Radiationt Chemistry

L I V R E S
FRO M B A S I C S TO
A PPL IC AT I ON S I N MAT E R I A L
A ND L I F E S C I E N C E S
%$)4 %$"9
Ml a n ie S POTHEIM -M AURIZOT,
Meh r a n M O STA FAVI,
Th ier r y D OUK I,
J a cqu el in e BELLO NI
L I V R E S
#/,,%#4)/.$)2%#4%$"9 Paul RIGNY
FRO M BA S I C S TO
A P P L IC AT IO N S IN MAT E R I A L
A ND L IFE SC I E N C E S
%$)4 %$"9
Ml a n ie SPOTHE IM -M AUR IZOT,
Meh r a n MOSTA FAVI,
Th ier r y D O UK I,
J a cqu el in e BELLO NI
AVENUEDU(OGGAR
0ARCD§ACTIVITmDE#OURTABOEUF"0
,ES5LIS#EDEX!&RANCE
Couverture, maquette intérieure et mise en page : Thierry Gourdin
Imprimé en France
ISBN : 978-2-7598-0024-7
Tous droits de traduction, d’adaptation et de reproduction par tous procédés,

réservés pour tous pays. La loi du 11 mars 1957 n’autorisant, aux termes des alinéas 2
et 3 de l’article 41, d’une part, que les « copies ou reproductions strictement
réservées à l’usage privé du copiste et non destinées à une utilisation collective »,
et d’autre part, que les analyses et les courtes citations dans un but d’exemple et
d’illustration, « toute représentation intégrale, ou partielle, faite sans le consentement
de l’auteur ou de ses ayants droit ou ayants cause est illicite » (alinéa 1er de l’article 40). Cette
représentation ou reproduction, par quelque procédé que ce soit, constituerait donc une
contrefaçon sanctionnée par les articles 425 et suivants du code pénal.
© EDP Sciences 2008

Contents
Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V I I
List of authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I X
Part I / Primary radiation-induced phenomena . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Chapter 1 An overview of the radiation chemistry of liquids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
George V. BUXTON
Chapter 2 Tools for radiolysis studies .................................................................. 17

James F. WISHART
Chapter 3 The solvated electron : a singular chemical species ................................. 35

Mehran MOSTAFAVI and Isabelle LAMPRE
Chapter 4 Water radiolysis under extreme conditions.

Application to the nuclear industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
Gérard BALDACCHINO and Bernard HICKEL
ÄIÄ
2!$)!4)/.#(%-)3429
Part II / Radiation chemistry mechanisms and applications . . . . . . . . . . . . . . . . . . . . . . . . 65
Chapter 5 Molecular formation in the interstellar medium ..................................... 67

Nigel J. MASON, Anita DAWES and Philip HOLTOM
Chapter 6 Water remediation by the electron beam treatment ............................... 79

Salvatore S. EMMI and Erzsébet TAKÁCS
Chapter 7 Metal clusters and nanomaterials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97

Jacqueline BELLONI and Hynd REMITA
Chapter 8 Water radiolysis in cement-based materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117

Pascal BOUNIOL
Chapter 9 Obtaining high performance polymeric materials by irradiation ............ 131

Xavier COQUERET
Chapter 10 Radiosterilization of drugs ............................................................... 151

Bernard TILQUIN
Chapter 11 Food irradiation: wholesomeness and treatment control . . . . . . . . . . . . . . . . . . . . . . 165

Jacques RAFFI et Jacky KISTER
III / Radiation damage to biomolecules,

radioprotection and radiotherapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
Chapter 12 Radiation-induced damage to DNA : from model compounds to cell . . . . . . 177

Thierry DOUKI and Jean CADET
Chapter 13 Mechanisms of direct radiation damage to DNA ................................. 191

Michael D. SEVILLA and William A. BERNHARD
Chapter 14 Charge motion in DNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203

Yuri A. BERLIN and Laurens D. A. SIEBBELES
Chapter 15 Genome maintenance mechanisms in response

to radiation-induced DNA damage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
Evelyne SAGE and Bertrand CASTAING
Ä>>Ä
#ONTENTS
Chapter 16 Pulse radiolysis studies of free radical processes
in peptides and proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
Chantal HOUÉE-LEVIN and Krzysztof BOBROWSKI
Chapter 17 Radiation-induced damage of membrane lipids and lipoproteins ......... 249

Monique GARDES-ALBERT
Chapter 18 Predicting radiation damage distribution in biomolecules .................. 265

Marie DAVIDKOVA and Melanie SPOTHEIM-MAURIZOT
Chapter 19 Chemical protection against ionizing radiation .................................. 277

Caroline PROUILLAC, Christine AMOURETTE and Ghassoub RIMA
Chapter 20 Advances in radiotherapy : new principles ........................................ 291

Nicolas FORAY and Jacques BALOSSO
Index ............................................................................................................... 301
ÄIIIÄ
Foreword
“L’Actualité Chimique” is a monthly scientific journal meant to convey information

on progress in the chemical sciences to a public endowed with a certain ability to master
scientific matters. The articles were written by scientists who took time out of their laboratories
to explain their studies and their knowledge with a pedagogy and an appeal suitable for
non-specialists. Mostly written in French, it creates a bond in the chemical community in
French-speaking countries where it is very much appreciated by scientists, teachers and
engineers.
However, the scope of the journal implies some limits that we want to erase
with this new collection “L’Actualité Chimique – Livres”, which will be complementary
in two directions: the first one is illustrated by the present book, as it addresses readers
more specialized than the journal usually does, and being written in English, it has the
ambition of attracting attention worldwide on a field of chemistry where recent progress
is noted. The second direction that will be found in the new collection is, in contrast, that
of disseminating the progress of chemistry for the benefit of a large, French-speaking, not
necessarily professional public. The first trend will produce books that we will find in many
laboratories; books produced according to the second trend will instead be largely found in
public libraries, in schools or even in the homes of scientifically curious people.
This first volume of “L’Actualité Chimique – Livres” is of the first kind and devotes itself
to Radiation Chemistry – From basic science to applications in biology and material science.
ÄVÄ
2!$)!4)/.#(%-)3429
This field of research is undergoing a true and fruitful rejuvenation. Already active in the mid
- 20th century, the development of this scientific field had been somewhat slowed down
by the high cost of short-pulse particle accelerators and specialized construction. Recent
progress in instrumentation e.g. : the shaping of picosecond radiation pulses, faster time-
resolved detection techniques, and powerful molecular structure determination techniques,
has coincided to enhance the capacity of radiation chemistry sufficiently to warrant new
investments and the start of new laboratories. Radiation chemistry today is responsible for
major progress in the understanding of the elementary chemical event and powerful enough
to unravel the mechanisms of the damage induced by radiation to living matter (a question
of great concern in the public) or the transformations induced in irradiated materials.
These aspects are developed in the book by international-level specialists and will
be of interest to scientists who are starting in the field, to more experienced ones, and also
to students and teachers; it will also be very useful to many professionals who apply or deal
with radiation in their activities to improve materials or to avoid radiation-induced damage
to them.
Paul RIGNY
Chief Editor of “L’Actualité Chimique”
March 2008
ÄK>Ä
Preface
Radiation chemistry deals with the chemical reactions resulting from the interaction
of high-energy photons or particles with matter. Such radiation possesses energy high
enough to induce ionisation of the components of the material and the breaking and
building of chemical bonds.
In the present volume, our purpose is to familiarise the larger communities of

students and chemists in other specialities with this relatively little-known but essential
domain of chemistry. The covered topics range from the basics (primary phenomena and
mechanisms) to the broad fields of their application. Understanding radiation-induced
chemical and biochemical reactions is essential for improving existing processes and
developing new ones.
Therefore we have called upon internationally recognized experts who kindly

agreed to contribute to this volume with clear, instructive and pedagogically presented
chapters abundantly illustrated with attractive colour figures.
The first chapters of Part I deal with primary radiolytic phenomena and describe
recent developments at the facilities used to create radiation-induced species, as well as the
most advanced methods for their detection and study. The mechanisms of radiation-matter
interactions and their consequences for the physical chemistry of liquids and solutions are
discussed.
ÄVIIÄ
2!$)!4)/.#(%-)3429
Part II describes specific mechanisms and key processes in space and nuclear chemistry,
as well as in material sciences and pharmaceutical and food chemistry. The high energy of
ionizing radiation offers the specific advantage of easy and homogeneous sample penetration.
Therefore, by targeting chemical bonds at room temperature via cost-competitive, chemical
additive-free processes, ionizing radiation can be used for many interesting purposes. For
example, thanks to the understanding of radiation-induced nucleation/growth processes, the
final size and properties of metal nanoclusters can be controlled for applications in catalysis,
electronics, and photography. High-performance polymeric materials, obtained using the
cleavage or the formation of chemical bonds by irradiation, have a multitude of uses in
everyday life. Remediation of waste-water requires the destruction of toxic chemicals, which
is efficiently accomplished by irradiation. The use of ionizing radiation for food treatment and
the sterilization of pharmaceuticals and medical devices operate via the efficient destruction
of micro-organisms, but they require systematic confirmation of the absence of any toxic
molecules that could be produced during irradiation.
The search for new means of improving the success of cancer radiotherapy motivates
an increasing interest in the chemical mechanisms underlying radiobiology. Part III of the
volume is devoted to this very active research domain, and in particular, to studies of the
damage induced by ionizing radiation in biomolecules (DNA, proteins, lipids). Answers
are given as to what are the mechanisms of the reactions in DNA and other biomolecules
following the initial ionization and excitation, how they can be simulated by computational
models, how radiation-induced lesions are repaired or prevented, and finally how this
improved knowledge is used to specifically eradicate tumours (cancer radiotherapy).
With no pretence of exhaustively covering in detail all the topics of radiation

chemistry, this volume will hopefully fulfil the expectation of the reader to learn about a
domain that we consider a most exciting and promising area of chemistry.
We cannot end this preface without addressing our thanks to Yann Gauduel and
Paul Rigny, respectively former and present Chief Editors of “ L’Actualité Chimique”, who
solicited and accompanied us in the realisation of this work. All the other members of the
editorial board of the journal and of EDP Sciences are equally warmly thanked.
Mélanie SPOTHEIM-MAURIZOT,
Mehran MOSTAFAVI,
Thierry DOUKI,
Jacqueline BELLONI
March 2008
ÄK>>>Ä
List of Authors
AMOURETTE Christine
Service de Santé des Armées / Centre de Recherches
24, Av. des Maquis du Grésivaudan - 38702 La Tronche / FRANCE camourette@crssa.net
BALDACCHINO Gérard
Commissariat à l’Énergie Atomique / Laboratoire de Radiolyse
Bât. 546 CEA/Saclay - 91191 Gif-sur-Yvette / FRANCE gerard.baldacchino@cea.fr
BALOSSO Jacques
CHU A. Michallon / Service de Cancérologie-Radiothérapie
BP 217 - 38043 Grenoble cedex 9 / FRANCE jbalosso@chu-grenoble.fr
BELLONI Jacqueline
CNRS-Université Paris-Sud / Laboratoire de Chimie Physique-ELYSE
Bât. 349 Université Paris-Sud - 91405 Orsay / FRANCE jacqueline.belloni@lcp.u-psud.fr
BERLIN Yuri
Northwestern University / Department of Chemistry
2145 Sheridan Road - Evanston, IL 60208-3113 / USA berlin@chem.northwestern.edu
ÄIXÄ
2!$)!4)/.#(%-)3429
BERNHARD William
University of Rochester / Department of Biochemistry and Biophysics
575 Elmwood avenue, Box 712
Rochester, NY 14642 / USA william_bernhard@urmc.rochester.edu
BOBROWSKI Krysztof
Institute of Nuclear Chemistry and Technology
Dept of Radiation Chemistry and Technology
Dorodna 16, 03-195 Warsaw / POLAND kris@ichtj.waw.pl
BOUNIOL Pascal
Commissariat à l’Énergie Atomique Saclay / Laboratoire des Bétons
Bât. 158 - 91191 Gif-sur-Yvette / FRANCE pascal.bouniol@cea.fr
BUXTON Georges
1A Hollin Crescent - Leeds LS16 5ND / UNITED KINGDOM george.buxton@talktalk.net
CADET Jean
CEA Grenoble / DRFMC/SCIB
17 rue des Martyrs - 38054 Grenoble cedex 9 / FRANCE jcadet@cea.fr
CASTAING Bertrand
CNRS / Centre de Biophysique Moléculaire
Rue Charles Sadron - 45071 Orléans cedex 2 / FRANCE castaing@cnrs-orleans.fr
COQUERET Xavier
Université de Reims Champagne-Ardenne / Réactions Sélectives et Applications
Europol’Agro - 51687 Reims cedex 2 / FRANCE xavier.coqueret@univ-reims.fr
DAVIDKOVA Maria
Nuclear Physics Intitute / Dept of Radiation Dosimetry
Na Truhlarce 39/64 - 18086 Praha 8 / CZECH REPUBLIC davidkova@ujf.cas.cz
DAWES Anita
The Open University / Department of Physics and Astronomy
Walton Hall - Milton Keynes MK7 6AA / UNITED KINGDOM a.dawes@open.ac.uk
ÄMÄ
,ISTOFAUTHORS
DOUKI Thierry
CEA Grenoble / DRFMC/SCIB
17 rue des Martyrs - 38054 Grenoble cedex 9 / FRANCE tdouki@cea.fr
EMMI Salvatore Silvano

CNR / ISOF
Via P. Gobetti, 101 - 40129 Bologna / ITALY emmi@isof.cnr.it
FORAY Nicolas
INSERM / European Synchrotron Radiation Facility
BP 220-38043 Grenoble cedex / FRANCE foray@esrf.fr
GARDES-ALBERT Monique
Université René Descartes Paris V / Laboratoire de Chimie Physique
45 Rue des Saints-Pères - 75270 Paris cedex 06 / FRANCE Monique.Gardes@univ-paris5.fr
HICKEL Bernard
Commissariat à l’Énergie Atomique / Laboratoire de Radiolyse
Bât. 546 CEA/Saclay - 91191 Gif-sur-Yvette / FRANCE b.hickel@wanadoo.fr
HOLTOM Philip
Walton Hall - Milton Keynes MK7 6AA / UNITED KINGDOM p.holtom@open.ac.uk
HOUEE-LEVIN Chantal
Université Paris-Sud / Laboratoire de Chimie Physique
bât. 350 - 91405 Orsay / FRANCE chantal.houee-levin@lcp.u-psud.fr
KISTER Jacky
Université Paul Cezanne Aix-Marseille / Systèmes chimiques complexes
CNRS-UMR Faculté de St-Jérôme 6171
13397 Marseille cedex 20 / FRANCE Jacky.kister@univ-cezanne.fr
LAMPRE Isabelle
Bât. 349 - 91405 Orsay / FRANCE isabelle.lampre@lcp.u-psud.fr
ÄXIÄ
2!$)!4)/.#(%-)3429
MASON Nigel J
Walton Hall - Milton Keynes MK7 6AA / UNITED KINGDOM n.j.mason@open.ac.uk
MOSTAFAVI Mehran
Physical Chemistry Institute, Centre ELYSE-CLIO
CNRS / University Paris-Sud, Orsay / FRANCE mehran.mostafavi@lcp.u-psud.fr
PROUILLAC Caroline
Université Paul Sabatier / Laboratoire Hétérochimie Fondamentale et Appliquée
118 Route de Narbonne - 31062 Toulouse / FRANCE prouilla@chimie.ups--tlse.fr
RAFFI Jacques
CEA Université Paul Cezanne Aix-Marseille III
Laboratoire de Radiolyse de la Matière organique
Avenue Escadrille Normandie-Niémen
13397 Marseille cedex 20 / FRANCE j.raffi@univ.u-3mrs.fr
REMITA Hynd
Bât. 349 - 91405 Orsay / FRANCE hynd.remita@lcp.u-psud.fr
RIMA Ghassoub
Université Paul Sabatier / Laboratoire Hétérochimie Fondamentale et Appliquée
118 Route de Narbonne - 31062 Toulouse / FRANCE rima@chimie.ups--tlse.fr
SAGE Evelyne
CNRS / Institut Curie
Centre Universitaire - 91405 Orsay / FRANCE evelyne.Sage@curie.u-psud.fr
SEVILLA Michael D.
Oakland University / Department of Chemistry
Rochester - 48309 Michigan / USA sevilla@oakland.edu
SIEBBELES Laurens
Delft University of Technology
Mekelweg 15 - Delft 2629 JB / THE NETHERLANDS L.D.A.Siebbeles@tnw.tudelft.nl
ÄM>>Ä
,ISTOFAUTHORS
SPOTHEIM-MAURIZOT Mélanie
INSERM senior scientist, Molecular Biophysics
Centre – CNRS, Orléans / FRANCE spotheim@cnrs-orleans.fr
TAKACS Erzsebet
Hungarian Academy of Sciences / Institute of isotopes and surface chemistry
PO Box 77 - 1525 Budapest / HUNGARY takacs@iki.kfki.hu
TILQUIN Bernard
Université Catholique Louvain 72-30
Unité d'Analyse Chimique et Physico-Chimique des Médicaments
72 Avenue E. Mounier - B1200 Bruxelles / BELGIQUE tilquin@cham.ucl.ac.be
WISHART James
Brookhaven National Laboratory / Chemistry Department
Upton, NY 11973 / USA wishart@bnl.gov
ÄXIIIÄ
Part I
Primary
radiation-induced
phenomena
Chapter 1
An overview of the radiation
chemistry of liquids
George V. BUXTON
Introduction
Radiation chemistry is the chemistry initiated by the interaction of high-energy

photons and atomic particles with matter, so-called ionising radiation. As a method of
generating free radicals for applications in general chemistry the most commonly used
sources of ionising radiation are 60Co G-rays, which are photons having energies of 1.17 and
1.33 MeV (1 eV = 1.6 × 10−19 J), or fast electrons from an accelerator with energies typically in
the range 2-20 MeV. The dose absorbed by the material is expressed in grays (1 Gy = 1 J kg–1)
and the dose rate in Gy s–1. In each case the result of the interaction of high energy particles
with molecules is the ejection of a single electron, called a secondary electron which itself
may have sufficient energy to cause further ionisations, but which rapidly (< 10−12 s) reaches
thermal equilibrium with the liquid and becomes trapped as a so-called solvated electron
(es–). In this way, stable molecules (M) are converted into solvated electrons and highly
reactive free radicals (Mt):
M MtFs– (1)
Pulse radiolysis experiments have provided clear evidence for solvated electrons
in both polar (water, alcohols, etc.) and non-polar (alkanes) liquids through their optical
absorption spectra.
An important characteristic of ionising radiation is that it is absorbed non-selectively
so that molecules are ionised according to their relative abundance in the medium of interest.
Ä(Ä
2!$)!4)/.#(%-)3429
Figure 1 : Initial non-homogeneous spatial distribution of ioni-
sation events in spurs along the track, and evolution with time by
diffusion and reaction up to homogeneous distribution at ≈10−7 s.
A qualitative, albeit rather crude, way of visualising the distribution
and evolution of spurs along the track is to recall the pattern of
splashes produced when you throw a small flat stone and make it
skip across the smooth surface of a pond. Initially the splashes are
well-separated (low LET) but as the stone loses its kinetic energy
they get closer together and finally merge (high LET) at the end of its
trajectory, which is generally not a straight line. With the elapse of
time the circular ripples produced by each splash expand and even-
tually overlap with those of their neighbours, which is analogous to
homogeneous distribution along the track of an ionising particle in
the radiolysis of water.
For example, in dilute solution ([solute] ≤ 0.1 mol dm−3)
the ionised molecules (M) are essentially those of the
solvent so that knowledge of the radiation chemistry of
the solvent is of paramount importance for chemistry
initiated by ionising radiation. However, the rate of
energy loss of an energetic particle passing through a
liquid is non-uniform ; it increases as the kinetic energy
of the particle decreases. The rate of energy loss per unit
length of track of the particle is known as the stopping power or linear energy transfer (LET).
When the ionising particle is a high energy electron, the ionisation events occur in small
clusters called spurs that are widely separated along the track and the radiation is classed
as low LET (Fig. 1). For energetic nuclear ions, e.g. H, He, Li, etc., the rate of energy loss is
much higher so that the spurs are much closer together and can overlap to form cylindrical
columns of ionised molecules in the liquid. In this case the radiation is classed as high LET.
Actually, low LET radiation has a high LET component due to low-energy secondary electrons,
and high LET radiation has a low LET component due to high-energy secondary electrons,
so it is common practice to classify types of radiation by their average LET over the length of
the particle track. In the case of 60Co G-rays it is the secondary electrons that cause most of
the ionisations but the G-photons, being uncharged, are very much more penetrative than
electrons of the same energy.
Radiation chemistry of water
It was in early studies of water radiolysis that the concept of spurs was developed [1].
The current state of knowledge can be summarised by the following reactions (the time by
which each event is estimated to be complete is indicated) (Fig. 2) [2]:
Ä)Ä
2ADIATIONCHEMISTR YOFLIQUIDS
H2O H2OtF− and H2O* 10−16 s (2)

t
H2Ot)2O 0))3O 10−14 s (3)
H2O* H OH, H20t
t t
10−13 s (4)
e−nH2O eaq− 10−12 s (5)
Reaction (2) represents ionisation and electronic excitation of water molecules ; this
occurs on the timescale of an electronic transition. The positive radical ion H2Ot is believed to
undergo the ion molecule reaction (3) in ~ 10−14 s. The electronically excited states H2O* are
known to dissociate in the vapour phase in reaction (4), and the electron released in reaction (1)
is known to be solvated by ~ 10−12 s. At this time, for low LET radiation such as 60Co G-rays and
fast electrons from an accelerator, the products of reactions (3) – (5) are clustered together in
small widely separated spurs, which on average contain 2 to 3 ion pairs. Next, these products
begin to diffuse randomly, with the result that a fraction of them encounter one another
and react to form molecular and secondary radical products, while the remainder escape
into the bulk liquid and effectively become homogeneously distributed with respect to
Figure 2 : Scheme of reactions of transient species produced by irradiation in water without or with a
diluted solute S acting as a radical scavenger. The processes in reaction (2) are shown in red.
Ä*Ä
2!$)!4)/.#(%-)3429
reaction with solutes acting as radical scavengers (Fig. 1). These spur processes of reaction
and diffusive-escape are complete within ~ 10−7 s, at which time the radiolysis of water for
low LET radiation ( e.g. 0.23 eV nm−1) can be represented as reaction (6) [2] (Fig. 2):
H2O 0.28 eaq−, 0.062 Ht, 0.28 tOH, 0.047 H2, 0.073 H2O2, 0.28 H3O (6)
Here the numbers are the radiation chemical yields (known as G values) in units
of μmol J−1. In the early literature (and sometimes still today) G values are quoted as
molecules/100 eV, or just as a number in which case the units molecules/100 eV are implied.
The conversion factor is:
1 molecule/100 eV = 1.036 × 10−7 mol J−1 (or 0.1036 μmol J−1)
For historical reasons the G values in reaction (6) are known as primary yields and it is
estimated that about 40% of the initial yields (Go) produced in reactions (3) – (5) are consumed
by the spur reactions, i.e. Go(eaq−) ~ 0.5 μmol J−1. The spur reactions are listed in Table 1.
Table 1. Spur Reactions in Water [2].
Reaction k (1010 dm3 mol−1 s−1)
eaq−Faq− H20)− 0.54
− t −
eaq OH OH 3.0
− t
eaq )3O H )2O 2.3
− t −
eaq ) H20) 2.5
Ht)t H2 1.3
t t
0) OH H2O2 0.53
t t
0)) H2O 3.2
−
H3O 0) 2H2O 14.3
The extent of these reactions increases with increasing LET due to overlapping spurs ; for
example, for LET = 108 eV nm−1 the primary yields (G values) become those in reaction (6'):
H2O 0.044 eaq−, 0.028 Ht, 0.056 tOH, 0.11 H2, 0.11 H2O2, 0.044 H3O (6')
Radiation chemistry of organic liquids
The complexity of organic molecules compared to H2O means that there is a greater
variety of products arising from reactions following the ionisation event in reaction (1).
Ä+Ä
Furthermore, the relative permittivity εr (or dielectric constant) of the liquid is an important
parameter. The probability P of an electron (es−) which is thermalised at a distance r metres
from its geminate positive ion (Mt) escaping recombination with it is exp(−rc/r) where rc is
the distance at which the Coulomb potential between es− and Mt is equal to thermal energy
(kT) and is given by the Onsager expression rc = e2/4πε0εrkBT, where e is the elementary
charge (e = 1.6 × 10−19 C), ε0 is the vacuum permittivity (ε0 = 8.854 × 10−12 J−1 C2 m−1), kB is the
Boltzmann constant (kB = 1.38 × 10−23 J K−1) and T is temperature in Kelvin. In water at 298 K,
where εr = 78.5, rc ~ 0.7 nm so that the probability of recombination of an electron with its
positive ion is small since r is a few nanometres. In contrast to this in a hydrocarbon such as
cyclohexane or benzene, for which εr ~ 2, rc ~ 30 nm, most of the electrons and ions produced
in reaction (1) recombine. Those ions that escape geminate recombination are called free ions.
The yields of free ions (Gfi) in these hydrocarbons are 0.005 μmol J−1 (benzene) and 0.015μmol J−1
(cyclohexane). More polar organic molecules such as alcohols show larger values of Gfi.
For example for ethanol with εr = 24.3, Gfi = 0.18 μmol J−1. In simple terms, one should
be able to estimate thermalisation distances r from Ggi, but other factors have to be
considered. Thus the values of Gfi for nonpolar liquids tend to increase with the sphericity of
the molecules, ranging from 0.005 μmol J−1 for benzene to 0.11 μmol J−1 for neopentane [3].
Aqueous solutions
The largest contribution of radiation chemical techniques to general free-radical

