see Figure 4.3b and 3.6), the
appropriate catalyst concentration is several hundreds
To unambiguously exclude the pH effect, the reagents
of milligrams per litre. An example of this is shown in
used to change the pH must contain counterions that
Figure 4.4, where open circles correspond to a
have no effect on the rate of water treatment. Sodium
photoreactor with a large diameter and solid triangles
hydroxide and hydrogen chloride or sulfuric acid have
to one with a smaller diame- ter, but both of several
generally been chosen to produce a basic or acidic pH,
centimetres. In this case, it is very clear that the
respectively. Organic buffers must be avoided because
optimum rate is attained at lower catalyst
they are potential con- sumers of •OH. For evident
concentrations when the photoreactor diameter is
reasons this is the same for toxic inorganic acids or bases.
In all of the cases described above, a ‘screening’ effect is
produced when the TiO2 concentration is very high.
54 Solar Detoxification

UV Lamp
UV path length


Reaction chamber water

UV UV Lamp



a) b) c)
FIGURE 4.3. Different laboratory photoreactor designs and zones of radiation penetration when illuminated in different ways.

reaction rate diminishes because of the excessive 1.2

opacity of the solution, which prevents the catalyst
further in from being illuminated. Besides, the larger 1.0
the size (see Figure 4.2), the less the opacity of the
suspension. When the radiation comes from a parabolic
Relative rate

trough collector, some- thing similar to what is shown
in Figure 4.3b occurs. In any case, these are only 0.6 TOC
approximations based on the results obtained by 1.0
Relative rate

different authors and, because of all of the above, it

is always necessary to find out experimentally the 0.5
optimum catalyst concentration for the plant studied.
That is, the minimum concentration at which the 0.0
maximum reaction rate is obtained. But it does not seem 0.0 0.2 0.4 –1 0.6
TiO2, g L

to be necessary to test a very wide range of 0.0

0.0 0.4 0.8 1.2 1.6 2.0
concentrations. This effect is shown in Figure 4.4. –1
TiO2, g L
Usually, the reaction rate increases very quickly with
TiO2 concentration, but only at low catalyst FIGURE 4.4. Influence of catalyst concentration on the rate
concentrations (usually below 100 mg/L). After that, of photocatalysis (normalized rates have been used to
the reaction rate stabilizes, and at very high catalyst make it more easily understood) in different reactors (see
concentra- tions the reaction rate decreases. When this text). The continuous line is only to clarify the tendency.
point is found, it is not really necessary to continue In the inset, relative reaction rates of degradation (two
checking any higher because no more useful different photore- actors) and mineralization (TOC) of
information is going to be obtained. As shown in Figure the same contaminant are shown. See text for comments.
4.4, initial substrate behaviour is not always the same as
that of total organic carbon (TOC). This is because of the during the photocatalytic reaction. That is, the screening
influence of intermediate products generated effect influences the decomposition of the original
product in
Fundamental Parameters 55

a different way than the other organic species present in According to the L–H model, the reaction rate (r) is
the reaction and, in the example presented, to mineralize proportional to the fraction of surface covered by the
the con- taminant it is unnecessary to use more than 0.2 g substrate (θx). In each case the following expression
L–1 of TiO . 2
can be obtained:
4.5. Initial contaminant dC k r KC
r  kr θx 
concentration influence dt 1  KC  K (4.4a)
s Cs

dC k KC
As oxidation proceeds, less and less of the surface of the r= = kr θ x = r (4.4b)
TiO2 particles is covered, as the contaminant is dt 1  KC
Evidently, at total decomposition, the rate of degradation is where kr is the reaction rate constant, K is the reactant
zero. A decreased photocatalytic rate is to be expected with adsorption constant, C is the concentration at any time, KS
increasing illumination time (or accumulated energy, see is the solvent adsorption constant and CS is its
Eq. 3.24). Most authors agree that, with minor variations, concentration (in water CS = 55.5 M). As CS >> C and CS
expres- sion of the rate of photomineralization of organic remains practically constant, the part of the catalyst covered
substrates with irradiated TiO2 follows the Langmuir–
by water is unalter- able over the whole range of C and the
Hinshelwood (L–H) law for the same saturation-type
previous equations can be integrated:
kinetic behaviour in any of four possible situations:
C0 K k K
1. The reaction takes place between two adsorbed ln + (C  C)  r (4.5a)
C 1  Ks Cs 1
substances. KsCs
2. The reaction occurs between a radical in the solution
the adsorbed substrate. C0
ln  0 C)  kr Kt (4.5b)
3. The reaction takes place between the radical linked to K(C C
surface and the substrate in the solution.
4. The reaction occurs with both species in solution. When C0 is very small, both equations can be reduced to an
order-one reaction rate equation:
In all cases, the expression of the equation rate is similar
the L–H model. From kinetic studies only, it is therefore not C0
ln = kt (4.6)
possible to find out whether the process takes place on the C
surface or in solution. Although the L–H isotherm has So, if ln(C0/C) is plotted against t (or the accumulated
been rather useful in modelling the process, it is energy, see Eq. 3.24), a line, the slope of which is the
generally agreed that both rate constants and orders are apparent reaction rate constant k', should be obtained
only ‘apparent’. They serve to describe the rate of (see Figure 4.5). Likewise, at higher concentrations, both
degradation, and may be used for reactor optimization, equations can be simplified by adjusting them to zero
but they have no physical meaning, and may not be used
order, (C0 – C) = krt, as might be the case at the
to identify surface processes. Thus, while not a useful
beginning of the experiment represented in Figure 4.5.
tool for describing the active species involved in
Using an L–H model, graphics similar to those
oxidation, engineers and solar designers seem to have a
depicted in Figure 4.6 may be obtained from the
common understanding of the usefulness of the
experimental data and from the linearization of the
unmodified L–H model.
previous equations. The effect of the initial concentration
As a consequence of the above, for L–H standard data on the degradation rate is shown in Figure 4.6a, where,
treatment, it is assumed that the reaction occurs on the due to the saturation produced on the semiconductor
sur- face, which is also the assumption most widely
surface as the concentration of the reactant increases, a
accepted as possible. Under these conditions, two point is reached at which the degrada- tion rate becomes
extreme situations are defined to illustrate the adsorption
steady. Figure 4.6b shows a linearization of Equations
on the catalyst surface: 4.4, where the slope of that straight line is (1 +
• Substrate and water compete for the active catalyst KSCS)/krK, but could also be 1/krK. Finally, Equations 4.7
are obtained from Equations 4.5 when the concentration
is half
• The reactant and the solvent are adsorbed on the
surface without competing for the same active
catalyst sites.
56 Solar Detoxification

of the initial concentration (C/C0 = 0.5): are held constant. Because of this, a series of tests at
different initial substrate concentrations has to be
0.693(1  KSC S ) (4.7a performed to demonstrate whether the experimental
t 0
0.5C  )
results could be
kr K kr adjusted in line with this model. The concentration range
has to be wide enough to allow correct fit of the L–H
0.693 0.5C0
t1 / 2   (4.7b lineariza- tion. This means, from the lowest
kr K kr ) concentration at which the initial rate could be
determined until the limit where the rela-
tionship between initial reaction and initial concentration
It should be emphasized that photodecomposition remains constant (see Figure 4.6). The results shown in
gives Figure 4.5 correspond to an example of each of the
rise to intermediates, which could also be adsorbed experi- ments to be carried out for this purpose. From the
compet- itively on the surface of the catalyst. The slope of the line corresponding to the starting points of
concentration of these intermediates varies throughout each of the exper- iments, the initial degradation rate can
the reaction up to their mineralization, and thus be calculated. Figure
Equation 4.4 may also take the following form: 4.7 shows the initial rate calculated in this way. From 0.2 to
0.4 mM of substrate the initial rate is steady. At this concen-
k r KC tration, catalyst saturation occurs and the reaction rate
r n
becomes constant.
1  KC   K i C i (i  The constants can be calculated from the graphic
1,n) inserted in Figure 4.7 using the L–H model.
where i is the number of intermediates formed during
degradation (the solvent is also included in the krKC0 1 1
summation). An understanding of the reaction rates and r0 = 1 + KC  1 = 1 +k K C
r k (4.9)
how the
0 0 r r 0
reaction rate is influenced by different parameters is
impor- tant for the design and optimization of an kr= 1.57 mM h–1
industrial system. The L–H reaction rate constants are K = 16.75 mM–1
useful for comparing the reaction rate under different
experimental conditions. Once the reaction constants kr With known kr and K, the time required to degrade a
and K have been evaluated, the disappearance of definite substrate concentration (C0) down to a certain
reactant can be estimated if all other factors level (C) may be found from Equation 4.5. With this,
reactor volume and surface area (see Eq. 3.2 or 3.3) can
be defined. If accumu- lated energy is used instead of
1.0 time, then Equation 3.24 should be used for obtaining
(C0 - C)=k t r the kinetic constants and, afterwards, for reactor design.
Normalized concentration

Ln(C0 /C) = k’t
ln (C0/C)

0.6 1.5

0.4 0.5

0.2 0 5 10 15

EUV, kJ/L or time, min

0 5 10 15 20 25 C0 C0 C0

EUV, kJ/L or time, min a) b) c)

FIGURE 4.5. Typical photocatalytic degradation. The FIGURE 4.6. Graphics related to the adjustment of data to an
insert shows data adjusted to Equation 4.6. L–H type kinetic model.
Fundamental Parameters 57

2.0 2.0 A  e   A (4.10c)

D h  D (4.10d)
1.5 3 1.5 A  D  Intermediates  (4.10e)
 Products

r0, mMh

1.0 
and the rate-limiting step is the reaction in the adsorbed

 phase (Eq. 4.10e). Therefore:
1 (kr K) = 0.0381

0.5 –1
(kr ) = 0.638
r  r  k A  D  (4.11)
0 e e   
0 25 50 75
 1/C0
0.0 0.0
0.0 0.2 0.4 0.6 0.8 1.0 In an n-type semiconductor such as titania, the photo-
C0, mM induced holes are much less numerous than electrons
(photo-induced electrons plus n-electrons): [p+]<<[e–].
FIGURE 4.7. Initial degradation rate as function of the Therefore holes are the limiting active species.
initial substrate concentration. The insert shows the Thence:
linear trans- formation of Equation 4.4, from which the
rate constant and the adsorption coefficient can be r  re  rd  kd[D]p 
estimated from the inter-
cept and the slope,   (4.12)
respectively. 
At any instant, one has:

   
d  p   r  r  r  0
4.6. Radiant flux influence A b d 
dt p  
 kA  kb e    p    kd [D]
     
Since 1990, the kind of solar technology that should be
involved in detoxification has been clarified. Initial   
experi- ments with parabolic troughs for water and :   
dishes or furnaces for gas-phase treatments have evolved   
to lower flux systems. The reason for using one-sun p   kA I (4.14)
systems for water
treatment is firmly based on two factors: first the   kb e    kd [D]
high percentage of UV photons in the diffuse
component of solar radiation and, second, the low-order
dependence of rates on light intensity. It has been and:
demonstrated by experiment that above a certain UV
photon flux, reaction rate dependency on
k k [D]I
intensity goes down from one to a half order (Mills et r  kb eA d  kd[D] (4.15)
any Herrmann, 1995). This does not seem to occur at
 
particular radiation intensity, as different researchers obtain
different results, but presumably is significantly affected
by experimental conditions. Some authors impute the From the above equation, it can be seen that the reaction
rate is directly proportional to light flux. In the case of
transition of r = f(I1.0) to r = f(I0.5) to the excess of
high fluxes, the instantaneous concentrations [e–] and [p+]
photogenerated species (e –, h+ and •OH). This can be
d than k [D] and [e ] = [p+].
become much larger –
demonstrated as follows. According to Section 1.4, the
five basic simplified equations are: Therefore Equation 4.14 becomes:
TiO2  hν  e h  2  k I
(4.10a kI
  h  N 
eenergy  p  k Ae ; therefor  p   A (4.16)
)  b  kb
(4.10b)  
e 
58 Solar Detoxification

The reaction rate becomes: effect should be measured experimentally in each device,
but this limit (several Suns) is usually accepted as a
 k I 1 / 2 general rule.
r  re  rd  kd [D] A  (4.17)
 kb 
4.7. Temperature effect
which means that r is proportional to I0.5 and that the rate
of electron-hole formation is greater than the rate of Because of photonic activation, photocatalytic
photocatalysis, which favours electron-hole systems do not require heating and operate at room
recombination. Therefore optimum use of light power temperature. The true activation energy Et is nil,
corresponds to the region where r is proportional to I. whereas the apparent activation energy EA is often very
At higher radiation intensities, another transition, low (a few kJ/mol) in the medium temperature range
from r = f(I0.5) to r = f(I0), is produced. At this moment, (20 ºC to 80 °C). However, at very low temperatures
the photo- catalytic reaction is no longer dependent on (–40 °C to 0 °C), activity decreases and activation
the radiation received, depending only on mass transfer energy EA becomes positive. By contrast, at ‘high’
within the reac- tion. Thus, the rate is constant although temperatures (>70 °C to 80 °C), for various types of
the radiation increases. This effect may be due to photocatalytic reactions, the activity decreases and
different causes, such as a lack of electron scavengers the apparent activation energy becomes negative. This
(that is, O2) or organic mole- cules in the proximity of the behaviour can be easily explained within the frame of
TiO2 surface and/or an excess of products occupying the Langmuir- Hinshelwood mechanism described
active centres of the catalyst. Actually, these phenomena above. The decrease in temperature favours
appear more frequently when working with a supported adsorption, which is a spontaneous exothermic
catalyst and/or at slow agitation speeds, which implies phenomenon. In Equation 4.4, θ tends towards unity,
less catalyst surface is in contact with the liq- uid and whereas KC becomes >>1. In addition, the lowering
there is less turbulence. It does not favour reactant temperature also favours adsorption of the final
contact with the catalyst or dispersion of products from reaction products, desorption of which tends to be the
the proximity of the catalyst to the liquid. rate-limiting step. To the contrary, when the temperature
Many articles on this aspect of photocatalysis increases above 80 °C, nearing the boiling point of
provide information on the light intensity at which the water, the exothermic adsorption of reactants is
change of order is produced. The values found are very disfavoured – this tends to become the rate-limiting step.
dissimilar. It may only be intuited that at an intensity of In addition to these mechanical effects, other conse-
several Suns (1 Sun = 22 W m–2), quantum yield quences of plant engineering must be considered. If
UV ThisUV temper- ature is high, the materials (see Section 3.2)
used for the plant should be temperature resistant (more
expensive), and oxygen concentration in water decreases.
Consequently, the optimum temperature is generally
between 20 °C and 80 °C. This absence of need for
r=k I½½
heating is attractive for photo- catalytic reactions carried
Reaction rate

r=kI out in aqueous media, and in par- ticular for

environmental purposes (photocatalytic water
purification). There is no need to waste energy heating
r=kI water that already possesses a high thermal capacity.

4.8. Quantum yield

Fotonic flux
In photochemistry, a concept called ‘quantum yield’ is used to
FIGURE 4.8. The relation between the photocatalytic evaluate the results obtained and to compare different
reaction rate and the intensity of the radiation received. exper- imental conditions. Recalling Equation 2.1 (Φ = An
N –1), A
Fundamental Parameters 59

the quantum yield of a photochemical reaction is defined organic pollutants and, like many of these, it is
with regard to the number of reacting molecules and the essentially degraded by oxidation rather than reduction.
number of photons absorbed.
In this book, a heterogeneous system made up of a rate of disappearance of
ζ (4.18)
sus- substrate
rate of disappearance of
pended solid (TiO ), a gas (O ) in bubbles and/or phenol
2 2
in an aqueous solution of a multitude of compounds
(initial where ζr is called ‘relative photonic efficiency’. When the
substrate, intermediates, H +, anions) is described. reaction rate for the test substances and phenol
Finding out the amount of photons absorbed by the (secondary actinometer) are obtained under identical
catalyst from the behaviour of the radiation incident on a experimental con- ditions, there is no need to measure
suspension such as this is very difficult. In order to the photon flux. The use of relative photonic efficiency
calculate this, if so desired, one would have to: renders comparison of process efficiencies between
studies carried out in different laborato- ries or pilot
• evaluate the light absorption of a very complex plants possible, because ζr is basically independ- ent of
reactive mixture, which, moreover, changes its the fundamental photocatalysis parameters (light
composition throughout the reaction; intensity, reactor geometry and TiO2 concentration for
• from this basis, determine the photon flux that arrives a given catalyst). However, it depends on the initial
at each particle of the catalyst to photoactivate it; and concentra- tion of substrate and on temperature. In any
• estimate the photons absorbed and dispersed. case, based on initial rates of degradation, ζr illustrates
only one aspect of photodegradation and is also useful to
Furthermore, it seems that this, for the moment, is a compare different photocatalyst materials for water-
difficult undertaking (Serpone et al., 1996). Remember treatment purposes.
that, in het- erogeneous catalysis, the reaction rate is An example of an application of ζr to degradation
usually expressed as a function of the grams of catalyst. of pesticides (lufenuron and propamocarb) under
In photocatalysis, it should include the number of active sunlight is shown in the following paragraphs. Phenol
centres, as well as the surface area of the catalyst. But as a (20 mg/L) has been used to calculate ζr following the
consequence of the above comments, the number of method proposed above. In Figure 4.9 the experiments
active centres is unknown and the surface of catalyst carried out with phenol are shown.
exposed to light is undetermined. The mineralization rate (measured by TOC analysis)
This is therefore simplified by considering only the is also included because efficiencies based on the
radia- tion of a certain wavelength (Eq. 3.21) incident on disappear- ance of organic carbon (ζr,TOC, Eq. 4.19)
the inside of the reactor in calculating NA. The value provide more practical information.
obtained from this is called the ‘estimated quantum yield’:
ΦE. No distinction is made between the photons ζr,TOC
corresponding to each wave- length, assuming that all
of them have the same effect on the
surface of the catalyst. This simplification has been rate of disappearance of substrate TOC (4.19)
accepted as valid by a multitude of authors and is widely 
rate of disappearance of TOC from
used in the bibliography. Consequently, quantum yields phenol
have some-
times been reported as lower limits, not allowing for Phenol and pesticide experiments have been performed
scat- tered light. A simple means of assessing process under exactly the same conditions. With sunlight, it is
efficiencies for equal absorption of photons is therefore not possible to work under constant illumination
desirable in het- erogeneous photocatalysis (Eq. 4.18). conditions. Therefore, Equation 3.24 is used to avoid
The initial photocon- version of phenol has been chosen this uncertainty and reaction rates used to determine ζr
as the standard process, and Degussa P-25 titania as the and ζr,TOC are calcu- lated using EUV instead of time.
standard photocatalyst. This compromise has been Relative photonic efficiencies of the two pesticides tested
adopted by a group of scientists (belonging to different are reported in Table 4.1. All the photonic efficiencies
research groups) considered among the most important in are lower than one, indicating that the initial
this field in the world (Serpone et al., 1996). The choice photocatalytic oxidative degradations of the test sub-
of phenol was dictated by the recognition that the stances, at the selected initial concentration, are less
molecular structure of phenol is present in many efficient than for phenol.
60 Solar Detoxification

1.0 1.0 equation that fits the degradation experimental data

accord- ingly. This equation is useful in a wide range of
initial con- centrations and is necessary for engineering
Normalized concentration

0.8 Average
Average 0.8
plant design. Experimentation at pilot-plant level is
Average essential to obtain these equations. Above a certain flux of
0.6 0.6
UV photons, the reac- tion rate changes depending on
intensity. It may only be intu- ited that at intensities of
0.4 0.4 several Suns, quantum yield diminishes. This effect
should be measured experimentally in each device, but
0.2 0.2 this limit is generally accepted. The optimum temperature
is generally between 20 °C and 80 °C. The use of relative
0.0 0.0 photonic efficiencies renders comparison of process
0 5 10 15 20 25 effectiveness between studies carried out in different
EUV, kJ/L reactors
FIGURE 4.9. Plots of normalized concentration as a possible.
function of accumulated energy for the
photodegradation and mineralization of phenol. C0 = 20
mg/L, TiO2 = 200 mg/L, pH0 = 5. Bibliography
TABLE 4.1. Relative photonic efficiencies for the two pesti-
HERRMANN, J.M. 1995. Heterogeneous photocatalysis: an
cides treated, with phenol as the standard reference (C0
emerging discipline involving multiphase system.
= 20 mg/L).
CATALYSIS TODAY, No. 24, pp. 157–64.
Substrate Rate µmol/kJ ξr ξr,TOC
Propamocarb 5.8 0.21 0.48 1995. Environmental applications of semiconductor
Lufenuron 8.8 0.32 0.73 photocatalysis. Chemistry Review, No. 95, pp. 69–96.


Summary Water purification by semiconductor photocatalysis.
CHEMICAL Society Review, No. 22, pp. 417–25.
Photolysis tests always have to be performed before photo-
catalysis tests in order to find out decomposition rates with- SERPONE, N.; SAUVÉ, G.; KOCH, R.; TAHIRI, S.; PICHAT,
out the presence of the semiconductor. The water to be P.; PICCINNI, P.; PELIZZETTI, E. AND HIDAKA, H. 1996.
treated must contain enough dissolved oxygen. To unam- Standardization protocol of process efficiencies and
biguously exclude the effect of pH, the reagents used to activation parameters in heterogeneous photocatalysis:
mod- ify the pH must contain counterions that have no relative photonic efficiencies ζr. JOURNAL of Photochemistry
effect on the rate of water treatment. The optimum catalyst AND Photobiology A: Chemistry, No. 95, pp. 191–203.
concentration always depends on the experimental device
used and, there- fore, must be always tested first. Direct TOMLIN, C.D.S. 1997. The Pesticide MANUAL, 11th ed.
application of the Langmuir–Hinshelwood model produces Farnham, Surrey, British Crop Protection Council.
an empirical
Fundamental Parameters 61

Self-assessment questions 3. What is the main difference between bubbling

oxygen or air inside a photoreactor?
Part A: True or false? 4. To unambiguously exclude the pH effect, what are
the reagents most often used to change the pH?
– +
1. Oxygen avoids recombination of e /h and promotes 5. Why are L–H constants called ‘apparent’ in
• OH formation. photocatalysis?

2. The reagents used to change the pH must 6. Calculate the L–H constant corresponding to
contain organic buffers. the following experimental data taken from a
series of photocatalytic experiments at different
3. The optimum concentration of the catalyst is initial concentrations:
always around 1 g /L.
TABLE 4.2.
4. Initial substrate behaviour is the only
t, min C, mg L–1 C, mg L–1 C, mg L–1 C, mg L–1 C, mg L–1 C, mg L–1
fundamental parameter for choosing [TiO2]OPT.
0 2.5 5 12 23 50 100
5. By applying the Langmuir–Hinshelwood (L–H) 5 2 4 10 20 45 90
rate law, it is possible to find out whether the
photocatalytic process occurs on the surface or in 10 1.5 3 8 17 40 80
solution. 15 1 2 6 14 35 70
20 0.5 1.5 5 11 30 60
6. When C0 is very small, L–H isotherm equations can
be reduced to an order-one reaction rate equation. 25 0.3 1 4 9 25 50
30 0.1 0.5 3 7 22 45
7. It has been demonstrated that above a certain 35 2 5 19 40
UV-photon flux, reaction rate dependence on intensity
40 1.5 4 16 35
changes from one to a half order.
45 1 3 13 30
8. Temperature does not affect the 50 2 11 25
photocatalytic reaction rate.
55 9 22
9. The amount of photons absorbed by the catalyst 60 7 19
must be known to compare photocatalytic reactors. 65 5 16

10. Simplifications must be employed to 70 4 13

calculate photocatalytic efficiencies.

7. What is the most accepted reason why
photocatalytic reaction rate dependence on intensity
changes order?
Part B
8. What is the usual UV power of light intensity at
1. Why is it necessary to carry out a blank test which the change of order is produced?
before putting TiO2 in a photoreactor?
9. What is the optimum temperature range
for heterogeneous photocatalytic reactions?
2. What are blank photocatalysis experiments mainly
for? 10. Why is a secondary actinometer necessary?
62 Solar Detoxification

Answers 5. Because with kinetic studies only, it is impossible

to determine whether the process is on the surface or in
solution, and therefore the constants obtained have no
Part A physical meaning, and may not be used to identify
surface processes.
1) True. 2) False. 3) False. 4) False. 5) False. 6) True.
6. kr = 2.19 mg L–1 min–1; K = 0.0192 (mg/L)–1
7) True. 8) False. 9) False. 10) True.
7. The transition of r = f(I1.0) to r = f(I0.5) is usually
Part B
1. Because readily sticks to the materials of attributed to the excess of photogenerated species (e–,
TiO2 the h+ and •OH).

2. This type of experiment evaluates the effect of the 8. Above 50 to 75 WUV per square meter of
treatment without catalyst, the increase in temperature photoreactor surface, the quantum yield diminishes.
in the reactor and the adsorption of the pollutant or its
metabolites in the materials of the pilot plant. 9. Between ambient temperature and 60 to 70 ºC.

3. The concentration of O2 inside the reactor. By 10. Because ζr is basically independent of fundamental
bubbling oxygen, the O 2 concentration is almost five photocatalysis parameters (light intensity, reactor
times greater than by bubbling air. geometry and TiO2 concentration for a given catalyst).

4. Sodium hydroxide and hydrogen chloride or

sulfuric acid.
Fundamental Parameters 63


Water Decontamination
by Solar Detoxification

Aims 2. Understand how it is possible to

combine photocatalytic and biological
This chapter describes the most important processes.
applications of photocatalysis for water
decontamination. It outlines the 3. See how photocatalysis can treat not
decomposition of organic and inorganic only organic but also inorganic
contaminants and the usefulness of biological contaminants.
testing in addition
to chemical analyses. Examples are also 4. Discover different ways to improve
given to help understand the degradation process efficiency by modifying the
pathways involved in the decomposition catalyst and/or adding other extra-electron
of the initial compounds. The use of acceptors.
modified catalysts and additional oxidants
to increase process efficiency by trapping 5. Review the most common analytical
the photogenerated electrons and/or by tools used to determine the degradation
producing extra oxidizing species is rate of the original contaminants.
discussed. Finally, an overview of the most
common sophisticated analytical techniques 6. Learn the most sophisticated
is presented. analytical techniques used to describe the
complicated mechanisms involved in the
Objectives photocatalytic degradation of organic
By the end of this chapter you will know
why photocatalysis for the treatment of 7. Understand how to determine the
contaminated water is important and will toxicity of water and why this is so
understand the best analytical techniques important for water treatment.
for determining the degree of
decontamination during treatment. You

1. Be able to describe the

stoichiometric parameters of the
reactions involved in the photocatalytic
64 Solar Detoxification


Symbol Description Units
API Atmospheric pressure ionisation
BOD Biological oxygen demand mg O2/L
CI Chemical ionisation
DPs Degradation products
Eo Oxidation potential V
ECD Electron capture detector
EI Electron impact
FID Flame ionisation detector
GC Gas chromatography
HPLC High pressure liquid chromatography
LC50 Concentration of a chemical that is lethal to 50% of the exposed population mg/L
LLE Liquid-liquid extraction
Poct Water/octanol partition coefficient
MS Mass spectrometry
NMR Nuclear magnetic resonance
NPD Nitrogen phosphorus detector
SPE Solid phase extraction
TOC Total organic carbon mg/L

5.1. Detoxification of pollutants of oxidative degradation reactions that may lead to

com- plete mineralization of the contaminant. In
addition, hydroxyl radicals may attack aromatic rings
Chemists play a significant role in devising new at positions occupied by a halogen, generating a
methods to solve environmental problems. Existing phenol homologue. The electrons of the conduction
methods deal with concentrated contaminants; what is band can also degrade organic compounds by
still needed is an effective method for handling toxic reductive pathways, as is shown in Figure 5.1 in the
case of tetrachloromethane.
materials that are widely dispersed in the
In general, the types of compounds that have been
environment. Irradiated TiO 2 efficiently degrades degraded include: alkanes, haloalkanes, aliphatic
nearly every significant functional group, including alcohols, carboxylic acids, alkenes, aromatics,
the most environmentally recalcitrant materials. The haloaromatics, poly- mers, surfactants, herbicides,
primary processes associated with semi- conductor- pesticides and dyes. A partial tabulation of organic
compounds successfully degraded by photocatalysis is
sensitized photodegradation of organic sub- stances provided in Table 5.1. Equation 5.1 generally holds true
are illustrated in Figure 5.1. As shown by Equations for an organic compound of general formula C nHmOp
1.21 to 1.24, the hydroxyl radical (•OH) is the major
intermediate reactive responsible for organic sub-
strate oxidation. The free radical HO • and its
conjugate O •– are also involved in degradation
processes, but these  (m  2p)  m
radicals are much less reactive than free hydroxyl C n Hm O p    n O2  nCO2  H2O (5.1)
 4  2
cals, which react strongly with most organic
substances by hydrogen abstraction or electrophilic
addition to dou-
ble bonds. Free radicals further react with molecular In the case of organic compounds containing halogens,
oxygen to give a peroxy radical, initiating a sequence Equation 5.2 shows how the corresponding halide is
Water Decontamination 65

CO2 h  TiIVOH A-
Inorganic acids
Oxidized TiIVOH
pollutant TiIIIOH+
IfA = O 2

O2 O2-  HO •, HO
- ,H O , OH
2 22 H2O
CCl4 +e -cbCCl •••3+Cl-
CCl •+ e -:CCl + Cl- Organic Oxidized
3 cb 2 pollutant intermediate

Cl Cl Cl Cl . Cl Cl Cl Cl Cl . Cl
. OH OH . OH
-H O -HCl OH
Cl Cl
Cl . Cl Cl Cl Cl Cl
Cl Cl
Cl Cl HO Cl O O

FIGURE 5.1. General processes for the photo-oxidative or photo-reduction degradation of organic compounds in
aqueous solution sensitized by semiconductor particles. Examples of photo-oxidation (PCP) and photoreduction
(CCl4) are shown.

TABLE 5.1. Some examples of TiO 2-sensitized photodegradation of organic substrates.

Class Example
Alkanes Methane, hexane, cyclohexane, dodecane, octane
Haloalkanes Trichloroethylene, carbon tetrachloride, dibromomethane,
chloroform Aliphatic alcohols Methanol, isobutanol, decanol, ethanol, butyl alcohol
Carboxylic acids Formic, acetic, malic, oxalic, lactic, salicylic, decanoic, chloroacetic
acids Alkenes Cyclohexene, isobutene, butadiene, propene
Aromatics Toluene, nitrobenzene, nitrotoluene, hydroxybenzoic acid, diphenylmethane,
xylene Haloaromatics chlorobenzoic acid, chlorobenzene, chlorophenols, bromophenols, PCBs
Polymers Polyethylene, polyvinylalcohol, PVC
Surfactants Dodecylbenzene sulfonate, dodecylsulfate, trimethylphosphate
Herbicides Atrazine, bentazon, simazine, propetryne, prometron, 2-4 D,
Pesticides Lindane, methamidophos, imidacloprid, DDT, oxamyl, endosulfan,
phenitrothion Dyes Methylene blue, methyl orange, red dye 79, cresol violet, congo red

CHOX  (m  2p)  5.1.1. Total mineralization

 n
O  
m  q 4
n m p q 2

 nCO  HO (5.2) In photodegradation, transformation of the parent organic

2 2
2 compound is desirable in order to eliminate its toxicity and
persistence, but the principal objective is to mineralize
Under photocatalytic oxidative conditions, sulfur is all pollutants. The effectiveness of degradation is not
recov- ered as sulfate in sulfur-containing compounds demon- strated purely by the entire initial compound
according to Equation 5.3 being decomposed. As observed in the HPLC
C n HmOpSr  xO2  nCO2  yH2O  (5.3 (Fig. 5.2), oxamyl (chromatogram A, retention time 4.45 min)
zH2SO4 )
66 Solar Detoxification


FIGURE 5.2. HPLC-UV chromatograms of photodegradation of oxamyl before photocatalytic treatment a) and
when the oxamyl has completely disappeared b).

disappears nearly completely (chromatogram B). In order to demonstrate that there are no product
However, many new organic compounds appear. losses, the molar ratio must be in accordance with the
Furthermore, the stoichiometry proposed for the organic substrate structure. For example, the
general reactions (Eqs. 5.1 to 5.3) has to be demonstrated pentachlorophenol decomposition reaction to which
in each case by a correct mass balance. Reactives and mass balances should be adjusted is shown in
products might be lost, which will lead to results about Equation 5.4. Therefore, [Cl –] = 5 [PCP]0, but this
which we cannot be confident. The mineralization rate is only occurs at the end of the experiment, when the
determined by monitor- ing inorganic compounds, such TOC is almost 0. During the degradation, the
as CO2, Cl–, SO 2–, NO – and 4 3
formation of intermediates impedes this, as these inter-
PO43–. When organics decompose, a stoichiometric mediates contain differing amounts of chlorides.
increase in the concentration of inorganic anions is
produced in the hν
water treated and, likewise, very often an increase in the
C Cl OH  O  2H O  
  5HCl
con- 6CO
centration of hydrogen ions occurs (decrease in pH). For 6 5
2 2 2 TiO 2
this (5.4)
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, save
in some cases, because it is influenced by other processes In Figure 5.3, a complete pentachlorophenol degradation
that take place in the medium: the effect of the TiO 2 test is shown. The products obtained are practically
suspen- sion, the formation of CO 2 and intermediates and stoichio- metrically correct. The slight difference may be
so on. attributed to
Water Decontamination 67

0.12 0.6 stable, which in part explains why total mineralization

takes much longer than dearomatization.
0.10 0.5
[PCP], TOC], mM

0.08 0.4 5.1.2. Degradation pathways
0.06 TOC H 0.3
As pointed out above, a variety of degradation products
PCP (DPs) are formed in photocatalytic processes.
0.04 0.2 Nevertheless, in most cases no attention is paid to the
possible formation of these DPs, although they allow the
0.02 0.1 degradation processes to be better understood and
evaluated and thereby enable comparison among the
0.00 0.0 degradation pathways of different types of compounds,
0 5 10 15 20 25 30 35
functional groups, and chem-
tR, min
FIGURE 5.3 Evolution of H and Cl– during pentachloro-
+ ical oxidation parameters. Cost-effective treatment to
phenol degradation. To more clearly demonstrate that reac- com- plete compound mineralization is usually not
tion 5.4 is completed, the concentration of TOC in mM is feasible, and the generation of by-products appears to
calculated assuming 1 mMol TOC = 6 mMol of C = 72 mg be unavoidable with photocatalytic degradation.
of C. Identification of those by- products is key to
maximizing overall process efficiency. Since hydroxyl
radicals react non-selectively, numerous by-products
the fact that not all the TOC is completely decomposed are formed at low concentrations. On the other hand,
and that results obtained with four different analytical some of the degradation products obtained are of
techniques were compared. interest, because they may be more toxic and persistent
The oxidation of carbon atoms into CO2 is relatively than the parent compound. In some cases of
easy. It is, however, in general, markedly slower than the photocatalytic TiO2 treatments, only traces of these
dearoma- tization of the molecule. Until now, the metabolites are detectable because they are degraded
absence of total mineralization has been observed only faster than the parent compounds and mineralization is
in the case of s-triazine herbicides, for which the final almost total in a short time. But in other cases, such as
product obtained was essentially 1,2,5-triazine-2,4,6, s-triazine herbicides (mentioned above), overall
trihydroxy (cyanuric acid), which is fortunately not conversion to the final degradation products takes longer.
toxic. This is due to the strong stability of the triazine For example, primary intermediates of the photocat-
nucleus, which resists most methods of oxidation. For alytic degradation of various aromatic pollutants detected
chlorinated molecules, Cl – ions are easily released into and identified correspond to hydroxylation of the
the solution. Nitrogen-containing molecules are benzene ring (see Figure 5.1). Maximum transient
mineralized into NH + and NO–. Ammonium ions are concentrations are much lower for these intermediates
than for the initial pollu-
tants, since CO , acetate and formate are formed in the initial
4 3 2
tively stable and the proportion depends mainly on the stages of degradation. The orientation of the hydroxylation of
ini- the aromatic ring depends on the nature of the
tial oxidation degree of nitrogen and on the irradiation substituents. For instance, for chlorophenols and
time. The pollutants containing sulfur atoms are dimethoxybenzenes, the para and ortho positions (with
mineralized into sulfate ions. Organophosphorous respect to OH for the chlorophenols) are favoured, as is
pesticides produce phosphate ions. However, expected. In contrast, for benzamide and nitrobenzene,
phosphate ions in the pH range used remained the hydroxylation occurs at all free sites, whereas a meta-
adsorbed on TiO2. This strong adsorption partially orientation is expected for elec- tron-withdrawing
inhibits the reaction rate, which, however, remains substituents. A list of compounds pro- duced during
acceptable. Until now, the analyses of aliphatic photodegradation of a pesticide (pyrimethanil) with TiO 2
fragments resulting from the degradation of the is shown in Figure 5.4. It will give an idea of the
aromatic ring have mainly revealed formate and complexity of the chemical reactions involved in
acetate ions. Other aliphatics (presumably acids, photocat- alytic methods.
diacids, and hydroxylated compounds) have been also
found. Formate and acetate ions are fairly
68 Solar Detoxification

Hidroxy derivatives Aliphatic

Pyrimidil derivatives
OH derivatives


N Pyrimethanil

Phenyl derivatives

FIGURE 5.4. Chemical structures of pyrimethanil and its degradation products obtained during a photocatalytic
treatment with TiO2.