chemistry has been made in aqueous solution because they provide a very convenient
way of generating an enormous variety of highly reactive species which cannot readily be
generated by thermal or photochemical methods. In particular, the technique of pulse
radiolysis has provided a wealth of kinetic and mechanistic information in inorganic, organic,
and biochemistry [4,5].
Properties of the primary radicals from water radiolysis
As shown in reaction (6), the principal primary radicals are the hydrated electron
−
(eaq ) which is a powerful reductant (standard reduction potential Eo = −2.87 V [6]) and the
hydroxyl radical (tOH) which is a powerful oxidant (Eo(tOH/OH−) = 1.90 V in neutral solution,
and Eo(H,tOH/H2O) = 2.72 V in acidic solution). The hydrogen atom (Ht) is not an important
species in neutral or alkaline solution, but it becomes the major reductant (Eo(H/Ht) = − 2.31 V)
in acidic solution through reaction (7):
eaq−) Ht (7)
Ä,Ä
2!$)!4)/.#(%-)3429
The hydrated electron may be visualised as an excess electron surrounded by

oriented water molecules (Chapter 3), although the precise details of its structure are not
yet settled. It reacts by transferring into a vacant orbital of an acceptor molecule or ion:
eaq−4n Sn−1 (8)
Here n is the charge on the acceptor S. Rate constants for reaction (8) range from
16 l mol−1 s−1 (S = H2O) up to the diffusion-controlled limit, although the measured activation
energy is invariably small (6-30 kJ mol−1), which suggests that the entropy of activation is the
dominant kinetic parameter. This can be rationalised in terms of the accessibility of a vacant
orbital on S for the electron to enter and explains why solvated electrons are sufficiently long-
lived to be observed by pulse radiolysis in liquids such as water, simple alcohols, ethers, amines
and alkanes whose molecules have no low-energy vacant orbitals. Organic molecules with
low-energy vacant orbitals (e.g. most aromatics, halides, thiols, disulfides, carbonyl- and nitro-
compounds) all react rapidly, eaq− acting as a nucleophile. Thus its reactivity is greatly enhanced
by electron-withdrawing substituents adjacent to double bonds or attached to aromatic rings.
Although the first step is electron addition to a vacant orbital, in some cases, notably with
organic halides (RX), bond breakage occurs very rapidly and the overall reaction essentially
appears as a dissociative electron capture process with elimination of the halide ion X−:
eaq−39 (RXt¦) Rt9− (9)
Inorganic ions react rapidly with eaq− when their reduction potentials are more
positive than − 2.9 V. In general, cations react faster than anions, reflecting the effect of the
negative charge of eaq−.
The hydroxyl radical readily oxidises inorganic ions and the reaction is often
represented as a simple one-electron transfer process:
t
0)4n Sn0)− (10)
However, there is abundant evidence from pulse radiolysis studies to indicate that
t
OH forms an adduct with the inorganic ion in the case of halides (X−) and simple metal
cations (Mn
aq) and it is suggested that this is the reaction path for most oxidisable inorganic
ions. A good example for halides is the reaction with chloride ion which only occurs in acidic
solution because Eo(Clt/Cl−) = 2.41 V. The mechanism comprises reactions (11) and (12) and
the intermediate HOClt¦ has been identified by its absorption spectrum [7]:
t
0)$M− HOClt¦ (11)
Ä-Ä
HOClt¦) Clt)2O (12)
Similarly, pulse radiolysis studies of the oxidation of simple aquo-complexes of

metal ions by tOH reveal that the reaction can be represented by the following steps :
t t
0).n
aq [Mn
aq , OH] M(naq
0)− (13)
The hydroxyl radical reacts with many inorganic anions at near-diffusion controlled
rates (k ~ 1010 dm3 mol−1 s−1 [5]), but with aquated metal ions (Mn = Tl, Ag, Cu, Sn,
Fe, Mn) the rate constants (k) have an upper limit of ~ 3 × 108 dm3 mol−1 s−1. There
is no correlation between k and the rates of exchange of water molecules coordinated to
Mn, which rules out ligand substitution as a general mechanism. Other possibilities are
abstraction of H from a coordinated water molecule or tOH entering and expanding the
coordination shell of the metal ion.
In alkaline solution, tOH is rapidly converted to its conjugate base Ot¦:
t
0)0)− Ot¦)2O (14)
Ot¦ is generally less reactive than tOH with inorganic anions and with Br−, CO32− and
Fe(CN)64− the reaction is immeasurably slow, a property which was used to measure the
equilibrium constant for reaction (14) and hence obtain pKa(tOH) = 11.9 [8].
In its reactions with organic molecules, tOH is an electrophile whereas Ot¦ is a
nucleophile. For example, in their reaction with aromatic compounds containing an aliphatic
side chain, tOH adds preferentially to the aromatic ring but Ot¦ abstracts H from the side chain :
t
0)$6H5CH3 (HOC6H5CH3)t (15)
Ot¦$6H5CH3 C6H5CH2t0)− (16)
This can result in a change of reaction mechanism when the pH is raised so that
O replaces tOH as the oxidant. tOH is rather unselective when it abstracts H from C−H
t¦
bonds because of the significantly larger energy of the H−OH bond being formed, e.g. with
1-propanol three different radicals are produced :
t
0)$)3CH2CH2OH CH3CH2tCHOH (53.4%)
t t
(CH3 CHCH20) CH2CH2CH2OH) (46%) (17)
Although it is a reducing radical, the reactions of the hydrogen atom with organic
molecules are similar to those of tOH. Thus it adds to C=C double bonds and aromatic rings
Ä.Ä
2!$)!4)/.#(%-)3429
and abstracts H from C−H bonds in saturated compounds ; in the latter case it is generally
less reactive and more selective than tOH. In its reactions with metal ions Ht is less reactive
than eaq− and in some cases it reacts via the formation of a hydride complex and effectively
becomes an oxidant :
[H]
t −
H 'F [Fe , H ] Fe3)2 (18)
Generation of secondary radicals
The radiolysis of water produces approximately equal yields of reducing (eaq–)t)
and oxidising (tOH) radicals, but in its applications in general chemistry it is desirable to
have either totally reducing or totally oxidising conditions. These conditions can readily be
obtained in the following ways :
(a) by interconversion of the primary radicals into a single kind ;
(b) by converting all the primary radicals into a single kind of secondary radical ;
(c) by removing the unwanted primary radical by reaction to form a relatively inert
secondary radical.
In this way the radiolysis of water provides a ready source of one-electron redox agents that
can be finely tuned in terms of reduction potential and electric charge, and this has been
widely used in obtaining kinetic and mechanistic data for free radical chemistry in aqueous
solution [5].
Oxidising conditions
A very convenient and widely used method of converting eaq– into tOH is to saturate
the aqueous solution with N2O so that reaction (19) followed by (14) occur :
eaq−/2O N20t¦ (19)
Under these conditions, the conversion eaq– to Ot¦ is essentially complete in ~ 15 ns but
the half-life of its protonation in reaction (14) is ~ 8 ns so complete conversion to tOH takes
~ 40 ns. In pulse radiolysis experiments designed to measure reactions of tOH one should take
care, therefore, to choose conditions that allow this radical to be fully formed before the
reaction of interest begins. The hydrogen atom reacts only slowly with N2O (k = 2.1 × 106 dm3
mol−1 s−1 [5]) so that approximately 90% of the radicals available in N2O-saturated neutral
solution aretOH and the remaining 10% are Ht.
Ä&%Ä
Oxidising radicals that are more selective than tOH in their reactions can be obtained
by converting tOH into another inorganic radical such as Br2t¦ (Eo = 1.66 V), I2t¦ (Eo = 1.05 V),
(SCN)2t¦ (Eo = 1.33 V), CO3t¦ (Eo = 1.5 V), N3t (Eo = 1.33 V). These radicals are particularly useful
for studying redox changes in metalloproteins and organometallic complexes because they
are more likely to react at the metal centre whereas tOH will also attack the organic moiety
to generate a reducing radical there, either by abstraction of H or by addition to an aromatic
ring or C=C bond.
A very useful way of generating oxidising radicals more powerful than tOH in neutral
solution is via reaction (20) :
eaq−42O82− SO4t−4042− (20)
SO4t¦ has Eo = 2.43 V and so can be used to generate other oxidants such as Clt
(Eo = 2.4 V) and NO3t (Eo = 2.5 V) through reaction with Cl− and NO3−, respectively.
Reducing conditions
A good method of obtaining totally reducing conditions is to convert the primary
radicals eaq−, Ht and tOH into a single kind of secondary radical ; for example in N2O-saturated
solution containing formate ion (HCO2−), CO2t¦ (Eo = − 1.9 V) is produced in reactions (21) :
t
OH/Ht)$02− H2O/H2$02t¦ (21)
A similar strategy can be used to generate a host of organic radicals as exemplified

in reactions (22) – (24) :
t
0)3) Rt)2O (22)
t
0)ɍ) )0ɍ
t (23)
t
0))33h) H(OH)R−tR'H (24)
)FSF3JTBOBMLZMHSPVQBOEɍJTBOBSZMHSPVQ
When reactions of eaq− are to be studied the usual method is to add 2-methyl-2-
propanol ((CH3)3COH) to convert tOH into the relatively unreactive radical tCH2C(CH3)2OH.
Reaction of eaq− with organic halides, RX, is also a useful source of organic radicals :
eaq−39 Rt9− (25)
Ä&&Ä
2!$)!4)/.#(%-)3429
The hydrated electron reacts rapidly with simple aquated metal ions, Mn
aq, to produce
0 0
hyper-reduced oxidation states such as Ag , Au , Cd , Co , Ni , Zn , etc. that are not easily
accessible by other methods :
eaq−.n M(n¦

(26)
These hyper-reduced metal ions are themselves powerful reductants and can be
used to reduce metalloproteins at negatively charged sites which are less accessible to eaq−.
One of the important fields of study opened up by reaction (26) is the production and
characterisation of noble and non-noble, mono- and multi-metallic clusters. Nano-particles
of many metals have been produced and their catalytic properties investigated in this way [9]
(Chapter 7).
One-electron reduction and oxidation in organic solvents
The advantage of organic liquids over water is the range of solvent polarities
available so that species that are insoluble in water can be studied in a suitable solvent or
mixture of solvents. A selection of illustrative examples is given in the next sections [10].
Alcohols
Methanol
The generation of free radicals in the radiolysis of methanol can be represented by
reaction (27), followed rapidly by (28) :
CH3OH es−, Ht, tCH2OH, CH3Ot (27)

Ht/CH3Ot$)3OH t
CH20))2/CH3OH (28)
Thus this system generates the reducing radical tCH2OH so that when the solubility of
a compound is higher in methanol than in water it is advantageous to effect its one-electron
reduction by the radiolysis of its methanolic solution. This technique has been utilised, for
example, for one-electron reduction of some organometallic complexes and quinones. As
in water, es− can be converted to tCH2OH via reactions (19) (with N2O) and (29) :
Ot¦$)3OH t
CH20)0)− (29)
Ä&'Ä
In alkaline solution tCH2OH deprotonates to form tCH2O− which is a stronger reductant than
t
CH2OH :
t t
CH20)$)3O− CH2O−$)3OH (K = 2.1 × 103 [11]) (30)
2 - Propanol
The radiation chemistry of 2-propanol is analogous to that of methanol ; the major
reducing radicals are es− and (CH3)2tCOH, and (CH3)2tCO− in alkaline solution. This solvent has
been used to investigate the one-electron reduction of fullerenes. For example, reduction
of C60, which is insoluble in water, by es− and (CH3)2tCOH generates C60t− which is stable for
hours in the absence of oxygen.
Chlorinated hydrocarbons
Radiolysis of solvents such as dichloromethane (CH2Cl2), tetrachloromethane (CCl4),

and 1,2-dicholoroethane (ClCH2CH2Cl) generates radical cations or radicals that are strong
oxidants. Thus, CH2Cl2t oxidises the fullerenes C60, C76 and C78 to the corresponding fullerene
radical cations Cnt (n = 60, 76, 78). The oxidising radical in CCl4 is CCl3t formed in reaction (30)
by elimination of chloride ion :
es−$$M4 CCl3t$M− (31)
The one-electron oxidation of a wide range of compounds has been investigated

in this way, including phenols and metallotetraphenylporphyrins. One-electron oxidation
of metallotetraphenylporphyrins is also achieved in 1,2-dichloroethane, but in this case
pyridine is added as a base to prevent demetallation of the complexes by HCl which is a
radiolysis product.
Acetonitrile
This is a versatile solvent in general chemistry because it is aprotic and has a high
dielectric constant. The main radical is the reductant CH3CNt¦ and this makes acetonitrile
a convenient solvent in which to study the one-electron reduction of transition metal
complexes that are not stable in hydroxylic media such as water and alcohols.
Ä&(Ä
2!$)!4)/.#(%-)3429
Hydrocarbons
In these nonpolar liquids, geminate ion recombination is the dominant process as

described earlier. The initial result of this recombination of electron and parent positive ion
is the formation of electronically excited states of the solvent molecule that are liable to
have quite high energies. Unless fragmentation of the molecule takes place through bond
breakage, the excited molecules fall to their lowest excited state through internal conversion
within ~ 10−12 s [12]. After this, the various processes known from photochemistry take
place, i.e. intersystem crossing, energy transfer, luminescence, bond rupture, etc. [12].
Aromatic compounds are much more resistant to radiation damage than alkanes and
alkenes, because the excitation energy is shared by all the π-electrons so that this energy
is not localised sufficiently in any bond to cause it to break. It is informative to examine
the radiation chemistry of two simple, symmetrical hydrocarbons, cyclohexane (C6H12) and
benzene (C6H6). In each case all C−H and C−C bonds are equivalent.
Cyclohexane
The major products and their G-values are H2 (G = 0.58 μmol J−1), cyclohexene (C6H10,
G = 0.33 μmol J−1) and bicyclohexyl ((C6H11)2, G = 0.18 μmol J−1), which can be understood
in terms of the decomposition of excited states of cyclohexane (C6H12*), formed by ion
recombination or directly by the ionising radiation, in the following ways :
t
C6H12* C6H11)t (32)
C6H12* C6H10)2 (33)
Ht$6H12 t
C6H11)2 (34)
2tC6H11 (C6H11)2 (35)
2tC6H11 C6H12$6H10 (36)
Reaction (32) in which a C−H bond is broken is the major fate of C6H12*. There are
only minor yields of products with less than six carbon atoms, indicating that rupture of C−C
bonds in cyclohexane is relatively unimportant.
Benzene
The resistance of benzene to radiation damage is reflected in G(H2) ~ 0.004 μmol J−1 ;
it is more than 100-fold less than the yield in cyclohexane. Other decomposition products
are formed in similarly low yields and the radiation chemistry of benzene is dominated by its
lowest singlet (1B*) and triplet (3B*) excited-state molecules. By adding suitable solutes the
Ä&)Ä
yields of these states have been measured as G(1B*) = 0.17 μmol J−1 and G(3B*) = 0.44 μmol J−1.
Since 3B* is a major primary species in the radiolysis of liquid benzene, pulse radiolysis
provides a ready method of producing and studying triplet excited states of other (solute)
molecules whose triplet energy lies below that of 3B* (3.56 eV) by triplet-triplet energy transfer
[12]. These triplet states are not always accessible by photo-excitation which predominantly
generates excited singlet states if intersystem crossing from singlet to triplet excited sate is
inefficient [12].
Concluding remarks
This brief description of the radiation chemistry of liquids shows how versatile is
the use of high energy ionising radiation in chemistry. Despite the large energy involved in
the radiolytic step, the radiolysis of most liquids produces solvated electrons and relatively
simple free radicals, and the technique can be tuned to solve specific problems. Two of the
more important applications are the following. The first is the identification of transient
intermediates, and the measurement of their reaction rates [13], which enables one to
unravel complex mechanisms initiated by free radicals in inorganic, organic and bio-organic
chemistry. The second is the synthesis of stable and mono-disperse nanoparticles of metals
that are very efficient catalysts (Chapter 7).
References
[1] Allen A.O., “The Radiation Chemistry of Water and Aqueous Solutions”, Van Nostrand, Princeton, N.J., 1961.
[2] Buxton G.V., Radiation Chemistry of the Liquid State: (1) Water and Homogeneous Aqueous
Solutions in “Radiation Chemistry : Principles and Applications”, Rodgers M.A.J., Farhataziz (eds),
VCH, Weinheim, 1987, 321-349.
[3] Holroyd R.A., The Electron: Its Properties and Reactions in “Radiation Chemistry : Principles and
Applications”, Rodgers M.A.J., Farhataziz (eds), VCH, Weinheim, 1987, 201-235.
[4] Rodgers M.A.J., Farhataziz (eds), “Radiation Chemistry : Principles and Applications”, VCH,
Weinheim, 1987.
[5] Ross A.B., Mallard W.G., Helman W.P., Buxton G.V., Huie R.E., Neta P., NDRL-NIST Solution Kinetics
Database:- Ver. 3.0, Notre Dame Radiation Laboratory, Notre Dame, IN and National Institute of
Standards and Technology, Gaithersberg, MD, 1998.
[6] Stanbury D.M., Reduction Potentials Involving Inorganic Free Radicals in Aqueous Solution, Adv.
Inorg. Chem., 1989, 33, 6-138. (All values of Eo quoted in the text are from this source).
[7] Hug G.L., Optical Spectra of Nonmetallic Inorganic Transient Species in Aqueous solution, Nat.
stand. ref. dat. ser.: NSRDS-NBS 69, 1981.
Ä&*Ä
2!$)!4)/.#(%-)3429
[8] Buxton G.V., Greenstock C.L., Helman W.P., Ross A.B., Critical Review of Rate Constants for Reactions
of Hydrated Electrons, Hydrogen Atoms and Hydroxyl Radicals (tOH/tO−) in Aqueous Solution,
J. Phys. Chem. Ref. Data, 1988, 17, 513-886.
[9] Belloni J., Mostafavi M., Remita H., Marignier J.-L., Delcourt M.O., Radiation-induced synthesis of
mono- and multimetallic clusters and nanocolloids, New J. Chem., 1998, 22, 1239-1255.
[10] Buxton G.V., Mulazzani Q.G., Radiation-Chemical Techniques in “Electron Transfer in Chemistry.
Volume 1: Principles, Theories, Methods and Techniques”, Balzani V. (ed), Wiley-VCH, Weinheim,
2001, p. 503-557.
[11] Johnson D.W., Salmon G.A., Pulse Radiolysis of Methanol and Ethanol: Acid-Base Behaviour of
Hydroxymethyl and Hydroxyethyl Radicals, J. Chem. Soc., Faraday Trans., I, 1975, 71, 583-591.
[12] Calvert J.G., Pitts J.N., “Photochemistry”, Wiley, New York, 1966.
[13] Buxton G.V., Measurements of Rate Constants for Radical Reactions in the Liquid Phase in “General
Aspects of the Chemistry of Free Radicals”, Z.B. Alfassi (ed), Wiley, Chichester, 1999, 51-78.
Ä&+Ä
Chapter 2
Tools for radiolysis studies
James F. WISHART
Since the first report of the chemical effects of radiation by Pierre and Marie Curie,
researchers have needed tools to deliver ionizing radiation for their scientific studies in
increasingly precise ways. In the earliest stages, this was accomplished by the development
of radioactive sources of increasing refinement and activity, and by the construction of X-ray
tubes of increasing power. In the middle of the 20th Century, particle (primarily electron)
accelerators took over as the primary tools of radiation chemists. At first, these accelerators
were employed as continuous radiation sources like their predecessors. However, the
development of pulse radiolysis techniques in the 1960s vastly increased the ability to study
radiation-induced chemical kinetics. Before long, time resolution was extended into the
picosecond regime [1-5]. In recent years, a new generation of radiolysis facilities has been
developed to extend temporal resolution to even shorter times, at the same time providing
a range of new transient detection capabilities [6-13].
Detailed accounts of the development of radiation chemistry and its tools can
be found elsewhere. The purpose of this chapter is to describe the basic characteristics of
continuous and pulsed sources of ionizing radiation for radiolysis studies, and to provide
a broad overview of the present and near-future status of radiolysis instrumentation
worldwide, for the benefit of readers who would like to use these powerful techniques to
advance their own research. It is inevitable under the circumstances that some facilities
may be missed and that future developments will soon render this overview out-of-date,
however the substantial progress that has been made in the years since the previous reviews
appeared [14-16] merits description here.
Ä&,Ä
2!$)!4)/.#(%-)3429
Types of ionizing radiation and their methods of generation
Ionizing radiation for chemical studies comes in many types with a vast range of
properties that allow adaptation of experimental design to the chemistry of interest. The
fundamental point to remember is that radiolytic energy deposition within the sample is
inhomogeneous on short time (nanosecond) and length (nanometer) scales, and that the
spatial pattern of reactive species produced from this energy deposition depends on the
type of incident radiation and its energy. Photonic radiations (X- and gamma rays) deposit
energy in well-separated interactions within the sample, while highly charged nucleons
produce much denser deposition patterns with a higher probability of overlapping regions
of ionized species. This effect is quantified as “Linear Energy Transfer” or LET. Photons and
electrons are considered to be low LET radiations, whereas protons and heavier nucleons
are high LET. There is also a dependence of LET on particle kinetic energy. Reactions of
(and between) primary radiation-produced species convert the spatial inhomogeneity into
variations in yield of radiolysis products on longer time scales (microseconds and beyond)
depending on the type and energy of radiation. Chapter 1 of this book and other publications
[14-17] discuss these effects in greater detail. The experimental choice of radiation type and
energy depends on the application. Electron beams and X-rays (gammas) are well suited
for general kinetics studies, while high LET radiations provide important details about the
physical mechanisms of radiolysis and its chemical effects, and they are also becoming quite
important in the study of clustered damage in nanostructured materials, including biological
systems (DNA and cells) and synthetic polymers and resists (e.g., ion beam lithography).
X- and gamma rays
Ionizing photonic radiation (X- and gamma rays) can be produced by the decay
of certain radioactive isotopes or generated by stopping or deflecting a particle beam
(bremsstrahlung and synchrotron radiation, respectively). X- and gamma rays, by virtue
of their low LET, can penetrate sample vessels of moderate thickness such as cryostats or
pressure vessels, however by the same token the deposited radiolytic dose (energy per
unit sample mass) is low compared to particle radiations of the same fluence. Radioactive
gamma sources based on 60Co or 137Cs isotopes are used in continuous radiation mode for
radiolysis product studies and competition kinetics measurements. There are two basic types
of source-based irradiators for chemical research. In the first, samples are placed within a
shielded vault containing a radioactive source in a shielded container. After the experimenter
exits the vault, the source is removed from the shielded container to expose the samples. In
a modification of this design, samples may be transported into and out of the irradiation
vault by a conveyor system. The second type of irradiator consists of a hollow cylindrical
Ä&-Ä
4OOLSFORRADIOLYSISSTUDIES
radioactive source permanently fixed in shielding. A mechanism is used to transport samples