5.1.3. Toxicity reduction en-route to complete mineralization. Toxicity testing of

the photocatalytically treated wastewater is therefore
The success of photocatalytic degradation in wastewater necessary, particularly when incomplete degradation is
treatment depends on how much the cost of illumination planned. Any number of whole-organism bioassays
(that is, energy) can be kept down. This can be achieved could be used for the assessment of water quality before
by improving the efficiency of photocatalytic disposal. Since it is not feasible to determine the specific
degradation, for example, by modifying the catalyst and toxicity of every toxic inter- mediary compound, whole-
the photoreactor design or by using additional oxidants. effluent toxicity testing using aquatic organisms is a
It is also possible to envisage incomplete photocatalytic direct, cost-effective and relevant means of determining
degradation to detoxify the wastewater. The products of effluent toxicity. Therefore, bioassays can provide a
incomplete degradation and their concentrations may be more direct and appropriate measure of mix- tures of
sufficiently innocuous to be dis- charged directly into the toxins than chemical analyses alone, which are not
environment or to undergo further biological treatment. sufficient to assess potential effects on aquatic biota
Biological treatment of biodegradable residual waters is (Tapp et al., 1996). A review of these tests is shown at
currently most compatible with the envi- ronment and is the end of this chapter.
the least expensive option. However, it is very difficult to Another useful technique for evaluating the feasibility
identify all the intermediary compounds of disposing of photocatalytically treated waters before
Water Decontamination 69

mineralization is to determine the biodegradability It has been discovered that ozonation (production of
(BOD) in sludge from biologically activated sludge drink- ing water) generates low levels of bromate ions,
plants (Pulgarín et al., 1999). This sludge is inoculated recognized as a suspected agent of cancer. Drinking
with the treated water and BOD tests are carried out. In water treated with ozone will typically have in the range
this case, the results are obtained in terms of ‘relative of tens of µg/L of bromate, and TiO 2 is able to sensitize
biodegradability’. This means that once biodegradability the photodecom- position of bromate (Eq. 5.7).
is achieved, it can only be assured if the same
biologically activated sludge used for the BOD
tests is used for disposing of the treated water. This test 2BrO h 
ν /TiO
 2Br   3O (5.7)
produces non-universal results, but could be very useful 3 2
the biological treatment plant that is to receive the Heavy metals are generally toxic and can be removed
effluent from the photocatalytic treatment is known prior from industrial waste effluents as small crystallites
to design of the photocatalytic plant. This permits deposited on the photocatalyst, as demonstrated by
disposal of the Equation 5.8:
pre-treated water into a biological (low cost) treatment. n n
M n  H O h 
ν /TiO
 M 0  nH   O (5.8)
2 2 42
5.1.4. Detoxification of inorganic Under identical conditions, the following reactivity
pollutants pattern has been found: Ag > Pd > Au > Pt >> Rh >> Ir
>> Cu = Ni
Photocatalytic processes may also transform inorganic = Fe = 0 for AgNO3, PdCl2, AuCl3, H2PtCl6 or Na2PtCl6,
substances from the environment. Specifically, H2S and RhCl3, H2IrCl6, Cu(NO3)2, Ni(NO3)2 and Fe(NO3)3. As
CN– can be converted into less-toxic materials, and the
various strate- gic and/or toxic metals can be removed photodeposition conversion increases, the metal
from waste effluents. The toxicity of H2S is comparable to particles form agglomerates, reaching several hundreds
that of HCN. In the pres- ence of light and a of nm. Silver photodeposition has been applied in the
semiconductor catalyst, the products of oxidation are recovery of Ag from used photographic baths in which
molecular hydrogen and sulfur. Free cyanides are the silver-thiosulfate complex is decomposed, Ag +
generated in large quantities in heat-treating operations being reduced to Ag°. Because of their favourable redox
and in metal-finishing industries. The greatest amounts potentials, only noble metals can be photodeposited. This
of cyanide-containing wastes are produced by precious- property has been used to selec- tively recover noble
metals milling operations and coal gasification processes. heavy metals. For instance, silver has been separated
CNO– is the first product of the photocatalytic oxidation from copper in solutions simulating indus- trial
of cyanides in the presence of polycrystalline TiO2 in an electrolytic baths. Other toxic, heavy non-noble metals
aqueous medium. The proposed mechanism implies the could be removed from water. Mercury, because of its
oxidation of cyanide by the photogenerated holes and the favourable redox potential, was photoreduced as a
reduction of oxygen by electrons according to Equation zero- valent metal. The cations Pb 2+ and Tl3+ have been
5.5. The reacting mixture must be at pH10 to avoid the deposited on UV-irradiated TiO2 powder as PbO2 and
formation of volatile HCN. Tl2O3. Similarly, uranium has been photodeposited on
TiO2 as U3O8 from uranyl solutions. From an
application point of view, the
recovery of silver from photographic baths seems the
CN  2h  2OH  CNO  (5.5 most promising issue, provided legislation on discharge
H2O ) water
containing Ag becomes stricter.
Cyanate ions are photo-oxidized to nitrates and a
satisfactory nitrogen balance is achieved according to the
5.2. Quantum yield improvement
overall reaction shown in Equation 5.6. At strong oxidant
conditions, volatile nitrogen-containing species are not
through additional oxidants
produced or are quickly
2 photo-oxidized
3 to NO– and NO–.
One practical problem in using TiO 2 as a photocatalyst is
CNO  4O2  2OH 3H2O  3CO2  3NO  electron/hole recombination (reverse of Eq. 1.21), which,
in the absence of proper electron acceptors, is extremely
70 Solar Detoxification

inefficient and thus represents a major energy-wasting peroxide reacts with conduction-band electrons (Eq. 5.9)
step, as well as limiting the achievement of a high to generate hydroxyl radicals, which are required for the
quantum yield. Oxygen has been chosen in most of the photomineralization of organic pollutants.
applications for this
purpose, although its role is not related only to electron
scavenging (see Chapter 4). But with only dissolved H2O2  e  OH  OH (5.9)
oxygen as an oxidant, low mineralization photo-
efficiencies (produc-
tion of CO2) are obtained (in the range of 1 to 5 per The following reactions (Eq. 5.10 and 5.11) can also produce

OH (reaction 5.11, which requires Z < 300 nm, does not
One strategy for inhibiting e –/h+ recombination is to take place with solar radiation).
add other (irreversible) electron acceptors to the
Outstanding enhancement of the rate of degradation of
various organic contaminants through the use of H2O2  O
 OH  OH  (5.10)
inorganic peroxides has been demonstrated (Pelizzetti et
al., 1991; Malato et al., 2000). The addition of other O2 H2O2  h  2OH (5.11)
oxidizing species
could have several different effects, including:
The effect of this electron acceptor deserves some further
• increasing the number of trapped e – in the e–/h+ pairs comment. In some cases, the addition has been found to
and, consequently, avoiding recombination; be beneficial, increasing the degradation rate. The
• generating more •OH and other oxidizing depends on the concentration, generally showing an
species; H2 O2
• increasing the oxidation rate of intermediate optimum range of concentration. At higher
compounds; concentrations the improvement starts to lessen. Whereas
• avoiding problems caused by a low O 2 concentration. this beneficial effect can easily be explained in terms of
prevention of elec- tron/hole recombination and
It must be mentioned here that in highly toxic additional •OH production through reactions 5.9 to 5.11,
wastewater where degradation of organic pollutants is inhibition could be explained in terms of TiO 2 surface
the major concern, the addition of an inorganic anion to modification by H2O2 adsorption, scavenging of
photoproduced holes (Eq. 5.12), and reaction with
enhance the organic degradation rate may often be hydroxyl radicals (Eq. 5.13).
justified. For better results, these additives should fulfil
the following criteria: they should dissociate into
harmless by-products and lead to the forma- tion of •OH
or other oxidizing agents. There is another advan-
tage related to the use of this type of oxidant when solar H2O2  2h  O2  2H (5.12)
energy is the photon source. Although scientific research 
H O  OH  H O  HO (5.13)
photocatalytic detoxification has been conducted for at used chemical and therefore very cheap. Being an electron
least the last three decades, industrial/commercial acceptor, hydrogen
applications, engineering systems and engineering design
methodologies have only been developed recently. In this
type of installation, the photoreactor is by far the most
expensive component and a barrier to commercialization.
Increase the photocatalytic reaction rate with these
additives would decrease photoreac- tor dimensions
proportionally and dramatically decrease overall costs.

5.2.1. Hydrogen peroxide

Hydrogen peroxide is the obvious candidate. It can

increase the efficiency of the process at high irradiance
(see Section 4.6) and it has been tested with a large
number of com- pounds. Also, it is a very commonly
Water Decontamination 71

2 2 2 2

The inhibition of adsorption not only depends on

the characteristics of the pollutant but also on the
hydrogen peroxide/organic concentration ratio. This
may be explained in terms of Langmuir–Hinshelwood
kinetics, rC = krKC/ (1 + KC) and competitive
adsorption (see Chapter 4). If pol- lutant concentration
(C) is too low and H2O2 concentration too high,
organic adsorption decreases because of adsorption of
hydrogen peroxide and, therefore, the additional
hydroxyl radicals generated by H2O2 do not react
efficiently. When C is higher, the radicals react more
easily and rC increases. When C is still higher, the
reaction rate is not as affected by adsorption (1 + KC =
KC; rC = kr). In the latter situation the reaction rate is
only dependent on •OH concentration related to
favourable (Eqs. 5.9 to 5.11) and unfavourable
reactions (Eqs. 5.12 and 5.13). There is an optimum
ratio of H2O2/C under these circumstances, whereby
organic material is
72 Solar Detoxification

sufficient to consume generated hydroxyls and to avoid HSO5  e   OH  SO42 or

detrimental reactions, if the peroxide level is not too high. HSO  e   OH  SO (5.21)
At high molar ratios an inhibition effect would be
5 4
expected, because the unfavourable reactions become
more and more
pronounced. All this may be summarized as: first, if Chlorate has been proven insufficient to improve effective-
pollu- tant concentration is low, the hydrogen ness. However, both 4IO – and bromate increase the
peroxide easily inhibits the degradation rate and, mineral- ization rate in all cases tested. Nevertheless,
second, if the molar ratio between H2O2 and pollutant is these additives are very expensive compared to
too high, the same is true. hydrogen peroxide and peroxydisulfate, and their
application would dramatically increase treatment cost.
5.2.2. Persulfate Even more importantly, they do not dissociate into
harmless products (Br – and I–), because hundreds of
mg/L of these anions are undesirable in water. Potassium
In homogeneous reactions, the persulfate ion accepts an peroxymonosulfate (commercially called ‘oxone’) was also
elec- tron and dissociates (Eq. 5.14). The sulfate radical examined as an irreversible electron acceptor. The
anion is also generated under adequate thermal and formula of this salt is 2KHSO5.KHSO4.K2SO4, written
photolytic (wavelength <270 nm) conditions. This in aqueous solution as HSO –. Upon accepting an electron
radical goes through the reactions explained below from the conduction band, HSO– would dissociate into
(Eqs. 5.16 and 5.17). Persulfate can therefore be a 5
two different pathways (Eq. 5.21). The disadvantage is
beneficial oxidizing agent in photocatalytical its high molecular weight. 5 Many grams of oxone are
detoxification because SO •– is formed from the oxidant necessary to achieve 1 active mol (HSO–).
compound by reaction with the 4 photogenerated
semiconductor electrons (e– , Eq. 5.15). In addition, it
can trap the photogenerated electrons and/or generate
5.3. Catalyst modification
hydroxyl radicals. The sulfate radical anion (SO•–) is a
very strong oxi- dant (E = 2.6 V) and engages in at least
three reaction modes with organic compounds: by The TiO2 band-gap represents only 5 per cent of the solar
abstracting a hydrogen atom from saturated carbon, by
spectrum (see Chapter 2). From the standpoint of solar-
adding to unsaturated or aromatic carbon, and by
removing one electron from carboxylate anions and collecting technology, it is therefore a rather inefficient
from certain neutral molecules. process even for a high added-value application. In
S O2  e  SO  SO2 (5.14) to other advanced oxidation technologies, photocatalysis has
2 8 aq 4 4
 eCB
 SO4 SO24 (5.15) the advantage of being solarizable and of being an environ-
mentally friendly technology. TiO2 is a cheap photostable
SO4  eCB SO24 (5.16
SO  H O   OH  SO2  H ) catalyst, and the process may be run at ambient temperature
and pressure conditions. Furthermore, the oxidant, molecu-
4 2 4
lar oxygen (O2), is the mildest one. TiO 2 is a mild
catalyst that works at mild conditions with mild oxidants.
5.2.3. Other oxidants However, as concentration and number of contaminants
increase, the process becomes more complicated and
challenging prob- lems such as catalyst deactivation,
Other compounds could also potentially increase the slow kinetics, low photo- efficiencies and unpredictable
reac- tion rate because they are also electron scavengers. mechanisms need to be solved. It is clear that naked TiO 2
The most frequently tested so far in heterogeneous needs extra help to under- take practical applications of
photocatalysis are described in Equations 5.18 to 5.21. industrial and environmental interest, and this could lead
IO   8H 
+ to the
of loss of
its mild some of Moreover,
operation. the charm even reactor set-ups using
4 8e 

4H 2 O I (5.18)
artificial light, which must bear the cost of running the lamps
ClO3  6e H 3H2O 
involved in them, will be much cheaper if visible radiant flux
Cl )

BrO3  6e 3H2O  (5.20 can be employed.

Br )
Water Decontamination 73

The redox process is based on the migration of Figure 5.6). Loading of Pt is optimum to achieve the
electrons and holes to the semiconductor surface, and on maxi- mum photocatalytic rate, affecting the distribution
two further oxidation and reduction steps (see Figure of elec- trons in the system. Above the optimum metal
5.1). Two basic lines of R&D have attempted to balance content, efficiency decreases, because once negatively
both half-reaction rates, one by adding electron acceptors charged, Pt particles become attractive for holes, which
(additional oxidants, already commented on above) and recombine with
the other by modifying catalyst structure and
composition. Both try to promote com- petition for
electrons and to avoid recombination of e –/h+ pairs. A
third approach has focused not only on increasing Metal
quantum yield but also on finding new catalysts with h
band- gaps that match the solar spectrum better. e-
Unfortunately, the choice of convenient alternatives for
substituting titanium dioxide in photocatalytic

detoxification systems is limited. An appropriate
semiconducting material should be:

• non toxic, 
• stable in aqueous solutions containing highly
reactive and/or toxic chemicals,
 h+
• not photo-corrodible under band-gap illumination,
• economical – that is – an increase in photocatalytic

FIGURE 5.5. Electron capture by a metal in contact with a
reaction rates must always be accompanied by a non- semiconductor surface.
proportional increase in overall process costs.
Metal oxides generally fulfil these criteria, but most
metal oxides are wide band-gap semiconductors or
insulators. Although iron (III) oxide is one of the few
exceptions (see Table 1.4), it has demonstrated
satisfactory activity in a limited number of cases. Other
non-oxide semiconductors (for example, CdS) are
usually unstable and photodegrade over time. A few
examples will be given in the following sections to
illustrate the large body of work conducted in the area of
photocatalyst modification.

5.3.1. Metal semiconductor modification

The addition of noble metals to a semiconductor could

modify the photocatalytic process by changing the
semicon- ductor surface properties. Figure 5.5 is an
illustration of the capacity of a metal in contact with a
semiconductor surface to capture electrons. After
excitation, the electron migrates to the metal, where it
becomes trapped and e–/h+ recombina- tion is avoided.
The hole is then free to migrate to the surface, where
oxidation of the organics can occur (see Figure 5.1).
The Pt/TiO 2 system (Pt deposited on titania by the FIGURE 5.6. Concentrating solar reactor with a platinum/
impregnation and reduction method) is the metal- titanium dioxide catalyst on ceramic saddles. Tested on
semiconductor system most commonly studied (see air contaminated by spray paint at Fort Carson Army
Base in Colorado, USA. (Courtesy of the US National
Renewable Energy Laboratory.)
74 Solar Detoxification


h Dye*
A A A-
EG= 3.2 eV EG= 2.5 eV
FIGURE 5.8. Steps of excitation with a sensitizer in the
presence of an adsorbed organic electron acceptor (A).

because of the large number of dye substances found in

TiO2 industrial textile wastewater.
FIGURE 5.7. The excitation process in a semiconductor-
semiconductor photocatalyst. 5.4. Recommended analytical
electrons into inefficient thermal energy. Care must
therefore be taken in studies conducted with modified
metal semi- conductors to use the optimum quantity of 5.4.1. Original contaminants
The analyses performed on samples collected during
5.3.2. Composite semiconductors photo- catalytic treatment have to be as complete as
possible in order to adjust the mass balance for the
Coupled semiconductors provide a way to increase the photocatalytic decomposition of the contaminants and
separation between charges and reduce the energy assure that the ana- lytical data are reliable and the
(increas- ing wavelength) necessary to excite the system. organic compounds have not disappeared in some other
Figure 5.7 shows CdS-TiO 2 as an example. The energy way (for example, through evap- oration, adsorption in
from light (λ < 497 nm) is large enough to cause an the reactor, or adsorption in the cata- lyst) besides
electron to leave the CdS valence band and to go over to photocatalysis. Liquid chromatography (HPLC) with UV
the conduction band. The hole remains in the CdS, while detection is the method of choice for analysis of
the electron is transferred to the TiO2 conduction band. samples from photocatalytic treatment, as it allows
The separate charges are then free to undergo electron direct injection of the aqueous sample into the
transfer with the species adsorbed on the surface. analytical column, avoiding the necessity of prior
extraction. Samples only require filtering to remove
5.3.3. Surface sensitization TiO2 or any other particles that could damage the
chromatographic column before they are injected into
the equipment. A device made up of a syringe and 0.22
Surface sensitization of a wide band-gap semiconductor µm membrane filters is used for this purpose. Gas
via chemisorbed or physisorbed dyes can increase the chromatography is only suggested if HPLC is not viable,
efficiency of the excitation process. The wavelength or when a pre-concentration is necessary as a result
range can be expanded by charge transfer from the of very low concentration of the parent compounds.
sensitizer to the semi- conductor. Figure 5.8 illustrates
the excitation, charge trans- fer and regeneration steps. If 5.4.2. Mineralization measurements (TOC)
the oxidative energy level of the excited state of the dye
is more negative than the semi- conductor conduction Total organic carbon (TOC) analyses of samples taken
band, then the dye can transfer the electron to the
during photocatalysis degradation experiments are vital
conduction band of the semiconductor. The electron in
for the following reasons:
turn can be transferred to reduce an organic acceptor
adsorbed on the surface. Without the presence of a redox • As identification of all the intermediates generated
couple, the dye-semiconductor system can also be used in during photodecompositon is not possible,
oxidative degradation of the dye itself. This is important pinpointing the moment at which only CO2
remains and the water is considered completely
decontaminated is crucial.
Water Decontamination 75

• Determination of the CO 2 produced might be Advantages of the GC-MS-based methods are:

reasonable, since this must be stoichiometric with the
• the large amount of structural information they yield
organic carbon present at the beginning in the
organic molecule. However, since the reactors are and the spectra libraries available either from
usually large and not airtight, loss of the carbon databases or from research papers that make DP
dioxide produced into the atmosphere prevents this. identification feasible;
For the same reasons, the samples might become • the durability and reliability of the GC-MS interface; and
contaminated by atmospheric CO2 and falsify results. • their highly efficient sensitivity and separation,
• The analyses provide a reliable, simple and rapid way which avoids overlapping compounds with similar
to establish the mass balance at any moment, to get an structures.
idea of the remaining amount of intermediates and to
check the extraction methods and analysis of However, GC-MS-based methods have important
intermediates. draw- backs because of their low capacity to analyse
very polar, less volatile and less thermally unstable
The basic techniques for the determination of TOC in compounds. Identification of DPs is usually based on
water have remained relatively unchanged for twenty their EI (Electron Impact) mass spectra, mainly by
years. Organic compounds are converted to CO 2 via a comparing the unknown compound spectrum with
combination of actions that may include chemical published spectra. A disadvan- tage of EI is that it does
oxidizing agents, ultra- violet radiation or high- not usually provide molecular weight information.
temperature combustion (Wangersky, 1993). The CO2 is Additional and very useful structural information on DPs
then measured using non-dispersive infrared can be obtained by chemical ionization (CI). From the
absorption, micro-coulometry or conductimetric molecular weight and interpretation of frag- mentation
techniques. As many water samples contain inorganic patterns, it is possible to hypothesize a molecu- lar
forms of carbon (HCO – and CO 2–), it is usually structure. However, even an EI mass spectrum does
necessary to 3 3 not provide enough information about the location of the
functional groups (for instance, the position of a
remove these, typically using a gas stripping technique group on a benzyl ring). Comparison with a standard
prior (when commercially available) is required for
to measurement of TOC. Some of the organic carbon unequivocal confirmation. When by-product standards
(VOCs) may also, very often, be removed by this are not com- mercially available, the laboratory
procedure. In the case of heterogeneous photocatalysis, it synthesis of DPs could be a solution.
is necessary to use a TOC analyser that can manage non- Because GC methods require compounds with high
filtered samples. When TiO 2 is filtered from the sample, vapour pressures, derivation of at least the acid fraction
there is a notable loss of organic compounds, because (BF3/MeOH or diazomethane) has become a typical
they are adsorbed onto the solid retained in the filter. procedure for identifying less volatile and more polar
compounds. How- ever, derivation is not easy and it is
also time-consuming. Because of this, LC-MS
techniques are gaining in impor- tance. LC techniques
5.4.3. Analysis of intermediates present several advantages over GC:

In order to analyse a reaction mixture containing • Little or no cleaning of the sample is required.
pesticides and their degradation products (DPs), it is • High polar, low volatile and thermally stable
necessary to have analytical screening methods available compounds are more easily analysed.
that permit separation and identification of compounds • Direct analysis of the samples is possible, avoiding
with very different hydrophilic–hydrophobic
characteristics and spanning very different polar DPs escaping from extraction procedures.
concentration ranges. The use of gas chromato-
The major role of LC-MS in the degradation processes
graphy with classical detectors, like ECD, NPD or
stud- ied is, first, to check DP molecular weight and,
FID, is obviously insufficient. But gas or liquid second to detect DPs that are not directly amenable with
chromatographic systems coupled with mass GC-MS techniques.
spectrometry (GC/MS and LC/MS) represent a good
The ready availability of atmospheric pressure
fast alternative for achieving very useful structural
ionization (API), electrospray (ESP)/ionspray (ISP) and
information about the compounds gener- ated in such
processes (Agüera and Fernández-Alba, 1998).
76 Solar Detoxification


+R 1 N CH CH
2 3 2 3

P = R1
R1 N
Pyrimiphos methyl CH2CH3
2 3


OCH3 = R2 OCH3
R1 N
R2 N
R2 N CH3
R1 N
R2 N 2 3
R1 N
R2 N R2 N


H R1 N
R2 N



2 3 P

FIGURE 5.9. Degradation pathway proposed for pirimiphos-methyl dissolved in water when illuminated in the
presence of TiO2.

pressure chemical ionization (APCI) LC-MS interfaces provide enough structural information for the
that provide general structural information and unequivocal structural elucidation of the DPs in many
sensitivity has expanded the applicability of LC-MS, cases, and this must be done by more complex, time-
allowing identi- fication of a large range of polar- consuming techniques such as LC/MS/MS or NMR. But
transformation products. LC-MS presents important the combination of both GC/MS and LC/MS is very
weaknesses, since the informa- tion they provide usually useful and, in many cases, enough structural information
lacks detailed structural data and is less sensitive and to evaluate the degradation process may be found
discriminating than that provided by GC- MS. These quickly. Other analytical measures, such as total organic
disadvantages can frequently prevent DP identifi- cation carbon (TOC) or ion chromatography (IC), can be of
as a consequence of the low detection threshold or great help in assessing the success of the degradation
overlapping peaks. As the DPs obtained are very process by evaluating the mineralization rate achieved or
complex (see Figure 5.9), mass spectrometry-based by establishing the mass balance of the whole process.
techniques do not
Water Decontamination 77

5.4.4. Extraction methods The OECD lists mortality in fish, impaired reproduction
in crustaceans and inhibition of growth in algae as
examples of ecotoxicity tests needed to predict the
Common handling procedures for the analysis of samples impact of released chemicals upon ecosystems. The
of chemically treated water involve the use of extraction two main invertebrate toxicity tests routinely used are
methods, because identification of degradation products the twenty-one-day Dapinia and the seven-day
has to be carried out at sub-mg/L level. Trace organic Ceriodaphinia survival and reproduction tests. Tests are
com- pounds in wastewater are still typically enriched by conducted by exposing the organisms to the toxins
liquid– liquid extraction (LLE) using an appropriate under control conditions. After the required incuba- tion
solvent. However, solid-phase extraction (SPE) is gaining period, live organisms are counted. Fish bioassays have
in acceptance (Chiron et al., 2000), mainly because SPE been conducted for decades. The distinct
generates less matrix inter- ference, and a wide range of physiological and behavioural responses of fish to low
new adsorbents (able to trap DPs with a wide range of levels of pollutants has been employed in the
polarities) are commercially avail- able, including: development of fish monitors that act as indicators of
alkylsilanes (modified silica, such as C-18 and end- water quality (Tothill and Turner, 1996). The tests are
capped C-18); porous polymers, such as poly (styrene- usually based on larval growth and survival, where
divinylbenzene) PRP-1 or PLRP-S; and carbon mod- ified newly hatched fish are exposed to a range of effluents for
materials, such as porous graphitic carbon (PGC). SPE up to seven days. The acute lethality test with fish
materials are expected to show different behaviours measures the concentration of a chemical that is lethal to
with respect to capacity and breakthrough volumes of 50 per cent of the exposed population after 96 h
both ana- lyte and matrix interferences. Because the (LC 50). Species such as rainbow trout (Oncorhynchus
retention of a com- pound is higher in its neutral form mykiss) and fat- head minnow (PIMEPHALES PROMELAS) are
than in its ionic form, phenol and benzoic acid commonly used. During recent years, research has been
extractions are better under acidic conditions while
carried out to reduce or replace acute fish tests with in
amines are best recovered at alkaline condi- tions.
vitro assays, using cultured fish cell lines. The use of algae
Problems may arise with multifunctional compounds.
in bioassays has proved useful in detecting metals,
The retention of hydrophilic compounds (Log Poct <0)
herbicides, pesticides and crude oil com- pounds.
such as aminophenols or cyanuric acid is poor, but can
However, the culturing and preparation of the algae
be high with porous graphitic carbon. On the other
suspension is a lengthy process and, further, it is difficult
hand, short- chain aliphatics such as oxalic or formic
to maintain an identical culture of algae each time the
acids (usually the final organic photo-oxidation
bioassay is conducted.
degradation products) may not be recovered at all. In
order to identify as many chemicals as possible, a Many of the above tests require specialized equipment
sequential extraction scheme involving different SPE and operator skills and are time consuming. The use of
adsorbents could be proposed. A C-18 phase is used higher organisms such as fish may also be ethically
initially to select all neutral hydrophobic compounds at undesirable. In the last few years there has been an
pH7. The C-18 filtrates are then passed through a increased interest in bacterial screening to assess toxicity.
polymeric adsorbent, also at pH7, where medium-polar Studies of effects on microbial function or activity
compounds are retained. Subsequently, filtrates are constitute a more direct, rapid and sensi- tive approach to
acidified to pH4.5 and pH2.5, respectively, to extract measuring chemical stress. These can be classified by
the majority of acidic compounds with a polymeric or the type of measurement used:
carbon-type adsorbent. Polymeric adsorbents can be • measurements monitoring the transformation of
used over pH2-13 without decomposition of the carbon, sulfur or nitrogen,
adsorbent material. • measurements that determine the activity of
microbial enzymes such as dehydrogenases,
5.4.5. Toxicity analysis adenosine triphos- phatases and other enzymes,
• measurements of growth, mortality and photosynthesis,
• measurements that determine glucose uptake
The test organisms in these assays include activity using radioisotopes,
representatives from four groups: microorganisms, • measurements of oxygen consumption using a
plants, invertebrates and fish. The Organisation for
dissolved oxygen electrode or respirometer,
Economic Cooperation and Development (OECD) has
proposed a minimum data set to assess the effects of • measurement of luminescence using a photometer.
chemicals on the environment.
78 Solar Detoxification

The development and applications of biological toxicity CHIRON, S.; FERNÁNDEZ-ALBA, A.; RODRÍGUEZ, A. AND
testing are rapidly increasing. Numerous bioassay proce- GARCÍA-CALVO, E. 2000. Pesticide chemical oxidation:
dures are now available, but it is difficult to assess the state of the art. WAT. Res., Vol. 34, No. 2, pp. 366–
sensi- tivity of these tests, and therefore a universal 77.
monitoring device for toxicity testing is unlikely to
become available in the near future. In any case, most FERNÁNDEZ-ALBA, A.R.; HERNANDO, L.; DIAZ, G. AND
recent studies have dealt with the use of the bacterial
luminescence assay for toxicity screening. The use of CHISTI, Y. 2001. Toxicity of pesticides in wastewater:
toxicity tests for evaluation of photo- catalytic treatments a comparative assessment of rapid bioassays. ACTA
is not very common for the moment, but several relevant CHIMICA ACTA, No. 426, pp. 289–301.
papers have already been published (Herrmann et al.,
1999; Fernández-Alba et al., 2002). HERRMANN, J.M.; GUILLARD, CH.; ARGÜELLO, M.; AGÜERA,
Summary Photocatalytic degradation of pesticide pyrimiphos-methyl:
determination of the reaction pathway and identification of
intermediate products by various analytical methods.
Photocatalytic degradation of organic and inorganic CATALYSIS TODAY, No. 54, pp. 353–368.
compounds follows a definite stoichiometry that has to
be determined by appropriate analytical techniques MALATO, S.; BLANCO, J.; MALDONADO, M.I.;
during photocatalytic experiments in order to close the FERNÁNDEZ-IBÁÑEZ, P. AND CAMPOS, A. 2000. Optimizing solar
mass balance between reactives and products. photocatalytic mineralization of pesticides by adding
Furthermore, degradation pathways and the by-products inorganic oxidizing species: application to the
formed are so complex that analysis of water toxicity recycling of pesticide containers. Applied CATALYSIS B:
during the treatment is recommended to assure that it ENVIRONMENTAL, No. 28, pp. 163–74.
decreases. The photocat- alytic degradation of
contaminants by photocatalysis can be enhanced by the PELIZZETTI, E.; CARLIN, V.; MINERO, C. AND GRÄTZEL, M. 1991.
use of additional oxidants such as hydrogen peroxide and Enhancement of the rate of photocatalytic degradation
persulfate and/or modification of the catalyst using on TiO2 of 2-chlorophenol, 2,7-dichlorodibenzodioxin
deposited metals, composite semiconductors or dye and atrazine by inorganic oxidizing species. New
sensitizers. JOURNAL of Chemistry, No. 15, pp. 351–9.
Preferably, the degradation of the original
contaminants should be monitored by HPLC, and the
final mineralization by TOC and Ionic Chromatography,
AND PÉRINGER, P. 1991. Strategy for the coupling of
but identification of intermediate products has to be
achieved through MS tech- niques (GC-MS and HPLC- photochemical and biological flow reactors useful in
MS). A combination of several MS techniques with mineralization of biorecalcitrant industrial pollutants.
several extraction methods is better in order to assure the CATALYSIS TODAY, No. 54, pp. 341–54.
identification and quantification of a sig- nificant number
of DPs. As the detection of all DPs is almost impossible, TAPP, J.F.; WHARFE, J.R. AND HUNT, S.M. 1996. Toxic IMPACTS
the use of different organisms for determining toxicity is of WASTES on the AQUATIC Environment. London, Royal Society
very important in order to guarantee that the treatment of Chemistry.
is correct.
TOTHILL, I.E. AND TURNER, A.P.F. 1996. Developments
Bibliography in bioassay methods for toxicity testing in water
treatment. Trends in ANALYTICAL Chemistry, No. 15,
pp. 178–87.
LC-MS evaluation of pesticide degradation products WANGERSKY, P.J. 1993. Dissolved organic carbon
generated from advanced oxidation processes in methods: a critical review. MARINE Chemistry, No.
waters: an overview. ANALUSIS, No. 26, pp.123–30. 41, pp. 61–74.
Water Decontamination 79

Self-assessment questions Part B

Part A: True or false? 1. What is the general stoichiometry for the

total mineralization of methamidophos
1. Electrons in the conduction band cannot
degrade organic compounds.
2. Why is the analysis of the initial compound not
enough to determine the efficiency of contaminant
2. When organics decompose, an increase in the degradation?
concentration of hydrogen ions is produced in the
3. Why is toxicity determination during a
3. The final objective of photocatalytic treatment is photocatalytic experiment very important?
always the total mineralization of the organic
contaminants. 4. What are the final products in the
photocatalytic treatment of CN–?
4. Inorganic compounds are degraded through
oxidative and/or reductive photocatalytic pathways. 5. What are the main advantages of an additional oxidant?

5. The addition of other oxidants to the 6. What are the most common oxidants used
photocatalytic process only produces additional for photocatalytic degradation of contaminated
oxidant species. water?

6. Hydrogen peroxide always enhances the 7. What metals are used for doping TiO 2, and how do
decomposition of organics. they increase photoefficiency?
7. Liquid chromatography (HPLC) with UV
8. What is the most common analytical technique
detection is the method of choice for analysing
for determining the disappearance of the initial
samples from photocatalytic treatment because it is compound during photocatalysis in water, and why?
the easiest.

8. TOC analysis is a method for measuring 9. What is the principal advantage of applying MS-
CO2 production during experiments. based analytical methods to the photocatalytic
degradation of organic compounds?
9. GC/MS and LC/MS are complementary methods.
10. What are the most common toxicity tests in
use today?
10. There is a universal monitoring device for
toxicity testing.
80 Solar Detoxification

Answers 5. Increasing the number of trapped e – of the e–/h+

pairs, generating more •OH as well as other
oxidizing species; increasing the oxidation rate of
Part A intermediate compounds; and avoiding problems
caused by a
1) False. 2) True. 3) False. 4) True. 5) False. 6) False. low concentration of O 2.
7) True. 8) False. 9) True. 10) False.
6. Hydrogen peroxide and persulfate.
Part B
7. Noble metals. After excitation, the electron migrates
to the metal where it becomes trapped and e –/h+
1. C2H 8NO2PS  7O2   2CO2  H 3PO 4
T iO2 / h
recombination is avoided.
H2SO4  HNO3 
H2O 8. Liquid chromatography (HPLC). This method
permits direct injection of the aqueous sample into
2. Because reactives and products could be lost, the analytical column, avoiding the necessity of
causing the results to be unreliable. extraction procedures.

3. Because the products of incomplete degradation 9. The identification of unknown compounds

and their concentrations may be sufficiently produced during photocatalytic treatment of
innocuous for discharge directly into the contaminated water.
environment or for further biological treatment.
10. Bacterial screening tests.
4. NO– and CO .
3 2
Solar Detoxification

Aims Objectives
This chapter discusses the basic factors When you have completed this chapter,
related to solar photocatalytic process you will have a basic knowledge and
technology and applications. The key understanding of the following areas:
issues related to the use of solar ultraviolet
radiation and their implications for 1. The key issues in collector
appropriate materials, and reactors for technology related to solar water
efficient light collection and use are detoxification.
discussed here.
2. The specific peculiarities of the use
of solar ultraviolet light.

3. The main characteristics of different

types of collectors for solar water
detoxification applications.

4. Advantages and disadvantages of

using non-concentrated sunlight.

5. Factors concerned the with the

main technological components: the
reactor, reflectors and catalysts.
82 Solar Detoxification


Symbol Description Units
CPC Compound parabolic concentrator
CPVC Chlorinated polyvinyl chloride
CR Concentration Ratio (solar collector)
C Geometric concentration ratio
Cε Effective concentration ratio
d Reactor tube diameter mm
D Aperture width (parabolic collector) mm
f Focal length (parabolic trough collector) mm
ECTFE Ethylenechloride tetrafluroethylene
ETFE Ethylenetetrafluoroethylene
FEP Fluorinated ethylenepropylene
IEP Isoelectric point
IFSH Institut für Solarenergieforschung GmbH (Hanover, Germany)
IR Infrared light (solar radiation)
NREL National Renewable Energy Laboratories
PSA Plataforma Solar de Almería
PTFE Polytetrafluoroethylene
PTC Parabolic trough collector
PVDF Polyvinylidene fluoride
PZC Point of zero charge
Re Reynolds number
UV Ultraviolet light (solar radiation)
TFE Tetrafluoroethylene
θA Semi-acceptance angle (parabolic collectors) degrees
ρs Specular reflectance
Optical error (reflective surfaces) mrad

6.1. Solar collector technology • non-concentrating or low-temperature, up to 150 °C,

• medium-concentrating or medium temperature,
generalities from 150 °C to 400 °C ,
• high-concentrating or high temperature, over 400 °C .
Traditionally, different solar collector systems have
been classified depending on the level of concentration This traditional classification considers only the
attained by them. The concentration ratio (CR) can be thermal efficiency of the solar collectors. However, in
defined as the ratio of the collector aperture area to the photocatalytic applications, the thermal factor is
absorber or reactor area. The aperture area is the area usually irrelevant (as already explained in Chapter 4),
intercepting radiation and the absorber area is the area whereas the amount of useful radiation collected (in the
of the component (whether fully illuminated or not) case of the TiO2 catalyst, with a wavelength shorter then
receiving concentrated solar radiation. This CR is 385 nm) is very important.
directly related to the working system temperature and, Non-concentrating solar collectors (Figure 6.1) are
according to this criterion, there are three types of static and non-solar-tracking. Usually, they are flat
collectors: plates, often aimed at the Sun at a specific tilt,
depending on the
Solar Detoxification Technology 83

geographic location. Their main advantage is their refracting surfaces (similar to convex lenses) that deviate
simplicity and low cost. An example is domestic hot- the radiation at the same time that they concentrate it
water technology. Medium-concentrating solar onto a focus.
collectors concentrate sunlight between five and fifty High-concentrating collectors have a focal point
times, so continuous tracking of the Sun is required. instead of a linear focus and are based on a paraboloid
Parabolic trough collectors (PTC) and holographic with solar tracking. Typical concentration ratios are in
collectors (Fresnel lenses) are in this group. The first the range of 100 to 10,000 and precision optical elements
have a parabolic reflecting surface (Figure 6.2) which are required. They include parabolic dishes and solar
concentrates the radiation on a tubular receiver located in furnaces.
the focus of the parabola. They may be one-axis tracking, Up to now, the solar collectors used for
either azimuth (east–west movement around a north– photocatalysis have been in the two first categories. In
south-oriented axis) or elevation (north–south order to illustrate the variation in performance
movement around an east–west-oriented axis), or two- between different orientations, Figures 6.4 and 6.5
axis tracking (azimuth + elevation). Fresnel lens show, respectively, a comparative analy- sis of
collectors consist of medium-concentrating and non-concentrating solar
collector efficiency with regard to direct incident
radiation. Direct radiation is the radiation that has no
interference from the atmosphere and, consequently,
has a known direction, and can therefore be
concentrated. Global radiation is com- posed of direct
and diffuse radiation. The data represented in Figure 6.4
correspond to direct radiation in an ideal cloud- less
year (based on average meteorological data on sunny
days at the Plataforma Solar de Almería), and show
the energy available from direct radiation on the
aperture plane of a one-axis parabolic-trough collector
with different orien- tations: elevation tracking,
azimuth tracking and azimuth tracking slightly tilted 8°
from the horizontal.
The calculations performed are geometric and based
on the cosine of the incident angle, this angle being the
one formed by the solar ray with the line normal to the
FIGURE 6.1. Non-concentrating solar collectors for aperture plane of the collector. They allow the amount of
domestic water-heating applications. direct radi- ation available at any given time for each

FIGURE 6.2. Medium concentrating solar collector:

recircu- lating parabolic trough reactor for water
purification using titanium dioxide slurry at NREL
(Courtesy of the US National Renewable Energy FIGURE 6.3. A high-concentration solar collector: a fixed-
Laboratory). focus solar reactor (PSA, Spain).
84 Solar Detoxification

FIGURE 6.4. Yearly efficiency of solar collectors in an ideal cloudless year: a PTC-one axis with different orientations.

FIGURE 6.5. Yearly efficiency of solar collectors in an ideal cloudless year: flat plate with different inclinations.
Solar Detoxification Technology 85

configuration to be known. In Figure 6.4, it may be

observed that the annual efficiency of azimuth tracking
(east–west movement around a north–south-oriented
axis) is about 10 per cent better than elevation tracking
(north–south move- ment around an east–west-oriented
axis). In the first case, this efficiency increases notably in
the summer and decreases in the winter (identical in the
Northern and Southern hemi- spheres), whereas it is
almost constant year-round in the second case. A slight 8°
tilt to the south in the Northern hemi- sphere and the
opposite in the south increases yearly efficiency about 5 per
cent in the azimuth-tracking configuration, due to the
reduction of the cosine factor over the year.
In the case of non-concentrating flat-plate
collectors, it may be observed that efficiencies are lower
FIGURE 6.6. The first engineering-scale outdoor solar
than with one- axis PTCs, attaining maximum efficiency
detoxi- fication reactor using a one-axis parabolic trough
with an inclination (to the south in the Northern
collector. Part of the 465 m 2 parabolic trough system at
Hemisphere and to the north in the Southern
Sandia National Laboratory. (Courtesy of the US
Hemisphere) from the horizontal equal to the local
National Renewable Energy Laboratory.)
latitude. This configuration – that is – the angle of tilt set
at the angle of latitude of the site, maximizes the annual
energy collected in a flat-plate collector. Although the tion with a Pyrex tube through which contaminated water
calcu- lations made here are for a specific location and could flow. Since then, many different concepts with a
latitude, the comparisons of solar radiation collection wide variety of designs have been proposed and
and the conclu- sions obtained are qualitatively valid developed world- wide, in a continuous effort to improve
for any other location. performance and reduce the cost of solar detoxification
systems. Among these different concepts, the most
important with regard to the definition of the overall
6.2. Collectors for solar water system are those related to whether or not radiation must
detoxification: features be concentrated, to the type of
reflective surface to be used, to the way the water circulates
through the reactor (tube, falling film or stirred vessel),
6.2.1. Specific features of and to the way in which the catalyst is employed.
solar UV light utilization One of the most important reactor design issues is
decid- ing between concentrating or non-concentrating
The specific hardware needed for solar photocatalytic collector systems. Concentrating systems have the
appli- cations has much in common with that used advantage of a much smaller reactor-tube area, which
for thermal applications. As a result, reactors and could mean a shorter circuit in which to confine, control
photocatalytic systems have both followed and handle the contami- nated water to be treated. The
conventional solar thermal collector designs, such as use of high-quality ultraviolet- light-transmitting reactors
parabolic troughs and non-concentrating collectors. At and supported-catalyst devices also seems more logical,
this point, their designs begin to diverge, as: from both an economic and an engi- neering point of
view, if concentrating collector systems are to be used.
• The fluid must be exposed to ultraviolet solar
Nevertheless, concentrating reactors have two
radiation, therefore the absorber must be UV-
important disadvantages compared to non-concentrating
ones. The first is that they cannot concentrate (that is,
• Temperature does not play a significant role in the
use) diffuse solar radiation, which is unimportant for
photocatalytic process, so no insulation is required.
solar thermal applica- tions because diffuse radiation is a
The first engineering-scale outdoor reactor for solar small fraction of the total solar radiation. However, solar
detoxifi- cation was developed by Sandia National photocatalytic detoxification with TiO2 as a catalyst uses
Laboratories (United States) at the end of the 1980s (see only the UV fraction of the solar
Figure 6.6) in which a parabolic-trough solar thermal
collector was modi- fied simply by replacing the
absorber/glazing-tube combina-
86 Solar Detoxification

spectrum, and since this radiation is not absorbed by solar photocatalytic systems is the need for a catalyst; in
water vapour, as much as 50 per cent of it, or even more the case of TiO2 it can be deployed in several ways, such as
in very humid locations or during cloudy or partly cloudy slurry or as a fixed catalyst (such as a fibreglass matrix
periods, can be diffuse. As non-concentrating solar inserted in the reactor tube).
collectors can make use of both direct and diffuse UV
radiation, their effi- ciency can be noticeably higher. The
second disadvantage of concentrating collectors is their 6.2.2. Parabolic trough collectors
complexity, cost and mainte- nance requirements. The
consequence of these disadvan- tages is that present Solar photoreactors for water detoxification were
state-of-the-art favours the use of non-concentrating originally designed for use in line-focus parabolic-trough
reactors for solar photocatalytic applica- tions. An concentra- tors. This was in part because of the historical
additional disadvantage of concentrating reactors is that emphasis on trough units for solar thermal applications.
their quantum efficiency is low, due to a square root Furthermore, PTC technology is relatively mature and
rather than linear dependence of rate on light flux, as existing hardware could be easily modified for solar
already explained in Chapter 4. photocatalytic processes. PTCs, considered medium
For many of the solar detoxification system concentrating collectors, are of two types:
components, the equipment is identical to that used for
other types of water treatment, and construction • one-axis parabolic trough, or
materials are commer- cially available. Most piping may • two-axis parabolic trough.
be made of polyvinylidene fluoride (PVDF), chlorinated
polyvinyl chloride (CPVC), or simply polyethylene. As explained previously, the first engineering-scale
Piping, as well as the rest of the materi- als, must be facility was developed from one-axis PTCs (Sandia
resistant to corrosion by the original contami- nants and National Labs, United States, 1989) and the second from
their possible by-products in the destruction process. two-axis PTCs (Plataforma Solar de Almería, Spain,
Neither must materials be reactive, interfering with the 1990, Figure 6.7). Both facilities are quite large pilot
photocatalytic process. All materials used must be inert plants (hundreds of square meters of collecting surface)
to degradation by UV solar light in order to be compat- and can be considered the first steps in industrialization
ible with the minimum required lifetime of the system of the photocatalytic process.
(ten years). Although one-axis tracking has been demonstrated
Optical material requirements are similar to other to be the most economically suitable for solar thermal
closed solar systems, but photocatalytic reactors must app- lications, certain peculiarities of photocatalytic
transmit UV light efficiently because of the process research make two-axis-tracking PTCs efficient for
requirements. In some cases, when the steam pressure of finding out exactly
contaminants in water is sufficiently low, a closed
system might not be required, in which case a
transmissive UV containment material could be avoided.
All pipes, reactor and connection devices must be
strong enough to withstand the necessary water-flow
pressure. Typical parameters are 2 to 4 bars for nominal
system pressure drop and a maximum of 5 to 7 bars.
Concentrating system materials must also be able to
withstand possible high temperatures that could result
from absorption of concentrated visible and infrared light
in the reactor.
With regard to the reflecting/concentrating
materials, aluminium is the best option because of its
low cost and high reflectivity in the solar terrestrial UV
spectrum. Commercially available film products
incorporate a thin aluminium foil with an acrylic coating.
The last peculiarity of FIGURE 6.7. A CIEMAT 384 m2 solar detoxification
facility using two-axis parabolic trough collectors, at
Plataforma Solar de Almería (PSA, Spain).
Solar Detoxification Technology 87

how much radiation reaches the photoreactor at any

given time and also, therefore, permit accurate evaluation
of the other parameters related to the photocatalytic
process. This accuracy allows comparison of
experiments carried out in such large photoreactors with
lab-scale photoreactors, where the calculation of incident
radiation is much easier. This also makes it possible to
reduce the number of variables during testing, using the
knowledge acquired by other authors.
The equation of the parabola is:
y (6.1)
4f FIGURE 6.8. Solar ray reflection on a one-axis parabolic
trough collector.
where f is the focal length. If D is the aperture width and
d, the reactor tube diameter, the geometric concentration
of the collector C is: Typical overall optical efficiencies in a PTC are in the
range of 50 to 75 per cent, with the following
D breakdown:
πd (6.2) • tracking system: 90 to 95 per cent
The basic components of a parabolic-trough collector • reflector/concentrator (reflectivity): 80 to 90 per cent
for photocatalytic applications are: the reflecting • absorber/reactor (transmittance): 80 to 90 per cent
concentrator, the absorber tube (photoreactor), the • mechanical collector errors: 90 to 95 per cent.
drive-tracking system and the overall structure. Of these,
the last two do not differ in photocatalysis from the Parabolic-trough collectors make efficient use of direct
applications for which they were originally designed, solar radiation. As an additional advantage, the thermal
and are identical to those existing for thermal energy collected from the concentrated radiation can be
applications. Reflective surfaces and photoreactor used in parallel for other applications. The size and
technology are specifically discussed in Section 6.4.1, as length of the reactor is smaller, receiving a large amount
they can be considered independently of the solar of energy per unit of volume, so handling and control of
collector used. The collector structure supports the the liquid to be treated is simpler and cheaper, and the
reflecting concentra- tor system, which reflects direct risk of leaks, which in many cases can be dangerous, is
insolation onto the receiver tubes. Two-axis PTCs lower. In general, this can also be translated into a
consist of a turret on which there is a platform reactor able to withstand higher pressures and able to
supporting several parallel parabolic trough collec- tors, employ potentially costly supported-catalyst
with the absorber in the focus. The platform has two configurations.
motors controlled by a two-axis (azimuth and
elevation) tracking system. Thus the collector aperture 6.2.3. One-sun (non-concentrating)
plane is always perpendicular to the solar rays, which collectors
are reflected by the parabola onto the reactor tube at
the focus through which the contaminated water to be One-sun non-concentrating collectors (CR = 1) are, in
detoxified circulates. One-axis PTCs have only one prin- ciple, cheaper than PTCs, as they have no moving
motor and a one-axis solar-tracking system; the reactor parts or solar tracking devices. They do not concentrate
tube (linear focus of the parabola) is then positioned in radiation, so their efficiency is not reduced by factors
the same plane as the normal vector of the collector associated with reflection, concentration and solar
aperture plane and the solar vector (see Figure 6.8). tracking. Manufacturing costs are cheaper because their
The angle formed by these two vectors is called the components are simpler, which also means easy and
incident angle of solar radiation. low-cost maintenance. Also, the non-concentrating
After all optical losses have been considered, the collector support structures are easier and cheaper to
effective concentrating ratio of PTCs is usually between install and the surface required for their
5 and 20.
88 Solar Detoxification

installation is smaller, because as they are static they do front or back end of a combined solar/microbiological
not project shadows on the others. treatment scheme (see also Section 8.1).
Based on extensive effort in the designing of small Any falling film or flat reactor must be covered to
non- tracking collectors, a large number of non- avoid direct contact with the atmosphere. The use of an
concentrating solar reactors have been developed for uncovered reactor is not recommended because of many
solar photocatalytic applications, which can be classified factors, includ- ing loss of volatile contaminants and dust
as follows: and dirt inside the reaction mixture.
• Trickle-down FLAT PLATE: based on a tilted plate Although one-sun designs possess notable
facing the Sun, over which the water to be treated falls advantages, the design of a robust one-sun photoreactor
slowly; the catalyst is fixed on the plate surface. is complex, as a result of the need for a large area of
weather-resistant and chemically inert ultraviolet-
• FREE-FALLING film: similar to the trickle-down flat plate,
transmitting reactors. The amount of materials
but with a higher flow rate and normally with a
required makes it necessary for them to be relatively
catalyst attached to the surface on which the liquid to
inexpensive. Those that best meet these requirements are
be treated circulates. It is usually open to the
certain types of plastics, for instance,
atmosphere, so it can be used only when volatile
compounds are not present.
• Pressurized FLAT PLATE: consisting of two plates
between which water circulates, using a separating wall
(which can be filled in with fibre) to which the catalyst
is attached.
• TUBULAR: this kind of collector usually consists of
many small tubes connected in parallel to make the flow
circu- late faster than a flat plate, but functioning
basically the same way.
• SHALLOW SOLAR ponds: this is a very interesting variety,
as pond reactors are easily built on-site, especially for
indus- trial wastewater treatment. Since
manufacturing indus- tries already use ponds for
microbiological treatment of wastewater, shallow
solar ponds can be used for the b)

FIGURE 6.9. Experimental set-up of a thin-film fixed-bed

reactor tested by ISFH at PSA installations. The housing is FIGURE 6.10. One-sun water treatment reactors with PTFE
made of Plexiglas and the catalyst is fixed on a flat glass tubes at NREL: early reactor a) and reactor with titanium
plate. (Courtesy of the Institut für Solarenergieforschung dioxide immobilized on fibreglass bundles b). (Courtesy of
GmbH, Hanover, Germany.) US National Renewable Energy Laboratory.)
Solar Detoxification Technology 89

fluoropolymer films, but these, although highly

versatile, possess lower tensile strength than the rigid
glass pipe and reduce the pressure capacity of the
photoreactor system. This combination of low pressure
capacity and large volume, cou- pled with the need to
either keep a catalyst slurry suspended or to ensure good
mass transfer to a supported catalyst, requires carefully
designed fluid control.
Non-concentrating systems require significantly more FIGURE 6.11. Solar reflection on a CPC collector.
pho- toreactor area than concentrating photoreactors and,
as a result most of the reactor tube circumference is
as a result, full-scale detoxification systems (hundred of
illumi- nated; but because of the ratio of CPC aperture to
square meters of collectors) must be designed to withstand tube diam- eter, no single point on the tube receives
the oper- ating pressures anticipated for fluid circulation
much more than one Sun of UV light. As a result, the UV
through a large field. As a consequence, the use of tubular
light incident on the reactor is very similar to that of a
photoreactors has a decided advantage because of the one-sun photoreactor and, as in the case of flat-plate
inherent structural efficiency of tubing; tubing is also
collectors, maximum yearly efficiency is normally
available in a large variety of materials and sizes and is a obtained at the same collector angle inclination as the
natural choice for a pressurized fluid system. Finally, its
local latitude.
construction must be economical and should be efficient
Performance is very close to that of the simple tubular
with low-pressure drop.
photoreactor, but only about one-third of the reactor
If a supported catalyst is used, the photoreactor has to tube material is required. As in a parabolic trough, the
be much larger than in a concentrating system. If the
water is more easily piped and distributed than in
catalyst is circulated in a slurry, the design would have to many one-sun designs. All these factors contribute to
avoid low- flow regions where the catalyst could settle
excellent CPC collector performance in solar
out of suspen- sion, which means that turbulent flow photocatalytic applications.
must be assured throughout the hydraulic circuit.
The explicit equation for a CPC reflector with a tubular
Containment of volatile organic contaminants to prevent
reactor can be obtained from Figure 6.11; a generic reflector
their escape into the atmos- phere is also of concern in a
point S can be described in terms of two parameters, angle
large reactor.
θ, subtended by lines originating at O (centre of the
reactor tube) to A and R, and distance ρ, given by segment
6.2.4. Compound parabolic
concentrator (CPC) θ  OA a OR (6.3)
ρ  RS (6.4)
CPC collectors are a very interesting cross between
concentrators and one-sun systems and are one of the RS being tangent to the reactor tube at R. One important
best options for solar photocatalytic applications. CPCs parameter for CPC definition is the angle of acceptance
have been found to provide the best optics for low 2θA, which is the angular range over which all or almost
concentration systems, and these non-imaging all rays are accepted (that is, reflected into the reactor
concentrators are exten- sively employed for evacuated tube) without moving the collector.
tubes. CPCs are static collec- tors with a reflective The solution is given in two separate portions, an
surface following an involute around a cylindrical ordi- nary involute from A to B and an outer portion from
reactor tube. This can be designed with CR = 1 B to C:
(or near 1), giving it the advantages of both PTCs and
ρ  rθ for θ  θA  π 2 part AB of the
one- sun collectors. (6.5)
Thanks to the reflector design, almost all the UV
radiation arriving at the CPC aperture area (not only θ  θ A  π 2  cos(θ  θ A )
direct, but also diffuse) can be collected and is available ρr
for the process in 1  sin(θ  θ A )
π 3π
for θA   θ   θ part BC of the curve (6.6)
the reactor. The UV light reflected by the CPC is more or
less distributed around the back of the tubular A
2 2
photoreactor and
90 Solar Detoxification



FIGURE 6.12. Obtention of CPC involute.

The CPC concentration ratio (CR) is given

1 (6.7)
sinθ A

In the special case of θA = 90°, CR = 1 and every CPC

curve is an ordinary involute (points B and C are
coincident). So, optimum CPC acceptance half-angles
(θA) for photocat- alytic applications are obtained from
60 to 90 degrees either side of the normal. This wide
acceptance angle allows the reflector to direct both
direct-normal and diffuse sunlight onto the reactor, as
UV light collection is not highly sensitive to these
acceptance angles. An additional advantage is that these
wide acceptance reflectors skip reflector-tube alignment
errors, which is an important virtue for a low-cost
photo- reactor array.
CPC reflectors are usually made of polished
aluminium. The structure can be a simple photoreactor
support frame with connecting tubing. Since such FIGURE 6.13. View of CPC shape a) and CPC
reflectors are significantly less expensive than tubing, photoreactor array b). (PSA, Spain).
their use is very cost-effective compared to deploying
non-concentrating tubular photoreac- tors without using
any reflectors, and preserves the advan- tages of using
tubing for the active photoreactor area.
Solar Detoxification Technology 91

6.2.5. Holographic collectors photons, the holographic concentrator could very well be
a good way not only to supply these photons while
filtering out those that are unnecessary, but also to
Another innovative idea is the holographic concentrator. minimize thermal heating of the photoreactor.
This concept has been extensively explored with regard Solar holographic concentrators can theoretically
to solar thermal applications as well as in the area of avoid the technological complexities associated with the
concentrators for photovoltaic systems. Holographic curved support of conventional reflecting concentrators.
surfaces are highly wave- length-selective and their However, many significant issues remain unresolved.
development for solar thermal applications, which One of the most important is whether the devices
require the full solar spectrum, has proved to be a very developed in research programs can be feasibly
difficult task. However, holographic tech- nologies could translated to realistic cost-effective scale for the
be very appropriate for narrow-wavelength- band collection of solar energy. Also, there are not many
processes such as photovoltaics and photochemistry. holographic materials in the UV wavelength region, and
Holography is basically a diffractive technology. It their ability to survive in an outdoor environment is
records the interference pattern between a reference unknown.
beam of highly coherent monochromatic light and an
object beam using the same light source. In the case of
solar holographic concentrators, the object beam is the
one in the focal region (point-focus or line-focus
6.3. Concentrated versus
concentrator) and the reference beam is the virtual image non-concentrated sunlight
of the Sun as a source. Once cre- ated, sunlight incident
on the holographic optic element will focus back to the The use of concentrating or non-concentrating collectors is
focal region by either transmission or reflection, based on laboratory and engineering-scale experiments
depending on whether the reference beam used to create carried out by many different research groups. As already
the hologram strikes the diffractive material from the mentioned, during these experiments it was found that non-
same or the opposite direction as the object beam. As a concentrating collectors have the advantage of collecting
holo- gram is a passive optical device, it is not possible to diffuse light. UV light is more susceptible to Rayleigh
track the Sun without moving the device. scatter- ing by atmospheric gases, mainly water vapour,
Normally, holographic elements are made with than visible light. The same mechanism scatters blue light
highly coherent monochromatic laser light, in order to more than red light, which is what causes the sky to appear
obtain the most efficient hologram at that wavelength. blue. Because of this scattering, as much as half of the UV
Efforts carried out with the Sun as the source of light radiation arrives at the Earth’s surface as diffuse light, even
have resulted in a maximum usable bandwidth of about on a clear day. Near- UV wavelengths (from 285 to 385 nm)
100 nm, obviously insufficient for thermal applications. comprise only 2 to 3 per cent of the energy in direct
However, as the photocatalytic process with TiO 2 uses sunlight, but they make up 4 to 6 per cent of combined
300 to 385 nm diffuse and direct sunlight.
Concentrating collectors focus only direct sunlight
and cannot collect diffuse light. Thin clouds, dust, and
haze reduce the direct-beam component of sunlight more
than the diffuse component. As a result, non-
concentrating collectors can use a resource that is not
only larger but also less variable than that available to
concentrating collectors, permitting, in many locations,
continual operation of the non-concentrating
detoxification system. Under cloudy conditions, non-
A Holographic optic element Holographic optic element concentrating devices can continue operating
B Holographic optic element
(although at lower rates), while a trough unit would
have to shut down. This fact has been successfully
demonstrated even in north- ern European locations
Focus with small solar detoxification pilot plants (Figure
FIGURE 6.14. Holographic concentration of solar light.
92 Solar Detoxification

Radiation distribution on absorber (o) of CPC type concentrator

Solar hour : 06:00 to 18:00 DER–ITE/INETI
Inc.angles : 90 to –90 degrees

FIGURE 6.15. Solar detoxification pilot plant in Cologne

FIGURE 6.16. Flux distribution on an absorber in the
(Courtesy of Deutsche Forschungsanstalt für Luft- und
course of the solar day (6:00 to 18:00). Simulation of
Raumfahrt – DLR).
CPC behav- iour with the following data: collector
An additional benefit of a non-concentrating orientation east–west; semi-acceptance angle 60°;
design is that the efficiency of the photocatalytic process truncation angle 80°; absorber radius 13.6 mm; optical
frequently decreases as light intensity increases (as gap 2 mm; concentration ratio 1.17. (Courtesy of the
explained in Chapter 4). This behaviour means that Renewable Energy Dept. INETI, Portugal.)
catalysts are unable to process all of the UV energy
available to the desired pollutant destruction reactions smaller reactor-tube area requirement (the smaller circuit
when subjected to high UV fluxes. Lower light intensities provides better control and handling of the contaminated
also slow down the rate of recombination, thus yielding water to be treated). Nevertheless, state-of-the-art
higher efficiency (see Figure 4.8). Nevertheless, technol- ogy cannot match the higher efficiency of non-
different researchers have obtained different results when concentrating systems. This has been demonstrated by
testing with low radiation intensities; so presumably, they many comparative efficiency studies, such as the one
are significantly affected by experimental conditions. performed by NREL (National Renewable Energy
Some authors attribute the transition of Laboratories, United States) with seven different small
r  f (I 0.5 ), to the excess of photogeneratedr species –
 f (I1.0 )(e ,
collectors (from 18 to 157 litres total capacity and from
+ •
h and OH). At higher radiation intensities, another 0.4 m2 to 53 m2 reflecting surface). Results showed that
 one-sun collectors were significantly more efficient than
transition from r  f (I0.5 )  r  f (I0 ) is produced. At
this moment, the photocatalytic reaction departs from concentrating collectors.
its dependence on radiation received, depending only As a consequence of the above, although it was
on mass transfer within the reaction, so the rate remains first thought that PTCs were the ideal technological
constant although radiation increases. This effect may be alternative, given their high cost and the fact that they
due to several causes, such as the lack of electron can only be oper- ated with direct solar radiation
scavengers (that is, O2), or organic molecules in the (implying location only in highly insolated areas), the
proximity of TiO2 surface and/or excess of products question has been decided in favour of the static non-
occupying active catalyst centres and so on. Actually, concentrating collector alterna- tive. Their intrinsic
this phenomenon appears to be more frequent with simplicity, low maintenance and operat- ing costs and
supported catalysts and/or at slow mixing speeds. This potential for reducing manufacturing costs, make one-
implies a small catalyst surface is in contact with the sun systems the natural choice for solar water
liquid and there is less turbulence, which does not favour detoxification.
contact of reactants with the catalyst and dispersion of Among the different technologies previously
products. described, CPCs seem to be one of the best-developed
Finally, one-sun photoreactors have additional advan- options for system design and implementation. In this
tages, such as decreasing optical losses from reduced or specific case, and since the absorber is illuminated
non-existing reflective surface. On the other hand, differently during the morning and the afternoon, if the
among possible advantages of concentrating systems is a radiation distribution is integrated over the solar day an
much almost regular distribution of light is
Solar Detoxification Technology 93

obtained along the reactor perimeter (Figure 6.16). This a light on baking flour, for example, you will not see a
agrees with the optical characteristic of low- direc- tionally shiny component. The powder will
concentrating CPC-type collectors, which can collect appear uni- formly bright from every direction. Many
within their accept- ance angle the radiation coming from reflections are a combination of both diffuse and
the hemisphere and place it on the absorber perimeter. specular components. One manifestation of this is
spread reflection, which has a dominant directional
component that is partially diffused by surface
6.4. Technical issues irregularities (Figure 6.17).
In the case of solar detoxification applications, the
strictest requirements are those of PTCs; for example, UV-
mirror materials need to have a specular reflectance of
In addition to the solar collector type, the most between 300 to 400 nm in order to achieve concentration
important ratios of from 1 to 20. For this configuration, the
technical issues related to solar detoxification hardware effective concentration ratio (Cε) can be related to the
are the reflective surface and the reactor tube, which optical per- formance parameters by:
are addressed here in more detail.
 σ  
6.4.1. Reflective surfaces C  C ρ sum  (6.8)
ε  s 
 σ 
The optical quality requirements of reflective surfaces for
solar applications are usually related to the concentration Where: C = concentration in the absence of surface
required by the particular application under and tracking errors, as defined in Equation 6.2; ρ s =
consideration. The higher the concentration desired, the specular reflectance; σsum = half angular extent of the
stricter the requirements for quality of parameters. Light Sun (Gaussian distribution) = 2.73 MRAD; and σ = total
reflected off a polished or mirrored surface obeys the optical error, which is a function of the slope, specularity
law of reflection: the angle between the incident ray and and tracking errors.
the normal to the surface is equal to the angle between The greater the errors, and particularly the reflective
the reflected ray and the normal. When light reflects off a sur- face errors, the lower the effective concentration
rear surface mirror, the light first passes through the ratio. The reverse is also true: the lower the effective
glass substrate, resulting in reflection losses, secondary concentration ratio is, the higher the optical errors may
reflections, refraction, absorp- tion and scattering of light be and therefore, the lower the quality of reflective
passing through the transparent substrate (second-surface surface required. This is an important additional factor
mirrors). Precision optical systems use first-surface in favour of low or non- concentrating systems, as these
mirrors that are aluminized on the outer surface to avoid lower-quality requirements (lower specular reflectance)
these phenomena. translate directly into lower manufacturing cost, as the
When light obeys the law of reflection, it is termed a reflector element can represent a considerable fraction of
‘specular reflection.’ Most hard polished (shiny) surfaces are collector cost.
primarily specular in nature. Even transparent glass Another important factor is the reflective base
specu- larly reflects a portion of incoming light. Diffuse material. For solar photocatalytic applications, the
reflection is typical of particulate substances like reflective surface must clearly be made of a highly
powders. If you shine reflective material for ultraviolet radiation. The
reflectivity between 300 and

Specular Diffuse Spread

FIGURE 6.17. Specular, diffuse and spread reflection from a surface.