into the center of the cylindrical source, where they receive uniform radiation by virtue of
the geometry. Due to the configuration, sample dimensions are limited to what will fit inside
the source cylinder. Some gamma irradiators can be fitted with flow systems that transport
fluids through the irradiation zone for controlled exposure times.
Particle accelerators can be used as continuous or pulsed X-ray sources by stopping

the particle beam (typically electrons) in a high-atomic-number material such as gold or
tungsten. The process of bremsstrahlung (“braking radiation” in German) produces a broad
continuum of X-radiation that peaks at one third of the incident particle energy. This method
can be used to produce nanosecond or picosecond X-ray pulses for time-resolved kinetics
studies. For example, accelerator-produced X-rays were used to initiate experiments to
measure the mobilities of excess electrons in non-polar liquids as functions of temperature and
pressure, using transient conductivity measurement techniques. The experiments depend
on the ability of X-rays to penetrate pressure vessel walls that would stop particle beams.
Other types of accelerators can be used as X-ray sources for specialized purposes.
Synchrotron facilities can provide very intense radiation over a wide but selectable range of
energies. Depending on the operating mode of the synchrotron it is possible to do time-resolved
studies. Extremely short, sub-picosecond pulses of X-rays can be generated by laser wake-field
accelerators (described below), which use terawatt electromagnetic fields from femtosecond lasers
to accelerate electrons but also produce intense pulses of X-rays. When a second femtosecond laser
pulse is used to interrogate the sample at various delay intervals with respect to the X-ray pulse,
it would be possible to follow the very earliest steps of the radiolytic process.
Because of their high intensity, X-ray tubes were commonly used as laboratory
radiation sources for radiation chemistry experiments until they were superceded by particle
accelerators during the middle part of the 20th Century. They still retain specialized uses in
research applications such as being used as the radiation source for MARY (MAgnetic field
effect on Reaction Yield) spectroscopy studies of radical cation lifetimes and reactivity in
alkane solvents [14,18]. MARY spectroscopy uses fluorescence to detect variations in singlet-
triplet dynamics in radical ion pairs as a function of magnetic field. It is particularly useful for
short-lived transients that are difficult to study by ESR.
Particle accelerators for radiolysis
There are several methods used to accelerate charged particle beams for pulse radiolysis.
Acceleration requires a force applied by an electric field. The field may be a continuous
Ä&.Ä
2!$)!4)/.#(%-)3429
gradient (or “DC” as in direct current) as produced by an electrostatic potential (Fig. 1a), or
oscillating in time and space as produced by radio frequency (RF), microwave, or optical
laser radiation (Fig. 1b).
Figure 1 : Schematic depictions of electrostatic (a) and oscillating electromagnetic (b) fields for charged
particle acceleration.
Examples of the electrostatic type of accelerator are the Van de Graaff generator and
capacitor-driven DC impulse generators. The characteristics of the two are quite different.
The Van de Graaff generator (Fig. 2) develops and maintains a continuous electrostatic
field by transporting charge (typically electrons) on a belt of non-conducting material that
passes between two terminals at ground and high potential, respectively. The terminals are
separated by stacks of insulating material alternating with metal plates. The metal plates
are each connected to a resistor chain that slowly drains the charge off the high potential
terminal so that the accelerating potential can be controlled by the belt charging rate. Typical
operating potentials of such accelerators are 2-5 MV. Standing electrostatic potentials higher
than 5 MV become progressively more difficult to sustain without extraordinary measures,
and beam energies of 2-5 MeV are adequate for many pulse radiolysis applications. An
evacuated beam tube runs from the high-potential terminal to ground. The insulating glass
spacers along the tube are interrupted at intervals by the metal plates connected to the
resistor chain, so that a uniform accelerating gradient is applied to the particle beam.
Electron beams are by far the most common usage of Van de Graaff generators
for pulse radiolysis, although ion sources may also be used. Typically, grid-gated cathodes
are used to produce electron pulses varying in length from several nanoseconds to
several microseconds with beam currents on the order of one Ampere. On some systems,
continuous emission is possible at significantly lower currents (~ 1 mA). The minimum pulse
width is determined by the response of the pulsing circuit. Specialized circuitry providing a
Ä'%Ä
Figure 2 : A Van de Graaff accelerator.
sub-nanosecond electron pulse was developed on the 3 MV Van de Graaff accelerator at the
Technical University of Delft [19]. Van de Graaff accelerators emit no electromagnetic noise
apart from the impulse generated from the electron beam itself, since they are electrostatic
devices. This is an advantage for the implementation of real-time detection systems such
as time-resolved microwave conductivity (TRMC) and transient electron paramagnetic
resonance (EPR). The Delft Van de Graaff has been the site of significant development of
Ä'&Ä
2!$)!4)/.#(%-)3429
the TMRC technique, which measures the migration and recombination processes of charge
carriers in materials [20]. Transient EPR detection is useful for identifying and observing
the reactions of radical species that are difficult to characterize by optical absorption
spectroscopy [21]. A 3 MeV Van de Graaff accelerator with EPR detection system is located
at the Notre Dame Radiation Laboratory (NDRL) in the United States. NDRL also hosts a
time-resolved resonance Raman spectroscopy (TRRR) system on a 2 MeV Van de Graaff.
The high specificity of vibrational spectroscopy provides information about the structure
of radical species [22]. The 4 MeV ELBENA Van de Graaff at the Hahn-Meitner Institute in
Berlin supports advanced transient absorption detection capabilities as well as AC and
DC conductivity detection methods with nanosecond time resolution (DC) [23] The 2 MeV
Van de Graaff at Brookhaven National Laboratory incorporates a sensitive transient
absorption detection system and time-of-flight charge carrier mobility measurements in
non-polar media by DC conductivity.
The second series of electrostatic accelerators are based on direct-current impulse

generators. The best-known examples of these are the Febetron units produced by the
Field Emission Corporation. These machines are based on a Marx-bank impulse generator,
whereby a high (0.6 to 2.3 MV) potential is generated by triggering a series of spark gaps to
switch a bank of capacitors from charging in parallel to discharging in series through a large
cathode tube. Very large electron currents (~ 7000 A) are passed in pulses lasting several to
tens of nanoseconds, delivering radiolytic doses up to 20 kGy. The large doses are well suited
for the study of radical-radical reactions and the radiation chemistry of gases. The electrical
noise generated by the spark gaps and the discharge of the Febetron is a major problem for
detection systems, requiring much signal averaging and background correction, however
the Febetron is unmatched in its ability to deliver large radiolytic doses in short pulse
durations. Although Febetron-based pulse radiolysis facilities were once located in many
laboratories, today only a few remain, notably a 600 keV unit at the Laboratoire de Chimie
Physique of the University of Paris XI, in Orsay, France.
Another kind of impulse generator-based accelerator based on a pulse transformer was

developed in Novosibirsk, USSR, in the 1970s. One such unit (ELIT-1, 1 MeV) has been operating
at the Time-Resolved Spectroscopy Laboratory of the University of Leipzig. It is equipped with a
beam deflection unit that can reduce the electron pulse width within the sample to less than a
nanosecond and it has been coupled to a Fourier transform EPR detection system.
Accelerators that use electromagnetic radiation to produce the accelerating field

can accelerate particles to much higher energies than electrostatic accelerators. Higher
beam energies provide more penetrating power for elaborate experimental setups such as
Ä''Ä
pressure vessels and cryostats, and the ability to better focus more charge into a smaller target
to produce a higher radiolytic dose. As a practical consideration however, beam energies
above 40 MeV (for electrons) are less desirable for pulse radiolysis because an increasing
fraction of the beam energy is converted into bremsstrahlung instead of being deposited in
the sample in radiolytic events. In addition, the bremsstrahlung produced by high-energy
beams can induce radioactivity in typical substances found in scientific equipment and
building materials, creating an additional safety hazard.
Linear accelerators (or linacs) are the most common type of oscillating-field
accelerator for pulse radiolysis. The accelerating sections consist of a series of resonant
cavities for radio-frequency or microwave radiation (100 MHz to 10 GHz, with 1-3 GHz being
most common). Charged particles are accelerated by the oscillating electric field to different
degrees depending on their position (phase) with respect to the RF cycle. The differential
acceleration collects the particles into bunches that are clustered around the optimal
acceleration phase in each RF cycle. This fact underscores a significant difference in beam
temporal profile between electrostatic and oscillating field accelerators. Beams emitted
from the former type are continuous for the duration of the emission, whereas for the latter
the beam is structured into a pulse train or “macropulse” consisting of a series of bunches
separated in time by the period of the accelerating radiation. Each bunch is inherently short
in time (~ 30 picoseconds) ; but significant technical effort is required to exploit the time
resolution offered by such a short pulse, as explained below.
The majority of linear accelerators used for pulse radiolysis work exclusively
in the pulse train mode and are typically referred to as “nanosecond” linacs, in contrast
to the “picosecond” linacs that can operate in the single bunch mode. Nanosecond linacs for
pulse radiolysis are distributed throughout the world, including the Notre Dame Radiation
Laboratory in the U.S., the Commissariat à l’Énergie Atomique (CEA) in Saclay, France,
the Institute of Applied Radiation Chemistry in Łódź, Poland, the Institute for Nuclear
Chemistry and Technology in Warsaw, Poland, the ISOF-CNR in Bologna, Italy, the Institute
of Isotope and Surface Chemistry in Budapest, Hungary, the Bhabha Atomic Research
Center (BARC) in Mumbai, India, the National Centre for Free Radical Research in Pune,
India, the Australian Radiation Protection and Nuclear Safety Agency in Melbourne, the
University of Auckland, New Zealand, and at the Shanghai Institute of Nuclear Research
in China. A linac for pulse radiolysis will be installed at the University of Manchester, UK,
as part of the establishment of a new research group in Radiation Chemistry. Transient
optical absorption spectroscopy is the standard technique for kinetics measurements
at each of these facilities. In addition, transient mid-infrared detection methods are under
development at the linac in Saclay [24].
Ä'(Ä
2!$)!4)/.#(%-)3429
Picosecond electron linacs require special techniques to generate single electron

bunches for ultrafast kinetic studies. The problem stems from the fact that it is not practical
to electrically gate a cathode electron source on and off in less than a nanosecond, thus
cathode gating is too slow to fill only one period (350 to 770 ps) of the RF cycle at the
accelerating frequencies of typical accelerators. In a clever development, a sweeping
beam deflection device was used on the nanosecond Novosibirsk linac to select a single
picosecond bunch from the pulse train [5]. Kinetics of charge recombination were followed
by using a streak camera to monitor emission from excited states produced via electron-
hole recombination. Aside from that special case, historically the standard method of
generating picosecond single pulses (Fig. 3) is to inject the electrons into a lower frequency
RF field operating at a sub-harmonic of the accelerating frequency (for example, 476 MHz
for an S-band, 2856 MHz linac), allow the electron packet to evolve into a compact single
bunch in the lower-frequency section, then to inject the compressed bunch into the higher-
frequency accelerating section. The system used for injecting the single bunches is called a
“sub-harmonic pre-buncher”.
Figure 3 : Schematic representation of an S-band pre-bunched picosecond linear accelerator.
Because it involves more accelerating structures with multiple RF frequencies and

phase relationships, and a high-performance electron gun source to inject the electrons in
about one nanosecond, a sub-harmonic pre-buncher adds a lot of cost and complexity of
operation of the linac facility. Consequently, only a few pre-bunched picosecond linacs were
built for pulse radiolysis studies. The first one was installed at Argonne National Laboratory
in the 1970s [2] ; subsequently other picosecond installations were built at the University of
Tokyo Nuclear Energy Research Laboratory (NERL) in Tokai-Mura [3,4] and Osaka University
[4]. Very recently, a new pre-bunched linac for pulse radiolysis has been built at the Shanghai
Institute of Applied Physics.
During the late 1980s and into the 1990s, a new technology arose for generating
picosecond electron beams [14,15]. Instead of using thermionic emission from a hot cathode
Ä')Ä
Figure 4 : Schematic
representation of the
BNL LEAF photocathode
electron gun accelerator
showing the laser-micro-
wave synchronization re-
lationship.
to produce the electron beam, a metal (Cu or Mg) or semiconductor (e.g., Cs2Te) cathode
is struck with an ultraviolet laser pulse to cause it to emit electrons via the photoelectric
effect. The electron pulse emitted from the photocathode is as short as the incident laser
pulse (although it tends to expand during beam transport). Thus it is possible to inject single
bunches of electrons directly into the high-frequency accelerating structure, if the laser
pulse is synchronized to the correct RF phase for proper acceleration. Commercial lasers
with the ability to synchronize themselves to external frequency sources became available
in the same time frame. The electron source and accelerator functions can be combined into
a single unit called a photocathode electron gun. A schematic of the photocathode electron
gun of the Brookhaven National Laboratory Laser-Electron Accelerator Facility (LEAF) [6,14]
is shown in Figure 4 and a picture of the accelerator with the photocathode back plate
removed is shown in Figure 5. While the LEAF accelerator consists of an integral structure of
3.5 microwave cavities that accelerates electron bunches to a final energy of 9 MeV when
Figure 5 : The LEAF accelerator

opened to show the 6 mm
diameter magnesium photo-
cathode in the center of the
back plate.
Ä'*Ä
2!$)!4)/.#(%-)3429
driven with 15 MW of microwave power, 1.5 cavity photocathode electron guns that produce
beam energies of 4-5 MeV are often used as injectors for systems with further accelerating
stages, including most of the photocathode radiolysis facilities listed further below.
Photocathode-based picosecond electron accelerators are conceptually simpler

than pre-bunched thermionic systems, although they require reasonably powerful,
multicomponent femtosecond or picosecond laser systems to drive the photocathode. In
addition, the availability of synchronized laser pulses allows the development of advanced
detection capabilities with unprecedented time resolution. The combination of ease of
use and powerful detection methods has stimulated strong interest in photocathode gun
systems. Since the installation of the first photocathode electron gun pulse radiolysis system
at BNL [5,13], four additional photocathode-based facilities have become operational and
two more are in progress. The operational centers include the ELYSE facility at the Université
de Paris-Sud XI in Orsay, France [7,8], NERL in Tokai-Mura, Japan [9,10], Osaka University
[11,12], and Waseda University in Tokyo [13]. Facilities under development are located at the
Technical University of Delft, the Netherlands, and the BARC in Mumbai, India.
The latest development in ultrafast electron sources for pulse radiolysis does away
completely with accelerating structures. Instead, femtosecond laser pulses at terawatt power
levels are focused tightly onto jets of helium atoms. The extremely strong electromagnetic
field of the laser radiation strips the electrons off the helium atoms and accelerates them to
very high energies, in a process called laser wake-field acceleration [25,26] but sometimes
referred to as “Table-Top Terawatt” or T3. Efforts are being made to characterize and control
the energy spectrum of the accelerated electrons [26]. Experiments have shown that it is
possible to use beams from these systems to do pulse radiolysis measurements [27],
although energy spread of the electron beam and geometrical constraints place the practical
time resolution in the few picosecond range despite the use of femtosecond laser pulses.
Laser wake-field acceleration is studied in several laboratories around the world, but work on
its development as a radiolysis method has occurred at the Terawatt Ultrafast High Field
Facility at Argonne National Laboratory (Fig. 6) [28], the Laboratoire d’Optique Appliquée, École
Polytechnique-ENSTA, Palaiseau, France [25-27], NERL in Tokai-Mura, Japan, and the Rutherford
Appleton Laboratory in the UK. T3 systems also hold promise as ultrafast X-ray sources.
Experimental Detection Techniques for Ultrafast Radiolysis
Transient absorption optical detection methods for picosecond and faster

resolution are subject to a number of considerations. First, the velocities of high-energy
electrons are approximately that of the speed of light in vacuum, velect ≈ c, whereas light itself
Ä'+Ä
is slowed by the refractive index n of the sample vlight = c/n. Therefore, visible light
falls about one picosecond behind the electron beam for each millimeter of transit
through water when the laser and electron beams are co-linear, thus the time resolution
is constrained by the path length through the sample and must be traded off against
signal magnitude. An arrangement that uses beams that cross at an angle can be used
to compensate for the difference in velocities [31]. Another factor is that real-time,
digitizer-based optical detection systems have response (bandwidth) limitations that
limit time resolution, although technology has improved over the years. Still today,
extremely fast biplanar phototube detectors and high-bandwidth oscilloscopes
(≥ 6 GHz) offer resolution in the visible region down to approximately 80 picoseconds, but
that is still slower than the capabilities of the accelerator to produce short electron pulses.
Detectors in other wavelength regions are slower and have complex responses that must be
deconvoluted from the kinetics.
Figure 6 : Vacuum chamber of the Terawatt Ultrafast High Field Facility (TUHFF) at ANL, configured for a
measurement of electron beam energy dispersion (courtesy of R. Crowell).
Another method of continuous detection is the use of a streak camera in

absorption mode. Measuring radiation-induced emission with a streak camera is relatively
straightforward, but sensitive transient absorption measurements are difficult to make
because of limited dynamic range of the streak camera. The first work in this area was done
in the 1970s, but little was done to follow up until the development of high-dynamic range
Ä',Ä
2!$)!4)/.#(%-)3429
streak cameras in recent years. One such streak camera has been used to great effect to
measure transient absorption spectra at the ELYSE facility in Orsay, France [8].
To overcome these limitations, stroboscopic detection methods have been used

since the earliest picosecond radiolysis measurements (Fig. 7). Analogous to laser pump-
probe experiments, in pulse-probe transient absorption spectroscopy a short pulse of light
is used to measure the absorbance of a sample as a function of time delay between the
electron pulse and the probe beam. Time resolution using this method is typically on the
order of the electron pulse width. Originally, C̆erenkov light generated from the electron
beam itself was used as a variable-delay probe beam [1,2]. Later, a free-running Ti-Sapphire
oscillator was used with a time-amplitude converter to measure spur decay of the solvated
electron. With the advent of laser photocathode electron gun systems that provide
picosecond-synchronized laser pulses, high-resolution pulse-probe kinetics measurements
have nearly become routine. Timing improvements [12] and streak-camera jitter detection
have made such measurements possible on thermionic, pre-bunched picosecond linacs
as well. For transient spectroscopy, two methods of probe pulse generation are used to
measure kinetics at wavelengths other than the fundamental of the gun-driving laser system
(~ 800 nm). At some facilities, a white light continuum is generated by focusing intense
laser pulses into a substrate such as a sapphire plate, fused silica or D2O [10,11,13], while
at others a specified probe wavelength is produced by an optical parametric amplifier
pumped by the gun-driving laser system [6,12]. Broadband detection of the white-light
probe is accomplished with a spectrograph/CCD or diode array combination [11], while
single-wavelength detection is done with a pair of diodes [6,8,11,13]. Broadband detection
works within the sensitivity range of silicon detectors (up to 950 nm), while NIR-sensitive
photodiodes (Ge and InGaAs) extend the range of single-wavelength detection to 1700 nm.
Pulse-probe transient spectroscopy has been used to follow the formation of Xe2* excimers
in the radiolysis of supercritical xenon [29], measure dissociation rates of aryl halide radical
anions, to re-evaluate the yield of hydrated electron at picosecond times [30], and to observe
the solvation of excess electrons in ionic liquids.
While affording high time-resolution, pulse-probe absorption measurements

typically require large numbers of shots to assemble a profile of the reaction by time and
wavelength. Since cumulative radiation damage of the sample can be a problem, sample
solutions are often flowed once-through or recirculated during experiments. Many
interesting studies are thus not practical to perform because samples are not available in
the necessary quantities or they cannot be flowed. To address this problem, techniques to
measure complete time profiles or spectra, in one shot or just a few shots, have recently
been developed. A group at ANL employed the Frequency-Domain Single Shot (FDSS)
Ä'-Ä
spectroscopy technique where a high-bandwidth femtosecond pulse is stretched to several

hundred picoseconds, passed through the sample, then dispersed by a monochromator
onto a photodiode array, where each wavelength corresponds to a different time. The Osaka
group has obtained spectra with high signal-to-noise in a single-shot, using a CCD to detect
broadband absorption by reducing the time interval between sample and reference shots to
1 ms [11]. At BNL’s LEAF facility, an optical fiber bundle is being used to create imaged probe
beams containing 100 different time delays in a single shot.
Figure 7 : Schematic of a general detection system for picosecond pulse radiolysis (courtesy of Dr. A. Saeki,
Osaka University).
Heavy ion sources
Radiolysis experiments using heavy ion beams (protons and heavier atomic nuclei
stripped of all electrons) occupy an important role, despite the relatively small number of
investigators in this area. Because heavy ions have much higher LET values, the pattern
of energy deposition, or track structure, is much denser than with “lighter” radiations. The
variation of yields of radiolytic products as a function of LET, and the effects of variable
Ä'.Ä
2!$)!4)/.#(%-)3429
concentrations of scavengers on those yields, have provided critical information for the
understanding of the radiolytic process [14,17]. The high local density of ionizations produced
by heavy ions creates localized areas of damage that can be useful, as in the ion beam
modification of polymers [14]. Ion beams have been used to fragment polymers to form
membranes, and conversely, to polymerize precursor substrates to form forests of polymer
rods on surfaces. The localized damage caused by heavy ions has important implications for
radiation biology and radiation medicine. The American space agency NASA has established
a heavy ion radiolysis facility at BNL (the NASA Space Radiation Laboratory) for in vitro and
in vivo studies to prolonged exposure to heavy ion radiation in interplanetary space. Proton
radiotherapy is a precise way of delivering large radiation doses to kill tumors. Heavy ion
radiolysis of water produces the hydroperoxyl radical HO2t, which is not formed by low-LET
radiolysis in the absence of molecular oxygen. It is an important means of causing oxidative
damage to hypoxic tumors.
Generally, heavy ion radiolysis experiments are performed at large, multipurpose