94 Solar Detoxification

The surfaces currently available that best fit these

require- ments are:
• electropolished anodized aluminium
(electrolytically formed aluminium oxide outer layer)
• organic plastic films with an aluminium coating (multi-
layer ‘sandwich’-type plastic–aluminium–plastic
Anodized coatings with thin oxide can provide good pro-
tection against some chemicals and good resistance to
abra- sion. Typically, thin (2–3 µm) oxide layers are
used to provide some measure of resistance to abrasion,
but they give little protection against moisture or
pollutants. Thicker oxide layers (up to 50 µm) are usually
FIGURE 6.18. Reflectivity of fresh metal coatings for specified when anodized aluminium is intended for
mirrors. engineering/marine applications
400 nm of traditional silver-coated mirrors is very but, unfortunately, such a coating results in considerably
low (reflected radiation/incident radiation) and lower reflectance in the UV range.
aluminium- coated mirrors are the best option in this An interesting alternative approach is to cover the
case (Figure 6.18). Aluminium is the only metal surface aluminium with a protective acrylic lacquer. Acrylic
that is highly reflec- tive throughout the ultraviolet lacquer- ing provides impressive outdoor resistance
spectrum. Reflectivities range from 92.3 per cent at 280 (more than 1,000 hours in a salt-spray fog chamber
nm to 92.5 per cent at 385 nm. Comparable values for without significant degra- dation), but also reduces UV
silver are 25.2 per cent and 92.8 per cent, respectively. reflectance. A compromise between outdoor resistance
A fresh deposited aluminium surface is fragile and and UV reflectance could be an optimum solution.
needs to be protected from weathering and abrasion, Another possible solution is aluminium-coated plastic
but the conventional glass cover used for silver- film. Several commercial coated plastic film products
backed mirrors has the drawback of significantly filtering have been used successfully in parabolic-trough
UV light (an effect that is duplicated due to the light applications.
path through the glass). The thin oxide layer that forms • ECP-244 (3M film, no longer manufactured)
naturally on aluminium is not sufficient to protect it
consisting of a 10-µm-thick aluminium surface
in outdoor environments. Under such exposure
conditions, the oxide layer continues to grow and covered by a 76-µm- thick acrylic surface. Average
UV reflectance drops off reflectivity of new film between 280 and 385 nm
thick is about 63 per cent.
dramatically. • SA-85P (manufactured also by 3M); this film has a
Various optical technologies require UV-reflective 50-µm-thick polyester backing, a 10-µm-thick alu-
ele- ments, such as UV mirrors for medical imaging, minium surface, and a very thin 2.5-µm-thick
astronomical telescopes, microscopy, UV curing, indoor acrylic covering. Average reflectivity in the same
lighting, microlithography, industrial micro-machining range is 87 per cent for new films. The superior
and UV laser reflection. Nevertheless, the specific reflectivity is due to the much thinner protective
requirements of these applications are very different from acrylic coating.
those of solar ones, mainly with regard to outdoor • ECP-305 (3M); a silver reflective film similar in
durability. The ideal reflective surface for solar composi- tion to SA-85.
photocatalytic applications must be:
• Other companies and even some research
• highly reflective in the UV range, institutions have developed similar coated films,
• acceptably durable under outdoor conditions for usually a combina- tion of over-100-µm-thick
extended service lifetimes, polyester backing on 10-µm aluminium foil with a
• reasonably priced, to permit the technology to be thin outer film of acrylic or a copolymer that is
com- petitive against alternative technologies. highly resistant to outdoor exposure and has good UV
Solar Detoxification Technology 95

100 D 100
90 90
80 80

Spectral total reflectance

70 70

60 60

50 50
40 40

30 A 3M ECP-244 aluminized film 30

B 3M SA-85 aluminized film
20 20
C Hostaflon ET aluminized polymer
10 D ALANOLD UV-MIRO aluminium 10
0 0
280 290 300 310 320 330 340 350 360 370 380
Wavelength (nm)

FIGURE 6.19. Reflectivity of different aluminium and plastic film surfaces. (Curve (d) courtesy of Alanold Aluminium-
Veredlung GmbH & Co., Germany).

Normally, because of their lack of rigidity, these films must residence times inside the reactor, resulting in decreased
be bonded over a stiff substrate and about 2 per cent per- formance compared to an ideal-flow situation. If the
specular reflectivity could be lost in this process. Also, catalyst is used in suspension (a slurry in the case of
the reflectivity of each film at the end of its lifetime TiO2), the Reynold number (Re) must always be over
(from five to ten years) would be only 88 per cent of the 4,000 in order to guarantee turbulent flow. This is
original bonded value. critical in avoiding catalyst settlement. Another
important design issue is that internal reactor materials
must not react with either the catalyst or the pollutants to
6.4.2. Photocatalytic reactor be treated or their by-products.
There is only a limited choice of materials that are
both transmissive to UV light and resistant to its
The requirements for the photocatalytic reactor are
similar to other advanced water or air oxidation destructive effects. Also, temperatures inside a one-
sun solar photocat- alytic reactor can easily exceed 40
processes, with the additional necessity of an
illuminated photocatalyst. The photocatalytic reactor °C, due to the absorption of the visible portion of the
must contain the catalyst and be transparent to UV solar spectrum. Therefore, a one-sun reactor must be
able to withstand summer temper- atures of around 60
radiation, providing good mass transfer of the
contaminant from the fluid stream to an illuminated to 70 °C to guarantee that no damage will occur that
photocatalyst surface, with minimal pressure drop across could reduce the flow. Finally, low pH resistance is
the system. needed, since the production of inorganic acids as
reaction by-products is quite normal (that is, the
As mentioned before, the square-root dependence on
destruc- tion of chlorinated hydrocarbons leads to the
light intensity provides better photo-efficiencies for
production of HCl).
one-sun designs, which leads to a flat-plate geometry. This
geometry is widely used for solar-powered domestic hot Common materials that meet these requirements are
water systems, in large part because of its simple design. fluoropolymers, acrylic polymers and several types of
Nevertheless, for water treatment, the reactor must be glass. Quartz has excellent UV transmission as well as
robust enough to work under usable water pressure, and good temperature and chemical resistance, but its slight
tube configurations clearly seem the most appropriate for advantage in terms of transmission in the terrestrial solar
fluid containment and pumping. Adequate flow spectrum does not justify its high cost, which makes it
distribution inside the reactor must be assured, as non- completely unfeasi- ble for photocatalytic applications.
uniform distribution leads to non-uniform
96 Solar Detoxification

FIGURE 6.20. Transmittance of different materials suitable for the manufacture of photoreactor tubes.

Plastics work well as long as they fulfil as fluoropolymer materials are poor IR-diffusers, but
transmittance, pressure and thermal resistance make excellent visible/UV diffusers (diffusion usually
specifications as well as maintaining their properties varies with wavelength).
during outdoor operation. Fluoropolymers are a good Acrylics could also potentially be used to enclose
choice of plastic for photoreac- tors due to their good UV the photoreactor. However, acrylics are very brittle and
transmittance, excellent ultraviolet stability and chemical would have to be employed in sheets, which increases
inertness. Several different types, such as ETFE their cost. On the positive side, acrylic polymer sheets
(ethylene-tetrafluoroethylene), PTFE (polytetrafluo- can be shaped with channels and flow patterns that could
roethylene), ECTFE (ethylenechloride- themselves be used as solar reactors. Other lower-cost
tetrafluroethylene), PVDF (polyvinylidene fluoride), polymers are available in tubular form, but none possess
FEP (fluorinated ethylene- propylene), PFA and TFE the necessary UV and chem- ical stability for
(tetrafluoroethylene), can be extruded into tubing and detoxification of water that may be contami- nated with a
used as a photoreactor. Tubular fluoropolymers are very variety of solvents or other pollutants. Also, low-cost
strong, possess excellent tear resist- ance and are polymeric materials are significantly more suscepti- ble to
flexible and lighter than glass. One of their greatest attack by pollutant molecules, and the dissolution of
disadvantages is that in order to achieve a desired organic contaminants in polymer materials could be a
minimum pressure rating, the wall thickness of a way of avoiding the degradation process.
fluoropoly- mer tube may have to be increased, which in Glass is another alternative for photoreactors.
turn will lower its UV transmittance. In addition, as a Standard glass, used as a protective surface, is not
result of the lack of rigidity, tube connections can satisfactory because it absorbs part of the UV radiation
withstand much lower pressures than glass tubes. ETFE that reaches it, due to its iron content. Borosilicate glass
and FEP are among the best candidates; ETFP has higher has good transmissive prop- erties in the solar range,
tensile strength (extrude-like) than FEP; this could mean with a cut-off of about 285 nm. Therefore, such a low-
thinner-walled tubes, resulting in cost savings (as less iron-content glass would seem to be the most suitable.
material is used) and higher UV trans- mittance – and Two undesirable effects reduce the performance of a
therefore higher photoreactor performance. The problem glass reactor for solar detoxification: increased
is that ETFE tubing is not as readily available as FEP absorption in the solar UV range between 300 and 400
tubing. FEP tubing with a 50 mm outer diameter and nm, and a further decrease of UV transmittance during
0.6 mm wall has a UV hemispherical transmittance (300 operation because of
to 400 nm), of 61.6 per cent. This light is transmitted as
Solar Detoxification Technology 97

the damaging impact of solar radiation in the same 50 ppm (Figure 6.22), but this incurs a corresponding
wave- length region (UV solarization). Both effects are increase in cost.
caused to a large extent by polyvalent ions that change Therefore, as both fluoropolymers and glass are
charge. The effect of the Fe ions in the glass, which valid photoreactor materials, cost becomes an important
change charge from Fe2+ to Fe3+ due to photo-oxidation issue. In large volumes, glass piping could be more
by photons with a wavelength below 400 nm, is expensive than fluoropolymer tubing, but from the
especially harmful. Furthermore, the oxidized Fe3+ ion perspective of perform- ance, the choice is the material
absorbs UV light. As a result, enhancement of that has the best combination of tensile strength and UV
transmittance in the 300 to 400 nm region could only be transmittance. On this basis, if a large field is being
accomplished by a strong reduction in iron content, designed, the large collector area means also a
down to considerable number of reactors and, as consequence, a
high system pressure rating. Fluoropolymer tubes are not
the best choice of material as high pressure is linearly
related to thickness and could result in higher cost. A
detailed analy- sis is recommended for any specific
One of the most important parameters in a tubular pho-
toreactor design is the tube’s diameter, as in both homoge-
neous or heterogeneous photocatalysis it must be
guaranteed that all arriving useful photons are kept inside
the reactor and do not go through it without intercepting a
catalyst par- ticle. The intensity of illumination affects the
relationship between reaction rate and catalyst
concentration. The disper- sion and absorption of light
causes photon density to dimin- ish almost exponentially
over the length of the optical path within a catalyst
suspension. At higher light intensity, catalyst
FIGURE 6.21. Glass-tube manufacture. Different composi- concentration can be higher.
tions mean that the glass can be used for a wide variety
of applications. (Courtesy of Schott-Rohrglas GmbH.)

FIGURE 6.22. Influence of iron on borosilicate glass light transmission (oxidative conditions). Samples: flat glass 3 -
mm in thickness. (Courtesy of Schott-Rohrglas GmbH.)
98 Solar Detoxification

1 cm particles suspended in the contaminated water, which

makes it necessary to recover them after treatment. This
process is addressed in the next section.
ID = 58 mm Supported catalyst configurations eliminate the need
for catalyst filtration (but with a significant reduction in
Illuminated ar–1ea system efficiency). The idea is to attach the catalyst to a
[TiO2] = 1 g L
support inside the reactor, as is done for gas-phase
stream treatment, which requires the catalyst to be
FIGURE 6.23. Zone of tubular reactor where light anchored onto some type of inert support. Desirable
penetrates if the catalyst concentration is 1 g L–1 (TiO2 characteristics of such a system would include being
heterogeneous photocatalysis). very active (comparable to slurries) and having a low
pressure-drop, a long lifetime, and a reasonable cost;
In the case of TiO2 heterogeneous photocatalysis, however, in the case of water treatments, this has not yet
when catalyst concentration is very high, a ‘screening’ been achieved.
effect pro- duces excessive opacity of the solution, Fixed-catalyst designs must address several problems.
preventing the cata- lyst particles farthest in from being As the catalyst must be exposed to sunlight and be in
illuminated and reducing system efficiency. The lower contact with the pollutant, the support must be configured
the catalyst concentration, the less opaque the to route the pollutant efficiently to the illuminated zone
suspension. One g L–1 of TiO2 catalyst reduces and, at the same time, maintain a high flow rate in the water
transmittance to zero in a 10-mm-inner-diameter – to ensure good mixing without significantly increasing
cylinder with concentrated light in a parabolic-trough system pres- sure, which usually means more power for
collec- tor (Figure 6.23). In a wider diameter tube, pumping, and thereby higher operating costs. Also, the
only an outer layer is illuminated. This means that a same criteria dis- cussed for photoreactor materials must be
larger inner reactor diameter permits use of lower kept in mind and applied when choosing a support.
optimum catalyst concentra- tions. Practical inner Supports tested so far have included fibreglass beads,
diameters for a tubular photoreactor would be in the metal fibres, steel mesh, alu- minium, many types of plastic
range of 25 to 50 mm. Diameters that are very small do (Figure 6.24) and ceramics such as alumina, silicon
not make sense because of the high associated pressure- carbide and silica, in the most diverse shapes. TiO 2
drop, and very large diameters imply a considerable dark coatings on tiny hollow glass beads called
volume, thus reducing overall system efficiency. ‘microballoons’ for catalyst removal by screening rather
than filtering have also been tested.
6.5. Catalyst issues Support of TiO2 on a stiff surface by adherence can be
achieved via several different techniques, such as dip
coating with solvents, deposits from precursors, vapour
The catalyst plays a major role, not only because of its deposition and sol-gel formation. Several important
impor- tance to the process, but also from a technological performance requirements are directly related to the
point of view. This is especially relevant in heterogeneous process used for cata- lyst fixation, such as the durability
photocatal- ysis, where Degussa P-25 titanium dioxide is of the coating and catalyst activity and lifetime. The
often the stan- dard particulate material against which surface area of the catalyst coating must also be
other catalysts have been, and continue to be, measured. considered, because a good contact of photons and target
TiO2 in aqueous-phase applications can be used in molecules with the catalyst is required for efficient
suspensions (slurry) or supported. In the first case, the performance.
catalyst must be recovered. The technological Studies performed to date have not yet identified a
implications of this are discussed below. fixed- catalyst system that performs as efficiently as
slurry systems. Several exceptions, such as a test
6.5.1. Slurry versus supported catalyst performed by NREL (United States) using silica beads
as the catalyst support, have approached the efficiency of
One of the major solar photoreactor system design issues the slurry – the high pres- sure-drop across the bed made
is whether to use a suspended or a supported catalyst. the silica-bead system imprac- tical. In aqueous systems,
The majority of experiments to date have used small compared to an unsupported
Solar Detoxification Technology 99

a) through the reactor and excellent fluid-to-catalyst mass


6.5.2. Catalyst recuperation and reuse

The need to remove the catalyst from the clean water
after treatment was initially considered a major
disadvantage of a slurry system; because ultrafiltration,
which was believed to be difficult and costly, would be
required for efficient sepa- ration of the submicron-sized
TiO2 particles from the slurry. Although it is true that
titanium dioxide powder particles are only about 0.03 µm
(30 nm) in diameter, or even smaller in some cases
(specially manufactured TiO2 can be up to
0.01 µm), once in water, the particles always
agglomerate into larger ones (from 0.3 to 0.6 µm),
simplifying the prob- lem considerably, since
microfiltration, which is much less expensive than
ultrafiltration, can be used. Carbon-graphite membranes
of 0.2-µm can effectively concentrate titanium dioxide
slurry from less than 1 g L–1 to up to 100 g L–1 with 2 to 3
bar trans-membrane pressure and permeate flow rates of
about 2,500 L h –1 per square meter of membrane.
Recovery rates of over 99 per cent have been obtained
0.5 µm membranes and no titanium dioxide has been
detected in water filtered with 0.2 µm membranes.
However, best performance can be obtained when
microfiltration is combined with previous titanium
FIGURE 6.24. a) Experimental concentrating solar dioxide sedimentation. About 90 per cent of the catalyst
reactor using titanium dioxide immobilized on glass can easily be recovered by sedimentation and the rest by
wool for treat- ing contaminated air streams. b) Parabolic microfiltration. This means a significant reduction in
trough reactor for water purification with immobilized time and electricity in the typical recovery process for
titanium dioxide. (Courtesy of the US National TiO2 concentrations of about 200 mg L–1. The lifetime
Renewable Energy Laboratory.) of membranes and the time between cleanings is also
increased considerably. This could be particularly
important with high volumes of water.
catalyst, immobilization of TiO2 results in a reduction in Titanium dioxide sedimentation is closely related to
per- formance around 60 to 70 per cent. An important colloidal stability and TiO 2 aggregation conditions. The
direct con- sequence of this fact is the necessity of suspension can easily be destabilized by adding an
multiplying by a factor of about three the size of the electrolyte (such as NaCl) and/or adjusting the pH to point
necessary solar collector field if similar efficiencies are of zero charge (PZC) and the isoelectric point (IEP) on
desired, making the overall sys- tem clearly less cost the surface of the catalyst particles, as both factors
efficient and less competitive than slurry systems. In modify the surface charge. Progressive particle
addition to the above, a key question is how long agglomeration (sizes from 1 to 10 µm) and settlement
supported catalysts will last in a real stream of water? A are then obtained. In the case of TiO2 (Degussa P-25),
short period of activity would mean frequent at concentrations of 200 mg L–1, the PZC is obtained at
replacement and, consequently, a marked rise in the about pH 7 (6.8 ± 0.2, in experiments at the Plataforma
overall system cost. Solar de Almería) when NaCl concentration is to the
In contrast to fixed catalyst configurations, slurry order of 10–6 molar.
configurations have the advantage of higher
throughputs (in the range of 200 to 400 per cent) a
low pressure-drop
100 Solar Detoxification

FIGURE 6.25. Sedimentation experiments at different pH. [TiO 2] = 0,2 g/L; [NaCl] = 0 M. The Y axis shows the
absorbance of the solution at 800 nm; no absorbance means absence of TiO 2 (see also Figure 4.2).

Therefore, more than 200 hours is needed for 75 per

cent of titanium dioxide to settle at pH 4.5, but at pH 7
only five hours of storage is needed to recover 90 to 95
Solar collectors are traditionally divided into three
per cent. This can be directly recovered in a conic-bottom
cate- gories: non-concentrating (or low temperature,
tank, and 5 to 10 per cent of the remaining catalyst can be
up to 150 °C), medium-concentrating (or medium
recovered by micro- filtration (see also Chapter 9).
temperature, from 150 °C to 400 °C ) and high-
Catalyst recovered, usually in highly concentrating (or high temperature, over 400 °C ).
concentrated slurry, can be reused – but not Concentrating solar systems make use of direct
indefinitely. Slurry lifetime has been tested with
radiation and need solar tracking mecha- nisms. Non-
satisfactory results under laboratory conditions (that
concentrating systems are much simpler as they do not
is, deionized water and only one contami- nant), but
need solar tracking and can collect direct and diffuse
with real water treatments catalyst lifetime would be
solar radiation with slightly lower annual efficiencies.
diminished because of contaminants. However, in
The specific hardware needed for solar
specific applications, tests have demonstrated that the
photocatalytic applications is very similar to that used
catalyst can be reused up to ten times, or even more,
for conventional thermal applications, with the
with- out any problem.
following major differences: the fluid must be exposed to
If part of the catalyst is lost in the drainage water, the the ultraviolet solar radiation (so the absorber must be
per- centage of catalyst that remains useful after each run
transparent to this radiation) and no thermal insulation is
is a rel- evant parameter. In field tests conducted by required, as the temperature does not play a significant
NREL in Florida, it was reported that approximately 10
role in the photocatalytic process. Besides, solar
per cent of the catalyst was washed away in the discharge
photoreactors for water detoxification were originally
water. In tests per- formed at the Plataforma Solar de designed to use line-focus parabolic-trough con-
Almería, 2.5 per cent of catalyst was lost. From this
centrators; non-concentrating collectors are used for solar
point of view, addition of 5 to 10 per cent more catalyst photocatalytic applications. They are more efficient
could be an interesting option to compensate possible
than concentrator-based systems because of the use of
loss and periodic replacement in case of catalyst
both direct and diffuse UV light, the square-root
poisoning. dependence between reaction rate and light intensity, and
their intrinsic simplicity.
Solar Detoxification Technology 10

CPCs (static collectors with a reflection surface BLANCO, J.; MALATO, S.; GARCÍA, G.; ARANDA, J.; SÁNCHEZ, M.;
following an involute around a cylindrical reactor tube) SICILIA, F. AND ROMÁN, M.I. 1991. FINAL CONFIGURATION of
are a very interesting cross between trough concentrators PSA SOLAR DETOXIFICATION Loop. Plataforma Solar de
and one-sun systems, and have been found to provide the Almería. (Technical Report: TR 06/91.)
best optics for low-concentration systems. CPCs designed
with a concentra- tion ratio of one, or near one, are FERNÁNDEZ, P.; DE LAS NIEVES, F.J. AND MALATO, S. 1999.
among the best options for solar photocatalytic TiO2 SEDIMENTATION Procedure. (Proceedings of the Second
applications. Aluminium is the only metal surface that Users European Workshop Training and Mobility of
offers high reflectivity values in the UV spectrum. Researchers Programme at Plataforma Solar de Almería,
Electropolished anodized aluminium and organic plastic Serie Ponencias), Madrid, CIEMAT.
films with an aluminium coating film are the most
appropriate reflective surfaces to use for solar
detoxification applications. Photocatalytic reactors must
be both transmis- sive and resistant to UV light. Reflector MATERIALS for SOLAR DETOXIFICATION of HAZARDOUS
Common materials that meet these requirements are WASTE, SERI/TP-257-4418. Golden, Colo.
fluoropolymers, acrylic polymers and borosilicate glass,
and tubular photoreactors designs are the best option. In KREIDER, J.F. 1979. Medium AND High TEMPERATURE SOLAR
TiO2 heterogeneous photocatalysis, sus- pended Processes. New York, Academic Press.
catalyst systems give efficiencies higher than sup- ported
catalysts. After use, TiO2 can be agglomerated and MAY, E.K.; GEE, R.; WICKHAM, D.T.; LAFLOON, L.A. AND
sedimented. Best recovery performances are obtained WRIGHT, J.D. 1991. Design AND FABRICATION of Prototype SOLAR
with a two-step process: sedimentation and RECEIVER/REACTORS for the SOLAR DETOXIFICATION of CONTAMINATED
microfiltration. WATER. Golden, Colo., Industrial Solar Technology
Corporation. (Final report for an NREL subcontract.)

PACHECO, K.; WATT, A.S. AND TURCHI, C.S. 1993. Solar

detoxification of water: outdoor testing of prototype
photoreactors. In: A. Kirkpatrick and W. Worek (eds),
BLAKE, D.M.; LINK, H.F. AND EBER, K. 1992. Solar ASME/ASES Joint SOLAR Energy Conference, pp. 43–9.
photocatalytic detoxification of water. In: K.W. Boer New York, American Society of Mechanical
(ed.), ADVANCES in SOLAR Energy, pp. 167–210. Boulder, Engineers.
Colo., American Solar Energy Society, Inc.
RABL, A. 1985. Active SOLAR Collectors AND Their APPLICATIONS.
BLAKE, D.M.; MAGRINI, K.; WOLFRUM, E. AND MAY, E.K. Oxford, Oxford University Press.
1997. WENDELIN, T. 1992. A survey of potential low-cost
Material issues in solar detoxification of air and water. concentrator concepts for use in low-temperature water
In: C.M. Lampert, C.G. Granqvist, M. Gratzel and S.K. detoxification. In: W.B. Stine, J. Kreider and K.
Deb (eds), OPTICAL MATERIALS Technology for Energy Efficiency Watanabe (eds), ASME INTERNATIONAL SOLAR Energy
AND SOLAR Energy Conversion XV, pp. 154–62. Bellingham, Conference,
Wash., SPIE – The International Society for Optical pp. 15–20. New York, American Society of
Engineering. Mechanical Engineers.
102 Solar Detoxification

Self-assessment questions Part B

Part A: True or false? 1. Why are concentrating solar systems more efficient
than non-concentrating ones when thermal applications
are considered?
1. The best yearly efficiency of a static non-
concentrating solar collector is obtained when the
2. Why are non-concentrating solar systems more
inclination from the horizontal is equal to the local
efficient than concentrating ones when UV
photocatalytic applications are considered?
2. Higher inclination than the local latitude gives
3. What must the diameter of a PTC reactor tube be if
higher efficiencies in winter than in summer (non-
the aperture width is 2.5 m and a geometric
concentrating solar collector).
concentration ratio of 10 is needed to be achieved?
3. Temperature plays an important role in
4. What is the geometric concentration ratio of a
solar photocatalytic degradation of water
CPC if the semi-acceptance angle is 70°?
5. A PTC with a geometric concentration ratio of 6 has
4. Diffuse UV solar light is a small portion of total
an effective concentration ratio of 4. What is the total
UV radiation, especially on a cloudy day.
optical error of the reflective surface if specular
reflectance is 85 per cent?
5. Non-concentrating solar collectors can use both
direct and diffuse solar light.
6. What is the main factor that limits the UV
transmittance of standard glass?
6. Concentrating solar collectors obtain better
efficiencies than non-concentrating ones in solar water
7. Why are reactor tubes of 200 mm inner diameter
detoxification applications.
not practical in TiO2 slurry systems?
7. Silver is not used in UV solar reflectors because of
8. Cite at least three of the main problems
its price.
of supported-catalyst systems.
8. Quartz is the best material for photocatalytic
9. How can titanium dioxide slurry particles be
reactors. aggregated and sedimented?
9. Supported catalyst systems have similar
10. Why must the Reynold number be higher than
efficiencies to slurry systems.
4,000 in TiO2 slurry systems?
10. Although the particles of titanium dioxide are very
small, once in water they agglomerate into ones that are
ten to twenty times larger.
Solar Detoxification Technology 10

Answers 4. Using Equation 6.7); C = 1/sin (70) = 1.06.

5. Using Equation 6.8); σ = 60.85 × 2.73/4 = 3.48 mrad.

Part A
6. The presence of Fe ions, especially Fe +3, which
1) True. 2) True. 3) False. 4) False. 5) True. 6) False. absorbs the UV light.
7) False. 8) True. 9) False. 10) True. 7. Because most of the inner volume of the reactor
would be dark due to the opacity of the catalyst
Part B
8. Less efficiency, much higher pressure-drop and the
1. Because they need a solar tracking system and need for periodic replacement.
the collectors are always directed towards the Sun.
9. TiO2 slurry suspensions can be destabilized by
2. Besides not having a solar tracking device, adding a small quantity of electrolyte, such as NaCl,
non-concentrating systems can collect UV solar light, and/or adjusting the pH to 7 in order to reach the point
which is an important contribution to their yearly of zero charge and the isoelectric point of the colloidal
efficiency. solution.

10. Because a turbulent flow regime must be assured

3. Using Equation 6.2); d = 2500/(10 × 3.1416) to guarantee adequate fluid mixture and to avoid
= 79.6 mm. possible catalyst settlement.
Solar Detoxification

Aim Objectives
This chapter describes the main proven When you have completed this chapter
and potential applications of solar you will have an appreciation of the
detoxification processes with following subjects:
demonstrated technical feasibility. The
advantages and limitations of these 1. The limiting factors and necessary
applications are also discussed. conditions of feasible solar
detoxification applications.

2. The main applications related

to contaminated water.

3. The technical considerations

for gas-phase treatment
106 Solar Detoxification

NOTATION on transferring laboratory research carried out by photo-

chemical groups to solar-engineered testing using
Symbol Description existing technology. These first results aroused
AOT Advanced oxidation technologies enthusiasm in the photochemical research community.
BTEX Benzene, toluene, ethyl benzene, and Their extrapolation to practical situations assumed an
xilenes BWTP Biological wastewater treatment plant ability to degrade almost any organic contaminant as
well as some metals. Later, more appropriate collectors
CVOC Chlorinated volatile organic compounds
and reactors were developed, but the need to understand
EDTA Ethylene diamine tetraacetic acid
the fundamentals of certain aspects of the reaction led to
EPA Environmental Protection Agency (USA) an increasing number of studies on kinet- ics,
HMX 1,3,5,7-tetranitro-1,3,5,7-tetrazacycloctane mechanisms, performance of mixtures and operating
IR Infrared light parameters, with varying results. It was a period of both
LLNL Lawrence Livermore National Laboratory promising and discouraging results. At present, a third
(USA) NREL National Renewable Energy Laboratory sce- nario seems to be underway, in which the boundary
(USA) PAH Polynuclear aromatic hydrocarbons condi- tions of applications have been determined and
the technology is focusing on a few specific initial
PCP Pentachlorophenol
applications, with the peculiarity that early
PCO Photocatalytic oxidation processes development and still- unsolved questions coexist with
PSA Plataforma Solar de Almería (Spain) near-commercial and indus- trial applications of the
PTFE Polytetrafluoroethylene technology. As a result, the environmental market,
PVC Polyvinyl chloride although very receptive to clean energy sources, is still
RDX Hexahydro-1,3,5-trinitro-1,3,5-triazine reluctant to adopt ‘high-risk’ initiatives with regard to solar
SNL Sandia National Laboratories (USA) detoxification processes.
TCE Trichloroethylene From the point of view of practical applications,
TNT Trinitrotoluene hetero- geneous photocatalysis with a semiconductor
VOC Volatile organic compounds (TiO2) is the process for which solar technologies are
most extensively employed today, although in some
specific applications; a well-known homogeneous phase
reaction, photo-Fenton, shows higher water-phase
7.1. Introduction degradation rates. In other solar detoxification options,
such as sensitized-photochemical oxidation by singlet
Detoxification is today the most successful oxygen, there is still very limited experi- ence to date.
photochemical application of solar photons, with a Photocatalytic oxidation processes (PCO) currently
number of installations and projects already in operation. under development are included in the same group of
This is due not only to the fact that solar detoxification is advanced oxidation technologies (AOT) as other
an outstanding demonstra- tion of how well suited solar radical- promoting processes. The main advantage of
energy is to environmental preservation, but also PCO over other AOTs is its potential for incorporating
because, unlike most photochemical processes, it is non- solar energy in the form of solar photons, whereby the
selective and can be employed with complex mixtures degradation process acquires significant additional
of contaminants. During the last decade, the number of environmental value. If solar photons can be collected
references and related patents on heteroge- neous and used efficiently at low cost, the number of
photocatalytic removal of toxic and hazardous com- opportunities for PCO may increase dramatically.
pounds from water and air can be counted in thousands, Although solar-driven PCO was at first considered a
and the number of applications and target compounds are univer- sal method of degrading organic pollutants, a
numerous. profusion of contradictory results in recent years
An analysis of the historical evolution of solar (positive results in almost real-life conditions together
photocatal- ysis clearly identifies three different stages of with other experimental results pointing out uncertainties
development. Initially, the efforts of solar-conscious and negative performance) has led to confused public
researchers focused perception. Solar PCO technology is currently viewed as
a very sensitive issue and scientific
Solar Detoxification 107

opinion is evolving towards a more conservative period useful is when contaminants are present in
of specific applications. concentra- tions that make conventional treatment
Within this context, treatment of industrial difficult.
wastewater, though difficult to develop, seems one of the
most promising fields of application for solar The above recommendations provide an indication of the
detoxification. The only general rule is that there is no type of industrial wastewater for which solar
general rule at all, each real case being completely detoxification can potentially be employed. Nevertheless,
different from any other. In some cases, the Photo- several additional conditions are needed before a
Fenton process has demonstrated higher degradation complete solar detoxification feasibility study can be
efficiencies than heterogeneous TiO 2 photocatalysis, but undertaken.
in others, the Fe cycle is affected by the contaminants
and Photo-Fenton does not work properly. As a • Throughputs should be REASONABLE: spatial velocities or
sur- face throughputs in the solar collector and solar-
consequence, preliminary research is always required to
assess potential pollutant treatment and to optimize the photon consumption related to the total volume to be
treated should be acceptable. The treatment capacity
best option for any specific problem, practically on a
case-by-case basis. must be high enough to make photocatalytic
degradation practi- cal. Many aqueous organic
It is not always easy to decide whether a specific
oxidation processes are too slow to be economically
water- contamination problem can be treated with solar
detoxifica- tion technology, since low efficiencies
• SOLAR photons must be used efficiently: the technology
produce hydroxyl radicals and slow kinetics may limit
economic feasibility. As mentioned above, preliminary to be applied must optimize the collection of solar
photons to be used. The overall energy needed per
tests are normally needed to assess process viability. In
an attempt to provide some guide- lines for the reader, the molecule destroyed must also be low enough to make
following general affirmations may be the process feasible. The use of external oxidants, such
as electron
scavengers, like S2O=8, H 2O ,2 O ,3or other substances,
made: could increase the quantum yield.
• –1
MAXIMUM ORGANIC CONCENTRATION of SEVERAL hundred mg L : DEACTIVATION): the degradation process must work
photodegradation processes work well in low or contin- uously without problems such as catalyst
medium concentrations up to a few hundred ppm of deactivation. Collector components, catalysts and
organics. The upper limit depends on the nature of the overall system must also be durable, guaranteeing
the contaminants, but concentrations over 1 gr L–1 are long periods of oper- ation without incident.
normally not suitable for solar photocatalytic • OPERATION AND MAINTENANCE processes must be simple: the
processes unless previously diluted. implementation of any real solar detoxification technol-
• NON-BIODEGRADABLE CONTAMINANTS : when possible, ogy application must require minimum operation and
bio- logical treatments are always the most cost- system supervision. As treatment processes are non-
effective processes. Only when the contaminants are productive, personnel costs associated must be reduced
persistent (non-biodegradable) do photocatalytic to a minimum. These considerations also apply to
processes make sense. maintenance.
• HAZARDOUS CONTAMINANTS present within complex mixtures of • BATCH system TREATMENT: it is clear that water
ORGANICS: one of the main advantages of solar treatment with solar detoxification should be run in
photocataly- sis is that it is non-selective, so non- recirculation mode with batch loads of contaminants
biodegradable con- taminants can be treated within a to guarantee complete destruction. This means that
complex mixture of other organic compounds. Solar the treatment must be independent of the process
detoxification also works very well with individual generating the wastewater and that on-line treatments
contaminants, but a mixture of contaminants increases normally are not feasible.
its utility. Hazardous contami-
nants also usually appear in concentrations Applications that fulfil both groups of requirements may
susceptible to photocatalytic treatment. be considered serious candidates for solar detoxification,
• CONTAMINANTS with no EASY CONVENTIONAL TREATMENT: and a detailed feasibility test study would be worth
an additional situation in which solar photocatalysis consideration.
will be
108 Solar Detoxification

7.2. Industrial wastewater

As previously indicated, solar detoxification technology
may be considered feasible for industrial effluents
containing highly toxic compounds for which
biological waste treat- ment plants are not feasible at
medium or low pollutant concentrations. This
technology is linearly dependent on the energy flux. It is
therefore also linearly dependent on the collector
surface, and investment in turn depends on the collector
surface. Land limitations must also be considered.
Normally, several hundred up to 2,000 m 2 is not a t
problem for a medium-sized factory. As a conclusion, a
FIGURE 7.1. Complete mineralization of a complex
reasonable order of magnitude of inflows for a solar
mixture of organic contaminants containing phenols
detoxification plant, using current technology, would be
using TiO2 (200 mg/L) and persulfate as an electron
in the range of several dozens up to a few hundred m 3
per day.
At the moment, and from the experience accumulated
by many scientists and researchers in the last ten years,
solar detoxification seems to be a good solution for BWTP. Figure 7.1 shows two degradation tests carried
destroying, among others, the following contaminants out on chemical industry wastewater containing a large
found in indus- trial wastewater: number of compounds, mainly phenols, and including
formol, phthalic acid, fumaric acid, maleic acid, glycol
• phenols compounds, xylene, toluene, methanol, butanol and
• agrochemical waste phenylethylene, among others.
• halogenated hydrocarbons As observed in Figure 7.1, complete degradation of
• antibiotics, antineoplastics and other chemical all initial substances and their intermediates is possible at
reasonable degradation rates. Table 7.1 shows examples
biocide compounds from the pharmaceutical
of industrial processes that generate wastewater
containing phenols that could be good targets for solar
• wood preservative waste (PCP, fungicides) photocatalytic processes.
• hazardous metal ions. One such application is the treatment of condensation
water from the manufacture of phenol-formaldehyde, one
of the oldest synthetic resins used in industry, for which
7.2.1. Phenols phenols are the raw materials. Phenolic resins are
obtained from the reaction of phenol or a phenol
substitute with formaldehyde. The reaction may be
These substances include phenol, cresol and other phenol portrayed as indicated in Figure 7.2.
substitutes. All the phenols are very toxic. Their The ingredients are boiled in a reflux condenser
maximum concentration in a biological wastewater (Figure 7.3) and condensation water is usually removed
treatment plant inlet (BWTP) is 1 to 2 mg/L. Even very by vacuum distillation. This condensation water contains
low concentrations of phenols (1 to 10 mg/L) in sev- eral mg/L of reagents that are very toxic.
freshwater produce unpleasant odours/tastes during the
Phenol-formaldehyde resins are usually prepared with
chlorinating process, so any dis- charge of phenols must
formaldehyde (40 per cent) – approximately 750 L of water
be avoided. The solar detoxification technology would
is removed for each ton of final product. When solid
therefore be useful in the treatment of water containing
formaldehyde is used, about 160 L of condensation water
this type of contaminant.
Phenols can easily be degraded by solar
photocatalysis. A few hundred square meters of solar
CPC collectors could crack or completely destroy the
phenols contained in small volumes of industrial
effluents prior to discharging to a
Solar Detoxification 109

TABLE 7.1. Example of processes producing phenol residues in wastewater.