accelerator facilities using cyclotrons or tandem Van de Graaff accelerators. The Grand
accélérateur national d’ions lourds (GANIL) in Caen is the primary facility in France for heavy
ion experiments. Major facilities in Japan include TIARA at the Japan Atomic Energy Agency,
the Heavy Ion Medical Accelerator (HIMAC) in Chiba, and the High-fluence Irradiation Facility
of the University of Tokyo. Heavy ion sources in the U.S. include ATLAS at ANL, the National
Superconducting Cyclotron Laboratory at Michigan State University, and the Triple Ion
Irradiation Laboratory at Oak Ridge National Laboratory. Because heavy ions are stopped in
very short distances within samples, special techniques and equipment configurations are
used to optimize transient signals.
Future trends
The past decade has been an encouraging period in the development of radiolysis
capabilities that has reversed an earlier trend of decline in number and accessibility. Two
new technologies, photocathode electron guns and laser wake-field accelerators, have
emerged and spawned a large new generation of ultrafast accelerator facilities. These
installations are developing advanced experimental techniques and making sophisticated
experiments available to a larger community of researchers than ever before. Earlier-
generation picosecond accelerators have been upgraded to high levels of performance.
New nanosecond linacs were installed at Notre Dame, Saclay and Pune, and the University
of Manchester has founded a program and Chair in Radiation Chemistry that will reinforce a
field that was in danger of disappearing from the U.K. These developments attest to the fact
that major research support establishments in several countries recognize that radiation
Ä(%Ä
chemistry and radiolysis experimentation have important roles to play in solving future
energy needs in many areas (including nuclear, solar and high-performance materials) and
in developing science to protect human health and the environment.
It is not taking a risk to predict that performance and capabilities of the new
radiolysis installations will markedly improve as these young facilities mature. But what
other developments can we look forward to? Certainly there is strong interest in bringing
additional spectroscopic tools into the pulse radiolysis laboratory. Efforts are underway to
adapt transient mid-infrared detection techniques to pulse radiolysis, to take advantage
of the specificity of vibrational spectroscopy [24]. Strong interest in nanoscience and the
mechanisms of reactions in heterogeneous systems will push the development of interface-
specific spectroscopies in radiation chemistry, for example surface-enhanced Raman
spectroscopy and second-harmonic or sum-frequency generation.
Acknowledgments
This work was carried out at Brookhaven National Laboratory under contract DE-AC02-
98CH10886 with the U.S. Department of Energy and supported by its Division of Chemical Sciences,
Office of Basic Energy Sciences. The author wishes to thank R. Crowell and A. Saeki for Figures 6 and 7.
References
[1] Bronskill M.J., Taylor W.B., Wolff R.K., Hunt J.W., Design and performance of a pulse radiolysis
system capable of picosecond time resolution, Rev. Sci. Instr., 1970, 41, 333-340.
[2] Jonah C.D., A wide-time range pulse radiolysis system of picosecond time resolution, Rev. Sci.
Instr., 1975, 46, 62-66.
[3] Kobayashi H., Ueda T., Kobayashi T., Washio M., Tabata Y., Tagawa S., Picosecond single electron
pulse for pulse radiolysis studies, Radiat. Phys. Chem., 1983, 21, 13-19.
[4] Tabata Y., Kobayashi H., Washio M., Tagawa S., Yoshida Y., Pulse radiolysis with picosecond time
resolution, Radiat. Phys. Chem., 1985, 26, 473-479.
[5] Grigoryants V.M., Lozovoy V.V., Chernousov Yu.D., Shebolaev I.V., Arutyunov V.A., Anisimov O.A.,
Molin Yu.N., Pulse radiolysis system with picosecond time resolution referred to Cherenkov
radiation, Radiat. Phys. Chem., 1989, 34, 349-352.
[6] Wishart J.F., Cook A.R., Miller J.R., The LEAF picosecond pulse radiolysis facility at Brookhaven
National Laboratory, Rev. Sci. Instr., 2004, 75, 4359–4366.
[7] Belloni J., Monard H., Gobert F., Larbre J.-P., Demarque A., De Waele V., Lampre I., Marignier J.-L.,
Mostafavi M., Bourdon J.C., Bernard M., Borie H., Garvey T., Jacquemard B., Leblond B., Lepercq P.,
Omeich M., Roch M., Rodier J., Roux R., ELYSE - A picosecond electron accelerator for pulse radiolysis
research, Nucl. Instr. Meth. A, 2005, 539, 527–539.
Ä(&Ä
2!$)!4)/.#(%-)3429
[8] Marignier J.-L., de Waele V., Monard H., Gobert F., Larbre J.-P., Demarque A., Mostafavi M., Belloni J.,
Time-resolved spectroscopy at the picosecond laser-triggered electron accelerator ELYSE, Radiat.
Phys. Chem., 2006, 75, 1024-1033.
[9] Muroya Y., Lin M., Watanabe T., Wu G., Kobayashi T., Yoshii K., Ueda T., Uesaka M., Katsumura Y.,
Ultra-fast pulse radiolysis system combined with a laser photocathode RF gun and a femtosecond
laser, Nucl. Instrum. Meth. A, 2002, 489, 554–562.
[10] Muroya Y., Lin M., Iijima H., Ueda T., Katsumura Y., Current status of the ultra-fast pulse radiolysis system
at NERL, the University of Tokyo, Res. Chem. Intermed., 2005, 31, 261-272.
[11] Saeki A., Kozawa T., Tagawa S., Picosecond pulse radiolysis using femtosecond white light with a high
S/N spectrum acquisition system in one beam shot, Nucl. Instr. and Meth. A, 2006, 556, 391-396.
[12] Saeki A., Kozawa T., Kashiwagi S., Okamoto K., Isoyama G., Yoshida Y., Tagawa S., Synchronization
of femtosecond UV-IR laser with electron beam for pulse radiolysis studies, Nucl. Instr. and Meth.
A, 2005, 546, 627-633.
[13] Nagai H., Kawaguchi M., Sakaue K., Komiya K., Nomoto T., Kamiya Y., Hama Y., Washio M., Ushida K.,
Kashiwagi S., Kuroda R., Improvements in time resolution and signal-to-noise ratio in a compact
pico-second pulse radiolysis system, Nucl. Instr. Meth. B, 2007, 265, 82-86.
[14] “Radiation Chemistry: Present Status and Future Trends”, Jonah C. D., Rao B.S.M. (ed.), Studies in
Physical and Theoretical Chemistry, Vol. 87, Elsevier Science, 2001.
[15] “Photochemistry and Radiation Chemistry: Complementary Methods for the Study of Electron
Transfer”, Wishart J.F., Nocera D.G., (eds), Adv. Chem. Ser. 254, American Chemical Society,
Washington, DC, 1998.
[16] “Radiation Chemistry: Principles and Applications”, Rodgers M.A.J., Farahtaziz (eds), VCH,
Weinheim, 1987.
[17] Mozumder A., “Fundamentals of Radiation Chemistry”, Academic Press, San Diego, 1999.
[18] Stass D.V., Sviridenko F.B., Molin Y.N., Magnetic field effect study of solvent hole deprotonation in
X-irradiated liquid n-alkanes, Radiat. Phys. Chem., 2003, 67, 207-210.
[19] Luthjens L.H., Vermeulen M.J.W., Hom M.L., de Loos M.J., van der Geer S.B., Revision of (sub)
nanosecond pulser for IRI Van de Graaff electron accelerator aided by field propagation
calculations, Rev. Sci. Instr., 2005, 76, 024702.
[20] de Haas M.P., Choffat F., Caseri W., Smith P., Warman J.M., Charge Mobility in the Room-Temperature
Liquid-Crystalline Semiconductor Poly(di-n-butylstannane), Advanced Materials, 2006, 18, 44-47.
[21] Werst D.W., Trifunac A.D., Observation of Radical Cations by Swiftness or by Stealth, Chem. Res.,
1998, 31, 651-657.
[22] Tripathi G.N.R., Time resolved resonance Raman observation of the extreme protonation forms of
a radical zwitterion in water, J. Chem. Phys., 2005, 122, 071102.
[23] Janata E., Lilie J., Martin M., Instrumentation of kinetic spectroscopy-11. An apparatus for
a.c.-conductivity measurements in laser flash photolysis and pulse radiolysis experiments, Radiat.
Phys. Chem., 1994, 43, 353-356.
Ä('Ä
[24] Le Caër S., Vigneron G., Renault J.P., Pommeret S., First coupling between a LINAC and FT-IR
spectroscopy: The aqueous ferrocyanide system, Chem. Phys. Lett., 2006, 426, 71-76.
[25] Malka V. et al., Electron acceleration by a wake field forced by an intense ultrashort laser pulse,
Science, 2002, 298, 1596–1600.
[26] Malka V., Faure J., Glinec Y., Lifschitz A.F., Laser-plasma accelerator: status and perspectives, Philosophical
Transactions of the Royal Society A, 2006, 364, 601-610 and references therein.
[27] Brozek-Pluska B., Gliger D., Hallou A., Malka V., Gauduel Y., Direct observation of elementary
radical events: low and high energy radiation femtochemistry in solutions, Radiat. Phys. Chem.,
2005, 72, 149–157.
[28] Crowell R.A., Shkrob I.A., Oulianov D.A., Korovyanko O., Gosztola D.J., Li Y., Rey-de-Castro R.,
Motivation and development of ultrafast laser-based accelerator techniques for chemical physics
research, Nucl. Inst. Meth. Phys. Res. B, 2005, 241, 9-13.
[29] Holroyd R.A., Wishart J.F., Nishikawa M., Itoh K., Reactions of Charged Species in Supercritical
Xenon as Studied by Pulse Radiolysis, J. Phys. Chem. B, 2003, 107, 7281-7287.
[30] Muroya Y., Lin M.Z., Wu G., Iijima H., Yoshii K., Ueda T., Kudo H., Katsumura Y., A re-evaluation of the
initial yield of the hydrated electron in the picosecond time range, Radiat. Phys. Chem., 2005, 72,
169-172.
[31] Yang J., Kondoh T., Kozawa T., Yoshida Y., Tagawa S., Pulse radiolysis based on a femtosecond
electron beam and a femtosecond laser light with double-pulse injection technique, Radiat.
Phys. Chem., 2006, 75, 1034-1040.
Ä((Ä
Chapter 3
The solvated electron :
a singular chemical species
Mehran MOSTAFAVI and Isabelle LAMPRE
Introduction
In this chapter, we pay close attention to the solvated electron, this so peculiar
“chemical species” already introduced in the first chapter as one situated at the origin of
most phenomena characterizing radiation chemistry. Its description, the analysis of its true
nature and of its main properties thus appears as essential in chemistry. Denoted by es−, it
can be visualized as a negative elementary charge surrounded by solvent molecules. The
peculiarity comes from the fact that the charge is delocalised but without a nucleus, and that
solvent molecules rally around it (Fig. 1). The solvated electron is a thermodynamically stable
radical but like most free radicals it has a short lifetime due to its great chemical reactivity.
As early as the nineteenth century, the solvated electron was observed but not identified
(Tab. 1). In 1808, Sir H. David and later, in 1864, W. Weyl reported the intense blue colouring
obtained by dissolving alkali metals in ammonia (NH3). Weyl also found that the resulting
solutions had reducing properties when used in organic synthesis, but he did not discover
the nature of these blue solutions. Nowadays it is known that the blue colour is due to the
solvated electron, as alkali metals dissolved in ammonia give the metal cation and a solvated
electron according to, for example in the case of sodium:
NH3 liq
Nametal NaNH3F NH3– (1)
Nearly one century later, in 1904, thanks to conductivity measurements, C. Kraus

showed that, in liquid ammonia containing dissolved alkali metals, there was a negatively
Ä(*Ä
2!$)!4)/.#(%-)3429
charged species with a conductivity

larger than that of anions, but
independent of the counter-
ion (Li, Na, K). He suggested
that this species was an electron
surrounded by ammonia molecules
behaving like an anion. In 1952, to
explain the bleaching of aqueous
solutions containing methyl blue
upon irradiation in the presence of
carbon dioxide (CO2), G. Stein also
proposed the transient formation
of a hydrated electron, similar to the
solvated electron in ammonia. The
direct spectroscopic observations Figure 1 : Schematic representation of the hydrated electron
obtained by molecular simulations [10]. The delocalised nega-
in 1962 by E.J. Hart and J.W. Boag
tive charge is surrounded by water molecules creating a cavity
of transient solvated electrons
with a radius of about 2.5 Å.
produced by pulse radiolysis of
aqueous solutions [1] opened the way to an extensive investigation on the properties and
theoretical models of the solvated electron. The solvated electron in water is known as the
hydrated electron, denoted by eaq−. After these first observations, the solvated electron was
soon detected in various solvents through its intense optical absorption band in the visible
or near infrared domain [2]. Owing to this property, the reactivity of the solvated electron
has been studied by transient absorption measurements in many solvents using pulse
techniques. In addition to pulse radiolysis, other methods can be used to produce a solvated
electron and allow its study in different environments (Inset). Due to the development of
ultrashort laser pulses, great strides have been made towards the understanding of the
solvation and short-time reactivity of the electron, mainly in water but also in other polar
solvents. So, despite its short life-time, the solvated electron is a unique chemical moiety
whose properties may be compared in many solvents. In this chapter, we consider the main
properties of the solvated electron, its reactivity and recent results concerning the solvation
dynamics of the electron.
Some physical properties of the solvated electron
The identification and the understanding of the chemical properties of the solvated
electron can be made through the knowledge of its physical properties. The properties of
the solvated electron depend on several factors such as solvent, temperature and pressure.
Ä(+Ä
4HESOLVATEDELECTRON
Table 1. Main dates for the discovery and study of the solvated electron.
Years Strides Authors
1808 Report on blue colouring of ammonia at the touch H. David a

of alkali metals.
1864 Blue colouring observed for solutions of alkali W. Weyl a,b

metals in liquid ammonia, methylamine and
ethylamine ; reducing properties of these solutions.
1908 Identification of a species with a negative C. Kraus a

elementary charge in ammonia solutions of alkali
metals, by conductivity measurements.
1952 Suggestion of transient formation of solvated G. Stein, R.L. Platzman b

electrons to explain the bleaching of aqueous
solutions containing methyl blue under irradiation
in the presence of CO2.
1962 First pulse-radiolysis experiments of aqueous E.J. Hart, J.W. Boag b [1]
solutions revealing the formation of solvated
electrons.
1971 Solvation dynamics of the electron in propan-1-ol J.H. Baxendale [23]

at −120 °C observed by pulse-radiolysis.
1987 Solvation dynamics of the electron in water at A. Migus et al. [26]

room temperature observed by ultrashort pulse
photolysis.
2007 Probing of the electron in water at attosecond time D. Nordlund et al. [34]
scale
See reviews: a) Edwards P.P., The electronic properties of metal solutions in liquid ammonia and related solvents, Adv. Inorg. Chem. Radiochem.,
1982, 25, 135-185. b) Boag J.W., Pulse radiolysis: a historical account of the discovery of the optical absorption spectrum of the hydrated elec-
tron, in “Early developments in radiation chemistry”, Kroh J. (Ed.), Royal Society of Chemistry, Cambridge, 1989, 7-20.
Volume
In 1921, by dissolving an alkali metal in liquid ammonia, C. Kraus and W. Lucasse
observed a volume expansion of the solution greater than that obtained for the dissolution of
ordinary salts [3]. They attributed this volume expansion to the formation of the solvated
electron with a cavity, regarded as a particle since the electron itself has a negligible volume.
For example, the dissolution of 3 moles of sodium in one litre of liquid ammonia induces an
increase in volume of 43 cm3 compared to the pure liquid. Assuming that all the metal is
dissociated, it may be deduced that in ammonia the electron occupies a spherical volume
with a radius of 0.18 nm. In fact, the cavity radius of the solvated electron in ammonia is
greater than that value and is about 0.3 nm.
Ä(,Ä
2!$)!4)/.#(%-)3429
Inset : Ways of producing the solvated electron
The solvated electron may be generated by different methods, more or less appropriate
according to the media and the detection systems.
Chemistry. Historically, the first method was the dissolution of alkali metals in amine solvents ;
nevertheless, this is useful only in media in which the lifetime of the solvated electron is long
enough (at least a few hours in the pure solvent). For example, in liquid ammonia, the solvated
electron may be formed from sodium metal :
Na metal/)3 liq m NaNH FNH −

(1)
3 3
Electrochemistry. Electrolysis reactions may also be used to produce the solvated electron in
similar media ; for instance, in liquid ammonia the reactions are:
At the anode :
4 NH3 – 2 e– m 2 NH2t/)4 m N2H4/)4 (2)
At the cathode:
NH3F− m eNH −

(3)
3
Radiolysis. Under high-energy radiations (G- or X-rays, beams of accelerated electrons or

positive ions), electrons may be ejected from the most abundant (solvent) molecules in the
medium. These ejected electrons have excess kinetic energy that is lost in collision with solvent
molecules, which may be electronically excited, or ionised to produce more electrons in a
cascade scheme. When their kinetic energy falls below the ionisation/excitation threshold of
the solvent, the electrons are “thermalised” and become “solvated” as solvent molecules get
reorganised around them.
Photolysis. Absorption of UV or visible light from a flash lamp or a laser is used to produce the
solvated electron. Two processes may occur: (1) the photon energy is sufficient to ionise a solute
and produce the electron, for instance, photo-detachment from an anion (Cl−, Fe(CN)64−, Na−…),
(2) the laser intensity is high enough to induce ionisation of the solvent by multiphoton
absorption (e.g. H20nhν mH2OtF−).
Ä(-Ä
4HESOLVATEDELECTRON
It is to be noted that the cavity models developed by theoreticians to describe the

solvated electron originate from this observed volume expansion. In these models a cavity
is occupied by the electron and the solvent molecules are organized around it (Fig. 1). The
great volume of the cavity accounts for the volume expansion of the solution.
Charge
Conductivity measurements performed by C. Kraus in ammonia solutions of alkali metals
gave evidence of a negatively charged species independent of the cation, the solvated electron.
However, it was to be wondered whether the same species was formed in the radiolysis of liquids.
By studying the rates of reaction of the reducing species produced by G- radiolysis of aqueous
solutions containing different neutral, cationic and anionic solutes, G. Czapski and H. A. Schwarz
found that this reducing species has a unit negative charge and is a hydrated electron [4].
Mobility
Most methods for determining the electron mobility use pulse radiolysis techniques
in which the concentration of electrons is followed during or after the ionizing pulse by
measurement of the change in conductivity. However, due to the inherent conductance of
polar liquids, direct conductivity measurements of solvated electrons are generally difficult
in these media. Therefore, the diffusion coefficient and the mobility of the solvated electron
in various solvents have been indirectly derived from the value of the rate constant of a
fast reaction thought to be diffusion controlled [5]. In 1969 it was discovered that excess
electrons in non polar solvents are from 100 to 100,000 times more mobile than ions in
the same liquids. The electron mobility depends on the solvent and on the structure of the
molecules. For example, the mobility in n-alkanes decreases with increasing chain length, but
it increases with branching ; at room temperature the value is 28.0, 0.15 and 70 cm2 V−1 s−1 in
ethane, n-pentane and 2,2-dimethylpropane, respectively [6]. In polar solvents the electron
mobility is generally slightly greater than ionic mobilities, which are about 7 s 10−4 cm2 V−1 s−1.
In water the measured mobility of the hydrated electron is 1.9 s 10−3 cm2 V−1 s−1. In a viscous
solvent such as ethane-1,2-diol, the value falls down to 2.8 s 10−4 cm2 V−1 s−1. The mobility
increases with temperature as the viscosity decreases. The viscosity effect shows that the
transport of the solvated electron is mainly due to diffusion and that transport via jump or
tunnel effect is not predominant for this species. This strengthens the idea of the solvated
electron as a “real” chemical entity and not as a free isolated charge.
Ä(.Ä
2!$)!4)/.#(%-)3429
Optical absorption
A major characteristic of the solvated electron is its optical absorption spectrum.
The optical absorption spectrum of the hydrated electron was first determined in 1962 by
Hart and Boag using transient absorption measurements in pulse radiolysis of pure water
and aqueous solutions of carbonates ; it appears as an intense broad structureless band with
a maximum around 720 nm in pure water [1]. Then, optical absorption spectra were reported
for the solvated electron in a large number of solvents. The position of the maximum and the
width of the absorption band depend on the medium. Figure 2 shows the optical absorption
spectrum of the solvated electron in various solvents at room temperature. The solvents
may be classified in three groups :
1/ Polar protic solvents with hydrogen bonds, like water and alcohols. The wavelength
at the absorption maximum of the solvated electron spectrum lies in the visible
domain, between 500 and 820 nm ; for instance, it is around 525 nm for glycerol
and 640 nm for methanol.
Figure 2 : Optical absorption spectra of the solvated electron in various solvents (from [2] with addition of
other data). Abbreviations: H2O: water; MeOH: methanol; EtOH: ethanol; 1PrOH: propan-1-ol; 2PrOH: propan-2-ol;
12ED : ethane-1,2-diol or ethylene glycol ; 12PD : propane-1,2-diol ; 13PD : propane-1,3-diol ; GLY : glycerol
or propane-1,2,3-triol; R4NNTf2: methyl-tributyl-ammonium bis[trifluoromethyl-sulfonyl]imide; EDA: ethane-1,2
-diamine ; 13PDA : propane-1,3-diamine ; NH3 : ammonia ; C3H8 : propane ; DME : dimethylether ; DEA: diethano-
lamine ; THF: tetrahydrofuran ; MeTHF: methyltetrahydrofuran ; DEE: diethylether.
Ä)%Ä
4HESOLVATEDELECTRON
2/ Liquid ammonia, amines and ammonium ionic liquids. The absorption spectrum of the
solvated electron is situated in the near infrared domain; in the case of liquid ammonia
at room temperature the wavelength of the absorption maximum is 1830 nm.
3/ Slightly or non polar aprotic solvents like ethers and hydrocarbons. The absorption
spectrum is observed in the infrared domain with a maximum at wavelengths
higher than 2000 nm ; for example, 2200 nm for tetrahydrofuran.
On the whole, the maximum of the absorption band of the solvated electron shifts
to higher wavelengths as the dielectric constant of the solvent diminishes. However, the
absence of a simple correlation between the maximum and the dielectric constant or the
dipolar moment of the solvent suggests that the position of the absorption spectrum is
mostly governed by the molecular structure of the liquids. Quantum simulations have
indicated that the electronic ground state of the solvated electron is a 1s-like localized state
and the excited states are three non-degenerate 2p-like states, also bound and localized,
followed by a band of delocalised states. Hence, the broad absorption band of the electron
corresponds mostly to an 1s m 2p transition with a contribution of the transition from the
bound state to the continuum at high energies [7]. The latter contribution accounts for the
asymmetry of the spectrum. As the energy levels of the exited states are generally close to the
conduction band of the solvent, the position of the absorption maximum gives information
on the depth of the potential energy well in which the solvated electron is localized. So, the
shorter the wavelength of the absorption maximum is (the higher the transition energy is),
the deeper is the well.
The absorption spectrum of the solvated electron depends on the nature of the
solvent, but is also particularly sensitive to parameters such as pressure- and temperature-
induced changes in solvent structure and properties (Chapter 4). The optical absorption band
shifts to higher energies (shorter wavelengths) with increasing pressure up to 2000 bar ; the
shift is larger in primary alcohols than in water and it correlates with the increase in liquid
density rather than with the rise in dielectric constant [8]. A rise in the temperature induces
a red shift of the solvated electron absorption spectrum. Thus, the absorption maximum in
water is located around 692 nm at 274 K and 810 nm at 380 K [9]. The asymmetric shape of
the absorption band of the hydrated electron at temperatures below the critical region is well
fitted by Gaussian and Lorentzian functions on the low- and high-energy sides, respectively [9].
Lately, quantum-classical molecular-dynamics simulations of an excess electron in water
performed for a wide range of temperature and pressure suggest that the observed red shift
of the optical absorption spectrum is a density effect rather than a temperature effect [10].
Indeed, by increasing the temperature the mean volume of the cavity of the solvated
Ä)&Ä
2!$)!4)/.#(%-)3429
electron increases, the electron is less confined in the cavity, the well becomes less deep.
The temperature dependence of the absorption spectrum of the solvated electron has been
recorded not only in water but also in alcohols (Figs. 3 and 4). The studies in ethane-1,2-
diol (12ED), propane-1,2-diol (12PD) and propane-1,3-diol (13PD) emphasize the significant
influence of the length of the aliphatic chain, and the effect of the distance between the two
OH groups on the behaviour of the solvated electron [11,12].
Figure 3 : Optical absorption spectra of the solva- Figure 4 : Energy of the maximum of the solvated
ted electron in propane-1,3-diol at different tempe- electron absorption band (Emax) as a function of
ratures (from [12]). The absorption spectrum shifts temperature in five different solvents (from [12]).
to longer wavelength with increasing temperature. By increasing the temperature, the absorption spectrum
shifts to longer wavelength so the transition energy
decreases. D2O : deuterated water ; CD3OD : deutera-
ted methanol ; 13PD: propane-1,3-diol ; 12PD: propa-
ne-1,2-diol ; 12ED: ethane-1,2-diol.
At room temperature, the wavelength of the absorption maximum is shorter in 12PD