Chemicals Process Waste description Quantity
Condensation water containing
1. Phenols & Phenol-formaldehyde
formaldehyde resins less than 2,000 mg/L of phenols and 500 L/ton of phenolic resin
2. Phenols Storage tanks, ships, Rinsing water containing less than 0.1% of the total production diluted
in lorries, piping 15,000 mg/L 60 volumes of water
0.1% of the total production diluted
3. Phenols Phenol production Spills, cleaning water
in 60 volumes of water

FIGURE 7.2. Formation of phenolic resins. (Courtesy of Ecosystem S.A., Spain).

removed from the final mixture. These phenolic residues

contain between 600 to 2,000 mg/L of phenol and
between 500 to 1,300 mg/L of formaldehyde.

7.2.2. Agrochemical compounds

This family includes a broad range of chemicals that
are extensively used in agriculture. Some of these
compounds are soluble in water, others are used as
suspensions; some are oil-based pesticides and some are
used as dusts. However, the majority of them are
dissolved, suspended or emulsified in water prior to
Phenolic spraying; the amount of wastewater gener- ated can vary
greatly depending on handling. Water-borne
contamination originates in many different processes,
such as cleaning and rinsing of spraying equipment,
dumping of leftover spray solutions, and plastic bottle
Pesticides are among the best candidates for the
applica- tion of solar detoxification technology. This is
because they are employed in very diluted solutions
or suspensions of multi-component formulas, in
small volumes and in batches. Very good results have
been obtained with organo- halogenated and
organophosphorus pesticides, carbamats, thiocarbamats,
2,4-D (2,4 DICHLOROPHENOXYACETIC ACID) and tri- azines. As
well as the large amount of pesticide waste gener- ated in
agriculture, there is also a huge amount of industrial
Final products
waste from factories producing the active ingredients,
FIGURE 7.3. Reactor condensation wastewater from the manu- partic- ularly in factories where the active components
facture of phenolic resins. (Courtesy of Ecosystem S.A., and other
110 Solar Detoxification

TABLE 7.2. Examples of processes that are potentially producers of wastewater containing agrochemical residues.
Chemicals Process Waste description Quantity
Wet scrubbing alkaline solution
Pesticides (NaOH) containing dissolved, 500 m3/year for a medium-sized
suspended and partially degraded company
Packagin pesticides
g (at high/medium concentration) 2,000 m3/year for a medium-sized
Floor-cleaning water, mixer- and reactor- company, with 0.05% of the active
Pesticides Manufacturing / cleaning and rinsing water components dissolved or suspended
Packaging in
100 parts of water
Water used to clean and rinse application
Pesticides Land application equipment. Formed by diluted —
spray solution (<100 mg/L)
Pesticides Land application Spraying solution prepared in excess —
An estimated 0.2 to 0.5 g of
Water used to clean and rinse plastic
pesticide Pesticides Plastic container recycling bottles prior to crushing and
per empty plastic bottle of 2 L average
volume, dissolved in 500 volumes of
(<100 mg/L)

additives are stored, weighed according to the product

for- mula, mixed and packed.
A complete description of an example of
photocatalytic degradation technology used for these
compounds can be found in Chapter 9. Another potential
example is the treat- ment of rinse water from cleaning
sprayers in agrochemical applications. This rinse water
contains pesticides and must be properly managed to
avoid contamination of the soil, ground- water and
surface water, which can occur if it is improperly
discharged. In many areas, farmers prepare the spraying
solu- tion at spring used by the whole town. Such places
are suit- able for the installation of a small or medium-
sized solar treatment plant. The technology could also be
of interest to large farm owners or pesticide application
FIGURE 7.4. Photocatalytic degradation of
7.2.3. Halogenated hydrocarbons dichloromethane, chloroform, trichloroethylene and
tetrachloroethylene using a titania catalyst manufactured
by ENEL (Italy). (Courtesy of University of Turin, Italy).
Solar detoxification has also been shown to be efficient
in the degradation of halogenated solvents. Halogenated halogenated wastes are factories of halocompounds
com- pounds are found in much pharmaceutical industry manu- facturing. The waste can be estimated as a low
waste, and increasing concerns about volatile organic percentage of the total production, dissolved at 100 to
compound (VOC) emissions and environmental 200 mg/L.
regulations and direc- tives are pushing industrial
managers to control VOC emissions. One of the VOC- Example: estimation of the required field
emission-control methods is wet scrubbing, and solar A feasibility study has determined that a 6 m2
detoxification can treat the contaminated water from the photoreactor can completely mineralize 250 L of
scrubbers. Other sources of wastewater containing 100 mg/L of chloroform in eight
hours of sunlight. Say
Solar Detoxification 111

25,000 L are to be treated. As all the parameters are The chemical industry also produces a great variety of
linear, assuming unchanged weather and wastewater biocides used as preservatives, especially in paint.
conditions, 600 m2 of collector field over eight hours, or Common examples of these compounds include: phenyl
300 m2 over sixteen hours is needed to treat the 25,000 mercuric acetate, dithiocarbamates and other sulfuric
L, while for a 100-m2 field, forty-eight sunny hours compounds, halogenated phenols, halogenated phenol-
would be necessary to complete the treatment. formaldehyde condensates and quaternary ammonium
Another potential halo compound application of salts. A medium- sized chemical company could
solar detoxification is treatment of the effluents from the typically produce 50 t/yr of this type of product. As
manufac- turing of PVC (polyvinyl chloride), which is equipment must be cleaned between each batch,
produced in huge quantities worldwide. The mass of approximately 500 to 1,000 m 3 of rinse water could be
these effluents is generated per year, with a normal concentration of 200 to
2.5 times greater than the final production of PVC (1 ton 50 mg/L of the biocide, for which solar treatment could
of PVC produces approximately 2.5 m 3 of effluent) and be feasible.
may be contaminated with short-chain polymers or vinyl
chloride, the PVC monomer.

7.2.4. Antibiotics, antineoplastics and 7.2.5. Wood preservative waste

other pharmaceutical biocide compounds
The most toxic and persistent of the wood preservatives
There are two main types of pharmaceutical industries: is pentachlorophenol (PCP). This insecticide is very
factories producing fine chemicals (that is, high-value well- known to solar detoxification researchers because
chemical products) and those that produce and mix raw it was one of the first compounds studied exhaustively in
materials and components to obtain a final product (a photocatalytic degradation. Although it is now banned,
chemical commodity). Since antibiotics are intrinsically PCP is still pro- duced and used in wood-protecting
bio- cides, they cannot be treated in biological treatment varnishes, because of its effectiveness against
plants. Antineoplastics are citotoxic (some of them woodworms and termites.
inhibit cellular division) and are commonly used in Other compounds in this group are creosote and
chemotherapy of leukemia and tumours. Both of them organic insecticides and fungicides. Although precise
are ecotoxic and should be degraded before dumping into information on the amount and chemical composition of
streams or sewage systems. Manufacturing equipment these residues is scant, the general impression is that they
(hoppers, conveyors/belt feeders, weighing and mixing produce a huge amount of waste. Wood-preserving
equipment, pill-forming or container-filling machines) is companies usually treat the wood (timber and sawed
normally cleaned thoroughly with a huge amount of pieces) in concentrated pesti- cide baths. When the baths
water. This can be stored and treated in batches, become less effective, their content is replaced with fresh
presenting ideal conditions for solar detoxifica- tion product. A conservative assumption is that 5 per cent of
treatment. the chemicals in the bath is discharged as waste.

TABLE 7.3. Examples of processes that are potentially producers of antibiotic and antineoplastic wastewater.
Chemicals Process Waste description Quantity
Pill forming and raw Cleaning water coming from pill and
1. Antibiotics material mixing process at container-filling equipment, 3–5,000 L /d at a medium sized factory
manufacturing of less than 1,000 mg/L of organic
pharmaceutical specialities compounds
0.05% of total production dissolved
2. Antibiotics Antibiotic production Fermentators and liophilisators cleaning
500 volumes of water
3. Antineoplastics Same as 1
4. Antineoplastics Same as 2
112 Solar Detoxification

FIGURE 7.5. Degradation of PCP at a PSA solar FIGURE 7.6. Cr+6 to Cr+3 solar photocatalytic reduction at
detoxification facility (Helioman’s collectors’ loop). See a PSA solar detoxification facility (CPC’s collectors’ loop).
also Figure 5.3 showing the solar photocatalytic See also Figure 7.7.
mineralization of PCP at the Helioman collectors.

7.2.6. Removal of hazardous as phenylmercury, or the removal of silver from black

metal ions from water and white photoprocessing waste. The presence of
organometal- lic compounds usually complicates
One additional advantage of the heterogeneous photo- wastewater treatment, as the metal becomes more inert
catalytic process is the presence of a reductive chemistry and the effectiveness of tradi- tional processes for
pathway that can be exploited to remove reducible aqueous metal removal, such as alkaline precipitation
species, such as hazardous heavy metal ions. It is and ion exchange, is reduced. Although the rate of metal
important to remember that both oxidation and reduction removal is generally slow, it remains attractive to solar
must occur simultaneously to maintain process activity, detoxification technology, as these applications are
so it is possible to use the presence of organics as normally widely distributed and can be processed in
reductants for the reduc- tion of metals and, at the same batches. When there are only metals, inexpensive and
time, the metals as oxidants for the oxidation of organics. non-hazardous sacrificial reductants (such as EDTA or
In general, the higher the con- centrations of organics, citric acid) can be added to enable photocatalytic
the faster the metal reduction rates. Similarly, an increase treatment.
in the concentration of metals increases the organics’ Another application of interest is the reduction of Cr +6
oxidation rate. to Cr+3. Photocatalytic reduction of Cr(VI) is very
Oxygen is usually the oxidant when there are no sensitive to pH, and is more efficient below pH 2. It
other oxidants or metals present, and water acts as the leaves the reduc- tion product, Cr(III), in solution in this
reductant element. Dissolved oxygen can compete range of pH. Around pH 5, Cr(III) forms a stable
with dissolved metals for conduction-band electrons precipitate. The Cr(VI) reduction rate is very sensitive to
and therefore can inhibit the rate of metal removal. the nature of the organic matter that is simultaneously
However, dissolved oxygen can be helpful for the being destroyed. Generally, the more easily oxidized the
degradation of the organics if the con- centration of organic compounds, the higher the photocat- alytic
metals is insufficient. The ability to remove metals reduction rate; this means that different waste streams
depends on the standard reduction potential for the may exhibit dramatically different treatment rates
reduction reaction. For example, it has been assessed depending on their specific chemical composition.
that Ag2+, Cr6+, Hg2+, and Pt2+ can be treated by
photocatalytic reduction, but Cd2+, Cu+2, and Ni+2 7.2.7. Other applications
Specific applications described in the literature are Many other potential applications of solar photocatalysis
the elimination of specific organometallic compounds may be found in the literature. Among them are the
(organic agents forming complexes with metal ions) destruction of organics in recyclable water from the
such microelectronics
Solar Detoxification 113

, , all over the world used for receiving, storing and

distributing raw chemicals and fuels to industry. Final
distribution of the stored product is in bulk by tank
, ,
trucks (except for a small amount in drums in the same
terminal enclosure).
, , Seaport tank terminals must carry out cleaning opera-
tions and also have to control emissions from the stored
products. These cleaning operations usually produce
, ,
huge amounts of water contaminated with low
concentrations of the chemical from tanks, pipelines,
, , jetty pipelines and hoses. Among the compounds
normally transported and distrib- uted this way, many
chemicals (such as phenol, metham sodium,
, ,
perchloroethylene, trichloroetylene, methylen chlo- ride,
sodium dichromate, and compounds such as styrene,
FIGURE 7.7. Simultaneous oxidation of phenol and toluene, nonil phenol and benzene chloride), may be suc-
reduction of Cr+6 to Cr+3 using solar detoxification cessfully treated by solar detoxification.
technology (PSA, Spain).
Example: Cleaning of phenol tanks
industry, the removal of explosives (such as TNT, RDX At the end of a contract for the rental of a 600-m3 tank
and HMX) from aqueous munitions waste, and the that had contained phenol, terminal personnel proceed to
oxidative degradation of cyanides. clean it for refill with styrene. This is done in a two-step
Because they are highly toxic, photocatalytic operation. First, saturated steam is injected to dissolve
degradation of cyanides is another potential industrial any phenol crys- tals remaining in the top and valves.
application, with the advantage that it produces neither 7,400 L of condensed water containing 6.4 per cent
highly toxic sub- stances such as cyanogen chloride nor phenol is then discharged into a waste storage tank for
sludges and avoids the use of additional chemical later recovery of the phenol by dis- tillation or its
reactants such as chlorine. destruction by thermal treatment. Then the
inside of the tank is rinsed out using a hose with a special
CN  2OH  2h+V  OCN  (7.1
spray nozzle. This rinse water contains 450 mg/L of
B H2O )
Furthermore, photocatalytic oxidation can transform CN – in a total volume of 16,000 L. The conditions (volume,
into less toxic substances, such as OCN –, and by concentration, transparency) of the waste generated in
carefully selecting reaction conditions, complete this second step are ideally suited to solar detoxification.
oxidation to CO2 and N2 can be obtained: Figure 7.8 shows an example of another potential appli-
cation: treatment of wastewater from cleaning a tank that
OCN  2H2O  V  CO2  12  4H (7.2 had contained metham sodium. Metham sodium (or
3h B N2 3 ) Vapam) is
In all the applications mentioned in this chapter, the a dithiocarbamate used as soil fumigant to control weeds,
presence of inorganic ions (such as chloride, phosphate, nematodes, fungi and insects (soil sterilizer). This
nitrate and sulfate) in water can have a negative effect on compound is a direct competitor of methyl bromide,
the destruction rates of the target compounds. High another soil disin- fectant and one of the ozone-depleting
concentrations of inorganic ions have been found to chemicals.
reduce the performance of the titania catalyst and, as a In a medium-sized seaport tank terminal (typically about
consequence, reduce signifi- cantly the feasibility of the seventy tanks with 28,000 m3 average storage capacity), at
overall treatment process. least fifteen cleaning operations are completed each year,
with huge amounts of residues generated annually in the
pipeline connection pit (‘snake pit’), hoses and cleaning of
7.3. Seaport tank terminals ships. There is also a potential application in any wet scrub-
ber these plants may use for VOC emission-control equip-
ment. All these residues can easily be managed by solar
A large percentage of the international chemical detoxification because they can be stored individually in
trade is shipped by sea. There are many tank terminals in
114 Solar Detoxification

240 Scrubber
Total organic carbon (ppm)

220 Date: 21 October 1993

Metham sodium
200 TiO2+ persulfate Contamination

Water table Vadose zone
140 zone Contamination plume

120 Recharging
well Well
0 10 20 30 40 50 60 70
FIGURE 7.9. Generic concept of contaminated groundwater
Residence time (min)
treatment. (Courtesy of ECOSYSTEM S.A., Spain).
FIGURE 7.8. Degradation of metham sodium wastewater
from a tank-cleaning process. Catalyst deactivation is also tends to have a high mineral and salt content,
observed after partial degradation of the initial TOC resulting in the need for pre- and post-treatment systems.
content. (PSA, Spain). Typically, contaminated sites are old chemical plants,
oil recycling plants (engine lubricant oil), refineries,
chemical weapon and explosive factories, pesticide plants,
small tanks or containers until the treatment process coke and gas factories, air force bases, harbours, railway
can be run. stations, power plants and substations. The contaminated
groundwater must be pumped out from a series of
7.4. Groundwater decontamination extraction wells located down- stream of the contamination
plume. After treatment, it is re- injected into the water table
Remediation of contaminated groundwater is a challenge through upstream injection wells.
in many parts of the world and especially in developed The first known treatment of contaminated
coun- tries. Typical substances found in contaminated groundwater by solar photocatalysis was carried out
groundwater are chlorinated hydrocarbons (including experimentally in the 1990s in the grounds of the
PCBs), aromatics from a variety of sources including Lawrence Livermore National Laboratory (LLNL) in
fuels (for instance, ben- zene and toluene), and a wide Livermore, California. This was also the first on-site
variety of other chemicals that includes pesticides, application of solar photocatalysis technology. During
solvents, phenols (PCP, creosote), TNT and DNT, dyes, the Second World War, a part of the grounds now
polynuclear aromatic hydrocarbons (PAHs) and even occupied by LLNL was a Naval Air Station training and
dioxins. main- tenance facility. Trichloroethylene (TCE) and
Solar detoxification can be considered a good solution other toxic chemicals were used extensively in normal
for in situ treatment and decontamination of operations to clean engine parts and other machinery. Over
groundwater con- taining substances where conventional the years, unconfined TCE and other volatile organic
biological treatment is not feasible due to dilution of compounds entered the local groundwater, where they are
pollutants. Also, groundwater is clear and transparent now slowly migrating off site. Today, TCE is present in
and there is always good storage capacity (the aquifer is the groundwater at concentrations ranging up to 500 ppb,
in itself a holding ‘tank’ for contami- nated water). Tests which is 100 times the acceptable EPA limit for drinking
have shown that the majority of pollutants found in water. The field experiment, developed by three US
contaminated groundwater are easily destroyed by government laboratories (NREL, SNL and LLNL), was
photocatalysis. Groundwater remediation also has the conducted at LLNL in 1991 using available trough tech-
advan- tage of a processing timescale of months or nology, and demonstrated the technical feasibility of this
years. application.
The difficulty is that standard discharge must The system consisted of two solar troughs, each 36.5
normally achieve drinking-water quality, which is m long, with an effective concentration ratio of
difficult and costly (for any treatment technology that is approximately
used). Groundwater
Solar Detoxification 115

FIGURE 7.10. Partial view of the 156 m2 parabolic-trough FIGURE 7.11. One-sun reactor built by American Energy
water treatment system tested on contaminated Technology, Inc. for treating contaminated groundwater in
groundwater at a site at LLNL (United States) in 1992. Florida (United States) in 1992. (Courtesy of the US
(Courtesy of the US National Renewable Energy National Renewable Energy Laboratory).

20 and a total solar collector area of 158 m 2 using a 7.5. Contaminated landfill cleaning
TiO2 slurry. It yielded outlet concentrations below 5
ppb, which meets the limits for a drinking-water Soil decontamination is another potential application of
process feasibility demonstration. One of the main solar detoxification. There are contaminated landfill sites
problems reported was the required acid pre-treatment, all over the world for which, depending on the nature of
which had a marked negative impact on the the con- taminants, water- or gas-phase photocatalytic
effectiveness of the process. treatments could be used. If the contaminants are
Another application that can be found in the literature soluble in water, water (or steam) can be used to extract
is the one-sun solar detoxification facility installed in them, and solar detoxification can then treat this water
1993 on a groundwater site contaminated by a former later. If the contami- nants are volatile organic
jet-fuel handling area at Tyndall Air Force Base in compounds, they can be desorbed from the solid as a
Florida (United States). Contaminants of interest that gaseous product using either vacuum extraction or heat
were present in the groundwater included benzene, and pumped into a gas-phase solar detox- ification
toluene, ethyl benzene, and xylenes (BTEX). Thirty one- system (see also Section 7.7).
sun collector modules (1.22 m ×
One example of this application is the
2.4 m) connected in series were used, each collector mineralization of lindane suspended in landfill
module made up of 66 UV transparent parallel tubes with leachate. The reaction between benzene and chlorine
a 0.64 cm inner diameter and 2.4 m in length (Figure under UV light produces six stable isomers of
7.11). Treatment of initial concentrations of 1 to 2 mg/L hexachlorocyclohexane that are carcinogens. Only one
of BTEX was per- formed in a batch system, obtaining of them, Lindane, which makes up 14 per cent of the
typical destruction lev- els of 50 per cent to 75 per cent final mixture, is actually an insecticide. The other
in three hours. The total volume treated during each run isomers, hazardous compounds and also very stable in
was 530 L. the environment, have been spread throughout the
One last example is the photocatalytic cleaning of environment for decades or dumped in badly managed
groundwater contaminated by gasoline from landfills (many sites are contaminated by lindane, some
underground storage tanks (Figures 7.12a and b). In this of them very well known in Europe). Lindane is one of
case the ground- water was contaminated with BTEX the oldest chlorinated insecticides and, because of its
(benzene, toluene, ethyl benzene and xylenes) at a stability in the environment, is banned for agrochemical
concentration of about 1 ppm. The test site was a gasoline and veteri- nary use in most countries. However, it is
station and the system consisted of six non-concentrating still used as ectoparasiticide to eliminate lice in
solar reactors of 3 m2, each using TiO2 slurry and a children and for other
1,900 L storage tank.
116 Solar Detoxification



FIGURE 7.12. Non-concentrating solar detoxification system for BTEX-contaminated groundwater at a commercial
site in Gainesville, Florida (United States), 1996. (Courtesy of the Solar Energy and Energy Conversion Laboratory,
University of Florida).

pharmaceutical and medicinal uses, with the approval of

health authorities.
7.6. Water disinfection
Figure 7.13 shows the photocatalytic degradation of Chlorine is the most commonly used chemical for water
Lindane (technical grade) suspension in water, dis- infection because of its ability to inactivate bacteria
simulating the final step of contaminated landfill and viruses. Nevertheless, the presence of organic
treatment and demon- strating the technical feasibility of impurities in the water can generate unwanted by-
the solar treatment process (maximum solubility of products of disinfec- tion, such as trihalomethanes and
lindane: 10 mg/L). other carcinogens. As a
Solar Detoxification 117


(100-280 nm) (280-315) (315-400
(280-315) (315-400nm)


* Definitions
* Definitions
based on
based on
effect Ozone- Germicidal
Ozone- Germicidal
producing Erythemal
producing Eryt hemal

Far UV Near UV
Near light”)
Far UV
(“black light”)

100 150 200 250 300 350 400

FIGURE 7.13. Simulation of contaminated landfill FIGURE 7.14. Common ultraviolet band designations
treatment using solar detoxification: mineralization of based on biological effects. (Courtesy of International
lindane (PSA Solar Detox Facility, Spain). Light, Inc., USA).

result, other water-disinfection technologies are components. Successful pilot plant experiments using
becoming increasingly important. Among them, this technique have been carried out in Israel and
ultraviolet irradiation with lamps is widely used to Germany.
destroy biological contami- nants, mainly using 254- Two major disadvantages of sensitized
nanometer wavelength. Solar ultra- violet, which is photodisinfection in comparison with conventional
primarily at 290 to 400 nm wavelengths, is much less techniques are the gener- ally slow kinetics and the lack
active as a germicide (see Figure 7.14). of residual disinfection capac- ity (after exposure to
Despite the broad spectrum of research, the potential light). On the positive side, the presence of biological
use of solar detoxification technology for water and organic contamination in surface waters is usually
disinfection is still largely unexplored. However, the highest in summer, when the greatest amount of solar
antibacterial effect of TiO2 has been demonstrated on radiation is available for the process. One possible
several microorganisms, including Escherichia Coli, realistic approach might be preliminary solar detox-
Lactobacillus, Streptococcus, and others. In all cases, the ification to partially disinfect water and reduce the level
photocatalytic oxidation effect of TiO2 particles was of organic contaminants, followed by limited
able to effectively sensitize bacteria to photo-killing by chlorination to maintain disinfection in distribution
solar exposure. This photo-killing action is associated pipelines and inhibit formation of undesired by-products.
with the disruption of the cell wall and membrane
through photocatalytically induced surface oxidation
result- ing in the disintegration of the cell. Disinfection 7.7. Gas-phase treatments
of viruses such as Phage MS2 and poliovirus 1 can also
be found in the literature.
Homogeneous methylene-blue-sensitized Gas-phase photocatalytic oxidation is one of the
photochemical disinfection has also been found to be advanced oxidation technologies (AOTs) with promising
highly effective with applications for end-of-pipe treatment of gaseous
E. Faecalis bacillus spores through singlet oxygen emissions and air purification. The first pilot
photo- oxigenation. Singlet oxygen is generated by experiments were also carried out in the 1990s. In the
absorption of a photon by a photosensitizer (dye) that is United States, the ambitious DOE Solar Industrial
excited and trans- fers its energy to a dissolved oxygen Program motivated development of the photocatalytic
molecule. Photo-killing occurs when this singlet oxygen oxidation technology. Some examples of applications
penetrates into the microor- ganism resulting in are cleaning of exhaust streams from the
oxidation and inactivation of cellular microelectronics industry and treatment of off-gases
118 Solar Detoxification

A4 . Cl
Cl C Cl CHCl2 CHCl
1 3

H Cl H Cl H Cl O2 H Cl O
C C Path A C C. C C
C C 6 7
A2 . A3
Cl Cl Cl Cl O O
(1) (2) Cl
Cl Cl Cl Cl O. A5
A1 2a 3 4 H .
C C Cl Cl
. . Cl Cl O
Cl H Cl
Cl Cl
Cl Cl 5
H Cl
.C C
Path B Cl
Cl Cl

FIGURE 7.15. Scheme of TCE gas-phase mineralization with PCO and a monolithic catalyst based on sepiolite/TiO 2/Pt.

The fundamental background of radical attack on

organic contaminants photocatalyzed with the assistance
of smog has been addressed by several tropospheric
chemistry specialists but, unfortunately, some crucial
questions related to the basic fundamentals of gas-phase
PCO still persist. A critical screening of the related
literature reveals conflicting results among the authors
regarding the influence of different process parameters.
The reaction mechanism in gas-phase, based on hydroxyl
radicals, has also recently been revised by other possible
options, resulting in a presently confusing stage for gas-
phase photocatalytic treatment processes.
Figure 7.15 shows the possible trichloroethylene
(TCE) PCO degradation pathways. Some intermediates
have been shown to form in the presence of chloride, and
it has been proposed that Cl could be the driving force
FIGURE 7.16. Destruction efficiencies of 400 ppm of TCE instead of OH. Other authors have proposed that both
using a MgSiO4/TiO2/Pt monolithic catalyst: mechanistic approaches via Cl and OH are essentially
thermocatalytic and photocatalytic processes at different correct and that the eventual pathway depends on the
temperatures (CIEMAT, Spain). characteristics of the cata- lyst (for example, internal
surface or porosity). TCE is one of the most studied
compounds because it is a major gas-phase air pollutant.
Its photocatalytic oxidation with titanium diox- ide has
extraction of contaminated soil vapour. Indoor air quality received good reports in the literature.
in houses, vehicles and spacecraft has also been a matter Figure 7.16 shows a lab experiment performed on
of study in the United States and Japan. In Europe TCE using a monolithic TiO 2-based catalyst at different
several ini- tiatives also address the development of gas- tempera- tures. Photoeffect was clearly differentiated by
phase solar PCO as an end-of-pipe technology for the using appro- priate filters. As it can be observed,
treatment of VOC emissions but, unlike water temperature played a negative role as a result of
detoxification, all the above- mentioned experiments adsorption–desorption phenom- ena. Nevertheless, apart
have been basically conducted with lamps. Very few from high photonic efficiencies (up to 95 per cent)
solar installations can be found in the liter- ature and reported with TCE, it is not easy to find appli- cations for
there continues to be a serious lack of solar-driven gas- pure photocatalytic gas-phase treatment processes,
phase PCO.
Solar Detoxification 119

as not many compounds have been found to be another way of promoting gas-phase applications.
significantly affected by photocatalysis, in most cases a Possible targets are the majority of hazardous air
combination of photocatalytic degradation and pollutants requiring abatement technologies:
destructive thermocatalysis is necessary. halogenated, aliphatic and aromatic hydrocarbons,
In addition, a major problem for solar gas-phase PCO alcohols, glycols, ethers, epoxides and phe- nols,
is the difficulty, or even impossibility, of working in normally present in air streams at concentrations of less
batch systems, implying an important additional than 5,000 ppm. Some applications for air purification or
difficulty for solar-driven processes because of the degradation of emissions in the semiconductor industry
natural uncontrol- lability of the Sun as the energy are also under development.
source. As a result, it is not easy to find a niche for Figure 7.17 shows a one-sun reactor with PTFE
solar-driven PCO applied to gas- phase and air window material and titanium dioxide immobilized on
purification. Hybrid solar–electric or pure electric polypropy- lene supports for gas-phase solar
devices using lamps are therefore envisaged for initial detoxification successfully tested by NREL (United
practical applications. States). For additional information on gas-phase
Table 7.4 summarizes NREL experience with gas- experimental treatment systems, see also Figure 6.24 and
phase photocatalytic degradation of different Figure 5.6.
contaminants. Aromatic compounds like BTEX have Unlike water phase, in which a catalyst may be used
slow or medium reac- tion rates and TCE is the fastest. in efficient slurry configurations, in gas phase this is not
BTEX tests demonstrate deactivation of the TiO2 catalyst possi- ble and all experimental systems developed to date
and the need to add ozone to enhance the process and use an immobilized catalyst. As previously discussed in
make PCO degradation possible. This and other possible Section 6.5, the catalyst must be anchored onto some
technology combinations could be type of inert support inside the gas-phase stream
treatment reactor.

TABLE 7.4. Qualitative degradation rates of different compounds screened for photocatalytic activity: slow (s),
medium (m) and fast (f ). (Courtesy of US National Renewable Energy Laboratory).
Compound Rate Compound Rate
Acetaldehyde m Methanol m
Acetonitrile s Methylene chloride s
Acetone s Methylethyl ketone m
Benzene s Methylisobutyl ketone m
Benzaldehyde m Perchloroethylene f
Butyl acetate m Propyl amine m
Butyl amine m Propylnitrile s
Decane m Tetraethylethylene m
Dichloroacetyl chloride m Toluene s
Dichlorobenzene s Trichloroethylene f
Diethyl amine m Xylenes m
Diisopropyl amine m Sec-butyl amine m
Ethane s t-butyl amine s
Ethanol m 1-butanol m
Ethyl acetate m 1-propanol m
Ethylbenzene m 1,1-dichloroethylene f
Ethylene m 2-propanol m
Fuel oil m 1,2-dichloroethylene f
Solar Detoxification 121

a) One interesting potential application of gas-phase

solar detoxification technology is the gas treatment from
contami- nated landfill recovery (see Section 7.5).
Chlorinated volatile organic compounds (CVOC) are
among the most pervasive subsurface ground
contaminants, and the remediation tech- nologies most
used for CVOC ultimately produce a gas- phase stream
containing CVOC, either from soil venting from the
unsaturated zone or groundwater recovery from the
saturated zone and subsequent air stripping. In the past,
direct discharge of gaseous CVOC to the atmosphere
was normal, but since the early 1990s many regulations
have required the abatement of these emissions using
b) the best available technologies.
Other possible applications are small air pollution
sources. These applications are not uncommon: examples
are dry cleaners, auto repair shops, bakeries and coffee
roast- ers. Additional potential solar applications include
remote sites, storage-tank vents or potentially explosive
waste streams. Finally, the regeneration of a carbon bed
or other adsorbent materials could be another promising
application for a solar system. In this application, the
adsorbent can per- form twenty-four-hour VOC removal,
and the solar system can be used to purge and destroy the
contaminants during daylight hours.
With regard to air disinfection, most of the work
carried out has been focused on the use of TiO 2 particles
as sensi- tizers for the destruction of bacteria and viruses
in air. In this field, a very interesting application is the
fabrication of self-disinfecting surfaces. Several studies
have reported the antibacterial effect of titanium dioxide
on indoor air in com- bination with UV fluorescent
lamps. The use of thin TiO 2 films as a method of
keeping surfaces free of biological material would be of
particular value where sterile surfaces are essential, such
as in operating theatres in hospitals. Although very few
demonstrations have been reported to date, it is expected
that more will take place in the future.
Finally, solar photocatalysis is the process that
FIGURE 7.17. Flat-plate reactor for treating contaminated
makes self-cleaning glasses work efficiently. Fouling of
air exiting from air-stripping units at McClelland Air
glasses is mainly due to dust and/or atmospheric particles
Force Base, California, 1997. (Courtesy of the US
stuck on the surface on greasy stains, mainly of fatty
National Renewable Energy Laboratory).
acids. Self-cleaning glasses could be coated with an
invisible thin layer of titania, which, under the
simultaneous presence of oxygen (air), atmospheric
Such a system must be being very active (a high water vapour and solar UV-light, is able to decompose
degradation efficiency), with low pressure-drop, a fatty acids by successive photodecarboxylation (photo-
long lifetime, and a reasonable cost. Very diverse kolbe) reactions and allow coated glasses to recover their
types of catalysts have been used in gas-phase systems, initial clearness (Figure 7.18).
from MgSiO4/TiO2/Pt monoliths to TiO2 immobilized
on glass fibres, ceramic substrates and inert plastic
materials. As in water-phase, residence times must be
as long as possible to increase degradation effi-
120 Solar Detoxification

Photochemical disinfection of effluents: pilot plant studies.
WATER RESEARCH, Vol. 24, No. 7, pp. 837–43.


FIGURE 7.18. Self-cleaning windows. Evolution of MARCHAND, E. 1993. Solar photocatalytic treatment of
photo- catalytic treatment processes on glasses coated groundwater at Tyndall AFB: field test results. In: SOLAR
with titanium dioxide. (Courtesy of Laboratoire de 1993: proceedings of the AMERICAN SOLAR Energy Society
Photocatalyse, Catalyse et Environnement (E.R. au ANNUAL Conference, pp. 235–9.
CNRS), Ecole Central de Lyon, France).
MALATO, S. 1999. SOLAR PHOTOCATALYTIC Decomposition of
Summary PENTACHLOROPHENOL Dissolved in WATER. Doctoral thesis.
Editorial CIEMAT.
Solar detoxification technologies could provide the
environ- mental waste management industry with a new MEHOS, M.S. AND TURCHI, C. 1993. Field testing solar
tool to destroy waste using clean energy from the Sun. photocatalytic detoxification on TCE-
Nevertheless, in spite of the large number of patents and contaminated groundwater. ENVIRONMENTAL Progress,
publications, industry is still reluctant to adopt the Vol. 12, No. 3, pp. 194–9.
process because of discouraging and contradictory
results, which has led research to focus on specific
potential applications. Solar detoxification is a useful NIMLOS, R.; WOLFRUM, E.J.; GRATSON, D.A.; WATT, A.S.;
technology for addressing non- biodegradable JACOBY, W.A. AND TURCHI, C. 1995. Review of RESEARCH Results
hazardous contaminants where no easy treatment by for the PHOTOCATALYTIC OXIDATION of HAZARDOUS WASTES in Air.
conventional technologies is available, at NREL /TP-433-7043.
concentrations up to several hundred mg L–1. Phenols,
agro- chemical wastes, halogenated hydrocarbons, NREL. 1995. Development of A Homogeneous Aqueous PHASE
biocide com- pounds from the pharmaceutical industry, PHOTOCATALYST for the SOLAR DETOXIFICATION of WATER.
wood-preserving waste, hazardous metal ions, cyanides Final Report to NREL. Markham, Ontario, Solarchem
and aqueous muni- tions waste are among the industrial Environmental Systems.
wastewater applica- tions. Other interesting
applications are the treatment of groundwater
contamination and seaport tank terminals, the cleaning of PRAIRIE, M.; EVANS, L.R. AND MARTINEZ, S.L. 1994.
contaminated landfills and water disinfection. Gas-phase Destruction of organics and removal of heavy metals in
applications are in an even more confusing situation as water via TiO2 photocatalysis. In: CHEMICAL OXIDATION:
some crucial questions related to the basic fun- Technology for the Nineties, Second INTERNATIONAL Symposium.
damentals still persist. In solar gas-phase PCO, the Lancaster, Pa., Technonomic Publishing Company.
catalyst must be used immobilized, and diverse types of (Symposium proceedings.)
supporting substrates, such as glass fibres, ceramic
substrates, inert plastic materials, and monoliths can be VINCENT, M. 2000. Solar detox market. In: TRAINING AND
used. Some potential gas-phase applications are gas
treatment from contaminated landfill recovery, Mobility of RESEARCHERS Summer School: INDUSTRIAL APPLICATIONS
degradation of emissions in the semicon- ductor industry, of SOLAR Chemistry, September 1998, Chapter 8.
air purification, air disinfection and self- cleaning Madrid, Editorial Ciemat, pp. 261–280
122 Solar Detoxification

AND HENRY, B.E. 1995. Photocatalytic TURCHI, C.; WATT, A.; WOLFRUM, E. AND PRAIRIE. M.
inactivation of coliform bacteria and viruses in 1996. Solar
secondary wastewater effluent. WAT. Res., Vol. photocatalytic processes for the purification of water:
29, No. 1, pp. 95–100. state of development and barriers to
commercialization. SOLAR Energy, Vol. 56, No. 5, pp.
Solar Detoxification 123

Self-assessment questions Part B

Part A: True or false? 1. Why is the environmental market still reluctant

to accept solar detoxification processes?
1. Preliminary research is required to assess 2. Indicate the three existing solar
potential treatment of pollutants in any specific photocatalytic processes.
problem on a case-by-case basis.
3. What is the main advantage of photocatalytic
2. Solar photocatalytic processes work well with
oxidation processes over other advanced oxidation
organic concentrations of over 1 gr/L.
3. Non-biodegradable organic contaminants are the
most logical target for solar photocatalytic processes. 4. In addition to the use of solar energy, indicate one
main advantage of solar photocatalysis.
4. There is no sense in addressing complex mixtures
of hazardous contaminants with solar photocatalysis. 5. Indicate at least three conditions indicative of a
viable solar photocatalytic process.
5. An indication that solar photocatalytic technology
may have a possible application could be the lack of an 6. Indicate why water treatment with solar
easy conventional treatment. photocatalysis must normally be in a batch system.

6. External oxidants must always be used as electron 7. As solar photocatalysis is an oxidation-reduction

scavengers, when feasible, to increase the quantum yield process, what is the normal reductant agent a) when
of degradation processes. only organic compounds are present, and b) when
metals are present?
7. There is usually no difference between the treatment
of contaminated water in a batch or on-line process. 8. From testing it has been found that a 4-m2
collector system can treat a 50-L sample of
8. Typical solar detoxification plants process from wastewater containing 200 mg/L of TOC from
several dozen up to a few hundred m 3 per day. organic compounds in four
hours of sunlight. How long would it take to treat 100
9. The presence of high concentrations of inorganic m3 of the same wastewater with a 500-m2 treatment
ions negatively affects the destruction rate and the system?
overall feasibility of solar detoxification processes.
9. What are the proposed mechanisms for gas-
10. It is difficult to find applications for pure phase photocatalytic degradation of chlorinated
photocatalytic gas-phase treatment processes as not organic compounds?
many compounds have been found to be affected by a
significant photocatalytic effect.
10. What is the worst situation for a water
treatment process based on TiO 2 photocatalysis?
124 Solar Detoxification

Answers 5. Reasonable throughputs, efficient use of solar

photons and reliable photocatalytic process with no
catalyst deactivation.
Part A
6. Because the degradation process is driven by solar
1) True. 2) False. 3) True. 4) False. 5) True. 6) True. photons, a source which cannot be controlled, making
7) False. 8) True. 9) True. 10) True. it necessary to separate the treatment from the
wastewater generating process.
Part B
7. When only organic compounds are present, oxygen is
the normal oxidant and water acts as the reductant
1. Because early successful experimental results element. When metals are present, oxygen is also the
presumed a wide range of practical applications, but oxidant and the metals themselves are the reductant
confusing and discouraging results later led to the elements.
current concentration on applications with specific
required boundary conditions.  100,000 
2. Heterogeneous TiO2 photocatalysis, homogeneous
8. R  4  500  64 hours of sunlight.
Photo-Fenton and homogeneous sensitized-
photochemical oxidation by singlet oxygen.  50 


3. Their potential solarization, introducing an important 9. Depending on the catalyst characteristics and the
additional environmental value in the pollutant boundary conditions, the main degradation
degradation process. mechanism can be based on the attack of ºCl or ºOH
or ºCl + ºOH radicals.
4. The fact that it is a non-selective process,
making possible the treatment of complex 10. The presence of high concentration of
mixtures of organic contaminants and inorganic ions
non-biodegradable organic contaminants. simultaneously.
Solar Detoxification 123


Economic Assessment

Aims Objectives
This chapter discusses the main factors After completing this unit, you will be able to:
that affect the cost of solar detoxification
systems and analyses the economic cost of 1. Determine the feasibility of a two-step
specific solar photocatalytic applications. photocatalytic–biological treatment process.
They are also qualitatively compared with
other treatment technologies. 2. Estimate the cost of solar
photocatalytic treatment.

3. Estimate the feasibility of solar

detoxification as a function of local
solar resources.

4. Compare possible alternative or

complementary treatment
126 Solar Detoxification


Symbol Description Units

AOPs Advanced oxidation processes
AST American Standards for Testing and
M Materials Atmosphere attenuation dimensionless
cλ coefficient at λ Chemical oxygen demand moles of O2 L–
COD Cosine of the incident angle of solar radiation to collector surface
cos p Compound parabolic concentrator
CPC Extraterrestrial irradiation
École Polytechnique Fédérale de Lausanne (Switzerland) W m–2 µm–1
EPF Cloud factor
L Fixed charge rate %
f Granular-activated %
FCR carbon Solar time angle
GA Henry’s Law constant degrees
C Total yearly hours of operation of a lamp-based system dimensionles
hs Total yearly hours of operation of a solar detoxification system s hours
H Intensity of solar irradiation over terrestrial surface at specific wavelength, W m–
HL λ 2
HS Intensity of extraterrestrial solar irradiation at the wavelength, λ m–2 W
Ib,λ Solar constant associated with a spectral interval (λ1, λ2) m–2 W
Io,λ Yearly average global UV irradiation m–2
I Local latitude degrees
– sc,λ dimensionless
I UVg Air mass ratio
Air mass ratio at sea dimensionless
L Moles L–1
level Molar
m Mili molar 10–3 moles L–
mo Molecular weight
gr mol–1
M Daynumber dimensionless
mM National Renewable Energy Laboratory
MW (USA) Number of electric lamps
N Total yearly collected UV photons (energy higher than 3.2 eV) photons
NRE Number of photons with energy higher than 3.2 eV per unit of time and photons m–2 h–
L NL surface Pentachlorophenol
NUV Hexahydro-1,3,5-trinitro-1,3,5-triazine
NΦ (E>3.2 Surface of solar collector field
Trichloroethylene m2
PCP Trinitrotoluene
RD Total organic
XS carbon Treatment mg L–1 or moles of C L–
TCE factor
TNT Mass treatment factor 1
Volumetric treatment g h–1 m–2
TO L h–1 m–2
C factor Ultraviolet
radiation A Elevation over
Tf m
sea level Solar altitude
Tfm degrees
Solar declination
Tfv Nominal average photonic flux electric
photons s–1
UV-A lamp Average oxidation state
z s
Economic Assessment 127

8.1. Photochemical and biological geometry and materials, the contaminants and their
reactors coupling concentration, UV-A irradiation level and catalyst
concen- tration. Nevertheless, this parameter does not
necessarily imply better or worse performance of a
The mass treatment factor (Tfm) of a specific solar specific system, because this depends on the discharge
detoxifica- tion system can be defined as the amount of objectives. Contaminant treatment, in its strictest
organic sub- stances the system is able to treat per unit of meaning, is the com- plete mineralization (TOC = 0) of
time and solar collector surface: the contaminants but, as previously indicated (Chapter
7), normally photocatalytic
m(TOC)  g  processes only make sense for hazardous non-biodegrad-
T   (8.1) able pollutants (toxic organic pollutants not treatable in
tS  2 conventional biological treatment plants). When feasible,
The volumetric treatment factor (Tfv) of a specific solar biological treatment of biodegradable residual waters is
detoxification system can be defined as the volume of the cheapest treatment technology and also the most
con- taminated water the system is able to treat per unit compati- ble with the environment. Therefore,
of time and solar collector surface: biologically recalci- trant compounds could be first
treated with photocatalytic
technologies until biodegradability is achieved, later trans-
V  L  ferring the water to a conventional biological plant. Such a
tS  hm2
combination of photocatalytic and biological treatments
could reduce treatment time and optimize the overall
Tfm and obviously depend on the specific solar system economics, since the solar detoxification system could be
Tfv and
the wastewater to be treated. The same system yields significantly smaller. As a consequence of the kinetic
differ- mech- anism (explained in Chapter 4), the first part of
ent treatment factors with different contaminants and the photo- catalytic process is the quickest, with higher
also varies depending on the solar irradiation, higher reaction rates producing the main part of TOC
on sunny days than on cloudy days. The most practical degradation, and the last phase is the longest, with lower
units are those indicated in Equations 8.1 and 8.2: the reaction rates and contri- bution to complete degradation.
grams of TOC degraded per hour and square metre of An example might be agrochemical compounds,
solar collector field, in the case of Tfm, and the litres of which are normally effectively cracked by photocatalytic
water treated per hour and square metres of solar degrada- tion processes. The active component of the
collector field in the case of Tfv. pesticide disappears after several minutes of irradiation,
but TOC
Example of calculation of treatment factors
If a specific solar detoxification plant having 300 m 2 of
col- lectors can treat 5 m3 of wastewater containing 50
mg L–1 of organic compounds in ninety minutes, the
treatment factors will be:
50 mg L1  5,000 3 g g
Tfm  
300 m2  1.5 10  0.55 (8.3)

5,000 L L (8.4)
Tfv   11.11
300 m  1.5
2 hm2
such as system
Treatment factors are a powerful tool for comparing
differ- ent detoxification systems and checking treatment
feasibil- ity for different wastewater problems. The
efficiency of solar detoxification systems is usually from
0.05 to 1 g of organic compounds degraded per hour and
square metre. This is a function of several parameters,
128 Solar Detoxification

FIGURE 8.1. Conceptual scheme of photocatalytic +

biologi- cal technologies coupling. (Courtesy of EPFL,
Economic Assessment 129

usually becomes negligible only after two or three methods alone has already been demonstrated at EFPL.
hours of irradiation. Design of a solar detoxification This strategy is based on the average oxidation state
(ΨAOS), which can be defined from the following
system to reduce the TOC (that is, disappearance of a
specific com- pound or compounds) only until toxicity
is low enough to be sent on to a biological treatment Average Oxidation State
plant might therefore  TOC 
be of interest.     4 
COD (8.5)
 
In the previous example (Eqs. 8.3 and 8.4), ‘treatment’ TOC
could mean the total degradation of organic contaminants where TOC (total organic carbon) and COD (chemical
or the degradation only until the achievement of oxy- gen demand) are expressed in moles of C L–1 and
biodegrad- ability. Therefore, the treatment factor may moles of O2 L–1, respectively. Average oxidation state
also depend on the system objective, and a very low can be between
treatment factor for complete mineralization could be +4 for CO2, the most oxidized state of carbon, and –4 for
transformed into a feasible one for integration with CH4, the most reduced state of carbon. The latter is given
biodegradation. by the reaction:
The feasibility of combining photocatalytic and
treatments to achieve highly efficient abatement of CH4  2O2  CO2  (8.6)
pollutants 2H2O
that cannot be treated by either biological or where, TOC = 1 (moles of C L–1) and COD = 2 (moles
photocatalytic O2 L–1) so, according to Equation 8.5, TAOS = –4. With
a) CO2, as oxidation is impossible, COD = 0 and TAOS =
+4. An example is oxidation of formic acid:

HCOOH  O2  CO2  2H2O (8.7)

In this case, TOC = 1 (moles of C L–1) and COD = 1

(moles O2 L–1), so ΨAOS = 0. When photodegradation is
applied, the average oxidation state of intermediates is a
function of photo-treatment time and could be a
good indication of when to switch from the advanced
oxidation process to biological treatment.
Oxidized substances are typical of biological
processes. Usually treatment performance is best for
organic com- pounds of low molecular weight.
Photodegradation treat- ment performance is also better
at the beginning of the process, when organic molecules
are broken up into simpler ones. As a consequence, it
makes sense to always check the feasibility of a two-step
process in which the photocatalytic treatment is
employed first to increase biodegradability and
biological treatment is applied afterwards. Figure 8.3
shows the typical evolution of ΨAOS when recalcitrant
toxic com- pounds are treated. Normally, a plateau is
observed after a time, suggesting that the chemical nature
of intermediates formed will no longer vary significantly.
As Figure 8.4 shows, there are two possible extreme
situations (A and B) with regard to the evolution of
FIGURE 8.2. Elimination of tensioactive compounds from Sometimes, there is a direct correlation between ΨAOS
agrochemical industrial wastewater after a few minutes and toxicity. This is the best situation (as reflected in
of photocatalytic treatment (PSA, Spain). point A of Figures 8.3 and 8.4), because it means that
the overall
130 Solar Detoxification

treat certain non-biodegradable toxic compounds has

been demonstrated at EPFL (Switzerland). A TiO 2
photocatalytic process followed by aerobic biotreatment
has also been suc- cessfully employed by NREL (United
States) for the treat- ment of wastewater from munitions
manufacturing plants (pinkwater) containing low
concentrations of high explo- sives, such as RDX
(hexahydro-1,3,5-trinitro-1,3,5-triazine), TNT
(trinitrotoluene), HMX and PETN.
It must be pointed out that photocatalytic degradation
of recalcitrant organic compounds quite often increases
the toxicity of the wastewater during its initial phase.
The rea- son is the proliferation at the beginning of the
process of many reaction by-products. Therefore, it is
important that the photocatalytic treatment be long
FIGURE 8.3. Photodegradation process: usual relationship enough to eliminate this increased toxicity. Also, with
between the average oxidation state and time. Points A and regard to the parameters of toxicity, it is very important
B are also related to Figure 8.4. (Courtesy of EPFL, that no intermediates formed be more toxic than the
Switzerland.) initial compounds, because this could disable the overall
feasibility of the treatment process.
The feasibility of such a photocatalytic–biological
process combination must always be assessed because it
could pro- vide an important cost reduction in the
treatment of haz- ardous compounds in wastewater, by
reducing the size of the necessary solar collector field. It
must be taken into account that, as with most solar
systems, the economics of water detoxification systems
are dominated by their capital cost. Thus, by reducing the
capital cost of the system, the process- ing cost is

FIGURE 8.4. Photodegradation process: normal

relationships between toxicity and time. Points A and B
8.2. Cost calculations
are also related with Figure 8.3. (Courtesy of EPFL,
Switzerland.) Cost is always a key topic when innovative technologies
are considered, and standard commercial procedures
indicate that any new technology should provide
significant reduc- tions in processing costs over
chemical status of the degradation process does not
competing technologies or significant new technical
progress (AOS remains constant, and toxicity
contributions for its successful marketing. In the case of
decreases, thereby increasing biodegradability).
solar detoxification, its main contri- bution is the
Nevertheless, even at low concentrations, toxic
environmental added value of using solar energy to
substances could be formed during the degradation
solve water contamination problems.
process (point B of Figures 8.3 and 8.4), increasing
This section addresses the main factors influencing
overall toxicity. This does not neces- sarily impede
initiation to biological treatment, but it does force the ‘final installed cost’, ‘operating cost’ and ‘treatment cost’ of a
solar detoxification system. The installed cost includes all
point of commencement to be moved to the right (point
costs associated with the design, construction, and start-
up of the facility, but it is assumed that normal services
As previously indicated, the feasibility of combining
and facilities such as water and electricity are available
Photo-Fenton (see Section 1.3.3) and a biological process
on-site and that
130 Solar Detoxification

land is also available at no cost. The four main The sum of the facility cost, contingencies, engineering and
components of installed cost are: spare parts is the ‘total installed cost’. It is important to note
that the cost of a previous feasibility study is not considered
• FACILITY cost: the total field cost includes the direct cost here because the extension and depth of this study depends
of equipment and system components, plus the on many factors, specific to each situation, which are
direct cost of installation labour. The major difficult to estimate. Table 8.1 shows the cost for different
components are: civil engineering, concrete typical facility sizes for a CPC solar field. These costs are
foundations, connection to power and water supply, based on real plant built recently (shown in Section 9.6).
solar collectors, tanks, manifolds, piping, pumps
In addition to the total installed cost, the second deter-
and valves, structures and supports, electrical
mining factor in the cost of water treatment is the
equipment, control system, instrumentation,
‘operating cost’. The operating cost must include all
catalyst separation unit, collector, and piping and
personnel and material costs required to operate and
con- nection assembly.
maintain the facility (supervision, procurement,
• Project contingency: since the general conceptual maintenance and so on), chemi- cal supplies and
design cannot include all the details of the specific final electricity costs. Usually, it comprises:
design, at this point, total field costs are still
necessarily incom- plete, and to properly account for any • Personnel: these costs are linearly dependent on
unidentified costs, a contingency item must be added facility size. Considering a solar detoxification facility
to the cost. This could be from 10 to 20 per cent of the normally operating seven days per week and fifty-two
facility cost, depending on the degree of design weeks per year, personnel requirements for a 500-m2
definition. The sum of facility cost and contingencies is solar plant are estimated to be only 0.1 person/year,
the ‘total facility cost’. since the plant is completely automated. Also, because
• System engineering AND ASSEMBLY : this includes of the simplicity of one-sun collectors, the total annual
system design and integration into existing maintenance require- ment is also estimated to be 0.25
installations, specifi- cations, procurement of system person/year, mainly for periodic cleaning.
components, scheduling, project management, system • Cost of MAINTENANCE MATERIALS: the annual cost of
assembly, training of opera- tors and engineering mainte- nance materials is mainly due to mechanical,
support. Engineering costs are esti- mated to be 50 electrical and instrumentation equipment, which could
per cent of the total facility cost. amount to about 2 per cent of the total facility cost.
• SPARE PARTS: 0.5 per cent of the total facility cost is
nor- mally included as a spare parts allowance.

TABLE 8.1. Estimated cost of typical sizes of solar detoxification facility (euros, 1999).

Collector area (m2) 100 200 300 500 1 000 10 000

Control, pumps and
instrumentation 23 228 25 190 27 900 32 048 50 793
20 221

Piping, valves,
structures and
Manifolds, tanks 22 201 28 315 38 471 58 311 232 139
14 647
Catalyst separation 3 507 4 473 6 078 9 212 36 673
recovery 2 314
Collectors and 21 220
reactor 42 440 63 660 106 100 168 556 1 338 891

Civil engineering 6 025 12 050 18 075 30 125 60 250 602 500
Equipment transport 601 911 1 162 1 579 2 393 9 525

TOTAL INVESTMENT 65 028 104 337 140 875 210 251 330 770 2 270 522
Cost/m2 650 522 470 420 331 227
Economic Assessment 131

• Electricity: the facility consumes electricity for twelve years with a depreciation period of ten years
pumps, control system, instrumentation and lighting: may normally be assumed. With this data and
the vast majority (around 90 per cent) by the main considering, as an example, that the plant is financed by a
system pump(s). The average annual consumption of debt with a 5 per cent interest rate and that the total
this auxil- iary power consumption, per m3 of treated annual expenses due to taxes, insurance and so on are 2
water, is diffi- cult to estimate as the ratio between total per cent of the total installed cost, an FCR of 17 per cent
treated volume and plant size is directly dependent on is obtained (0.1 from the ten year depreciation + 0.05 +
the nature of the contaminants. It could more easily be 0.02 = 0.17).
estimated as a func- tion of the solar field size, as this
determines total pres- sure-drop and, consequently, 8.2.1. Example A: TiO2-based solar
pump specifications. In view of this, yearly power detoxification plant
consumption can be estimated as between 40 and 80
kWh per m2 of installed solar field.
• CHEMICAL supplies: although one of the main factors in A 300-m2 solar detoxification plant has been designed
the yearly operating cost is the consumable chemicals, (see also Chapter 9) for the yearly treatment of 6,000
an estimate cannot be given here because it depends m3 of wastewater contaminated by pesticide, using a
heavily on each specific treatment process. The TiO2-persulfate (see Section 5.2) photocatalytic system:
chemicals used
by Photo-Fenton processes are completely different TiO2  hν  e  h (λ  390 (8.10)
from TiO2 processes; they also depend on the nature
and com- position of the wastewater, for which V (8.11)
some pre or post- treatment might be needed, with B
specific chemicals in each case. A case-by-case h+V  H2O  OH  H+

cost study must therefore be performed. Two B (8.12)

examples can be found in Sections 8.2.1 and 8.2.2.
S2O82  eCB
  SO  SO2
4 4

SO  H O  OH  SO2  H
4 2 4
The sum of costs for personnel, maintenance materials,
elec- tricity and chemical supplies is the ‘total operating It is estimated that the plant would operate about 3,000
cost’ of the solar detoxification system. To obtain the hours per year. An estimate of the yearly treatment cost,
‘annual treat- ment cost’, per m3 of treated wastewater, total in 1999 euros, can be performed as in Table 8.2.
installed costs are converted to an annual levellized cost To obtain the treatment cost, an FCR of 17 per cent,
considering a fixed charge rate (FCR), which is obtained by with ten-year plant depreciation, is estimated (Table 8.3).
calculating all the fixed costs (excluding operation) for the It should be noted that a 4 per cent reduction in the
life of the plant. The annual levellized cost and treatment FCR (from 0.17 to 0.13) has a greater effect on the
cost can be calcu- lated as: annual treat- ment cost than a 50 per cent reduction in the
solar collector field. Also, about 50 per cent of the
Annual Levelized Cost annual operating cost is the cost of the persulfate itself, so
any reduction in this affects
 Total Installed Cost  FCR  Operating (8.8) the treatment cost very significantly. In this way, it must
Cost be taken into account that addition of consumable
reagents (such as persulfate) may not necessarily be
Treatment Cost (Euros/m3 ) required by spe- cific TiO2 photocatalytic applications.
Annual Cost ) 8.2.2. Example B: Photo-Fenton based

Annual Treatment
The ‘fixed charge rate’ represents the equivalent revenue that solar detoxification plant
must be generated annually to meet all the charges on each
euro of facility investment. FCR is normally equal to the A 200-m2 solar detoxification plant has been designed for
sum of the return on debt, taxes, depreciation, insurance and the yearly treatment of 6,000 m 3 of wastewater
so on. For solar detoxification treatment plants, a plant contaminated by pesticides, using Fenton’s reagent (H2O2
life of and Fe2+) irradiated
132 Solar Detoxification

TABLE 8.2. Estimated operating cost of a TiO 2-persulfate solar detoxification plant (euros, 1999).

A Facility cost Data from table 8.1 (300 m 2 of solar collector system
aperture area) 140 875
B Project contingency 15% of the facility cost is estimated 21 131
C Engineering and setup 50% of A+B (total facility cost) 81 003
D Spare parts 0.5% of A+B (total facility cost) 8 100


F Personnel cost 0.25 person/year @ € 20 000/person/year. 5 000

G Maintenance material cost 2% of A+B (total facility cost) is estimated 3 240
H Electricity 5 kW/hour average consumption, 3 000 operation
hours/year = 15 000 kWh @ 0.1 €/kWh. 1 500
I Chemical supplies – TiO2 catalyst at 200 mg–1 concentration, re-used up to
10 times = 120 kg per year. 1 080
– S2O8Na2 (MW 238), at concentrations of 5–10 mM, overall
consumption per batch cycle of 15mM = 21 420 kg per year. 53 550
– Sodium hydroxide, NaOH (MW 40), 50% solution, for
pH adjustment and neutralization of the treated
water = approximately 4 800 kg per year. 1 344
J ABLE 8.3. Estimated annual
TOTAL OPERATING COSTtreatment cost of a TiO 2-persulfate solar detoxification plant (euros, 1999).
F+G+H+I 65 714

K Annual levelized cost E × FCR + J (formula 8.8) 108 402

L Annual treament cost K is divided by 6 000 (yearly treated volume) 18.0/m3

in the UV-V range (Photo-Fenton method): 0.02 = 0.12) and this would mean a very significant
reduction in treatment cost.
Fe2  H2O2  Fe3  OH  OH (8.14 • Operation is a significant portion of the cost of treated
) water. Reduction is likely to be difficult due to the
low capacity factor of the facilities and the costs of
Fe3  H2O  hν  Fe2  H  OH (λ  580 nm) consum- able supplies, which are proportional to
) the volume of
As in the previous case, 3,000 operating hours per year treated water.
are estimated. The yearly treatment cost would then be Two additional examples of solar photocatalytic
that shown in Table 8.4. treatment cost found in the literature are the following:
Also as in the previous example, an FCR of 17 per cent
with a ten-year plant depreciation period is estimated • Groundwater remediation at Livermore, California,
(Table 8.5). by NREL (United States). Average treatment
Two additional conclusions from this example may be capacity of
added: 4.4 L/s of water, with a peak flow of 30 L/s, inlet
con- centration of 400 ppb TCE and maximum outlet
• For a depreciation period of twenty years (longer concen- tration of 5 ppb. Total treatment cost was
facility lifetime), the FCR would be 12 per cent reported as US$4.07/m3.
(0.05 + 0.05 +
Economic Assessment 133

TABLE 8.4. Estimated operating cost of a Photo-Fenton solar detoxification plant (euros, 1999).

A Facility cost Data from table 8.1 (300 m 2 of solar 104 337
collector system aperture area)
B Project contingency 12% of the facility cost is estimated 12 520
C Engineering and setup 50% of A+B (total facility cost) 58 429

D Spare parts 0.5% of A+B (total facility cost) 5 843


F Personnel cost 0.20 person/year @ € 20 000/person/year. 2 337

G Maintenance material cost 2% of A+B (total facility cost) is estimated 3 240
H Electricity 4 kW/hour average consumption, 3 000 operation hours/year
= 12 000 kWh @ 0.1 €/kWh. 1 200
– FeSO4 7H2O (MW 278) catalyst at 1mM concentration,
= 1 668 kg per year. 1 080
– Hydrogen peroxide, H2O2 (MW 34), 30% solution, overall
I Chemical supplies consumption per treatment cycle of 33.75 mM, = 22 950 kg per year. 17 442
– Sodium hydroxide, NaOH (MW 40), 50% solution, for batch mode
neutralization of the treated water = approximately 3 036 kg per year. 849


TABLE 8.5. Estimated annual treatment cost of a Photo-Fenton solar detoxification plant (euros, 1999).

K Annual levelized cost E × FCR + J (formula 8.8) 57 349

L Annual treatment cost K is divided by 6 000 (yearly treated volume) 9.5/m3

• Study for a site at the Rocky Flats plant near Boulder, becomes cheaper than photons generated by electricity.
Colorado (Bechtel Corp., 1991) to treat an annual Obviously, when only a small amount of solar irradiation
volume of 2,246 m3 with a peak water flow of 6.3 L/s. is available, solar-driven processes make no sense, so
The levellized cost of treated water was estimated at the measurement of the solar resource available at a
US$10.57/m3. This cost was dominated by the specific loca- tion might permit a reasonable initial
expensive system required to treat inorganic estimation of the advis- ability of a more in-depth study
components of the water at Rocky Flats. on solar photocatalytic processes.
Electric ultraviolet lamps are currently available on
8.3. Solar or electric photons? the market for a variety of applications, such as lighting,
food processing, medical treatment, tanning, lacquer
drying, photochemical synthesis, photo-polymerization
Another important aspect to bear in mind in an economic
processes and attracting insects. They can also be used
assessment is the threshold at which the collection of
for photo- catalytic degradation processes.
solar photons required to drive the photochemical
134 Solar Detoxification

The most suitable of the different types of electric and 30 per cent of UV-photon production and an average
lamps available for UV-photon production are the low- efficiency (lifetime) of about 20 per cent. Average lamp
pressure mercury fluorescent lamps (the same common life is around 20,000 hours.
fluorescent light normally used for illumination), which With the above data, it is easy to obtain the
are based on the generation of an electric arc through a equivalence between UV-lamp and solar systems. As an
sealed chamber con- taining mercury vapour. The result example, we can calculate how many standard 40-W
is 245-nm radiation that is absorbed by a phosphorous UV-fluorescent lamps are equivalent to a 500-m2 TiO2-
wall coating producing a fluo- rescent radiation which, based solar plant with an average yearly global UV
depending on the coating, can be adjusted to any irradiation of 20 W m–2. The parameter used to compare
spectrum. As for any arc lamp, ballast is required to them is the number of useful UV photons generated or
provide the appropriate current and voltage to start. collected. In the case of the solar plant, this can be
Several studies have demonstrated that these lamps are obtained directly from the standard ASTM solar
the easiest, simplest and cheapest way to produce UV spectrum (assuming constant spectral distribution) using
pho- tons and, in fact, they are already being used in the following equation (see also Section 8.4):
commercially available TiO 2 photocatalytic lamp
Another type is the germicidal lamp, which is N  5.8  1021 I  (8.16)
basically photons
the same except that the 254-nm light passes directly Φ(E  3.2
eV)  m2h 
UVg  

through transparent glass. As their peak light emission is

centred at the mentioned wavelength, they produce fewer Where
valid photons and are less efficient for photocatalytic N8 (E>3.2 eV) = number of photons with energy over 3.2 eV
appli- cations. Medium and high-pressure mercury lamps (wavelength up to 387 nm) per unit of time
are also similar to the low-pressure lamps, but the higher and surface.
pressure generates higher intensity radiation and also –
IUVg = yearly average global UV irradiation (W m–2).
tends to shift the spectrum towards the visible light.
Finally, metal halide lamps, high-pressure sodium lamps The total amount of UV photons collected yearly by the
and incandescent lamps cannot be used for UV light solar plant would be:
NUV  NΦ (E > 3.2 eV) HSS  5.8  1021
Figure 8.5 shows two typical low-pressure 40-W
fluorescent lamps with dominant emission spectrum (8.17)
centred at 313 and 340 nm. According to manufacturer  20  3,000  500  1.74 
data, the ini- tial standard efficiency of these lamps is 1029
between 25 per cent :
NUV = total UV photons (with energy over 3.2 eV) collected
1.0 1.0 yearly.
PSA spectrum (25 WUV/m ) 0.9
HS = total yearly hours of operation of a solar
ASTM spectrum detoxification system. This value depends on the
geographic loca- tion, but it may be estimated at
Irradiance (W m nm )

0.7 UVA–340
0.7 3,500 (near the equa- tor to parallel 20) to 2,500
(40th to 50th parallels). In the example, 3,000 is

0.6 UVB–313
0.5 used.
S = Surface area of solar collector field (500 m 2 in the

0.3 example).