(565 nm) than in 13PD (575 nm), indicating that the two neighbouring OH create deeper
electron traps in 12PD compared to 13PD. However, the traps in 12PD appear less deep than
in 12ED since the energy of the absorption maximum of the solvated electron measured at
a given temperature is lower in 12PD than in 12ED. This shows an influence of the additional
methyl group on the solvent structure, in particular on the three-dimensional networks of
hydrogen-bonded molecules. Moreover, an increase in temperature affects more greatly
Ä)'Ä
4HESOLVATEDELECTRON
the absorption spectrum of the solvated electron in 12PD and 13PD than in 12ED. The
effect is also greater in 12PD compared to 13PD. Indeed, the temperature coefficients are
– 2.9 s 10−3, – 2.8 s 10−3 and – 2.5 s 10−3 eV K−1 for 12PD, 13PD and 12ED, respectively (Fig. 4).
The decrease in viscosity versus temperature is also larger in 12PD compared to 13 PD and
12ED. The different behaviour for the three solvents results from larger modifications of the
solvent structure and molecular interactions for 12PD than for 13PD and 12ED. So, the solvated
electron interacting strongly with the solvent is very sensitive to solvent molecular structure.
Chemical reactivity of the solvated electron
Soon after the discovery of the absorption spectrum of the solvated electron in
pulse radiolysis experiments (Chapter 2), the rates and mechanisms of its reaction with a
wide variety of solutes was studied. Although it is a transient species, the solvated electron
is a very important reducing agent. Indeed, its reduction potential is very negative ; the
value of E°(H2O/es−) for the solvated electron in water is equal to – 2.8 V with respect to the
standard hydrogen electrode. The reaction rate constant and the probability of encounter
with another species decide the so-called lifetime of the solvated electron, which therefore
depends on the experimental conditions.
Geminate recombination and spur reaction

In radiation chemistry one of the most important reactions of solvated electrons
is recombination with positive ions and radicals that are simultaneously produced in close
proximity inside small volumes called spurs. These spurs are formed through further ionisation
and excitation of the solvent molecules. So, in competition with diffusion into the bulk,
leading to a homogeneous solution, the solvated electron may react within the spurs (Chapter 1).
Geminate recombinations and spur reactions have been widely studied in water,
both experimentally and theoretically [13-16], and also in a few alcohols [17,18]. Typically,
recombinations occur on a timescale of tens to hundreds of picoseconds. In general, the primary
cation undergoes a fast proton transfer reaction with a solvent molecule to produce the stable
solvated proton and the free radical. Consequently, the recombination processes are complex
and depend on the solvent. The central problem in the theory of geminate ion recombination is
to describe the relative motion and reaction between the two particles with opposite charges
initially separated by a distance r0. In water, the primary products of solvent radiolysis are the
hydrated electron eaq–, the hydroxyl radical vOH and the hydronium cation H3O:
GnhN H2O
H2O H2OtF– eaq–t0))3Oaq (4)
Ä)(Ä
2!$)!4)/.#(%-)3429
Interestingly, it has been shown that the recombination of the hydrated electron
is greatly dominated by reaction with vOH radical, because the reaction with H3O is not
diffusion controlled despite the Coulombic attraction. Geminate ion recombination is usually
considered to be negligible in water because the Onsager radius is small (0.7 nm) compared
to the radius of the distribution of eaq– in the spur (~ 2.3 nm) (Chapter 1).
Figure 5 depicts the decay of the solvated electron due to spur reactions in two
different solvents, water and tetrahydrofuran. In both liquids, the solvent relaxation is very
fast (less than 1 ps), therefore, the absorption signals on the picosecond time scale are due
to the fully solvated electron. As the dielectric constant of tetrahydrofuran is low (E = 7.6
compared to 80 for water), the electrostatic attraction is not screened by the solvent and
geminate recombination between the solvated electron and the cation can occur over long
separation distances in contrast to water. Moreover, the mobility of es– in THF is roughly
three times higher than that in water. That explains why the decay of the solvated electron
is more important in tetrahydrofuran compared to water [19].
Figure 5 : Decay at 790 nm of the solvated electron in water (H2O) and in tetrahydofuran (THF) (from [19]).
The decay of the solvated electron due to spur reactions is faster in a slightly polar solvent like THF than in water
which is a very polar solvent.
It is to be noted that, after geminate recombination, when diffusion takes place on the
nanosecond and longer time scale, reactions between the radiolytic species still occur and,
in the absence of any other solutes, those reactions are responsible for the disappearance of
the solvated electron. Besides, the metastability of the “blue” solutions of alkali metals in
liquid ammonia is due to the fact that the solvated electron does not react with another solvated
electron and that it reacts extremely slowly with the protonated form (NH4) which is at an
extremely low concentration.
Ä))Ä
4HESOLVATEDELECTRON
Reactions with solutes

During the last 35 years, the reactivity of the solvated electron has been widely
investigated, mainly by pulse radiolysis (Chapter 1). Indeed, because of its intense optical
absorption in the visible and near infrared spectral regions, the solvated electron is most
conveniently and reliably observed directly using pulse radiolysis or photolysis. So, the
majority of rate constants of reactions between the solvated electron and solutes or electron
scavengers have been measured by the method of decay kinetics. The reaction is evidenced
by the shortening of the lifetime of the solvated electron as the concentration of the solute is
increased (Fig. 6). If the reaction between the solvated electron and the solute is dominant,
the decay approaches a simple exponential form from which the reaction rate constant can
be deduced. The scavenging factor is the product of the scavenger concentration by the rate
constant ([S] s k4F –).
s
Figure 6 : Decays of the solvated electron recorded at 575 nm in pulse-radiolysis of ethane-1,2-diol in the
presence of silver cations, Ag+, at various initial concentrations in mol l−1. In pure ethane-1,2-diol and at the
lowest concentration of Ag+, the decay is mostly due to reactions of the solvated electron with other species
produced by radiolysis. By increasing the concentration, the decay becomes faster as the solvated electron reacts
predominantly with the Ag+ cation.
Many reactions of the solvated electron with different solutes, such as aliphatic,
aromatic or heterocyclic compounds, and also anions and cations, have been studied.
Ä)*Ä
2!$)!4)/.#(%-)3429
A wealth of information on the reduction of metal ions in aqueous solutions has been
obtained and a compilation was published in 1988 [20]. However, alkali or alkaline earth
metal ions such as Li, Na, or Mg, cannot be reduced by the hydrated electron in aqueous
solution but can form an ion pair with the solvated electron in polar liquids. Among the various
reactions of the solvated electron, the reduction of halogenated hydrocarbons is often used
in radiation chemistry to produce well-defined radicals because of the selective cleavage of
the carbon-halogen bond by the attack of the solvated electron. This reaction produces the
halide ion and a carbon-centered radical, and is of great interest for environmental problems
related to the destruction of halogenated organic contaminants in water and soil [21,22].
Figure 7 : Schematic representation of the formation of the solvated electron following solvent ionisation.
The fate of the electron starts with its ejection from a molecule upon ionisation by radiolysis or photolysis.
The ejected electron progressively loses its excess kinetic energy by collisions with solvent molecules (thermali-
sation step). Then the electron is localised, trapped in a solvent site or cavity and becomes solvated when the
solvent molecules have obtained their equilibrium configuration after relaxation.
Solvation dynamics of the electron
A simplified view of the early processes in electron solvation is given in Figure 7.

Initially, electron pulse radiolysis was the main tool for the experimental study of the
formation and dynamics of electrons in liquids (Chapter 2), first in the nanosecond time
range in viscous alcohols [23], later in the picosecond time range [24,25]. Subsequently, laser
techniques have achieved better time resolution than pulse radiolysis and femtosecond
pump-probe laser experiments have led to observations of the electron solvation on the
sub-picosecond to picosecond time scales. The pioneering studies of Migus et al. [26] in
water showed that the solvation process is complete in a few hundreds of femtoseconds
and hinted at the existence of short-lived precursors of the solvated electron, absorbing in
the infrared spectral domain (Fig. 8). The electron solvation process could thus be depicted
by sequential stepwise relaxation cascades, each of the successive considered species or
Ä)+Ä
4HESOLVATEDELECTRON
states of the electron having a fixed, individual spectrum [27,28]. However, subsequent
studies on electron solvation following the photoionisation of water have favoured the so-
called “continuous shift” model in which the existence of these IR-absorbing species is not
strictly required [29,30]. In the latter model, only one localised electron is considered but
its spectrum undergoes a continuous evolution, shifting to the blue (shorter wavelengths)
during the solvation. In alcohols, the formation of the solvated electron is slower than in
water [25,27]. The necessity to include both stepwise mechanism and continuous relaxation
to successfully interpret experiments has also been suggested [31].
Figure 8 : Absorption spectra of the electron at different delays after photo-ionisation of liquid water at
21 °C (from [26]). The solvation process is very fast as the hydrated electron (λmax = 720 nm) is observed within
1 ps from precursors absorbing in the infrared domain (λ > 1200 nm).
A continuous shift of the absorption spectrum of the solvated electron may be

interpreted in two ways. First, this shift would be governed by solvent molecular reorientation
around the electron (solvation), which has been suggested by a correlation between the
relaxation time of the solvated electron and the dielectric relaxation time of the solvent
[24]. Second, as suggested by Madsen et al. [30], the shift could be due to a “cooling” of the
solvated electron i.e. a “cooling” of water around the solvation cavity since the spectrum
of the solvated electron is known to shift to lower wavelengths at lower temperatures.
This means that the absorption spectrum of the electron at any time during the relaxation
process is identical to the spectrum of the solvated electron in the state of equilibrium
with the solvent at some higher temperature. In this approach, the electron stabilisation is
Ä),Ä
2!$)!4)/.#(%-)3429
viewed as a succession of quasi-equilibrium states that are fully characterised by the time
evolution of the local temperature. These two approaches of the “continuous shift” model
are phenomenological: no explanation is given, on theoretical grounds, as to why such a
picture of the electron solvation might be correct.
Figure 9 : Transient absorption spectra of excess electrons in liquid ethane-1,2-diol for five time-delays after
two-photon ionisation of the solvent by 200 fs pulses at 263 nm from [18]. The basic features of the solvated
electron are already there after a few picoseconds.
We recently studied the formation of the solvated electron in pure ethane-1,2-

diol by photo-ionisation of the solvent [18,32]. The results showed that the excess electron
presents a wide absorption band in the visible and near-IR domains at short delay times
after the pump pulse, and that the red part of the absorption band drops rapidly in the first
few picoseconds while the blue part increases slightly (Fig. 9). The time resolved spectra
were fitted correctly by either one of two solvation models: a stepwise mechanism involving
several distinct species and a continuous relaxation model. In Figure 10 are reported, as an
example, the kinetics and spectra of the three successive species (the weakly bound ewb–,
the strongly bound esb− and the solvated electron es−) involved in the electron solvation
dynamics according to the stepwise model.
Ä)-Ä
4HESOLVATEDELECTRON
Figure 10 : Time evolution of the concentrations and absorption spectra of the three electron species invol-
ved in the two-stepwise mechanism ewb− m esb− m es− used to fit the experimental data obtained for the
electron solvation in ethane-1,2-diol (Fig. 9) (from [32]).
However, the fact that the time-evolution of the absorption spectrum of the solvated electron
can be accurately described by the temperature-dependent absorption spectrum of the ground
state solvated electron (Fig. 11) suggests that the spectral blue shift would be mostly caused
by a continuous relaxation, or “cooling” of the electron trapped in a solvent cavity. To conclude,
this analysis clearly indicates that it is not obvious to select a unique model to describe the
solvation dynamics of electron in ethane-1,2-diol, and in other solvents.
Figure 11 : Similarity between the absorption spectra recorded in liquid ethane-1,2-diol at early time after
two-photon ionisation (Fig. 9) and the absorption spectra of the solvated electron obtained at high tempe-
rature by pulse radiolysis (from [11] and [32]). These results suggest that the electron solvation is a continuous
process similar to the continuous shift of the absorption spectrum of the solvated electron with temperature.
Ä).Ä
2!$)!4)/.#(%-)3429
Moreover, in a recent study on the of liquid water, Lian et al. [33] observed two
distinct regimes of the spectral evolution of the electron. In the first picosecond after
photoionisation, while the main absorption band of the electron progressively shifts to the
blue (from 820 to 720 nm), a new peak at 1.15 μm and a shoulder at 1.4 μm are observed,
which fully decayed in one picosecond. These features are attributed to the coupling
between the excess electron and the modes of the water molecules lining the solvation
cavity. At later delay times (t > 1 ps), the position of the band maximum is fixed, but the
spectral profile continues to change by narrowing on the red side and broadening on the
blue side. It appears that these results cannot be fitted by the presently available models
and that more advanced models are needed. Very recently, by probing the electron
delocalisation in water at attosecond time scale, D. Nordlund et al. [34] reported that an excess
electron propagating along the H-bond network is trapped at a weakly H-bonded OH group.
This trapping lasts long enough (> 20 fs) for electron solvation to occur, leading to the
hydrated electron.
Conclusion
The solvated electron is a transient chemical species which exists in many solvents.
The domain of existence of the solvated electron starts with the solvation time of the precursor
and ends with the time required to complete reactions with other molecules or ions present
in the medium. Due to the importance of water in physics, chemistry and biochemistry, the
solvated electron in water has attracted much interest in order to determine its structure and
excited states. The solvated electrons in other solvents are less quantitatively known, and
much remains to be done, particularly with the theory. Likewise, although ultrafast dynamics
of the excess electron in liquid water and in a few alcohols have been extensively studied
over the past two decades, many questions concerning the mechanisms of localisation,
solvation, and thermalisation of the electron still remain. Quantum and molecular dynamics
simulations are necessary to unravel the structure of the solvated electron in many solvents
and to better understand its properties.
References
[1] Hart E.J., Boag J.W., Absorption spectrum of the hydrated electron in water and in aqueous
solutions, J. Am. Chem. Soc., 1962, 84, 4090.
[2] Dorfman L.M., Jou F.-Y., Optical absorption spectrum of the solvated electron in ethers and in binary
liquid systems, in "Electrons in fluids", Jortner J., Kestner N.R. (eds), Springer, New York, 1973,
Ä*%Ä
4HESOLVATEDELECTRON
447-459. Wishart J.F., Neta P, Spectrum and reactivity of the solvated electron in the ionic liquid
methyltributylammonium bis(trifluoromethylsulfonyl)imide, J. Phys. Chem B, 2003, 107, 7261-7267.
[3] Kraus C.A., Lucasse W.W., The conductance of concentrated solutions of sodium and potassium in
liquid ammonia, J. Am. Chem. Soc., 1921, 43, 2529-2539.
[4] Czapski G., Schwarz H.A., The nature of the reducing radical in water radiolysis, J. Phys. Chem.,
1962, 66, 471-474.
[5] Delaire, J.A., Delcourt, M.O., Belloni J., Mobilities of solvated electrons in polar solvents from
scavenging rate constants, J. Phys. Chem., 1980, 84, 1186-1189.
[6] Holroyd R.A., The electron: its properties and reactions, in “Radiation Chemistry, principles and
applications”, Farhaziz and M.A.J. Rodgers (eds), VCH, 1987, 201.
[7] Rossky P.J., Schnitker J., The hydrated electron : quantum simulation of structure, spectroscopy,
and dynamics, J. Phys. Chem., 1988, 92, 4277-4285.
[8] Jou F.-Y., Freeman G.R., Shape of optical spectra of solvated electrons. Effect of pressure,
J. Phys. Chem., 1977, 81, 909-915.
[9] Jou F.-Y., Freeman G.R., Temperature and isotope effects on the shape of the optical absorption
spectrum of solvated electrons in water, J. Phys. Chem., 1979, 83, 2383-2387.
[10] Nicolas C., Boutin A., Levy B., Borgis D., Molecular simulation of a hydrated electron at different
thermodynamic state points, J. Chem. Phys., 2003, 118, 9689-9696.
[11] Mostafavi M., Lin M., He H., Muroya Y., Katsumura Y., Temperature-dependent absorption spectra
of the solvated electron in ethylene glycol at 100 atm studied by pulse radiolysis from 296 to 598 K,
Chem. Phys. Lett., 2004, 384, 52-55.
[12] Lampre I., Lin M., He H., Zan Z., Mostafavi M., Katsumura Y., Temperature dependence of the
solvated electron absorption spectra in propanediols, Chem. Phys. Lett., 2005, 402, 192-196.
[13] Green N.J.B., Pilling M.J., Pimblott S.M., Clifford P., Stochastic modeling of fast kinetics in a radiation
track, J. Phys. Chem., 1990, 94, 251-258.
[14] Goulet T., Jay-Gerin J.-P., On the reactions of hydrated electrons with OHt and H3O. Analysis of
photoionization experiments, J. Chem. Phys., 1992, 96, 5076-5087.
[15] Crowell R.A., Bartels D.M., Multiphoton ionization of liquid water with 3.0-5.0 eV photons, J. Phys.
Chem., 1996, 100, 17940-17949.
[16] Thomsen C.L., Madsen D., Keiding S.R., Thögersen J., Christiansen O., Two-photon dissociation
and ionization of liquid water studied by femtosecond transient absorption spectroscopy,
J. Chem. Phys., 1999, 110, 3453-3462.
[17] Kloepfer J.A., Vilchiz V.H., Lenchenkov V.A., Germaine A.C., Bradforth S.E., The ejection distribution
of solvated electrons generated by the one-photodetachment of aqueous I− and two-photon
ionization of the solvent, J. Chem. Phys., 2000, 113, 6288-6307.
[18] Soroushian B., Lampre I., Pernot P., De Waele V., Pommeret S., Mostafavi M., Formation and
geminate recombination of solvated electron upon two-photon ionisation of ethylene glycol,
Chem. Phys. Lett., 2004, 394, 313-317.
Ä*&Ä
2!$)!4)/.#(%-)3429
[19] De Waele V., Sorgues S., Pernot P., Marignier J.-L., Monard H., Larbre J.-P., Mostafavi M., Geminate
recombination measurements of solvated electron in THF using laser-synchronized picosecond
electron pulse, Chem. Phys. Lett., 2006, 423, 30-34.
[20] Buxton G.V., Greenstock C.L., Critical review of rate constants for reactions of hydrated electrons,
hydrogen atoms and hydroxyl radicals in aqueous solution, J. Phys. Chem. Ref. Data, 1988, 17,
513-886.
[21] Mackenzie K., Kopinke F.-D., Remmler M., Reductive destruction of halogenated hydro-carbons in
liquids and solids with solvated electrons, Chemosphere, 1996, 33, 1495-1513.
[22] Sun G.-R., He J.-B., Pittman Jr, C.U., Destruction of halogenated hydrocarbons with solvated
electrons in the presence of water, Chemosphere, 2000, 41, 907-916.
[23] Baxendale J.H., Wardman P., Direct observation of solvation of the electron in liquid alcohols by
pulse radiolysis, Nature, 1971, 230, 449-450.
[24] Chase W.J., Hunt J.W., Solvation time of the electron in polar liquids. Water and alcohols,
J. Phys. Chem., 1975, 79, 2835-2845.
[25] Kenney-Wallace G.A., Jonah C.D., Picosecond spectroscopy and solvation clusters. The dynamics
of localizing electrons in polar fluids, J. Phys. Chem., 1982, 86, 2572-2586.
[26] Migus, A., Gauduel Y., Martin J.L., Antonetti A., Excess electrons in liquid water: first evidence of a
prehydrated state with femtosecond lifetime, Phys. Rev. Lett., 1987, 58, 1559-1562.
[27] Shi X., Long F.H., Lu H., Eisenthal K.B., Electron solvation in neat alcohols, J. Phys. Chem., 1995, 99,
6917-6922.
[28] Assel M., Laenen R., Laubereau A., Dynamics of excited solvated electrons in aqueous solution monitored
with femtosecond-time and polarization resolution, J. Phys. Chem. A, 1998, 102, 2256-2262.
[29] Hertwig A., Hippler H., Unterreiner A.-N., Transient spectra, formation and geminate recombination
of solvated electrons in pure water UV-photolysis: an alternative view, Phys. Chem. Chem. Phys.,
1999, 1, 5633-5642.
[30] Madsen D., Thomsen C.L., Thögersen J., Keiding S.R., Temperature dependent relaxation and
recombination dynamics of the hydrated electron, J. Chem. Phys., 2000, 113, 1126-1134.
[31] Pépin C., Goulet T., Houde D., Jay-Gerin J.-P., Femtosecond kinetic measurements of excess
electrons in methanol: substantiation for hybrid solvation mechanism, J. Phys. Chem., 1994, 98,
7009-7013.
[32] Soroushian B., Lampre I., Bonin J., Pernot P., Pommeret S., Mostafavi M., Solvation dynamics of the
electron produced by two-photon ionisation of liquid polyols. 1. Ethylene glycol, J. Phys. Chem. A,
2006, 110, 1705-1717.
[33] Lian R., Crowell R.A., Shkrob I.A., Solvation and thermalization of electrons generated by above-the-
gap (12.4 eV) two-photon ionization of liquid H2O and D2O, J. Phys. Chem. A, 2005, 109, 1510-1520.
[34] Nordlund D., Ogasawara H., Bluhm H., Takahashi O., Odelius M., Nagasono M., Petterson L.G.M.,
Nilsson A., Probing the electron delocalization in liquid water and ice at attosecond time scale,
Phys. Rev. Lett., 2007, 99, 217406.
Ä*'Ä
Chapter 4
Water radiolysis under
extreme conditions. Application
to the nuclear industry
Gérard BALDACCHINO and Bernard HICKEL
Introduction
Water radiolysis in the nuclear industry comes from the interaction of radiation
(alpha, beta, gamma, recoil nuclei) with enough energy to interact with liquid water. As
water is used as coolant in current Generation 2 and 3 nuclear reactors (PWR, BWR, EPR…),
water decomposition is expected. This phenomenon was already recognized a century
ago, and since then numerous chemists have succeeded in interpreting the physical and
chemical mechanisms involved (Chapter 1) [1]. Nevertheless, water radiolysis under extreme
conditions of temperature, pressure, pH and high Linear Energy Transfer (LET represents the
loss of energy per unit length, LET = –dE/dx) is not fully understood due to its inherent rich
and complex chemical reactivity.
In addition to understanding the interaction of radiation with water, the nuclear

industry must obviously also take into account the excess production of molecular hydrogen
and hydrogen peroxide, and control this excess in order to avoid explosive conditions and
corrosion of the water circuitries. Due to the working conditions of the current reactors
(T > 310 °C, P > 100 atm in Pressurized Water Reactor, PWR), it is mandatory to predict the
evolution of the chemistry when submitted to high temperature and pressure. Nevertheless,
a few experiments have shown that the linear Arrhenius law model is not applicable at
temperatures above 250 °C. Hydrogen production overestimates have been necessary in
Ä*(Ä
2!$)!4)/.#(%-)3429
order to render the safety control of these reactors as efficient as possible. Only in recent
years have experiments been performed at high temperature, and fast (picosecond)
processes after the ionization step have begun to be observed under almost supercritical
conditions (T > 374 °C, P > 21 MPa). Temperature is one of the parameters suspected to affect
the complex chemistry in water under radiation. Type and energy of the ionizing particles
(which determine the LET), pressure and pH, are also known to affect the fast processes
of recombination because they change the initial distribution and concentrations of the
reactive species in the medium. Additionally, extreme conditions are also encountered
in storage conditions of radioactive materials (high pH values, high pressure in porous
materials...)(Chapter 8). These factors mean the fundamental aspects of water radiolysis
represent a great and important challenge in the future of the nuclear energy industry.
The reactions of free radicals generated by the radiolysis of water (Chapter 1), such
as hydrated electrons, hydroxyl radicals, superoxide radicals, hydrogen atoms, are now
studied under extreme conditions using heavy ion beams of high LET generated by high
energy cyclotrons and a high temperature up to and over the supercritical state (T > 374 °C,
P > 21 MPa).
In future generations of nuclear reactors – especially supercritical water reactors