0.2 0.2 IUVg = yearly average global UV irradiation during the 3,000
0.1 0.1
estimated hours of operation. 20 W m –2 in this
exam- ple (see also Chapter 8.4).
0.0 0.0
270 290 310 330 350 370 390 410
So if the lamps must supply the same amount of UV energy:
Wavelength (nm)
FIGURE 8.5. Ultraviolet spectra: solar (standard and PSA
measured spectrum) and two low-pressure 40 W mercury NUV  1.74  1029  ΦPH HL NL
lamps (QUV fluorescent lamps: UVB-313 and UVA-340).  5.04  1022  8,760  NL (8.18)
Economic Assessment 135

Where: 5 × 1029 (5.E+29)

φPH = Nominal average photonic flux of the lamp. 1030 (1.E+30)
Manufacturer’s information usually places this
value between 1.3 to 1.4 × 1019 photons per second Using Equations 8.16 and 8.17, the solar field necessary
(stan- dard mercury UV lamps with emission (with different possible yearly average UV global
spectrum centred at 360 nm). In the example, this irradiation) can be calculated to collect the targeted
has been translated to photons per hour. amounts of photons. Once the solar fields have been
HL = Total yearly hours of operation of a lamp-based calculated, their cost can be esti- mated from the data in
detox- ification system, which in this case is the Table 8.1 (using a polynomial fitting). Finally, with the
full year (8,760 hours). estimated cost of the solar field, the annual cost can be
NL = Number of lamps of the system. calculated using Equation 8.8, considering an FCR of 17
per cent, as in the previous examples. No operating costs
So, according to Eq. 8.18, in the example considered, the are considered. It should be noticed that the only cost
equivalent number of lamps would be 394. included here is the cost of the solar collector, including
From this particular example, the following general the reactor tubes, as it is the only hardware related to UV-
equa- tion for estimating the equivalent number of any photon collection. The results are shown in Tables 8.6 to
type of electric lamps can be obtained: 8.10.
 I UVg   HS  The comparison of electric and solar UV photon
N  5.8  1021 S collec- tion is only affected by the reactor system, since
the rest of the system components are quite similar in
design and cost. In the case of a solar system it is
necessary to include the cost
 ΦPH HL 
(8.19 of the solar collector field, which does not exist in the case of
) a lamp system. In the case of a lamp-based system, however,
Based on these equations, the cost of collecting UV in addition to the electrical installation itself, one must
photons with solar technology may be compared to the include the cost of the lamps and their periodic
generation of the same amount of UV photons using replacement, the cost of electricity and the higher cost of
electric lamps. This is done considering the following the reactor, which has to be more complex than a simple
amounts of photons to drive the photochemical process: tube so as to either allocate the lamp inside the reactor or
situate it exter- nally. On the other hand, reactors based
1028 (1.E+28) on lamps can be more compact than solar collectors, thus
5 × 1028 (5.E+28) needing much less space. All these factors make the cost
comparison of both systems a difficult task. Using
1029 (1.E+29) Equation 8.18 the equivalent

TABLE 8.6. Estimated yearly cost of collecting 1.E+28 solar UV-photons at different yearly average UV global
irradiation (Costs are indicated in euros, 1999).

I UVg Hs S Yearly cost Photon cost
(WUV/m2) (hours) (m2) Investment cost (FRC = 17%) (€/1.E+25 photons)
40 3 500 12 13 345 2 269 2.27
35 3 500 14 13 638 2 319 2.32
30 3 500 16 14 029 2 385 2.39
25 3 000 23 15 124 2 571 2.57
20 3 000 29 16 082 2 734 2.73
15 2 500 46 18 953 3 222 3.22
10 2 500 69 22 779 3 872 3.87
5 1 500 230 49 457 8 408 8.41
136 Solar Detoxification

TABLE 8.7. Estimated yearly cost of collecting 5.E+28 solar UV photons at different yearly average UV global
irradiation (Costs are indicated in euros, 1999).
(WUV/m2 ) Hs
(hours) S(m ) Yearly
(FRC = 17%) Photon
(€/1.E+25 photons)
2 Investment
UVg cost cost
40 3 500 62 cost
21 550 3 663 0.73
35 3 500 70 23 013 3 912 0.78
30 3 500 82 24 963 4 244 0.85
25 3 000 115 30 419 5 171 1.03
20 3 000 144 35 187 5 982 1.20
15 2 500 230 49 457 8 408 1.68
10 2 500 345 68 404 11 629 2.33
5 1 500 1 149 198 524 33 749 6.75

TABLE 8.8. Estimated yearly cost of collecting 1.E+29 solar UV photons at different yearly average UV global
irradiation (euros, 1999).
(WUV/m2 ) Hs
(hours) S(m ) Yearly
(FRC = 17%) Photon
(€/1.E+25 photons)
2 Investment
UVg cost cost
40 3 500 123 cost
31 782 5 403 0.54
35 3 500 141 34 701 5 899 0.59
30 3 500 164 38 589 6 560 0.66
25 3 000 230 49 457 8 408 0.84
20 3 000 287 58 942 10 020 1.00
15 2 500 460 87 262 14 835 1.48
10 2 500 690 124 709 21 201 2.12
5 1 500 2 299 376 772 64 051 6.41

TABLE 8.9. Estimated yearly cost collecting 5.E+29 solar UV photons at different yearly average UV global
irradiation (euros, 1999).
(WUV/m2 ) Hs
(hours) S(m ) Yearly
(FRC = 17%) Photon
(€/1.E+25 photons)
2 Investment
UVg cost cost
40 3 500 616 cost
112 712 19 161 0.38
35 3 500 704 126 990 21 588 0.43
30 3 500 821 145 945 24 811 0.50
25 3 000 1 149 198 524 33 749 0.67
20 3 000 1 437 243 928 41 468 0.83
15 2 500 2 299 376 772 64 051 1.28
10 2 500 3 448 546 036 92 826 1.86
5 1 500 11 494 1 479 332 251 486 5.03
Economic Assessment 137

TABLE 8.10. Estimated yearly cost of collecting 1.E+30 solar UV photons at different yearly average UV global
irradiation (euros, 1999).
(WUV/m2 ) Hs
(hours) S(m ) Yearly
(FRC = 17%) Photon
(€/1.E+25 photons)
2 Investment
UVg cost cost
40 3 500 1 232 cost
211 554 35 964 0.36
35 3 500 1 407 239 321 40 684 0.41
30 3 500 1 642 276 016 46 923 0.47
25 3 000 2 299 376 772 64 051 0.64
20 3 000 2 874 462 527 78 630 0.79
15 2 500 4 598 706 316 120 074 1.20
10 2 500 6 897 999 924 169 987 1.70
5 1 500 22 989 2 048 971 348 325 3.48

number of lamps (NL) to generate the same amount of even – depending on how compact the system assembly
UV photons can be obtained directly: is – a cooling system. As this cannot be estimated with
any precision, Table 8.7 attempts to cover a complete
1.E+28 photons → 23 lamps (40 W low pressure
range of possible costs by presenting two conservative
mercury fluorescent tubes)
estimates, one in which the estimate of per-lamp reactor
5.E+28 photons → 113 lamps (40 W low pressure cost is low (€100) and another in which it is high (€200).
mercury fluorescent tubes)
The main conclusions arrived at from a comparison of
1.E+29 photons → 226 lamps (40 W low pressure
Table 8.11 and Tables 8.6 to 8.10 are the following:
mercury fluorescent tubes)
5.E+29 photons → 1,132 lamps (40 W low pressure mer- • The available solar UV-irradiation for a specific
cury fluorescent tubes) location is the main factor in determining the cost of a
1.E+30 photons → 2,265 lamps (40 W low pressure mer- solar-based system.
cury fluorescent tubes). • The cost of collecting photons with solar technology
decreases proportionally as the yearly total increases
With these data, Table 8.11 shows the estimated cost of
because of the reduction in collector production cost
UV photon generation with electric lamps. The main
for larger numbers. This is not the estimated case of
advantage of an electric system is its total availability lamp- based systems as they are much more compact.
(twenty-four hours a day), which is on average three
• When fewer photons are needed (1.E+28), lamp-
times higher than solar systems. The total yearly cost is
based systems are cheaper than solar.
obtained by using Equation 8.8 with an FCR of 17 per
• Solar-based systems with collector apertures of over
cent as calculated above for the solar-based system. The
100 m2 are clearly cheaper than lamp-based systems
operating cost includes only electricity, replacement of
when the average yearly global UV irradiation is
lamps and the cost of labour involved in replacement.
over 15 W/m2 (this must be considered a qualitative
To obtain these values, a 20,000- hour lifetime has been
approach rather than quantitative).
assumed for the lamps. The cost of the lamp system may
be considered as linearly dependent on the number of • The threshold UV radiation (below which lamp-
based systems are cheaper than solar) decreases as the
yearly amount of photons increases.
The main uncertainty in Table 8.11 is the cost of the
reac- tor in the lamp-based photocatalytic system. It is All data included in Tables 8.6 to 8.11 are
clear that this item would necessarily be an expensive summarized in Figure 8.6, which shows the
component of the electric system as it must include comparative cost of solar and electric technologies for
piping, external UV- reflectors to avoid loss of UV UV-photon collection and
photons, supports, wiring, and
138 Solar Detoxification

TABLE 8.11. Estimated yearly cost of UV photon generation with electric lamps. Electricity cost = 0.15 Euros/kWh;
FCR = 17 per cent. In Table 8.11A the cost per lamp is €100; in Table 8.11B the cost per lamp is €200 (euros,


Number of 1.00E+2 5.00E+2 1.00E+2 5.00E+2 1.00E+3

photons Number Unit
8 8 9 9 0
of lamps
Lamp, ballast and 23 113 226 1 132 2 265
6.80 154 770 1 540 7 701 15 402
Reactor cost 100.00 2 265 11 325 22 650 113 249 226 499
A) Investment cost 106.80 2 419 12 095 24 190 120 950 241 900
Electricity cost
52.56 1 190 5 952 11 905 59 524 119 048
(0.15 €/kWh)
Lamp replacement 2.98 67 337 675 3 373 6 746
Labour cost of
3.15 71 357 714 3 571 7 143
B) OPERATION cost 58.69 1 329 6 647 13 294 66 468 132 937
TOTAL COST: 76.85 1 741 8 703 17 406 87 030 174 060
A × FCR + B
Cost per 1E+25 UV photons 1.74 1.74 1.74 1.74 1.74


Number of 1.00E+2 5.00E+2 1.00E+2 5.00E+2 1.00E+3

photons Number Unit
8 8 9 9 0
of lamps
Lamp, ballast and 23 113 226 1 132 2 265
6.80 154 770 1 540 7 701 15 402
Reactor cost 200.00 4 530 22 650 45 300 226 499 452 997
A) Investment cost 206.80 4 684 23 420 46 840 234 199 468 399
Electricity cost 52.56 1 190 5 952 11 905 59 524 119 048
(0.15 €/kWh)
Lamp replacement 2.98 67 337 675 3 373 6 746
Labour cost of
3.15 71 357 714 3 571 7 143
B) OPERATION cost 58.69 1 329 6 647 13 294 66 468 132 937
TOTAL COST: 93.85 2 126 10 628 21 256 106 282 212 564
A × FCR + B
Cost per 1E+25 UV photons 2.13 2.13 2.13 2.13 2.13

generation, respectively. Three different amounts of Figure 8.6 enables easy calculation of the solar
photons are plotted (1.E+29, 5.E+29, 1.E+30) against the threshold that determines the advantage of solar over
average global solar irradiation by adjusting the data in lamp-based tech- nology, as the yearly UV-photon
Tables 8.8, requirement of the majority of possible applications will
8.9 and 8.10 (yearly cost). be in the range of 1.E+29 to
Economic Assessment 139

(electrical lamps)

FIGURE 8.6. Comparative cost of UV photon FIGURE 8.7. Comparative cost of UV photon
collection/gener- ation with solar technology (CPCs) and collection/gener- ation with solar technology (CPCs) and
electric lamps (elec- tricity cost = €0.15/kWh), electric lamps (elec- tricity cost = €0.05/kWh),
respectively. Data from tables 8.6 to 8.11 (Costs are respectively. Data from tables 8.6 to 8.11 (Costs are
indicated in 1999 euros). indicated in 1999 euros).

1.E+30. If a figure of €0.15/kWh is assumed for

electricity cost, a threshold between 10 and 12 WUV/m 2
is obtained (see Figure 8.6). If a different cost is
assumed, Table 8.11 must be updated. This is the case of
Figure 8.7, which corre- sponds to a cost of €0.05/kWh.
In this case, the threshold for solar advantage over
lamp-based systems is logically higher than before.

8.4. Solar resources assessment

Because the dimensioning of a solar photocatalytic

system is directly related to the amount of UV light
available, accurate assessment of the solar resource is
important. Solar UV-light (300 to 400 nm) comprises
roughly 2 to 3 per cent of the direct beam energy and
4 to 6 per cent of the combined direct beam and diffuse FIGURE 8.8. Average direct and global UV irradiance
irradiation (global). Non-concentrat- ing systems have (sunrise to sunset) at PSA (Almería, Spain).
an advantage with regard to resource assessment, as Meteorological data records from 1991 to 1995.
there are extensive data on the total energy (direct and
diffuse) available from sunlight and the 4 to 6 per cent
30 W m–2. However, the best situation is when a
conversion factor makes resource assessment a straight-
database of historical global UV irradiation is available,
forward process for non-concentrating systems.
such as the case in Figure 8.8.
The cost of a specific solar detoxification treatment
As noted in Section 2.4, to reliably estimate the annual
plant may be easily estimated from the indications in
average UV radiation, the so-called ‘cloud factor’ (f ) must be
Sections 8.1,
found. This is the average percentage of global UV
8.2 and 8.3, once the degradation process has been assessed
radiation that is expected to be lost due to clouds throughout
and treatment factors have been identified. Nevertheless,
the year. The cloud factor is also intended to be a
the yearly average UV irradiation at the specific location
measurement of atmospheric transparency and is affected
of potential solar detoxification plants must be known. In
by all the atmos- pheric components that can absorb or
the so-called ‘solar belt’ (between the Earth parallels 40
scatter solar radiation. Figure 8.9 shows the results obtained
North and 40 South), this irradiation is usually between
at PSA, following the procedure indicated in Section 2.4.
15 and
Economic Assessment 141

α = Solar altitude, defined as the angle formed by the solar

vector to the horizontal surface of the Earth
Solar altitude can be obtained as follows:

sin α  sin(L) sin(δS )  cos(L) cos(δS ) (8.23)

cos(hS )
 2π(10,5  N) 
δ  sin(23, 45) cos 
s  
 36,525 

 Where:
 L = Local latitude.
 δs = Solar declination.
 hs = Solar time angle, which is zero at noon, positive
FIGURE 8.9. Average ‘cloud factor’ for global UV irradiance before and negative after noon. It can be calculated
(sunrise to sunset) at PSA (Almería, Spain). Meteorological at any time by adding 15º of each hour of difference
data records from 1991 to 1995. with respect to the noon (Example: at 15.00 hours
of solar time, hs = –45º).
Another interesting alternative for estimating the N = Day number; 1 at 1 January and 365 (or 366) at 31
annual UV radiation availability is through the Lambert- December. 23.45 is the angle formed by the Earth’s
Beer Law (also called the ‘Bourguer Law’), which axis and the plane formed by the Sun and the elliptical
indicates that light attenuation through a continuous tra- jectory of the Earth.
medium is proportional to the flux radiation and the
distance covered: Equations 8.21 to 8.24 allow solar irradiation at any specific
Ib,λ  Io,λ e c m
(8.20) location on Earth to be detailed; must be measured at the
specific location and Io,λ can easily be calculated. Due to the
where: Earth’s slightly eccentric elliptical orbit, solar
Ib,λ = Intensity of solar irradiation on the Earth’s surface at extraterrestrial radiation varies ±3.4 per cent throughout the
a specific wavelength, λ. year depending on the day number:
Io,λ = Intensity of extraterrestrial solar irradiation at the
 2πN  
same wavelength, λ. Io,λ  Io,λ (N)  Isc,λ 1  0.034 cos  
cZ = Atmosphere attenuation coefficient at 365  (8.25)
m = Air mass ratio (see Chapter 2), defined as the  

oblique optical path described by a photon in the
atmosphere relative to the minimum vertical path to λ2 
arrive at the same terrestrial location. If the Isc,λ   E(λ) dλ (8.26)
elevation of the Sun is 90°, then m = 1. λ1

The air mass ratio can be obtained for any moment by Where:
following equations: Isc,λ = Solar constant associated with the spectral
0.0001184 Z
interval (λ1,λ2) to which is intended to obtain the
m  mo e (8.21 atmospheric attenuation coefficient (cλ). This
) interval must also be coherent with the measure of
mo  1,229  (614 sin  614 sin Ib,λ (i.e., if we are using a solar radiometer which
α measures within the interval 295 to 400 nm, Isc,λ
α)2 (8.22
must be calculated to the same interval).
where: )
mo = Air mass ratio at sea level E(λ) = Extraterrestrial irradiation, which can be found in
z = Elevation above sea-level in Table 8.12.
140 Solar Detoxification

TABLE 8.12. Extraterrestrial solar irradiation.

λ Eλ λ Eλ λ Eλ
(µm) (W m–2 µm–1) (µm) (W m–2 µm–1) (µm) (W m–2 µm–1)
0.115 0.007 0.43 1 639 0.90 891
0.14 0.03 0.44 1 810 1.00 748
0.16 0.23 0.45 2 006 1.2 485
0.18 1.25 0.46 2 066 1.4 337
0.20 10.7 0.47 2 033 1.6 245
0.22 57.5 0.48 2 074 1.8 159
0.23 66.7 0.49 1 950 2.0 103
0.24 63.0 0.50 1 942 2.2 79
0.25 70.9 0.51 1 882 2.4 62
0.26 130 0.52 1 883 2.6 48
0.27 232 0.53 1 842 2.8 39
0.28 222 0.54 1 783 3.0 31
0.29 482 0.55 1 725 3.2 22.6
0.30 514 0.56 1 695 3.4 16.6
0.31 689 0.57 1 712 3.6 13.5
0.32 830 0.58 1 715 3.8 11.1
0.33 1 059 0.59 1 700 4.0 9.5
0.34 1 074 0.60 1 666 4.3 5.9
0.35 1 093 0.62 1 602 5.0 3.8
0.36 1 068 0.64 1 544 6.0 1.8
0.37 1 181 0.66 1 486 7.0 1.0
0.38 1 120 0.68 1 427 8.0 0.59
0.39 1 098 0.70 1 369 10.0 0.24
0.40 1 429 0.72 1 314 15.0 0.0048
0.41 1 751 0.75 1 235 20.0 0.0015
0.42 1 747 0.80 1 109 50.0 0.0004

The solar constant for the whole solar spectrum (total that the atmosphere filters UV radiation relatively more
solar radiation) is 1,367 W m –2 (value adopted by the than the overall spectrum.
World Radiation Center, with an uncertainty of 1 per With this background, by measuring Ib,λ, the
cent). In the case of the UV spectrum (295 to 400 nm), atmospheric attenuation coefficient (cλ) can be calculated.
using Table 8.12, it is: This parameter includes all the factors relevant to solar
irradiation, such as Raleigh dispersion, ozone absorption,
atmospheric turbidity
400 2 and spray, absorption by water molecules, clouds and so on.
 E(λ) dλ  118.1  13.6  104.4 W (8.27
) The knowledge of at a specific location makes it possible to
 cλ
forecast solar irradiation during the year. Variations in cλ
This value of 104.4 W m–2 represents 7.72 per cent of total over the year also enable conclusions to be made about
extraterrestrial irradiation. As the percentage on the Earth’s changes in atmosphere performance (with regard to solar
surface is lower (normally from 4 to 6 per cent), this means light) at the specific location.
142 Solar Detoxification

8.5. Comparison with other depend on the contaminants to be destroyed. Natural gas
technologies or propane is the usual energy source. Electricity can also
be used in the primary chamber, as in infrared furnaces.
Incineration is the most frequently used treatment for
A general qualitative approach to the range of highly concentrated wastes or contaminated soil. The
application of different treatment technologies standard per- formance requirement is 99.9999 per cent
(concentration of contaminant versus flow rate to be (required by law in many countries).
treated) is provided in Figure 8.10. Incineration is a source of debate, with opposition to
The technologies discussed below may be considered the construction of new incinerators coming from the
the main competitors of aqueous-phase solar local pop- ulations because of the increasing perception
detoxification for the treatment of hazardous water that they cause health and safety problems. Incineration
contaminants. As only non-biodegradable contaminants is also an energy- intensive process subject to stream size
are considered here, bio- logical treatment technologies and concentration constraints, with significant capital
are not discussed. equipment and operating costs. As a consequence, it is
expensive compared to other treatment processes.
8.5.1. Thermal oxidation Typical operating costs are also high, in the range of
€200 to €1,000 per ton, depending on the nature of the
contaminants and the facility utilization factor. Fuel is
Most of the thermal oxidation technologies (also known as
the primary factor affecting operating cost.
‘thermal incineration’) are used to destroy hazardous wastes
The advantages of incineration are that the volume of
at high temperatures (normally higher than 650 °C). At such
waste is greatly reduced, often by as much as 90 per cent.
temperatures, chemical bonds are effectively broken and
Furthermore, the resulting ash is generally more stable
transformed into other more benign substances. There are
and less likely to leach into groundwater than the parent
different technologies (rotary kilns, infrared furnaces, con-
com- pounds. On the other hand, poor combustion during
ventional incinerators and so on), but almost all of them are
incin- eration often produces by-products that are at least
variations of the same process. Incineration is typically a
as toxic as the parent material. Metals such as lead,
two-step process in which a primary combustion chamber
mercury, and chromium could be released into the air.
burns the waste material at temperatures of about 1,000 °C
The burning of some plastics can produce hydrogen
and the exhaust gases containing desorbed oxidized or par-
chloride, which becomes hydrochloric acid rain if
tially oxidized organic contaminants are transferred to the
combined with the mois- ture in air. Benzene,
secondary combustion chamber (afterburner) where these
chloroform, and TCE have frequently been found in stack
gases are burned to produce carbon dioxide, carbon monox-
emissions. Finally, generation of organic compounds and
ide, inorganic acids and water. Afterburners generally oper-
dioxins as a consequence of incomplete destruction has
ate at about 1,200 °C and residence times in both
always been a concern.

8.5.2. Catalytic oxidation
WAO & incineration recovery
1,000 Incineration
Wet advanced oxidation (WAO)
Chemical oxidation Catalytic oxidation (also known as ‘catalytic incineration’) is
WAO & biological
Advanced oxidation processes very similar to thermal oxidation. Contaminants are pre-
TOC (mg L)

100 heated, mixed, and combusted at high temperatures to
form carbon dioxide and water. The main difference is the
presence of a catalyst inside the combustion unit that
10 Ozone reduces the acti- vating energy necessary for combustion.
Solar detoxification
H2 O2
Therefore, combus- tion occurs at a lower temperature
Biological than for thermal oxidation and, as a result, fuel costs for
1 catalytic oxidation are usually lower than for an
0 10 20 30 40 50 60 70 80
3 –1 equivalent thermal oxidation system.
Flow rate (m h )
Catalytic materials, such as platinum or palladium,
FIGURE 8.10. Technology fit map: range of application of and metal oxides such as chrome-alumina or cobalt oxide
different water treatment technologies. are
Economic Assessment 143

introduced into the combustion unit in either a The Henry’s Law Constant H can then be expressed as
monolithic or beaded configuration. The average catalyst a dimensionless ratio of a compound’s concentration in
lifetime is nor- mally from two to five years, after which air to its concentration in water. The application of Henry’s
deactivation by inhibitors, blinding by external particles Law Constant makes it possible to assess the maximum
and thermal ageing, render it ineffective. efficiency of an air-stripping system theoretically
The most relevant characteristics of catalytic in possible under ideal conditions. A compound with high
compari- son with thermal oxidation systems are the H values is more likely to be removed from water and
following: soil by air-stripping. Normally, compounds with an H
over 0.001 can be removed effectively.
• Not every organic contaminant may be treated by H is over 0.01 for many organic compounds, which is
catalytic oxidation systems, which are not effective the reason the air-stripping process is so widely used.
for streams containing lead, arsenic, phosphorus, Conventional air-stripping treatments of hazardous
bismuth, antimony, mercury, iron oxide, tin, zinc or organic compounds in water are based on packed or tray
other catalyst deactivators. columns in which water and air usually flow downward
• Catalytic oxidation systems are usually applied to and upward respectively, with a very high contact area
low concentration streams, since high concentrations between the two phases (spray chambers, venturi
may be associated with high heat content that can scrubbers, plate or tray towers, packed towers and so
generate enough heat from combustion to deactivate on). The technology has been applied to a wide range of
the catalyst. organic compounds with varying degrees of
• Associated with the above-mentioned characteristics, effectiveness at contaminant concentrations between
air is usually not required for the combustion process. 250 and 10,000 mg L –1. The initial investment in air
• Temperature and pressure through the catalyst bed stripping may be considered medium and operating costs
should be monitored to preserve catalyst activity. are low.
Catalytic oxidation systems normally operate at The main environmental objection to air stripping is
tempera- tures between 250 °C and 500 °C, since the undeniable fact that it does not really destroy
excessive heat can deactivate most catalysts. hazardous compounds, but simply transfers them to
• Catalyst poisoning from metals or halogens and/or another medium, as the organic compounds stripped
bind- ing from particulate matter over time can from water are released directly into the atmosphere.
decrease destruction efficiency. Periodic catalyst Tighter restrictions on emis- sions from air strippers are
replacement or reactivation is then required. limiting their applicability and increasing their cost, as
Catalytic incinerators have a higher initial cost than secondary air-phase treatment systems (that is, real
thermal incinerators, but their lower fuel costs and treatment processes) are required afterward.
thereby lower operating cost offset this.
8.5.4. Adsorption

8.5.3. Air stripping Adsorption of organic compounds on a solid adsorbent is

a widely used technique for treating hazardous organic
Air stripping is a liquid-to-gas transfer process whereby com- pounds in water. The water is forced through tanks
pol- luted water is sprayed into the air, allowing volatile
organics to escape. It is also widely used for contain- ing activated carbon and contaminants are
contaminated soil con- taining VOCs. The basic principle retained by weak intermolecular forces. Many organic
of air-stripping technology is the process of mass transfer and some inorganic compounds are efficiently removed,
of diffusion and volatilization. The natural tendency of including chlorinated hydrocarbons, organic phosphorus,
any organic compound to volatilize from the aqueous to
carbonate-based pesti- cides, polychlorinated biphenyls
gaseous phase is given by Henry’s Law, which gives the
partial pressure pA produced by a com- pound of (PCB) and trace metals. Carbon, derived from wood,
concentration cA dissolved in a liquid solvent. coal, or other carbonaceous raw materials, is the most
commonly used adsorbent. Other adsorbents include
silica gel, alumina and zeolite. For carbon
pA  HcA (8.28) adsorption, the organic compounds are transferred to a
144 Solar Detoxification

carbon canister that must be later transported for disposal likely to be affected by different contaminants,
or regeneration. concentration and plant size. These studies indicate that
Three types of carbon adsorbents are common: activated cost does not increase linearly with concentration, but
granules, activated powders and fibres. Granular activated much more slowly, because as concentration increases
carbon (GAC) is currently the most common type of carbon contaminant loads of car- bon also increases. Therefore,
adsorbent because of the significant surface area provided GAC processing costs are lower in large plants as a
by the granules. Powdered activated adsorbents are result of the economies of scale.
generally cheaper. They are of lower quality than granular
activated carbon and they cannot be regenerated. They are 8.5.5. Membrane technology
used in packed columns with high-pressure drops. Powder
coatings are used exclusively in batch operations. Carbon
Membrane technology refers to the use of a semi-
fibres are used in a honeycomb structure to maximize
permeable membrane to separate contaminants from a
surface area with adsorption directly on the fibre’s surface.
wastewater stream; this separation technique is often
Carbon regeneration and the recovery and disposal used to purify materials or to pre-concentrate waste
of contaminants are the primary factors in the operating streams prior to treat- ment, so it is not a real treatment
cost. Periodical regeneration in which the carbon is technology. A waste stream (typically aqueous) is passed
cleaned and reactivated, restoring its capacity for across a porous membrane, where diffusion of specific
adsorption and further use, is necessary. A thermal species takes place, separating and concentrating species
regeneration process fuelled by electricity, natural gas or of interest. The membrane is a composite of a polyamide
oil is the most commonly used. The adsorbed materials barrier on a polysulfone support, which normally rejects
are pyrolyzed, forced off, and finally oxidized at over organic molecules with a high molec- ular weight.
90 °C.
The main membrane processes are reverse osmosis,
As with air stripping, the main objection to adsorption ultra- filtration and microfiltration, all of which are
by GACs is that they only displace contamination to a pressurized. Ionic exchange by resins can also be
large sor- bent phase, and a regeneration or disposal considered a membrane process. Nevertheless, the most
problem remains as the contaminants are not really common process is reverse osmosis, which has been used
treated. The use of active carbon adsorption is also for years to treat and purify drinking water. Polluted
feasible only as long as current leg- islation allows water is forced through semi- permeable membranes
storage of this type of waste. In the European Union, (about 150 micrometers thick) under pressures as high as
regulations on waste generation are becoming stricter. 100 bars, separating contaminants from the clean water.
The common practice of using carbon only once is Membranes (normally ceramic or polymeric) are
expensive and polluting because the used carbon is manufactured in different configurations such as plate-and-
usually destroyed by open burning. Carbon regeneration frame, spiral-wound, tubular, capillary tube and hollow
by indirect firing is feasible only for about four cycles, fibre.
after which the carbon must be destroyed. Furthermore,
Many types of organic compounds can be
each time it is regenerated, only about 50 per cent of the
separated by membranes, including salts and
carbon is reused. The remainder is too fine and restricts
chlorinated and organophosphoric pesticides. Reverse
the flow in filtration units. Often, the regenerated carbon
osmosis requires peri- odic cleaning and rejuvenation of
also loses some of its activity and is not as efficient as
filters. Residues that con- tain high concentrations of
when new. Finally, increas- ing transportation and
salts, heavy metals and toxins must be disposed of
carbon regeneration costs are reduc- ing its cost
either in landfills or by incineration. Additionally,
effectiveness. Consequently, GAC processes are
reverse osmosis is an especially intensive energy
expensive to operate and generate undesirable
consumer, because of the high hydraulic pressures
secondary waste gases, and changing environmental
needed to offset osmotic pressure, which require
regulations indi- cate that open burning of these
substantial pumping power. Depending on membrane
materials may be prohibited in the near future.
porosity and the nature of water contaminants, 2.5 to 5
GAC is a mature technology with small initial kWh are needed per cubic meter of water treated. As a
investment and moderate operating costs, so significant result, the major draw- backs of the membrane
cost reduction is not anticipated. Sensitivity studies have technology are its relatively high cost for large-scale
shown how costs are applications and the fact, again, that it is not a real
treatment process but only a separation technology.
Economic Assessment 145

8.5.6. Wet oxidation with UV oxidation, similar to the synergistic effect that
can be observed between UV radiation and an oxidizer
Wet oxidation, by which organic and inorganic such as hydrogen peroxide in wastewater treatment. This
compounds present in water (or any other liquid) are treatment process can be used selectively and reportedly
oxidized with oxy- gen or air in the presence of water, acts on chlo- rinated hydrocarbons faster than on other
has been commercially available since the 1970s organic com- pounds. Halogenated organic compounds
(Zimpro Passavant Environmental Systems, Inc.). This are oxidized to simpler organic forms and, in some cases,
process is normally used with water con- taining 1 per to carbon diox- ide, water and innocuous salts. The
cent to 30 per cent hazardous organic matter (that is, 10 combined UV-ozone system has also been observed to
to 300 g L–1) and treatment conditions vary from 180 °C precipitate heavy metals, such as oxides or metals,
at 20 bars to 400 °C at 280 bars. During the process, although UV oxidation is not usu- ally used to remove
water favours the dissolution of oxygen and heat transfer metals. The concentration of inorganic chemicals in
to the compounds to be treated. wastewater must be low so as not to absorb or shield the
When treatment conditions are below the critical UV rays. Investment costs of ozone technology are
point of water, 374 °C and 225 bars, reactions take place medium, its operating cost being in the range of €5 to
in the water and the process is denominated ‘subcritical wet €15/m3.
oxidation’ or ‘low pressure wet oxidation’; otherwise, it is
called ‘supercrit- ical wet oxidation’. Low-pressure wet 8.5.8. Advanced oxidation processes
oxidation achieves 99 per cent efficiency in the
destruction of contaminants with residence times of In addition to all the conventional technologies indicated
around one hour; supercritical wet oxida- tion can above, a new type of treatment, commonly referred as
achieve 100 per cent efficiency with a residence time of ‘advanced oxidation processes’ (AOPs), is emerging. AOPs
five minutes, but higher temperatures and pressures are are increasingly being considered as alternatives to
required. The main disadvantage of wet oxidation more conventional technologies because they destroy
systems is their very high installation and operating costs. hazardous organic compounds rather than transferring
them to other media. They also have a potentially lower
8.5.7. Ozone oxidation cost and greater effectiveness. AOPs are generally
characterized by their abil- ity to generate hydroxyl
Ozone (O3), used as an oxidizing agent since the radicals, some examples being UV/hydrogen peroxide,
beginning of the twentieth century, is considered an UV/ozone, UV photocatalytic oxida- tion and solar
effective treatment process. Despite some controversial detoxification.
reports, ozone is not known to produce toxic or Solar detoxification has some unique advantages over
mutagenic substances. It can be generated by a high other AOPs, such as:
voltage discharge in the presence of oxygen. After
generation it must immediately be diluted in the water to • The use of sunlight as the photon source, which
be treated, as it is a very unstable gas. This means that means it is a ‘green’ technology.
ozone must be generated at the point of treatment. • The process can be either heterogeneous (TiO 2) or
Ozone treatment has become very common in recent homo- geneous (Photo-Fenton), with the possibility of
years thanks to improvement in ozone generators, and providing chemical pathways and surface interactions
the technology is quickly replacing many traditional not available in other treatment systems.
oxidation processes based on chlorine, hydrogen • A reductive chemical pathway is used to remove
peroxide, perman- ganate and so on. Ozone is used to reducible species, such as heavy metal ions and some
oxidize low concentrations of organic contaminants as organic compounds.
well as inorganic compounds and water-soluble metals, • It can be operated in the liquid or gas phase, in
converting them to their insoluble form to permit contrast to processes using ozone or hydrogen
separation of metallic ions (for instance, Fe++ and peroxide, which are generally applied only in the
Mn++). Another important application of ozone is liquid phase.
Solar detoxification is a modular technology offering the
Ozone treatment is reported to be more effective for advantage of system flexibility, which is important when
haz- ardous organic contaminants when used in
146 Solar Detoxification

treating low to moderate flow rates. However, this community acceptance and absence of undesirable by-
same attribute decreases the possibilities for the products are commonly considered carefully in addition
economies of scale that other pollution control to cost. It is expected that, because it is so attractive in
technologies benefit from. Thus, while the cost of solar these areas, the solar detoxification system should be
detoxification remains relatively con- stant as flow rate able to build a significant market share in the near future.
increases, the normalized costs for other technologies
usually drop.
From the economic perspective, to date, solar Summary
detoxifica- tion cost estimates have been based on very
limited field expe-
rience. However, this experience and projections of capital
and operating costs show that solar photocatalytic
treatment of water could cost from a few euros up to €20 As photocatalytic processes only make sense for hazardous
or €30 per m3. This cost is higher than air stripping, non-biodegradable pollutants and, when feasible, biological
adsorption or membrane technologies, but with the treatment of residual water is always the cheapest, it makes
important advantage that solar detoxification is a real economic sense to combine the two processes. Biologically
treatment process, while the others are only contaminant recalcitrant compounds can be treated with photocatalytic
separation technologies. There is also a pos- sibility of technologies until biodegradability is achieved, transferring
optimizing the cost of a specific treatment process by the water to a conventional biological plant later. Such a
combining two (or more) different technologies; Figure combination, based on the average oxidation state parame-
8.11 shows the combination of solar detoxification and ter, could reduce treatment time and optimize the overall
GAC for removal of pentachlorophenol (PCP). economics. Operation and investment costs of solar
In addition, the investment and operating costs of detoxifi- cation plants are estimated to be lower than
solar detoxification are lower than for technologies such alternative tech- nologies such as incineration or wet
as incin- eration or wet oxidation, while other AOP oxidation, in the same range as ozone, and higher than air
technologies have similar investment and operating stripping, adsorption or membrane separation technologies.
costs. The latter, however, are not real treatment technologies.
It is unclear at present how much of the hazardous- Solar collecting of UV pho- tons is clearly feasible at
waste treatment market could be captured by solar locations where yearly average solar UV irradiation is
–2, which means almost the entire ‘Sun
detoxification technology, but potentially it is very large. higher than 15 W m
Environmental remediation field studies have belt’.
demonstrated that technology selection is clearly not
based on cost alone. Factors such as complete on-site
treatment (which limits owner liability), Bibliography
Photocatalysis National Laboratory Report. SAND91-7005,
Albuquerque, N.M.