(SCWR), 4th generation nuclear reactors and the ITER project (International Thermonuclear
Experimental Reactor) – water should still be considered as a suitable coolant fluid, but it will
be submitted to more extreme conditions of temperature and LET (high flux of neutrons).
All contemporary studies show that it will be beyond reach to extrapolate the existing
simulations to these new conditions without experimental determinations of essential
parameters such as radiolytic yields and rate constants.
This chapter will focus on the influence of two essential parameters on the chemical
mechanism of water under radiation : Linear Energy Transfer and temperature.
Influence of the type of radiation
In contrast to the situation that prevails in photochemistry, there is no compound

transparent to high energy radiation. In water for example, the LET of high energy photons
(such as gamma rays) is 0.2 keV per micrometer, while for visible light, it would be zero. In
dilute aqueous solutions, the chemical changes observed on the solute are due indirectly to
radicals generated from the solvent water, and not due to the direct radiation effect on the
solute (Chapter 1).
Ä*)Ä
7ATERRADIOLYSISUNDEREXTREMECONDITIONS
The LET value is used to calculate the energy delivered to the material (the dose)
but a similar value of LET can be attributed to different types of particles at different values
of energy, with various resultant chemical effects. In order to identify the production yield of
free radicals or molecules related to the type of ionizing particle (gamma rays, alpha rays or
high energy heavy ions), other parameters should be used, such as the specific energy and
the charge of the particle. The geometrical structure of the track takes greater and greater
importance in the interpretation of the experimental and simulation results (Fig. 1 gives
an example of a track structure for an alpha ray of a few MeV). It has also been proposed
that the initial energy density transferred to the matter might be an adequate parameter
to predict the concentration level of molecular species (H2 for example), as shown in recent
publications based on pulse radiolysis experiments (Fig. 2 is an example of the experimental
set-up) [2].
Figure 1 : Simulated ionization track due to an alpha ray of a few MeV of energy coming from the left side in
the «blue» water continuum. Each red circle is an ionization event. This local distribution, in the nanometer ran-
ge, is the beginning of a complex chemistry. Ionizations occur mainly around the trajectory axis of this incident
ion, and this area is named “core track”. Some high energy electrons can be ejected and they can form their own
track named “delta ray”. When delta rays are sufficiently numerous (that depends on the incident ion energy and
charge) a new area around the core can be named “penumbra”. The penumbra has the characteristics structure
of a “low LET area” because the ionizations are produced by high enrgy electrons.
Decomposition and relaxation
When radiation interacts with matter, the energy involved is much higher than the
bond energy of the molecules. Intuitively one would thus expect all chemical bonds to be
Ä**Ä
2!$)!4)/.#(%-)3429
Figure 2 : Picture of the experimental set-up using a red laser light to detect the hydrated electron in a flow
cell by using the time-resolved absorption spectroscopy. High energy heavy ions were provided by the cyclotron
GANIL, Caen, France. The safety protections against the laser light were taken off for the photography needing.
immediately broken. This is however a common misconception since the duration of the
interaction is less than 10−15 s and within this time span only ionizations and electronic
excitations can occur. The bond scissions take place at a later stage during the relaxation
of the system. All liquids are expected to be partly decomposed by radiation, and indeed
they are, except some exotic species such as liquid sodium (where there is no bond to be
broken after ionization) which is used to cool the breeder reactors. In the case of liquid
hydrogen fluoride, the resultant compounds of decomposition (H2 and F2) probably react
spontaneously to return to the state of HF.
Radiolysis of water : a puzzling case
The case of water decomposition under radiation is rather complex (Chapter 1). It
has been known for a long time that, when carefully degassed and irradiated with low LET
radiation such as X-rays or 60Co gamma rays, pure water gives rise to almost no products,
the amount of which remains at a very low steady state concentration. It seems that water
decomposes at the very beginning of the irradiation and then becomes inert. But for
Ä*+Ä
high LET radiation such as alpha rays, water decomposition occurs with the formation of
molecular hydrogen, hydrogen peroxide and molecular oxygen. When the water contains air
or other solutes, water decomposition is always observed whatever the LET. In the presence
of dissolved H2, water decomposition is completely suppressed at low LET, but still occurs
under alpha rays irradiation.
The explanation of this behavior, which has paramount importance to the nuclear
industry, was given by Allen [3]. After World War II, Allen depicted a model of water
decomposition under radiation that considers the production and consumption of H2. The
key role of OHt, H2 and O2 involved in the chain as a carrier or a breaker is clear (see Inset ).
Within this chain reaction, the reaction between H2 and O2 (which is thermodynamically
favorable) takes place in water at high temperature only in the presence of a catalyst such
as copper or silver cations. In the radiolysis of water, the reaction can take place at room
temperature in the presence of free radicals which form the molecular products H2, H2O2 and
O2 at the first step inside the tracks or the spurs. Subsequently in the bulk of the solution,
the free radicals which have escaped recombination in the tracks recombine as molecular
products into water. The molecular products are formed in the nanosecond range and their
recombination takes place in the millisecond range.
Inset : The series of reactions involved in the chain reaction

depicted by Allen
First note that the global water radiolysis reaction can be written :
radiation
H 2O Ht, OHt, eaq–, H2, H2O2)02t )3O (0)
We observe that O2 is not the primary species in the water radiolysis ; it is formed by secondary
reactions. The key reaction which explains the unique role of H2 is :
OHt)2 ¬ H20)t (1)
This reaction converts an oxidizing radical OHt into a powerful reducing species, the Ht atom.
The first step is the reduction of O2 in H2O2
Ht02 ¬ HO2t (2)

HO2t )02t ¬ H2O202 (3)
Ä*,Ä
2!$)!4)/.#(%-)3429
The second step is a chain reaction between Ht and H2O2
Ht)2O2 ¬ OHt)2O (4)
followed by reaction 1 which gives back an Ht atom.

Whilst O2 is still present, the chain reaction does not take place, because the rate constant of
reaction 2, which is a chain breaker, is 200 times larger than the rate constant of reaction 4
which is a chain carrier.
There is another chain termination which is important to explain the paradoxical behavior of
water radiolysis.
OHt)2O2 ¬ Ht)02t (5)
We can see that H2O2 is both a chain carrier (reaction 4) and a chain breaker (reaction 5). When
its concentration increases too much it prevents the chain reaction by forming O2 (reaction 5
followed by reaction 3).
The rate constants k and k have the same order of magnitude, so it is necessary to add a
5 1
sufficient concentration of H2 to insure that reaction 1 is always faster than reaction 5.
The net results of these reactions is
O2)2)t ¬ 2H2O (6)
Reaction 6 shows that we need at least two Ht radicals for one H2 and O2.
At low LET radiation like gamma and X-rays, this is easily achieved and the water behaves as
if it was completely inert under irradiation. At high LET (alpha rays) the ratio of the molecular
products over the free radicals increases and there are comparatively too few Ht to O2 atoms,
The recombination of H2 and O2 is thus incomplete, the chain reaction between H2 and H2O2
cannot start and water decomposition is observed.
This set of reactions also clearly explains why water radiolysis takes place when some solutes
are present. Most of the chemical species react with the Ht atom or the OHt radical and prevent
the chain reaction.
The special case of HOt2
The particular case of superoxide radical (HOt2/O2t −) in water radiolysis comes from
the property that its formation yield increases with LET, a behavior contrary to fast and
natural recombination of other radicals in dense ionization tracks [4]. Logically, this increased
Ä*-Ä
radical recombination rate accelerates the production of molecular species (H2 and H2O2)
and decelerates the production of radicals. The superoxide species is probably formed in
the track-core. However, it can easily escape the track to survive in the water bulk since its
reactivity in water is low, mostly by a slow disproportionation (reaction 7),
H2O
HOt2 0t2– H2O2020)– k = 9.7 x 107 l mol−1 s–1 (7)
This reaction has several consequences in radiobiology concerning the sensitization

of living cells submitted to heavy ion beams for radiotherapy (hadrontherapy). The efficiency
of hadrontherapy with carbon ions mainly results from the property of local energy deposition
in the Bragg peak region but can also be the consequence of molecular oxygen generation
(reaction 7) in tumors which are known to be hypoxic cells. Some studies have tried to explain
how the superoxide radical is formed in the tracks of heavy ions, and have proposed several
possible mechanisms. A high local concentration of ionizations can favor reactions which can
then interfere with geminate recombination. The production of Ot atoms has for a long time been
suspected in the core track. Multiple ionizations of water molecules can be a source of Ot atoms
because the huge energy deposited in the medium is considerably greater than the total
energy needed to ionize the total number of water molecules along the ion track. This model
has recently been exploited in Monte Carlo simulations [5] for which Figure 3 presents a
comparison between experimental and calculation results. There is a fair agreement between
the few experimental results concerning HO2t/O2t¦ and H2O2. Nevertheless, more exhaustive
experiments are necessary to validate this model for values of LET higher than 1000 eV/nm.
Figure 3 : Radiolytic yields of HO2t

as a function of LET. Experimental
data include the yield of O2 due to
a particular experimental method
for determining the yield. Carbon
ions (black line) had various energies
and then various LET values [4]. The
red lines are from recent simulation
results for HO2t + O2 [5] ; the red dash-
line represents only HO2t for which
two experimental data are available
(red symbols obtained for Argon and
Sulfur high energy ions, see [2] and
references therein).
Ä*.Ä
2!$)!4)/.#(%-)3429
Application to nuclear reactors
In Pressurized Water Reactors (PWR) which are the type of reactors used by EDF
for electricity production in France, it is important to avoid the formation of O2 (and H2O2)
arising from water radiolysis, since this leads to corrosion of the primary loop and increases
the radioactivity of water. Minimization is achieved by adding H2 and removing O2. The other
additives, boric acid and LiOH, used to absorb neutrons and regulate the water pH, do not
react with the free radicals and cannot interfere with the chain reaction. The system works
well owing to the intense flux of gamma radiation (low LET). The other types of radiation,
fast neutrons and especially the alpha particles generated by the nuclear reaction between
a thermal neutron and boron
10
#Om 7-J4He (4He has an energy of 2.2 MeV) (8)
have a much higher LET, and in the absence of gamma flux the water will be decomposed.
This phenomenon is clearly visible in Figure 4. Water decomposition begins at a particular
threshold of A/G ratio. At a A/G ratio of 8, a sharp change in the chemistry between alpha-ray
radiolysis and gamma ray radiolysis is observed [6].
Figure 4 : Water decomposition yield, G(-H2O), as a function of the flux ratio alpha /G irradiation (mixture of
high and low LET particles). Increasing this ratio means increasing the LET value. The observed decomposition
threshold is shifted to the right (higher LET) and disappears betwen 30 °C and 200 °C. The scale corresponds also
to the concentration in 10B. The difference between circle and triangle symbols is the initial enrichment in 10B in
the chemicals. Adding 10B that absorbs neutrons under neutron/gamma irradiation produces alpha rays having
high LET (200 eV/nm) [6].
Ä+%Ä
Water in the ITER reactor
A similar problem occurs for the future fusion reactor prototype, ITER, which is to
be built in Cadarache, France (Project view in Figure 5). The first reactor wall will be cooled
by water and again it is mandatory to avoid the formation of O2 and H2O2. The nature of the
radiation is different from that of a fission reactor and the radiolysis will be more severe, since
90% of the irradiation comes from high LET radiation (the 14 MeV fusion neutrons) and only
10% from low LET radiation (the gamma flux). We have performed experiments under a mixture
of low and high LET radiation types to check whether the water radiolysis remains inhibited
under the ITER conditions. By continuously varying the ratio of low to high LET radiation, we
have observed a threshold at which water starts to abruptly decompose (Fig. 4). The position
of the threshold depends on the temperature (Fig. 4, at 200 °C the decomposition disappears).
Other conditions influence the water decomposition: LET, dose rate, temperature and the
presence of impurities in water, such as copper cations which shift the threshold toward lower
LET. In the ITER reactor, the irradiation conditions are fixed, and the only variable parameters
are the concentrations of hydrogen and impurities in the cooling water.
Figure 5 : View of the ITER project. Water circuity will be submitted to a high flux of 14 MeV neutron and G-rays.
These radiations come from the fusion reaction in the plasma generated in the tokomak.
Ä+&Ä
2!$)!4)/.#(%-)3429
We have shown that impurities such as copper cations have a great influence on the
decomposition as it is presented in Figure 6 by the production of hydrogen peroxide.
Figure 6 : Influence of the dose rate and the copper cation concentration under irradiation with 1 GeV C6+
(LET = 26 eV/nm) at GANIL. O : [Cu2+] = 0 ; Q : [Cu2+] = 1 μmol l–1 ; V : [Cu2+] = 6 μmol l–1.
Towards supercritical conditions
By increasing the temperature, the rate constants of the reactions occurring in

the nanosecond/microsecond stage change according to the Arrhenius law up to 200-
300 °C. The radiolytic yields of the radical species are similarly increasing with temperature.
Recent publications tend to show that this increase is due to the acceleration of diffusion
with temperature of the free radicals in the heterogeneous step in the pico-nanosecond
range following ionization. Figure 7 shows the main result of the single picosecond high
temperature experiment yet performed. These experimental validations were obtained by
pulse radiolysis (Chapter 1), using picosecond pulses of high energy electrons, delivered by
the ELYSE picosecond electron accelerator in University of Paris-Sud, Orsay [7]. Below 300 °C
the Arrhenius plot becomes non linear and the rate constants are no longer predictable. The
challenge in 4th Generation nuclear plants of the SCWR project (Supercritical Water Reactor,
see the scheme of the project in Figure 8) is to understand the chemical mechanisms in
water irradiated under the extreme conditions of supercritical water (T > 374 °C, P > 21 MPa).
Recent works presenting experimental results show a strong pressure dependence of the
radiolytic yields [8]. We do not yet clearly understand whether this dependence is related to
the change of the density of the medium or to the pressure activation of reactions.
Ä+'Ä
Figure 7 : Picosecond kinetics of hydrated electron recombination in the ionization spurs as a function of the tem-
perature [7]. Hydrated electron is produced in 15 ps pulse of 8 MeV electrons delivered by the electron accelerator ELYSE.
The absorbance of 2 cm of pure water is analyzed by a laser at 790 nm. Due to the red-shift of the hydrated electron
spectrum with increasing temperature, the absorbance maximum value at 790 nm decreases from 23 ° C to 350 ° C. In the
time range of 3 ns, the recombination of hydrated electron appears less and less efficient with increasing temperature.
Figure 8 : Scheme of the SCWR project. Supercritical water at about 510 °C and 25 MPa will be used as the
coolant material.
Ä+(Ä
2!$)!4)/.#(%-)3429
Challenge for the future
The challenge for the future is to explain radiation-matter interactions under a

wide range of conditions of temperature, pressure, LET and pH. This will be approached
by simulating and comparing interactions with experimental time-resolved results in the
picosecond range. The realization of this challenge requires the development of increasingly
sensitive detection methods and increasingly accurate simulations. It should be noted that,
besides the nuclear industry which stands to benefit directly from the application of this
research, the fields of radiobiology, medicine and astrobiology are also highly concerned by
a better understanding of the mechanisms of this exotic chemistry.
References
[1] Mozumder A., Hatano Y. (eds), “Charged particle and photon interactions with matter. Chemical,
physicochemical, and biological consequences with applications”, Marcel Dekker, New York, 2004.
[2] Baldacchino G., Vigneron G., Renault J.-P., Pin S., Abedinzadeh Z., Deycard S., Balanzat E., Bouffard S.,
Gardès-Albert M., Hickel B., Mialocq J.-C., A nanosecond pulse radiolysis study of the hydrated electron
with high energy ions with a narrow velocity distribution, Chem. Phys. Lett., 2004, 385, 66-71.
[3] Allen A.O., “The radiation chemistry of water and aqueous solutions“, Van Nostrand, New York, 1961.
[4] LaVerne J.A., Radiation chemical affects of heavy ions, in “Charged particle and photon interactions
with matter. Chemical, physicochemical, and biological consequences with applications”,
Mozumder A., Hatano Y. (eds), Marcel Dekker, New York, 2004, 403-429.
[5] Gervais B., Beuve M., Olivera G.H., Galassi M.E., Rivarola R.D., Production of HO2 and O2− by multiple
ionization in water radiolysis by swift carbon ions, Chem. Phys. Lett., 2005, 410, 330-334.
[6] Pastina B., Isabey J., Hickel B., Radiolyse de l’eau dans les réacteurs nucléaires à eau pressurisée.
Effet de l’hydrogène, J. Chim. Phys., 1997, 94, 226-229.
[7] Baldacchino G., De Waele V., Monard H., Sorgues S., Gobert F., Larbre J.P., Vigneron G., Marignier
J.L., Pommeret S., Mostafavi M., Hydrated electron decay measurements with picosecond pulse
radiolysis at elevated temperatures up to 350 °C, Chem. Phys. Lett., 2006, 424, 77-81.
[8] Lin M., Katsumura Y., Muroya Y., He H., Wu G., Han Z., Miyazaki T., Kudo H., Pulse radiolysis study
on the estimation of radiolytic yields of water decomposition products in high-temperature and
supercritical water : use of methyl viologen as a scavenger, J. Phys. Chem. A, 2004, 108, 8287-8295.
Ä+)Ä
Part II
Radiation chemistry
mechanisms
and applications
Chapter 5
Molecular formation
in the interstellar medium
Nigel J. MASON, Anita DAWES and Philip HOLTOM
Introduction
Ever since the development of consciousness human beings have been asking
themselves that most basic question : where they came from ? The “answer” to this question
has been a fundamental tenant of most religions, with each having a “genesis” story to
explain how humanity came into being. With the emergence of science and our ability to
explore our own solar system and look out into the universe (Fig. 1) the question of whether
we are “alone” or if life has developed elsewhere has intrigued scientists and the public alike,
with science fiction becoming one of the most popular genres of both literature and film.
Hence, one of the greatest challenges of modern science is to determine how life began on
Earth and whether the conditions for life to appear and develop are common across the
Universe. Indeed, the discovery of life (even in its simplest form) on a planet other than Earth
would be a defining moment in human history, while understanding the origins of life itself
would be a crowning achievement of human scientific endeavour.
Today, for the first time, it is possible to develop a realistic, rigorous scientific
programme to address these questions and begin to provide answers to these fundamental
questions. Such a research programme has led to the development of a new scientific
discipline, astrobiology, a necessarily interdisciplinary field of research bringing together
chemists, physicists, biologists, as well as astronomers, geologists, engineers and even
philosophers all with the common aim of both determining the origins of life and developing
instrumentation to explore our own solar system and beyond for evidence of life.
Ä+,Ä
2!$)!4)/.#(%-)3429
Figure 1 : The Heart and Soul Nebulas. Is the heart and soul of our Galaxy located in Cassiopeia? Possibly not,
but that is where two bright emission nebulas nicknamed “Heart” and "Soul” can be found. The Heart Nebula,
officially IC 1805 on the right, has a shape reminiscent of a heart . The nebulas are brightly coloured due to red
light emitted from excited hydrogen gas. Deep in the nebulae forming from the dust and gas several clusters of
young stars are visible. Such nebulae are also the chemical factories of many molecules including perhaps the
precursors of life itself. [Credit : Richard Powell, Digitized Sky Survey, Palomar Observatory, STScI]
The Origins of life
The fundamental laws of chemistry are believed to be universal, that is they are
the same in any part of the universe. Thus in seeking to determine whether life has evolved
elsewhere in the Universe we must seek to understand the chemical origins of life. The
chemical ingredients of life are well known, the formation of self replicating molecules
capable of expressing a genetic code (e.g. DNA) and molecules necessary for the containment
of such molecules in a single entity, e.g. lipids forming the membrane of a cell. In addition,
since life needs a medium in which to exchange chemicals, the presence of water is essential
as a “universal solvent”. The mechanisms by which such “prebiotic” molecules are formed is
Ä+-Ä
-OLECULARFORMATIONINTHEINTERSTELLARMEDIUM
the topic of the rest of this brief review. Our understanding of how such prebiotic molecules
were subsequently assembled on Earth to form the larger more complex cellular structures
remains unknown with several theories having been developed but no experimental
evidence to support any of these hypothesis has yet been provided. Indeed we remain a
long way from being able to “create life in the lab”.
In exploring the origins of life, we are faced with one of two possibilities : either
the molecules of life were assembled on Earth or the building blocks of life were formed
as part of the more general astrochemistry that gave birth to the solar system. The former
leads to a more “Earth Centric” view of the universe suggesting that the Earth fulfilled some
rather special criteria to allow molecules to assemble while the latter is more conducive to
life being a universal phenomenon equally probable in any other solar system.
Knowing that organic molecules and microorganisms can be preserved as fossils,

a record of the appearance of life on Earth and its early evolution, as well as the biosphere
environment, should be preserved in rocks dating from the earliest epochs. The most salient
problem, however, is to find such rocks. There are no rocks dating from the first 500 million
years of Earth history because they have been destroyed by plate tectonics and other
thermal processes. However there are examples of very well preserved rock formations e.g.
3.5 Ga-old in Australia and South Africa and the 3.8 Ga-old rocks of Isua, Greenland. These
rocks, formed just one billion years (1 Ga) after the consolidation of the Earth, are believed
to contain the fossilized remains of microorganisms, their colonies, biofilms and/or the
geochemical signatures left by their metabolic activities.
Such fossil evidence itself provides scientists with a major challenge since it suggests
that life was able to emerge very quickly from the earliest chemical conditions. Thus if the
ingredients of life were formed on the Earth the chemical processes must have “simple” and
capable of being sustained in what we believe to have been a very inhospitable regime, with
no oxygen atmosphere and hence no ozone layer to protect the surface from biologically
harmful UV irradiation, a surface prone to impact by meteorites and with active volcanism.
Alternatively, the basic ingredients of life might have appeared formed as a result of the
chemical processes that established our solar system and that are inherent in the formation
of planetary systems.
In the 1950s in a pioneering experiment (Fig. 2) Stanley Miller and his student
Harold Urey suggested that the molecules of life might have been formed in the early Earth’s
atmosphere. Placing a mixture of methane (CH4), ammonia (NH3), hydrogen (H2), and water
(H2O) in a glass bulb – a mixture they thought was a good representation of the chemical
constituents of the Earth’s early atmosphere – they ran a continuous electric current through
Ä+.Ä
2!$)!4)/.#(%-)3429
the system to simulate lightning storms,

believed to be common on the early Earth.
Analysis of the products of the discharged
gas was performed using chromatography.
After running the discharge for a week 10-
15% of the carbon had been converted into
new organic compounds including some
2% transformed into amino acids which in
biology may be used to make proteins. This
experiment demonstrated conclusively that
molecular compounds, such as amino acids,
which are essential to cellular life, could be
made relatively easily by providing energy to
a “chemical soup”.
However we now know that the