BLANCO, J. AND MALATO, S. 1994. Photocatalitic treatment of


hazardous wastewater; cost comparison between solar

GAC and electrical technologies. In: A. Roy and W. Grasse
(eds), INTERNATIONAL Conference on COMPARATIVE Assessments
of SOLAR Power Technologies. Jerusalem, Israel, pp. 217–30.


SOLAR. Almeria, Instituto de Estudios Almerienses and
% Eliminated PCP CIEMAT (eds), SolarPACES (International Energy
FIGURE 8.11. Optimization of solar detoxification and
GAC technologies to PCP degradation.
Economic Assessment 147

AND WRIGHT, J.D. 1991. Design AND FABRICATION of Prototype Technologies for HAZARDOUS WASTE REMEDIATION. New York,
SOLAR RECEIVER/REACTORS for the SOLAR DETOXIFICATION of International Thomson Publishing.
CONTAMINATED WATER. Industrial Solar Technology
Corporation. NREL report, Golden, Colo. PULGARIN, C.; INVERNIZZI, M.; PARRA, S.; SARRIA, V.;
POLANIA, R. AND PÉRINGER, P. 1999. Strategy for the
MALATO, S.; BLANCO, J.; RICHTER, C.; BRAUN, B. AND coupling of photochemical and biological flow
MALDONADO, M.I. 1998. Enhancement of the rate of solar reactors useful in mineralization of biorecalcitrant
photocatalytic mineralization of organic pollutant by industrial pollutants. CATALYSIS TODAY, No. 54, pp.
inorganic oxidizing species. Applied CATALYSIS B: 341–52.
ENVIRONMENTAL, Vol. 17, pp. 347–56.
MEHOS, M.; WILLIAMS, T. AND TURCHI, C.S. 1994. and the Photo-Fenton reaction-comparison of advanced
oxidation processes for wastewater treatment.
Overview Chemosphere, Vol. 28, No. 8, 1994, pp. 1447–54.
of SOLAR DETOXIFICATION Activities in the United STATES,
NREL/TP-471-7262. Golden, Colo., National SCHERTZ, P.; KELLY, D. AND LAMMERT, L. 1992. ANALYSIS of Cost
Renewable Energy Laboratory.
WATER DETOXIFICATION Systems. Golden, Colo., Solar
MARKET Study. Final Report for NREL., Golden, Colo.
148 Solar Detoxification

Self-assessment questions Part B

Part A: True or false? 1. Operating 2,800 hours per year, a 300-m2 solar
detoxification plant treats 5,000 m 3 of water
containing 250 mg L–1 of hazardous contaminants
1. The average oxidation state of any organic yearly. If, once 50 per cent of TOC degradation has
compound is a number between 0 and +4. been attained, the wastewater is biodegradable and it
is transferred to a biological treatment plant, what
2. When contaminated water is treated by oxidation, would be the mass and
its average oxidation state increases. volumetric treatment factors of the photocatalytic facility?
3. If toxicity increases with oxidation, biodegradability
2. Why may toxicity of wastewater increase at
will not be achieved.
the beginning of photocatalytic treatment?
4. If toxicity decreases during photocatalytic
3. What are the main factors in the operating cost of a
treatment, biodegradability increases.
solar detoxification treatment plant?
5. The solar UV irradiation available for a specific
4. What is the fixed charge rate? How can it be calculated?
location is the main factor in determining the cost of a
solar system.
5. What is the standard average efficiency of UV
photon production of mercury fluorescent lamps?
6. The threshold below which electrical systems are
cheaper than solar ones decreases when the total
yearly amount of photons needed increases. 6. A solar TiO2 detoxification facility can treat 3,500
m3 of contaminated water yearly with an average
global solar irradiation of 20 W m 2. How much water
7. The air mass ratio is equal to 1 at solar noon at
could be treated at a different location with an average
any latitude.
solar irradiation of 25 W m2?
8. Standard required performance for thermal
7. How many mercury fluorescent lamps (40 W
oxidation technologies is 99.9999 per cent.
tube) would be equivalent to a 350-m2 solar
collector field working 3,500 hours per year where
9. Catalytic oxidation processes usually require average solar UV irradiation is 26 W m2?
higher temperatures than non-catalytic oxidation
8. What are the main differences between catalytic
and non-catalytic thermal oxidation?
10. Ozone oxidation can normally be applied to
high concentrations of organic contaminants.
9. What is the main environmental objection to
using air-stripping, absorption and membrane
technologies to treat hazardous wastewater?

10. When is an oxidation process defined as supercritical?

Economic Assessment 149

Answers 5. 20 per cent.

6. As solar irradiation is proportional to the

Part A amount of useful photons and this is linearly
dependent on the reaction rate:
1) False. 2) True. 3) False. 4) True. 5) True. 6)
True. V  3,500  4,375
7) False. 8) True. 9) False. 10) False. m3
Part B
7. Using Eq. 8.19:
1. Using Eqs 8.1 and 8.2:
 IUVg 
∆m(TOC) (250  5,000) 2 g N  5.8  1021   HS 

  

 

Tfm    0.74 L
PH   HL 
tS 2,800  hm2
 
 S  5.8  10 21 3,500
∆V 5,000,000 L 1.35  1019  8,760
Tfv    5.9  434 lamps
tS 2,800  300 hm2

The total mass of organic substances is divided in 8. Catalytic oxidation a) does not treat every organic
two, as only 50 per cent is really degraded, but not the contaminant; b) is typically applied to low
total volume, which may be considered to be completely concentrations;
treated. c) does not usually require air for combustion; d) requires
monitoring of temperature and pressure to preserve
2. Due to the presence of not only the initial catalyst activity; e) requires periodic replacement or
toxic compounds but also the many by- reactivation of the catalyst.
products of the degradation reaction at the
beginning of the process. 9. The air-stripping, absorption and membrane
technologies only transfer the contaminants to a different,
3. Costs of personnel, maintenance materials, more easily managed, medium. They cannot be considered
electricity and chemicals. real hazardous-wastewater treatment technologies.

4. FCR is the factor that converts total installed cost (or 10. When it takes place at temperatures and pressures
total plant investment) into annual treatment cost, higher than the critical point of water (225 bars and 374
considering the volume of wastewater to be treated. FCR °C). If a fluid other than water is used, these
is obtained by calculating all the fixed costs (except conditions are defined by the specific critical point of
operation) for the life of the plant. the working fluid.
150 Solar Detoxification


Project Engineering

Aims Objectives
This chapter describes the systematic After completing this chapter, you will
process of feasibility study, preliminary have a basic knowledge of the following
design, final design and construction of a subjects:
solar detoxification plant, using the
knowledge explained in the previous 1. Feasibility studies for solar
chapters. Some important aspects of detoxification applications.
project management are also indicated.
2. Initial pre-design of an
engineering system.

3. Implementation and management of

solar detoxification projects.
152 Solar Detoxification


Symbol Description Units
CIEMAT Centro de investigaciones energéticas, medioambientales y tecnológicas (Spain)
COD Chemical oxygen demand mg L–1
CPC Compound parabolic collector
DG-XII Directorate General XII of European commission
EC European Commission
EUV,n Accumulated UV energy incident on the reactor, since the start of the degradation
process up to tn , per unit of volume k J L–1
GAC Granulated active carbon
Hs Total yearly hours of operation of a solar detoxification system h
LLNL Lawrence Livermore National Laboratory (USA)
mM Mili molar 10–3 moles L–1
PLC Programmable logic controller
ppm Parts per million (equivalent to mg l–1)
PSA Plataforma Solar de Almería
Re Reynolds number dimensionless
S Solar collector field area m2
T Experimentation time min
Tfm Mass treatment factor g h–1 m–2
Tfv Volumetric treatment factor L h–1 m–2
TOC Total organic carbon mg L–1
tR Residence time min/h
tT Total elapsed time (since beginning of photocatalytic process) min
UVG Global ultraviolet solar irradiation WUV m–2
UVG, Average incident radiation on the collector surface for each At interval (At = t – t ) W m–2
n n n–1 UV
VR Illuminated reactor volume L
VT Total volume of hydraulic circuit L
VTOT Total solar detoxification plant volume L

9.1. Feasibility study more properly called a ‘feasibility study’, as any new potential
application of solar detoxification must be tested before
going ahead with the design and implementation of the
Solar detoxification projects, just as any other treatment plant. When complex mixtures of hazardous
engineering project, must follow the logical sequence of contaminants are present, degradation pathways and
pre-design, pre- liminary design and final design with the reaction kinetics could be very different even from those of
final objective of designing and building facilities of other wastewaters of a simi- lar origin. This means that the
feasible construction and satisfactory operation, for design of a specific solar detox- ification system is a
which contractors can confidently make bids. Without complex procedure requiring a prior feasibility study
competent design, implementation and operation, no following the method defined below:
engineering project can succeed.
The pre-design phase must define the scientific and • identify targeted recalcitrant hazardous compounds
engi- neering data required before the final design can • identify any pre-treatment that could enhance photo-
proceed. In solar detoxification projects the pre-design catalytic efficiency
phase might be
Project Engineering 153

engineering-scale pilot plant tests, at a later stage of the

project, it is important to know whether the differences
are a consequence of the reactor and system
configurations or of variations in the wastewater.

9.1.2. Identification of possible


The objective of pre-treatment, if necessary, is to

condition contaminated waters for organic
destruction. This can be achieved by addressing any
constituents of the contaminated water that interfere
with the process. The presence of initial impurities in
the wastewater to be treated could produce a
significant reduction in process efficiency because of:
FIGURE 9.1. Experimental mobile pilot plant for solar
water detoxification (CIEMAT, Spain, 1996). • unexpected chemical reactions due to the presence
of specific compounds
• incomplete destruction of one or several components
• identify the best photocatalytic process (e.g. TiO2/TiO2
• presence of substances that inactivate the catalyst
• presence of substances that reduce the oxidation kinetics
persulfate/TiO2 + other oxidant/Photo-Fenton)
• presence of substances that block UV solar radiation.
• determine optimum process parameters (e.g.
catalyst concentration, additional oxygen) These possible impurities, mainly inorganic ions, could
• assess toxicity and biodegradability to determine the even lead to the non-viability of the degradation process.
best moment to stop the photocatalytic process The only way to avoid this is by designing some form of
• identify possible post-treatment processes pre-treatment to adjust the chemistry of the water to
• determine treatment factors. conditions suitable for the detoxification process. It
should be noted that such pre-treatment increases capital
All this can be done in the laboratory using indoor and system operating costs and may also affect the control
reactors with simulated solar radiation, as described in strategy of the treatment plant.
Section 3.1. Nevertheless, it is always much better to Examples of pre-treatment are the following:
carry out the feasi- bility study using experimental solar
devices, as the results obtained will necessarily be more • Filtering, in the case of turbidity or cloudy water that
reliable when extrapolated to the engineering level. could block incoming solar UV-irradiation.
Section 3.2 describes several exper- imental solar • Acidification, to eliminate substances such as bi-
systems. carbonate.
• Adjustment of pH to reduce the concentration of total
9.1.1. Identification of target recalcitrant dissolved solids by precipitation.
• Addition of hydrogen peroxide to precipitate
hazardous compounds substances such as dissolved iron ions that react to
form a ferric oxide precipitate, which may then be
The first step must be a complete analysis of the
removed by filtering.
wastewater to be treated in order to identify the target
chemicals and their initial and desired final • Cation bed exchange, in the case of impurities such as
concentrations. This is very important to avoid calcium ions that would adversely affect the process.
inadequate data in the following phases of the project. In the end, the specific impurities for each case must be
Also, depending on the specific application, the nature of iden- tified and, depending on their effect on the
the contaminants and/or their concentration may change overall photo- catalytic process, a particular pre-
before the system is actually constructed, as the time treatment should be designed if necessary. Other pre-
between the study and its implementation could take treatments are independent
months or even years. As some differences normally
arise when laboratory experiments are compared with
154 Solar Detoxification

of the presence of impurities. This is the case for oxygen, which came from an ultrafiltration process, leaving no inor-
hydrogen peroxide, persulfate or any other oxidant ganic impurities present; thus the homogeneous process
injected into the process stream before the start of the works quite well. By contrast, heterogeneous degradation
photocatalytic process to act as hole receptors on the does not work because of the high presence of organic matter.
titanium dioxide cata- lyst. In many cases reduction of In other cases, the Photo-Fenton Fe cycle may be
initial pH to between 5 and 6 is also recommended to affected by the presence of impurities that produce the
enhance the initial degradation rate. opposite result. In the case of Photo-Fenton, pH must
also be around
9.1.3. Identification of most suitable 2.5 to avoid the formation of non-soluble iron
photocatalytic process hydroxides. It is clear that this first research step is
absolutely necessary in order to decide on the most
appropriate process for further in-depth study.
With regard to solar photocatalysis, when complex
mixtures of organic contaminants are present, no general
indication can be given at all and each case is completely 9.1.4. Determination of the optimum
different from process parameters
every other. It is therefore recommended that different pho-
tocatalytic processes be tested in the beginning to identify
the one that is the most adequate. Until both This phase seeks the definition of the essential scientific and
heterogeneous TiO2 photocatalysis and homogeneous engineering data required before proceeding with system
Photo-Fenton processes have been extensively analysed, design. This can be done by defining, implementing and
no advantage can be ascribed to any particular conducting an experimental plant to determine the
application. following parameters:
A good example is shown in Figure 9.2, where the • catalyst load
Photo- Fenton photocatalytic process is seen to be highly
efficient in degrading about 6,000 mg L–1 of initial TOC, • pH required
compared to the relative inefficiency of TiO2 • oxygen
heterogeneous photocatalysis. The reason for this is the • additional oxidant load, if any
wastewater, taken from a cat- aphoretic painting process • contaminant concentration, if this can be varied.
at an automobile assembly plant,

Pure wastewater
20,000 100

16,000 80
H2O2 consumption (mg L )
COD and TOC (mg L )

TOC Photo-Fenton
12,000 COD Photo-Fenton 60
TOC Titanium dioxide
COD Titanium dioxide
H2O2 consumption (Photo-Fenton)
8,000 40

4,000 20

0 0
0 100 200 300 400 500 600 700 800
Residence time (min)

FIGURE 9.2. Treating wastewater from the painting section at a car assembly factory (ultrafiltration from cataphoresis
process). (CIEMAT, Spain, 1999).
Project Engineering 155

9.1.5. Post-treatment process


Post-treatment processes adjust the water chemistry to

conditions suitable for discharge. In the case of industrial
wastewater that is to undergo further, biological
treatment, post-treatment may be reduced to:
• pH adjustment
• catalyst separation/precipitation and recovery.
Depending on the nature of the wastewater treated and
its later use, and whether solar detoxification is used for
water purification (as in groundwater
decontamination), the post- treatment system could
then require additional processes to meet the specific
local regulatory requirements before discharging
water. As an example, possible treatment processes
• carbon dioxide removal
FIGURE 9.3. Catalyst mixing system for the LLNL water
treatment system. (Courtesy of the US National Renewable • elimination of possible residual hydrogen
Energy Laboratory). peroxide (if used in the photocatalytic process)
• using a cation-exchange bed to remove specific
existing inorganic compounds
It will also assess the photocatalytic degradation process • using GAC filters to remove trace quantities of
over time: organic materials.
• total organic carbon of the wastewater Another possible factor that could require post-
• individual concentration of key hazardous compounds treatment is the water temperature. Temperature has a
identified during the first step (Section 9.1.1) negligible influence on the photocatalytic process and
• toxicity the use of non- concentrating solar technology
• average oxidation state. considerably reduces the increase in water temperature
during the treatment process. However, the usual
It might also be strongly advisable to individually test the recirculation or batch system design may cause the water
pho- tocatalytic degradation of the key recalcitrant to become too hot for the process in summer, volatilizing
contaminants identified in order to check whether there compounds with low boiling points, or too hot for the
are significant dif- ferences from those obtained with the system, as in the case of a cation exchanger at the inlet of a
real wastewater prob- lem. Reaction rates are normally post-treatment system; as this limits the inlet tempera-
higher when individual compounds are tested, but if ture, a water-to-air heat exchanger must be installed
there are significant differences, it might be worthwhile before the post-treatment system, or inserted in the water
repeating the pre-treatment process to try and identify circuit in the solar collector field.
possible additional pre-treatments that could speed up the
detoxification process. Obviously, potential gains in
overall reaction rate must always be weighed against
9.1.6. Determination of treatment factors
the cost of adding additional processes and reagents.
The disappearance of hazardous contaminants must After all the above factors have been analysed and the
nor- mally be parallel to the reduction of toxicity and the steps indicated have been systematically followed, a
increase in biodegradability. These and the evolution of feasibility study to calculate the treatment factors for the
the average oxidation state would determine the optimum best system configuration proposed may be carried out.
point at which to stop the oxidative process and transfer These treatment factors – mass treatment factor (Tfm) and
the wastewater to a conventional biological treatment volumetric
plant (see Section 8.1).
156 Solar Detoxification

treatment factor (Tfv) – have already been defined in Section were consumed in the region, producing around 1.5
8.1 (Eqs. 8.1 and 8.2) and must be associated with the million empty bottles and containers with an average
average solar UV irradiation recorded during the tests volume of
performed. These values will permit the necessary solar 1.9 L, 99 per cent of which are plastic. A small amount
field size to be established. of pesticide residue always remains in the used bottles
All the information assembled during the feasibility and they are therefore hazardous waste, which cannot
study must be documented in a report, which may also be handled in the same way as conventional garbage. To
include any additional relevant data collected and date, there has been no way to dispose of these bottles
recommendations for future activities. and recover them for re-use; most of them are thrown
away with the rest of the regular agricultural waste or
just dumped anywhere. In order to solve this
9.2. Feasibility study example environmental problem, a process was designed to
recycle the high-quality plastic in these bottles into a
One example of a feasibility study could be the treatment valuable raw material. The recycling process shreds the
of pesticide-contaminated water from the recycling of plastic, which is then washed, leaving a relatively small
agro- chemical plastic containers. amount of water contaminated by a few hundred mg L–1
of total organic carbon content of persistent toxic
9.2.1. Background compounds. As the water must be re-used, those con-
taminants have to be treated and eliminated. Solar photo-
catalytic detoxification was proposed to treat the
In Mediterranean countries, intensive agriculture in wastewater and a feasibility study was carried out.
green- houses has become a very important part of the The experimental research was performed at the
economy, especially in recent years. In southeastern PSA Solar Detoxification facility. Definition of the work
Spain, the province of Almería alone has about 25,000 necessary
hectares of such greenhouses, and intensive agriculture is
currently the most important economic activity in the Nematocides (2,314)
area. However, important environmental problems
associated with it have also arisen,
Molluskicides (23)
such as the extensive and intensive use of pesticides Insecticides (827) Herbicides (134)
(require- ments are about 200 times greater than for Phytoregulators (698)
conventional agri- cultural methods). Acaricides (52)
Others (55)
One particular problem is caused by the enormous Fungicides (1,115)

ber of empty pesticide bottles disposed of each year. FIGURE 9.4. Tonnes of pesticides used in the Almeria region
According to 1995 data, about 5,200 tonnes of pesticides (1995 data) (Source: AEPLA, Spain).
TABLE 9.1. Selected pesticides for the feasibility study assessment, CIEMAT (Spain), 1996.
Commercial name Producer Active ingredient Formula
Rufast Rhône-Poulenc Acrinatrin C26H21F6NO5
Vertimec Merck Abamectin C48H72O14
Thiodan AgrEvo Endosulfan α-β C9H6Cl6O3S
Dicorzol AgrEvo Formetanate C11H16ClN3O2
Confidor Bayer Imidacloprid C9H10ClN5O2
Match Ciba-Geigy Lufenuron C17H8Cl2F8N2O3
Tamaron 50 Bayer Methamidofos C2H8NO2PS
Vydate DuPont Oxamyl C7H13N3O3S
Scala AgrEvo Pyrimethanil C12H13N3
Previcur AgrEvo Propamocarb C9H20N2O2
Project Engineering 157

started with research on the pesticide market in Almería good solar irradiation. Residence times (tR) were calculated
(Figure 9.4), which provided the qualitative and as:
quantitative V
distribution of pesticides consumed in the region. tR  tT (9.1)
From these, ten pesticides (with all the main chemical V T
present) were selected as representative of those most
Where tT is the total time elapsed since the beginning of
used by the greenhouses in order to carry out a complete
solar degradation test program. Table 9.1 shows the list the experiment, VR the illuminated reactor volume and
of pesti- cides used. VT the total volume of the hydraulic loop.

9.2.2. Experimentation: a) 80
Two–axis parabolic trough system
TiO2-persulfate tests –2
TiO2+ S 2O8= (10 M) TEST

TOC concentration (mg L )

The test programme included individual and combined 60 Active compound: Abamectin
tests of selected compounds under different conditions.
In order to simplify the study, and the number of tests to
be carried out, the following was assumed:
40 4
• Although pesticides, which are usually sprayed in TOC evolution
con- centrations between 200 ppm and 3,000 ppm pH evolution
depending on the product, are diluted in water before

20 2
their applica- tion, pesticides in wastewater from the
washing process are always more diluted than they
are in normal use.
• The ten pesticides selected are considered 0 0
representative of the entire local market (several 0.0 0.5 1.0 1.5 2.0
hundreds of such com- pounds are used in the area). Residence time (h)
• The same numbers of empty plastic containers are
gener- ated by each of the ten selected products. b) 125 125
CPC system. TOC evolution
• The same amount of residue remains in all the =
TiO2+ S 2O8 (10
TOC concentration (mg L )

containers. 100 100

Under these hypotheses, a mixture of the ten selected

pesticides, with a total TOC content of 100 mg L–1, 75 75
was considered representative of the wastewater to be Acrinathrin
found after washing. Wastewater was artificially Lufenuron
generated by mixing identical concentrations of the 50 50
selected pesti- cides and tested in solar degradation
experiments in both Helioman (Figure 7.5) and CPC
(Figure 7.6) systems. 25 25
Figure 9.5 a) shows mineralization of total organic
carbon from the insecticide abamectin in the parabolic
0 0
trough system. The TOC is found to practically 0 1 2 3
disappear after one Residence time (h)
hour of exposure to sunlight (with good weather FIGURE 9.5. a) Solar mineralization of TOC from the
conditions). Photocatalytic degradation tests performed insecticide abamectin; test performed in a parabolic
in a CPC system show a very similar pattern. Figure 9.5 trough system; average direct UV light: 38.1 watts m–2. b)
b) shows degradation of TOC from the insecticides Samples of photocatalytic degradation in CPC system;
acrinathrin, methamidophos and lufenuron (three average global UV solar radiation was: 26.2 watts m–2
different experiments on sunny days). Each one was (acrinathrin test); 33.6 watts m–2 (methamidophos) and
individually tested and similar degradation was 33.7 watts m–2 (lufenuron test). TiO2 (Degussa P25): 200
observed (see Table 9.1): degradation of about 100 ppm mg L–1. (CIEMAT, Spain, 1996).
of TOC in a residence time of from one to two hours,
158 Solar Detoxification

120 1.5
Mixture of ten Persulfate conc.: 10 mM (constant)
selected pesticides
TOC Persulfate conc.: 8.5 mM (constant)

EC50 (%) P. phosphoreum

Total organic carbon (mg L )

TOC concentration (mg L )

Persulfate conc.: 5 mM (constant)

Persulfate conc.: 20 mM (initial)
80 1.0


40 0.5


0 0.0
0.0 0.5 1.0 1.5 2.0
Residence time (min)
Residence time (h)
FIGURE 9.7. TOC mineralization of a mixture of ten selected
FIGURE 9.6. TOC mineralization of a mixture of ten selected
pesticides (CPCs). TiO2 (Degussa P25) concentration: 200
pesticides (parabolic troughs); direct UV sunlight: 36.3
mg L–1, slurry. (CIEMAT, Spain, 1998).
watts/m2; TiO2 (Degussa P25) concentration: 200 mg L–1;
persulfate addition: 0.01 molar; treated volume: 250 l.
EC50 toxicity measured by Microtox®. (CIEMAT,
Spain, 1996). All titanium dioxide tests were conducted in 200-mg
slurry concentrations with 5 to 10 mM persulfate, L–1
which has been found to be the optimum concentration in
However, the experiments of most practical interest PSA experiments. Na S O was added at the start of the
2 2 8
are those in which the mixture of all ten of the selected
com- pounds was used. These were performed in the tests and at regular intervals to assure continuous
CPC system presence through- out tests by measuring the S O2–
as well as in parabolic troughs, assuming the four hypotheses consumed. This addition 2 8
previously considered. For these experiments, 10 ppm of persulfate does not imply any environmental problem
of each pesticide was added to 250 L of distilled since it only produces small concentrations of sulfates
water and homogenized for thirty minutes. Samples were that increase the salinity of the water treated (maximum –1
taken period- ically and the total organic carbon in the sulfate permissible in drinking water is 250 mg L ; there
suspension was analysed; pesticide concentration was is no limit for wastewater). In the experiments indicated
monitored by liquid chromatography. Figure 9.6 in Figure 9.7, 80 per cent of the TOC was removed in a
shows some of the results obtained, which included period of two to four hours and 90 per cent in a period of
toxicity measurements throughout the experiment. As the two and a half to five hours.
complete mineralization of all organ- ics present in the
water was not intended, it should be observed that 9.2.3. Photo-Fenton tests
no highly toxic compound was generated. Microtox®
(a widely accepted toxicity measurement system) is
usually expressed in terms of the EC50, which is the Photo-Fenton experiments were also conducted to
effec- tive concentration causing a 50 per cent deter- mine the effect of the initial TOC and iron
reduction in light from a luminescent marine concentration on the degradation rate. Initial
bacterium (PHOTOBACTERIUM phosporeum), indicating that concentrations were 100 to 500 ppm of TOC and 0.25
50 per cent of the bacteria pres- ent have been killed. mM to 2 mM of iron (Fe) concen- tration. After initial
homogenization for thirty minutes, the pH was adjusted
In the case displayed in Figure 9.6, a TOC reduction
to 2.5 or 2.8 by addition of sulfuric acid. Ferrous sulfate
of 90 per cent is followed by a marked toxicity reduction
was added immediately afterwards and H 2O2 was added
(to obtain the EC50, a 40 per cent concentration of the
in portions of 10 to 20 per cent of the stoichiometry
initial sample is required, while the final concentration
(from previous COD measurements) until the experiment
required is 135 per cent).
was completed. Additional iron concentration
Project Engineering 159

increases the degradation until a maximum rate is One of the difficulties in both the Photo-Fenton and
reached. This occurred when iron compounds absorbed TiO2-persulfate experiments that gave rise to possible
all the irradiated photons. Figure 9.8 shows the error was the measurement of the initial amount of
degradation curves for different experiments with pesticide, because the highly viscous liquids adhered
approximately 100 ppm of pesticides. strongly to the glass-measuring cylinder. Furthermore,
The field experiences showed that an increase in the the amount of dis- tilled water to be added could not be
concentration of iron did not improve the degradation rate calculated accurately. Both of these factors caused
as much as had been expected from previous laboratory fluctuation in the concentration of total TOC. Since not
experi- ments. The poor performance of the 0.5-mM iron all the pesticide could be dissolved at once in the distilled
test may have been caused by the high initial TOC and a water, TOC increased during the first sixty minutes of
slight devia- tion in pH adjustment. Nevertheless, 80 per the reaction.
cent of the initial TOC was removed in less than three
hours. With an excess of H2O2, degradation can be
improved up to 90 per cent. 9.2.4. Conclusions and treatment factors

The total volume treated was the same for all the
experiments (250 L), and the CPC collector field (2 ×
60° semi-aperture angle, 1 Sun concentration and total
irradiated volume of 110 L) was formed by three
Total organic carbon (mg L )

stationary modules with an overall aperture area of 9

m2. Collectors were orientated to the south and tilted 37°,
which is similar to the local latitude, to obtain the
maximum yearly efficiency. With these data and the
results shown in figures 9.7 and 9.8, the mass treat- ment
factor, Tfm (Eq. 8.1), and volumetric treatment factor, Tfv
(Eq. 8.2) for both the TiO2-persulfate and Photo-
Fenton processes can be obtained. Table 9.2 shows
the results.
Treatment factors are observed to be significantly
higher with Photo-Fenton, and the last 10 per cent of
Residence time (min)
degradation cannot be achieved with this process. This
FIGURE 9.8. Pesticide degradation by Photo-Fenton fact is an important issue in plant design at the
process: comparison of different iron concentrations for preliminary design stage.
100 ppm of pesticides in wastewater (CIEMAT, Spain,

TABLE 9.2. Degradation of pesticide mixture: mass treatment factor (Tfm) and volumetric treatment factor (Tfv)
obtained for TiO2-persulfate and Photo-Fenton processes, CIEMAT (Spain), 1998.

TiO2-persulfate (90% Photo-Fenton (80%

degradation) Tfv T
fm Tfv
Test 1 0.81 6.44 2.28 18.52
Test 2 0.59 4.69 3.88 31.45
Test 3 0.86 7.41 3.11 26.88
Test 4 0.89 7.65 2.43 22.83
Test 5 0.60 6.06 3.07 28.74
Test 6 1.18 13.89 — —
Test 7 0.93 11.26 — —
AVERAGE 0.84 8.20 2.95 25.68
160 Solar Detoxification

9.3. Preliminary design parameter for calculating the collector field area,
however, because the solar radiation, which obviously is
an essential parameter, is not included in the calculation
The goal of preliminary design is to develop specific
of the treatment factors, as Table 9.2 above (obtained
param- eters for the siting, layout, and dimensions of the
from Figures 9.7 and
solar detoxification facility, on the basis of the feasibility
study. The preliminary design is complete when it 9.8 in the previous example) shows.
provides all of the necessary basic elements for the final Treatment factors, Tfm and Tfv, are based on the use of
design and con- struction, often considered as equivalent res- idence time, tR, or the time the water has been
to 30 to 40 per cent of the complete final design. No exposed to radiation, as the unit of calculation for
additional data should have to be developed after the analysis. This could lead to erroneous conclusions when
completion of the preliminary design because this should there are important differences in the radiation incident
include sufficient engineering details to proceed rapidly in the reactor because of clouds or time of day. One way
to the final design. This step typically formulates a to avoid this problem is to use a relationship between
schematic process diagram and a site layout plan. experimental time, plant volume, collector surface and
The final figures for volume of water to be treated, the radiant power density, UVG, as measured by a UV
hazardous contaminant and its concentration, solar radiometer. The amount of energy collected by the
necessary collector surface and pre- and post-treatment reactor (per unit of volume) from the start of the
processes, must be assessed at this point. A typical list of experiment until each sample is collected may then be
the data required for found by Equation 9.2:
the preliminary design might be the EUV,n = UV,n +
tn VTOT ;
following: E (9