Urey-Miller experiment is not in fact a good
representation of the conditions of the
early Earth since it is now believed that the
early Earth’s atmosphere was not reducing.
Instead, scientists believe the atmosphere
was full of oxidants, such as CO2, and N2. An
oxidizing atmosphere is essentially neutral,
and does not permit organic chemistry to
occur. Furthermore while it is believed lightning
storms were extremely common on the
primitive Earth, they were not continuous
as the Miller-Urey experiment portrayed.
Thus it has been argued that, while amino Figure 2 : The Urey-Miller Experiment. The glass
acids and other organic compounds may vessel to the lower left of the picture is used for the
have been formed, they would not have water reservoir and mimics oceans on the early Earth.
been formed in the amounts which this An electrical discharge of a gas mixture mimicking
experiment produced. Hence, the time ranges the composition of the early Earth’s atmosphere ta-
necessary for the development of sufficient kes place in the large bulb at the top right. Molecules
formed in the discharge may be condensed and are
concentrations of such prebiotic molecules
collected at the bottom of the system.
are too long to explain the emergence of life
as witnessed by the fossil record – indeed some estimates suggest that, if the Urey-Miller
hypothesis was to be adopted, then we would still be waiting for life to evolve on Earth.
Ä,%Ä
So if these compounds were not created in a reducing atmosphere here on Earth

as Miller and Urey suggested, then where did they actually come from ? Perhaps they are
formed in space itself as part of the process by which planets and stars are formed?
The Interstellar medium as a chemical factory
The interstellar medium (ISM) is the region of space between the stars (Fig. 1) that
appears dark since gas and dust in these regions absorb the light from stars behind them.
However, these regions still have very low densities (they constitute a vacuum far better than can
be produced artificially on the surface of the Earth). From recent astronomical observations we
now know that the interstellar medium is rich in molecules (Table 1) including vinegar (acetic
acid), benzene and even simple sugars such as glycolaldehyde. Larger more complex organics,
such as the simple amino acid glycine, may also be present but require the development of
better instrumentation before they can be detected. The ISM is therefore a rich depository of
prebiotic molecules and it is from the ISM that stars (and the planets) are formed.
.
Table 1 Molecules detected in the Interstellar medium.
2 atoms SiN HNC C2H2· HC3N 7 atoms (CH3)2O
H2 SiO HNO CH2D ? HC2NC C6H CH3CH2OH
AlF SiS MgCN HCCN HCOOH CH2CHCN HC7N
AlCl CS MgNC HCNH H2CNH CH3C2H C8H
C2·· HF N2H HNCO H2C2O HC5N 10 atoms
CH SH· N 2O HNCS H2NCN HCOCH3 CH3C5N (?)
CH HD NaCN HOCO HNC3 NH2CH3 (CH3)2CO
CN FeO ? OCS H2CO SiH4· c-C2H4O (CH2OH)2 ?
H2NCH2-
CO O2 ? SO2 H2CN H2COH H2CCHOH COOH
Glycine ?
$0 3 atoms c-SiC2 H2CS 6 atoms 8 atoms CH3CH2CHO
CP C3· CO2· H3O C5 H CH3C3N 11 atoms
CSi C 2H NH2 NH3 l-H2C4 HCOOCH3 HC9N
HCl C2O H· c-SiC3 C2H4· CH3COOH 12 atoms
·
KCl C 2S H2D, HD CH3· CH3CN C7H C6H6 (?)
NH CH2 SiCN 5 atoms CH3NC H 2 C6 13 atoms
NO HCN AlNC C 5· CH3OH CH2OHCHO HC11N
NS HCO 4 atoms C 4H CH3SH l-HC H· (?)
6
NaCl HCO c-C3H C4Si HC3NH CH2CHCHO (?)

OH HCS l-C3H l-C3H2 HC2CHO 9 atoms
PN HOC C3N c-C3H2 NH2CHO CH3C4H
SO H2O C 3O CH2CN C5N CH3CH2CN
SO H2S C 3S CH4· l-HC4H*···· (?)
Underlined molecules are solid. AINC : Aluminium Chloride. Those molecules marked with « ? » indicate molecules whose obervation is still to
be confirmed.
Ä,&Ä
2!$)!4)/.#(%-)3429
The suggestion that the molecular building blocks of life could be formed in space
is intriguing since such regions would seem to be rather unlikely places for the development
of chemistry. The ISM is cold (temperatures of 10-30 K) and “empty” with pressures of less
than 10−12 torr such that the probability for a collision between two compounds is low and,
at such low temperatures, the “reaction rate” would be expected to be very low (hence in
most industrial chemistry the reactants are heated to increase their reactivity). Nevertheless
the detection of such molecules within the ISM makes it clear that these are chemically active
zones. The solution to this apparent paradox is that the chemistry in the ISM is somewhat
different from the conventional chemistry we observe on Earth, much of it being induced by
radiation. The ISM contains several different sources of radiation, namely ;
t-JHIU (produced from stars) of which the ultraviolet is the most important for
inducing chemistry .
tCosmic rays – high energy ions about 89% of which are protons, 10% helium nuclei
(alpha particles) and about 1% ions of the heavier elements. It is believed that
most galactic cosmic rays derive their energy from supernova explosions, which
occur approximately once every 50 years in our Galaxy.
t5IFSFBSFBMTPTPVSDFTPGgamma and X-rays, the latter often being a signature of
neutron stars.
Such radiation sources all have sufficient energy to break the chemical bonds
of molecules in the ISM and thence produce both reactive radicals and ions capable of
inducing further chemistry. In the gas phase much of the chemistry in the ISM is driven by
ion-molecule reactions (Fig. 3). Such reactions are barrierless that is they require no energy
to start the reaction rather once the reactants are brought together the reaction appears
to occur spontaneously. Such barrierless reactions are also prevalent if one of the reacting
species is a free radical (e.g. the hydroxyl radical tOH). Such reactions can therefore occur at
low temperatures, indeed it has been noted that the reaction rate may actually increase at
low temperatures.
Despite the success of ion-molecule and radical chemistry in explaining some of

the molecular formation in the ISM, the wealth of molecules formed and the time scales
upon which they develop can not be explained by gas phase alone. For example molecular
hydrogen is the most abundant molecule in the ISM but there is no gas phase reaction
scheme that can explain its formation. Furthermore, although complex hydrocarbons can be
formed in such reactions, they may equally be destroyed by the radiation field. Indeed the
more complex the molecular species is, often the more prone to destruction by ultraviolet
light and cosmic rays it turns out to be.
Ä,'Ä
Figure 3 : Schematic of ion molecule reactions induced by radiation in the interstellar medium. "e" indicates
an electron-induced reaction.
Thus, it is necessary to consider the role of the dust in the ISM. Dust comprises about
1% of the mass of the material in the ISM and can act as a surface upon which chemistry can
occur (Fig. 4). Although we have yet to actually examine a piece of interstellar dust we believe
it is similar to that found in the Solar System and recently collected by the Stardust mission.
The dust is either carbonaceous or silicate in nature, comprising of small particles, typically
sub-microns in size, probably with an irregular (fractal?) structure. Being so cold (around 10 K)
the dust grains act as a depository for any gaseous molecules which “stick” to the surface.
Hence Ht atoms may collide with the surface, and subsequent reaction between such Ht
Ä,(Ä
2!$)!4)/.#(%-)3429
atoms is now believed to be the main mechanism by which molecular hydrogen is formed in
the ISM, the product H2 being desorbed back into the ISM as a result of subsequent heating
of the grains (e.g. by grain-grain collisions). Other gaseous atomic and molecular species in
the ISM may also collide with the dust slowly forming a complex multi component ice layer.
It is this ice that provides the “high density” target within which chemistry may be induced
by interstellar radiation allowing more complex molecules to be formed (Fig. 4).
Figure 4 : Molecular formation on interstellar dust grains.1) An ice layer is formed by accretion onto dust grains.
2) The icy layer is exposed to irradiation from the ISM. 3) Further ice is accreted on top of the irradiated layers.
4) Icy material and molecules are ejected from the ice into the gas phase through heating events such as shocks
and grain–grain collisions.
Molecular formation in astrochemical ices
Experimental studies of the formation of molecules within astrochemical ices are

now being conducted in several international laboratories. A typical apparatus is shown in
Figure 5. An ultrahigh vacuum chamber is required (typically less than 10−9 torr) to mimic
the ultralow pressures found in space within which is mounted a cryostat, cooled by liquid
helium to temperatures typical of the ISM (10 K). Gases are then introduced into the vacuum
Ä,)Ä
chamber and deposited on a sample mounted off the cryostat. Irradiation of the ice may
be provided by a variety of sources : ultraviolet light may be produced by ultraviolet lamps
or by synchrotron radiation (such as that to be provided by the French Soleil synchrotron ;
synchrotrons are also a bright source of X-rays (e.g. the ESRF at Grenoble) while cosmic
rays may be produced by ion accelerators (e.g. Van de Graaff accelerators). After irradiation
the products of the subsequent chemistry may be detected using a variety of techniques
including : (i) infrared spectroscopy, which can detect molecules in the ice itself and (ii) mass
spectrometry heating the ice after irradiation until the products are desorbed.
Figure 5 : A typical experimental apparatus used to explore astrochemistry, comprising an ultrahigh vacuum
chamber to mimic low pressures in space, a cryostat to prepare surfaces at very low temperatures, an irra-
diation source to provide energy to induce chemistry and a detection system (e.g. Fourier transform infrared
spectroscopy or FTIR) to determine what molecules are formed in the astrochemical processes.
Ä,*Ä
2!$)!4)/.#(%-)3429
Recent research has shown that the structure of the ice (morphology) depends not
only on the deposition temperature but also on the deposition time with slow deposition
forming crystalline structures at temperatures where during fast deposition amorphous
structures are produced. Ice morphology plays a vital role in the chemistry that can occur after
irradiation of the ice mixture. A densely packed ice may slow migration of radicals through the
ice whilst an ice full of pores and voids may allow faster diffusion and provide a larger surface
area upon which reactions can occur. At present we do not know the morphology of the ice
on an interstellar dust grain and this highlights the difficulty in undertaking experiments
that aim to mimic conditions in the ISM. However, it should be remembered that in space
deposition rates are very slow – a few molecules per year or less – far slower than any rates
we can achieve in the laboratory. Similarly, in all the laboratory experiments undertaken to
date, the bulk ice surfaces used are much larger than those found in space being > 1 cm2,
whilst the typical area on a dust grain will only be a few μm2 such that thermal effects and
the role of surface compared to bulk chemistry may be quite different. Nevertheless using
such experimental systems made it possible to demonstrate that in the ISM irradiated ices
may induce sufficient chemistry to form complex molecules that can form the “building
blocks of life”.
In 2002 two different research groups, one at the NASA AMES laboratory in the USA
(M P. Berstein et al.), the other at Leiden University in the Netherlands (G.M. Muñoz-Caro et al.),
studied the irradiation of a mixture of ices using UV light. The NASA Ames group used a
20:2:1:1 ice mixture of H2O: CH3OH:NH3:HCN, the Leiden group a 2:1:1:1:1 ice mixture of
H2O:CH3OH:NH3:CO:CO2. In both cases, after several hours of irradiation at temperatures
of 10 K, amino acids, 6 of which are protein constituents, glycine, alanine and serine were
formed. More recently experiments using electrons to irradiate simpler binary mixtures of
CO2 and CH3CN (methylamine) or CO2/NH3 have also shown that glycine is readily formed
in simple astrochemical ices (P.D. Holtom et al.). These experiments demonstrate that the
formation of amino acids in the cold regions of space is not only possible but may occur
with a high probability. Thus the ISM may form a large depository of molecules that may
subsequently be assembled to form the building blocks of life such as the DNA bases.
Meteorites and panspermia
Demonstrating that it is possible to form molecules in the ISM that may subsequently
assemble to form the building blocks of life on Earth (or any other planet) is not in itself
evidence that the origins of life lie in the ISM. It is also necessary to know how such molecules
might be transported to the planet. Our knowledge of how stars and planets form relies on
physical/chemical models but we believe we know the basic mechanisms. Within the ISM
Ä,+Ä
some disturbance (e.g. a shock wave from a nearby exploding star) triggers the collapse of
the dust cloud and its surrounding gases to form a nebula within which dust coalesces and
eventually a star is born. Close to this new Sun solar heat may vaporise ices and prevent
lightweight elements, like hydrogen and helium, from condensing but further from the Sun
it may still be cold enough for ices to remain intact in the form of comets or meteorites.
Hence comets and meteorites may be a repository of molecules formed in the ISM capable of
then impacting planets and depositing their “prebiotic” material on an early Earth. Evidence
for this is found in the chemical composition of meteorites found on Earth. On September
28, 1969, a meteorite fell over Murchison, in the state of Victoria, Australia (Fig. 6). While only
100 kilograms were recovered, analysis of the meteorite has shown that it is rich with amino
acids. Indeed over 90 amino acids have been identified by researchers to date. Nineteen of
these amino acids are used in biological processes on Earth. The Murchison meteorite (and
others) has therefore demonstrated that the Earth may have acquired some of its prebiotic
material by “planetary infall”, a hypothesis that is often described as “panspermia”.
Figure 6 : The Murchison meteorite which landed at Murchison, Australia on September 28, 1969. Over 100
kilograms of this meteorite have been collected. Classified as a carbonaceous chondrite, type II (CM2), this me-
teorite is likely to have come from a comet. More than 92 different amino acids have been identified within the
Murchison meteorite, nineteen of these are also found on Earth such that many believe that such meteorites
provided the ‘seeds’ of life on the early Earth.
Conclusion
How life evolved on Earth and whether life can (and has) evolved elsewhere in
the universe remains one of the greatest unanswered questions of modern science. Our
understanding of how irradiation in the interstellar medium can induce chemistry in the
Ä,,Ä
2!$)!4)/.#(%-)3429
most inhospitable regions of the universe and form complex molecules is an essential first
step into our understanding where the molecules of life were (are) formed. A combination of
experiment, modelling and astronomical observations provides us with much information
on the chemical processes in the ISM, and suggests that life's molecular ingredients may be
formed throughout the universe. What remains unknown is how these prebiotic molecules
(amino acids, hydrocarbons, sugars) were assembled to form proteins, the nucleotide
bases – guanine adenine, cytosine, thymine and uracil – and thereafter the self replicating
molecules RNA and DNA. This represents today's great challenge for those interdisciplinary
scientists who are developing the new research field of astrobiology.
Acknowledgements
The authors would like to take this opportunity to thank several research agencies for their
own work in astrochemistry : in the UK the CCLRC, EPSRC, NERC and PPARC Research councils, The
Royal Society and The British Council as well as the Open University. Such research is necessarily
international and we therefore whish to thank the EU Framework programme and the European
Science Foundation for their support in conducting research and for integrating the European
Community in common research projects.
References
Gilmour I., An Introduction to Astrobiology, Cambridge university Press, 2004.

Shaw A., Astrochemistry: From Astronomy to Astrobiology, John Wiley and Sons Ltd., 2006.
Chela-Flores J., The New Science of Astrobiology : From Genesis of the Living Cell to Evolution of
Intelligent Behaviour in the universe, Kluwer Academic Publishers, 2004.
Bernstein M.P., Dworkin J.P., Sandford S.A., Cooper G.W., Allamandola L.J., Racemic amino acids from the
ultraviolet photolysis of interstellar ice analogues, Nature, 2002, 416, 401-403.
Holtom P.D., Bennett C.J., Osamura Y., Mason N.J., Kaiser R.I., A combined experimental and theoretical
study on the formation of the “aminoacid” glycine (NH2CH2COOH) and its isomer (CH3NHCOOH) in
extraterrestrial ices, Astrophysical J., 2005, 626, 940-952.
Muñoz Caro G. M., Meierhenrich U. J., Schutte W. A., Barbier B., Arcones Segovia A., Rosenbauer H.,
Thiemann W. H.-P., Brack A., Greenberg J.M., Amino acids from ultraviolet irradiation of interstellar ice
analogues, Nature, 2002, 416, 403-406.
Ä,-Ä
Chapter 6
Water remediation
by the electron beam treatment
Salvatore S. EMMI and Erzsébet TAKÁCS
Introduction
Water shortage and security are becoming a major national and regional priority
because of water-intensive lifestyles, rapid industrialization, urbanization, and agricultural
intensification worldwide. Therefore the treatment of municipal and industrial wastewater
as well as of drinking water is a prime task for environmental engineering. Let's consider,
for example, the problems encountered with halogenated organic compounds. The
ubiquitous nature of halogenated organics, from trichloroethylene to the wood preservative
pentachlorophenol, has spawned considerable discussion about the impact of such
halogenated organics in the environment. Chlorophenols are a serious health concern in
drinking water treatment, as they can be present in the source water or be formed as by-
products of chlorine-based disinfection methods. An ample literature reports on effective
degradation of chlorophenols and other toxic pollutants such as dioxins, polychlorinated
biphenyls (PCBs), and benzofuran derivatives [1-3]. Another class of compounds, emerged
as a focus of attention because of their impact on the environment, is that of dyes and
pigments, which are the single largest group of industrial chemicals. This chapter briefly
surveys the electron beam (EB) Advanced Oxidation Process, reporting some recent results on
the decompositions of a dye (Apollofix-Red) and of a pesticide (carbofuran). Two successful
applications of the technology on the industrial scale are also presented.
Ä,.Ä
2!$)!4)/.#(%-)3429
Advanced Oxidation Processes (AOPs)
A large class of Advanced Oxidation Processes (AOPs), such as UV-peroxide,

ozonation, Fenton and Fenton-like processes, photocatalysis, electrokinetics and sonolysis
have been effectively employed in order to decompose hazardous compounds in water [3-5].
All these methods are based on the generation and use of hydroxyl radicals as the primary
oxidant for the degradation of organic pollutants (Fig. 1). Table 1 reports the reduction
potentials of popular radical species relevant to wastewater remediation. From that it can
be seen that tOH has the highest oxidative power. Being also a very fast reactant, tOH can
attack and oxidize most of the hazardous compounds dissolved in water. Electron beams in
the MeV range, and to a lesser extent G-rays, are excellent tools to produce tOH radicals by
operating the lysis of water induced by radiation (radiolysis).
Table 1. Reduction potentials of radical species relevant to wastewater remediation* (Ox + e– m Red).
Species (alphabetic order) Couple pH Std Red Pot (1)
Chlorine (Cl2) Cl2/Cl2to 0.42 .. 0.60
Chlorine dioxide (ClO2t) ClO2t, H/HClO2 ~ 1.30
Chlorine radical anion (Cl2to) Cl2to/2Cl− 2.09(2)
Hydrogen peroxide (H2O2) H2O2, H/ H2O, tOH 7 0.46 .. 0.87
Hydroxyl radical (1) (tOH) t

OH / OH− 1.9
Hydroxyl radical (2) (tOH) t

OH, H/ H2O acid 2.7
Hypochlorous acid (1) (HClO) HClO, H / H2O, Clt - 0.46

− t
Hypochlorous acid (2) (HClO) HClO/ Cl , OH - 0.04
Oxide radical ion (Oto) Oto, H2O/ 2 OH− 1.78
Ozone (O3) O3 / O3to 11-12 1.01
Perhydroxyl radical (1) (HO2t) HO2t, H / H2O2 0 ~ 1.50
Perhydroxyl radical (2) (HO2t) HO2t / HO2− 0.79
Superoxide radical (O2to) O2to, H/ HO2– 1.00
* A more positive reduction potential for the couple Ox/Red, means that Red is a more powerful oxidant. When protons are involved, the reduc-
tion potential of the half-cell varies with pH. 1) Wardman P., J. Phys. Chem., Ref. Data, 1989, 18,1637-1755. 2) Atkins P.W., Physical Chemistry,
Fifth Edition, Oxford University Press (1994).
Ä-%Ä
7ATERREMEDIATION
Figure 1 : Hydroxyl radicals decompose pollutants. Typical tOH attack to an aromatic ring.
Water radiolysis by electron beam
A beam of electrons, travelling through water, forms a swarm of highly reactive

radicals: hydrated electron (eaq−), hydroxyl radical (tOH) and hydrogen atom (tH) [6] (see G.V.
Buxton, Chapter 1). To perform an AOP, eaq– and H atoms should be neutralized to poorly
reacting species or, better still, transformed into a further amount of tOH radicals. This is
achieved by adding one or more scavengers to the system. Figure 2 outlines a variety of
systems which characterize different EB operating systems. Among these, a straightforward
oxidative degradation process is represented by the saturation of water with N2O above pH 3:
N2O captures the strong reducing eaq− and converts it stoichiometrically into tOH. As a result
of this method, practically only tOH remains, except for a negligible quantity of Ht. The
plainness of this system is also suitable for mechanistic studies with pulse radiolysis, i.e. the
chemical kinetics application of EB (see next page).
The rapidity with which the EB treatment initiates the decomposition of pollutants
may be promptly checked by reasonably assuming that i) tOH reacts with rate constants in
the range 106-109 l mol−1 s−1 , and i i) pseudo-first order conditions are maintained throughout
the process (i.e. tOH reacts quantitatively with the pollutant). Therefore a pulse of electrons
depositing 10 Gy of energy in N2O saturated water triggers the decomposition of ~ 5 x 10−6 mol/l
of pollutant. In an ordinary accelerator facility, sweeping a flow of water with pulses of
1 microsecond duration at a repetition rate of 100 pulses per second, 0.5 millimol/l of
compound per second are immediately modified. Certainly, the calculation of the production
speed of the oxidizing species allows only a rough assessment of the efficiency of an AOP.
Ä-&Ä
2!$)!4)/.#(%-)3429
In fact, the whole rehabilitation is affected by a wide range of factors, among which a
primary role is played by the variety and nature of pollutants, and the required quality of
water. As a matter of fact, industrial plants demonstrate the ability of EB to treat and recycle
millions of litres of wastewater per day (see later), a goal well above the reach of the other
technologies.
Figure 2 : Homogeneous and heterogeneous aqueous systems used to produce tOH and other oxidizing
radicals in typical E-beam AOP experiments. Any system can be employed separately or in association with the
others. Oxidation chain carriers are red tagged.
To exemplify the destruction of a pollutant, the decomposition of a hydrocarbon is

illustrated in Figure 3 : the right side represents the radical production systems, while on the
left the most important reactions following tOH attack are given. In Figure 3, tOH operates its
oxidizing action by abstracting a hydrogen atom from the hydrocarbon chain. Besides water,
an alkyl radical is formed from the pollutant : this radical starts a series of reactions, such as
peroxidation, fragmentation, further H abstraction, disproportionation, molecular growth etc.
t
OH may also initiate an oxidative pathway by accepting an electron (electron transfer), or by
adding to a π electron system (bond breaking). The latter is the most common pathway in the
presence of an aromatic ring, and it will be described later by means of two representative
pollutants.
Ä-'Ä
7ATERREMEDIATION
Figure 3 : Oxidative degradation of a hydrocarbon by E-beam AOP. Irradiated water produces tOH both
directly and by the mediation of dissolved scavengers (right side). tOH oxidizes the hydrocarbon by H abstraction.
The resulting alkyl radical starts a recursive chain degradation giving rise to new carbon- and oxygen-centered
radicals, carbonyls, carbocations, hydroperoxides, polymers, etc. (left side).
Pulse radiolysis
The simultaneous presence of pollutants of different nature in wastewater gives

rise to a very complex kinetic degradation scheme. The kinetic parameters of each reaction
constitute the basic data needed to design an efficient degradation process. A suitable tool
to make such kind of investigation is pulse radiolysis (Chapter 2) coupled with a kinetic
spectrometry detection system [7]. Aiming to achieve meaningful results, reactions with
water radicals are studied separately as outlined below (Chapter 1) :
– In N2 or Ar saturated solutions, between pH 3 and 11 the reactions take place with

t
OH, Ht and eaq .
– Below pH 2 in N2 or Ar saturated solutions, nearly equal amounts of the tOH radicals

and Ht atoms are present. This is because aqueous electrons are converted to Htby
reacting with the hydroxonium ions:
eaq )3O+ m Ht)2O (1)
Ä-(Ä
2!$)!4)/.#(%-)3429
– Ht atoms reactions with solutes can be carried out alone below pH 2 in N2 or Ar

saturated solutions containing 0.2 – 1 mol l–1 tert-butanol.
In this system, tert-butanol scavenges tOH and forms poorly reactive radicals :
t
0) $)3)3COH m tCH2(CH3)2$0))2O (2)
(The reaction between Ht and tert-butanol is slow).
– Reductions by eaq are performed in N2 or Ar saturated solutions above pH 3 and in

the presence of 0.2 – 1 mol l–1 tert-butanol (there is a small contribution from the
Ht atom reactions). Radicals formed from tert-butanol are also present.
– In air or oxygen saturated solutions, the reactive species are the tOH radical, and
the couple O2t /HO2t (superoxide radical anion/perhydroxyl radical). The couple is
originated by scavenging eaq and Ht with O2 :
eaq 02 m O2t (3)

Ht02 m HO2t (4)
and is in acid/base equilibrium with a pKa value of 4.8
O2t )3O m HO2t)2O (5)
– Solutions saturated with O2 and containing tert-butanol are used for studying the
reactions of the O2t /HO2t acid-base radical pair.
The degradation of pesticides : carbofuran
Carbofuran (C12H15NO3), is a carbamic insecticide and nematocide vastly employed

in North America and in some European countries to protect maize, rice, alfalfa, onion, garlic,
potatoes, etc. However, it is toxic by contact and ingestion, and lethal over the level of 11 mg/
kg (LD50, rats). The Maximum Contamination Level (MCL) for drinkable water, as established
by EPA (USA), corresponds to 0.18 Mmol/l.
Several AOPs have been proposed to operate carbofuran decomposition, and many
of them have demonstrated their effectiveness. The choice of the more appropriate system
will consider parameters like the rapidity of action and economical factors.
Ä-)Ä
7ATERREMEDIATION
Carbofuran
Figure 4 : Gamma-Radiolytic degradation of a pesticide carbofuran at 200 μmol l–1 and the synergistic effect
of a metal oxide (TiO2 ). Left) : carbofuran in deaerated atmosphere and in the absence of catalyst.
Right) : carbofuran in oxygenated atmosphere and in the presence of TiO2. In the latter system the dose neces-
sary to decompose 200 μmol l–1 carbofuran decreased by an order of magnitude.
The example below refers to a recent study by pulse and gamma radiolysis [8]. A pH z 6
has been generally used. Hydroxylation mainly concerns the aromatic ring of carbofuran
(Fig. 4 ) ; in fact, the first intermediate shows the absorption of a cyclohexadienylic type
radical in the region 280-330 nm. The hydroxylation attack to carbofuran
t
0)DBSCPGVSBOm tCF-OH (6)
proceeds with a rate constant k0)$' = 6.6 x 109 l mol−1s−1. In the presence of oxygen, the
carbofuran cyclohexadienylic radical adds oxygen (peroxidation) with a rate constant kperox
z 107 l mol−1s−1. As measured after gamma radiolysis, oxygen participation makes the global
decomposition three times more efficient than in anaerobic solutions; the efficiency is further
enhanced by adding a 1% amount of TiO2. Typically, under these conditions a sample containing
carbofuran 200 Mmol l–1 is completely decomposed by 1.25 kGy of G-radiation dose, i.e. an order of
magnitude more efficiently than without O2 and TiO2 (Fig. 4).
TiO2 catalysis
TiO2 catalysis proceeds through the production of tOH radicals at the solid-liquid
interface. The mechanism of action lies on the separation of charge upon absorbing a radiation of
energy higher than 3.2 eV, i.e. below 380 nm (UV photons, E-beam and G-rays): holes (h) are left in
the valence band (VB) and electrons (e −) are promoted to the conduction band (CB).
Ä-*Ä
2!$)!4)/.#(%-)3429
It appears that TiO2 particles in water are widely hydroxylated. Therefore holes
escaping annihilation may migrate to the surface and oxidize adsorbed water molecules
and hydroxyl ions. Oxygen adsorbed at the surface captures electrons preventing their
recombination with h, and therefore favouring the yield of tOH radical (Fig. 5).
Figure 5 : Formation of oxidizing radicals at the

solid-liquid interface of TiO2 and water. Radia-
tion promotes a separation of charge in the cata-
lyst. H2O, OH–, and O2, adsorbed at the surface are
converted to radicals by electrons and holes which
migrate to the interface. Oxidative degradation is
greatly enhanced (see also Fig. 4).
Decoloration of dyes : Apollofix-Red
In dilute dye solutions all the intermediates of water radiolysis, i.e. hydrated electron
(eaq ), hydroxyl radical (tOH) and hydrogen atom (tH) induce the decomposition of the solute.
The degradation of the dye can be easily followed by taking decoloration curves.
This is usually done by preparing appropriate solutions, irradiating them by increasing the
dose stepwise and by taking the UV-VIS absorption spectra after each irradiation [9]. The
decoloration is due to the destruction of the color centers. The color fading with dose is
illustrated in Figure 6.
The upper part shows the spectra of aqueous N2O saturated solutions of a reactive dye
(Apollofix-Red), before and after EB irradiation with different doses (0-3 kGy), while the lower
part shows the photographs of samples treated in the same way. The chromophore center of
this dye is an azo group connecting a benzene ring with a naphthalene part. Decolored products
are formed when the extensive conjugation is destroyed, for instance when the conjugation
through the –C–N=N–C– bridge is lost. End-product analysis by High Performance Liquid
Chromatograph (HPLC, diode array detection was used for this measurement) also showed a
sharp decrease in the concentration of the starting compound (Fig. 7).
Ä-+Ä
7ATERREMEDIATION
$SROORÀ[5HG
Figure 6 : Dye decoloration. Aqueous N2O saturated solutions of a dye (Apollofix-Red 0.25 μmol l–1) are
E-beam irradiated with increasing doses: the reaction with tOH destroys the color centers. This is shown both
by a progressive weakening of the absorbance of samples (spectra) with dose (0-3 kGy) and also by a progressive
increasing of their transparence (photographs below the spectra). The inset shows the decoloration-dose curve
at 530 nm.
Figure 7 : Three-dimensional HPLC chromatogram showing the separation of degradation products in N2O
saturated dye (Apollofix-Red AR-28 at 0.25 mmol l–1) solution irradiated with 0.6 kGy dose. Destruction of AR-
28 molecules and formation of degradation products in tOH radical reactions as measured by HPLC with diode
array detection. AR I and AR II are used for AR-28 and for its hydrolysed form, respectively, PI and PII are products.
Ä-,Ä
2!$)!4)/.#(%-)3429
The mechanism is not as straightforward as it may appear: in the first step, in fact,
t
after OH radical addition to any aromatic ring, a new colored product is formed, i.e. the
cyclohexadienylic type radical may reform the conjugated system in disproportionation
reactions [7,10]. The process is exemplified with a substituted benzene (Scheme 1). Using high
doses, and especially in the presence of O2 or O3, aldehydes and organic acids are formed from
the tOH-dye adducts and finally decomposed to CO2 and H2O (see upper part of Scheme 1).
Scheme 1 : Decomposition outline of the aromatic part of Apollofix-Red dye, modeled with a substituted
benzene (lower part) and achieving its complete mineralization (upper part). A disproportionation reaction
between two cyclohexadienylic radicals regenerates the aromatic ring structure in a phenol derivative. The R-group
can be an alkyl group or a halogen atom.
Ä--Ä
7ATERREMEDIATION
Electron accelerators and operative conditions
At present radiation technology is a popular method for sterilization, polymerization,

polymers and semiconductors modification, but not for environmental remediation. Actually,
while various new features of irradiated materials were worthy of industrial investments, the
improvement given to the quality of water and air was not economically rewarding. In the
past indeed, the requirement of high initial investments constituted a restraint to the E-beam
implementation in the environmental field. Nowadays, however, accelerators are becoming
cheaper while freshwater is more and more precious, such that radiation rehabilitation
and recycling of wastewater can be profitably considered by both public authorities and
investors.
To see that, let us focus on a couple of factors that mainly determine the economy
of the EB remediation process: i.e. a high power (by which high volumes of water can be
rehabilitated daily), and the efficiency (which establishes the transfer rate of the accelerator
nominal power into useful energy). To this respect, it is instructive to learn from Table 2
that DC and UHF machines (Chapter 2) best match with these requirements. In the DC
type machines, electrons are accelerated by a direct-current field, while in the UHF type,
acceleration occurs across an electromagnetic field oscillating at few hundreds MHz. These
accelerators achieve high powers coupled with moderately-high efficiencies, and therefore
represent the best choice for wastewater treatment. At present, the linear accelerators,
which are based on a microwave field, show efficiencies and power below the others, such
that they are not suitable for environmental purposes.
processing.
Table 2. Features of accelerators for environmental processing.
Efficiency Energy
Type Accelerating field Power (kW) (up to)
range (%) range (MeV)
DC electrostatic 60-80 400 0.1-5
UHF 100-200 MHz 25-50 700 0.3-10
Linear 1.3-3 GHz 10-50 150 2-10
Considering some practical aspects, it has to be remarked that the EB is best used
as a pre-treatment step to integrate a conventional bio-, and physico-chemical plant. Under
such conditions a dose of 2 kGy allows the destruction of most organic pollutants with
unmatched success even in case of bio-refractory compounds. Simultaneously, disinfection
of wastewater is achieved. When only disinfection is required, the dose may be reduced to
0.2 kGy, resulting in more than 95% inactivation of the total coliform bacilli without any
Ä-.Ä
2!$)!4)/.#(%-)3429
previous chlorination. An appropriate machine for these processes should produce 1-2 MeV
electrons, while the output power can be within the 40 to 400 kW range, depending on the
flow rate required. Studies showed that uniform dose distribution is achieved by transversally
pumping a thin layer of effluent under the beam area, as illustrated in Figure 8.
Figure 8 : Optimal flow system irradiator geometry. Penetration of 2 MeV electrons in water is 1 cm.
Industrial and municipal treatments
Data on the reactivity of various classes of toxic substances and of dyes with tOH
and other radicals were made available by means of pulse radiolysis (see above). Oxidative
degradation in complex systems could then be simulated with numerical methods and
tested on pilot plant scale. Finally, some fully operating industrial plants demonstrated the
viability of EB processing both on the environmental and economical aspect.
Among these plants, the wastewater rehabilitation plant of the Voronezh synthetic
rubber plant (Russia) is worth to mention [2]. Here, two electron accelerators (0.7-1 MeV
energy and 50 kW power), can treat up to 2,000 m3 of wastewater per day with the specific
goal to transform the non-biodegradable emulsifier Nekal into a biodegradable form. A
further decomposition is achieved with biological methods.
Another industrial plant was put into operation in December 2005 for the EB
treatment of dyeing wastewater (Fig. 9) in Korea at Daegu Dyeing Industrial Complex (DDIC).
DDIC includes about hundred factories, the majority of which has equipment for dip dyeing,
printing, and yarn dyeing [11, 12]. The production requires high consumption of water and
emits large amount of highly colored industrial wastewater. The whole facility treats up to
Ä.%Ä
7ATERREMEDIATION
80,000 m3 wastewater /day. From this volume, 10,000 m3/day are pre-treated presently by
EB, using a 1 MeV - 400 kW - accelerator combined with bio-treatment facilities.
Figure 9 : Combined wastewater treatment at DDIC (from Bumsoo Han originals, EB-TECH Co., Ltd., Daegu,
Korea, on permission of the author).
An instructive case study, illustrating different operational conditions in the

Mediterranean area, was carried out by Lopes [13]. The author made a predictive modelling
on the implementation of EB to the urban wastewaters of the city of Palermo (Sicily). At
present the city of Palermo is served by two depurator plants. One of them, called “Corsairs
water”, is under development to accomplish the treatment of 300,000 m3/day. Goal of Lopes's
investigation aimed at assessing the disinfection line of the depurator by EB. The process
was then modelled on a daily treatment of 33,000 m3 with 10 MeV electrons and a dose of
0.2 kGy, produced by one or two 190 kW accelerators of the UHF type, i.e. the Rhodotron
TT300 (IBA).
Cost evaluation
Cost evaluations for EB treatment may vary, mainly depending on accelerator

voltage and power, on the cost of electricity, and on the reuse assigned to restored water
(i.e., on the degree of pollutant decomposition or disinfection). Among other factors, the
cost for generating a unity of power is of primary importance: 1 kW is cheaper for high
power accelerators, i.e. it may cost 20 k$ for a 40 kW accelerator, but only 5 k$ for a 400 kW
machine. Nonetheless, a realistic and updated evaluation is available from the experience
at the Daegu plant above [12] and from the mentioned study at the University of Palermo
[13]. Let us prospect two typical cases at Daegu, both using a 1 MeV-400 kW-accelerator
(cost 2 – 2.5 million US$), running 365 days/year. Case A: 100,000 m3/day of wastewater from
Ä.&Ä
2!$)!4)/.#(%-)3429
a municipal plant are rehabilitated for irrigation and industrial use with a dose of 0.2 kGy
(disinfection only). Case B: 10,000 m3/day of textile wastewater are rehabilitated for self-
recycling with a dose of 1-2 kGy (dyes decoloration and disinfection). In Case A, the capital
investment per m3 of effluent would be around 0.11-0.12 US$, while the operating cost
would be ca. 0.03 US$. In case B, costs increase 7-8 times, being respectively 0.82 US$/m3
and 0.24 US$/m3 (Tab. 2). These financial quotations include investments for the industrial
plant (accelerator, shielded room, water reactor, design, and others), interest, depreciation,
and electricity (80% efficiency at a cost z 0.05 US$/kWh) and do agree with the actual costs
after 12-15 months of operation. Costs for land, research and development, and for authority
approval are not included. At the same site, EB treatment combined with bio-plant leads to
z 20% savings in operating costs, compared to bio-plant alone, reduces the residential time,
and moreover yields water of better quality.
In the Palermo case [13], one accelerator (cost z 4 million €) working almost at
full capacity (180 kW), or two accelerators at half-power, are conceived. A comprehensive
economical model reached conclusions matching with the Daegu budget plan: precisely, in
Palermo, each m3 could cost 0.17 € employing one accelerator, or 0.28 € with two accelerators
(Tab. 3). The latter system, although more expensive, is still acceptable as it could minimize
risk of service interruption for possible failure. However, a pilot plant with only one accelerator
should be a wise starting system, as has been done in Daegu.
Table 3. Operational features and cost of EB treatment at DDIC (Daegu, Korea) and Corsairs Water (Palermo,
Sicily).
Accelerator
Plant Model Type Energy, N. of Operating Aim Status Dose, Water Total cost
(MeV) Acc power (kGy) treated (/m3)
3
(kW) (m /day)
tested in
DDIC ELV-12 DC 1 1 400 disinfection pilot 0.2 100 ,000 0.15 US$
plant
industrial
1 1 400 recycling operative 2.0 10,000 1.06 US$
Corsairs Rhodotron UHF 10 1 180 disinfection study 0.2 33,000 0.17 %

Water TT300
10 2 90 each disinfection study 0.2 33,000 0.28 %
Ä.'Ä
7ATERREMEDIATION
Concluding remarks
In the preceding sections, the reader got acquainted with a few basic concepts on
the remediation of water pollution by injecting energy in water and its latest developments.
The Electron beam treatment is a physico-chemical method which pursues the rehabilitation
of polluted waters by the oxidative action of the tOH radical, and we saw that the methods
for tOH generation distinguishe various Advanced Oxidation Processes.
The EB treatment is founded on the excitation and dissociation of water by high

energy electrons. Energetic electrons are produced by devices which operate similarly to a
TV set and, in the same way, are switchable on and off at the touch of a button. Exactly like in
the TV operation, no radiation is stored in the environment. Wastewater can be rapidly and
safety rehabilitated to be reused, for example, in industrial cooling and washing cycles, fire
fighting, street washing, green park and horticultural irrigation. EB is particularly suitable
for treating medium-large volumes of effluents from industries, hospitals, municipal plants,
and animal-breeding. It does not give rise to concern for the environment, as it minimizes
or rules out the addition and stock of chemicals. Furthermore, it works at room temperature
and atmospheric pressure. Electrons penetrate in the depth of water even in case of turbidity
(not possible by UV methods), and generate very high concentration of tOH in a fraction of a
microsecond.
The versatility of EB is further experienced by the ease with which radical kinetics is
governed ; in fact, the concentration of tOH can be modulated on that of pollutants by easily
tuning the beam current.
Conventional rehabilitation treatments are inadequate when wastewaters include

bio-resistant chemicals, e.g. the carcinogenic polychlorobiphenyls (PCB’s) and aromatic
compounds. In this case a pre-treatment with EB converts bio-refractory compounds to bio-
degradable ones and increases their solubility in water, finally reducing cost and residential
time. Therefore, the design of a cost effective rehabilitation plant conceives the usage of
EB as a method for improving the biodegradability of hazardous compounds, rather than
for achieving their complete mineralization (i.e. production of CO2 and water). Complete
mineralization is, in fact, a rather costly objective and may not always prove to be necessary for
remediation. On the other end, conversion of a pollutant into a useful non-toxic product (e.g.,
fuel or polymer) may be alternatively a desirable goal. However, if complete mineralization is
the ultimate purpose of wastewater treatment, radio-catalysis may be associated with other
radiolysis methods, granting a synergistic effect to the process. An environmental friendly
metal oxide, such as TiO2 for instance, may be used in this case. Besides the benign chemical
Ä.(Ä
2!$)!4)/.#(%-)3429
action, the EB kills viruses and bacteria and should be regarded as the method of choice for
water security, i.e. a rapid rehabilitation of bio-contaminated waters, in the eventuality of
environmental accidents, acts of terrorism or vandalism.
Recent industrial achievements and projects showed the EB method to be both

economically rewarding and effective in protecting the environment, and as such was
granted credit by the IAEA [14]. Finally the Electron Beam technology should be regarded
both as a trustworthy and an innovative process to fulfil a sustainable chemistry.
References
[1] Getoff N., The Role of Peroxyl Radicals and Related Species in the Radiation-Induced Degradation
of Water Pollutants, in “Environmental Applications of Ionizing Radiation”, Cooper W.J., Curry R.D.,
O’Shea K.E. (eds), Wiley, New York, 1998, 231-246.
[2] Pikaev A.K., Electron Beam Purification of Water and Wastewater, in “Environmental Application
of Ionizing Radiation”, Cooper W.J., Curry R.D., O’Shea K.E., Wiley (eds), New York, 1998, 495-506.
[3] Pera-Titus M., Garcia-Molina V., Banos M.A., Gimenez J., Esplugas S., Degradation of chlorophenols
by means of advanced oxidation processes: a general review, Appl. Catal. B. Environment, 2004,
47, 219-256.
[4] Legrini O., Oliveros E., Braun A.M., Photochemical process for water treatment., Chem. Rev., 1993,
93, 671-698.
[5] Kamat P.V., Meisel D., Nanoparticles in advanced oxidation processes. Current Opinion in Colloid
and Interface, Science, 2002, 7, 282-287.
[6] Buxton G.V., Greenstock C.L., Helman W.P., Ross A.B., Critical review of rate constants for reactions
of hydrated electrons, hydrogen atoms and hydroxyl radicals (tOH/tO–) in aqueous solution, J. Phys.
Chem. Ref. Data, 1988, 17, 513-886. Updated version : www.rcdc.nd.edu.
[7] Wojnárovits L., Takács E., Emmi S.S., Reactivity differences of hydroxyl radicals and hydrated
electrons in destructing azo-dyes, Radiat. Phys. Chem., 2005, 74, 239-246.
[8] Emmi S.S., De Paoli G., Takács E., Pálfi T., Electron beam remediation of wastewater. The
hydroxylation of carbofuran, 24th Miller Conference on Radiation Chemistry, Le Londe les Maures,
France, 10-15 September 2005, 64.
[9] Solpan D., Güven O., Takács E., Wojnárovits L., Dajka K., High-energy irradiation treatment of
aqueous solutions of azo dyes: steady-state gamma radiolysis experiments. Radiat. Phys. Chem.,
2003, 67, 531-534.
[10] Wojnárovits L., Takács E., Irradiation treatment of azo dye containing wastewater : An overview,
Rad. Phys. Chem., 2008, 77, 225-244.
Ä.)Ä
7ATERREMEDIATION
[11] Han B., Ko J., Kim J., Kim Y., Chung W., Makarov I.E., Ponomarev A.V., Pikaev A.K., Combined
electron-beam and biological treatment of dyeing complex wastewater. Pilot plant experiments,
Radiat. Phys. Chem., 2002, 64, 53-59.
[12] Han B., Kim J.K., Kim Y.R, Salimov R.A., Kuknasov N.K., Nemytov P.I., High power accelerators for
environmental application, IAEA-TECDOC-1473, IAEA, Vienna , 2005, 119-124, and personal
communication, April 2007.
[13] Lopes V., Nuclear Technologies Case Studies in the Civil and Industrial field, Laurea magistralis
Thesis, 2005-2006, University of Palermo, 106 pp.
[14] IAEA, Radiation processing of wastewater, in Radiation processing: environmental applications,
IAEA, Vienna, 2007, 25-44, and references quoted therein.
Ä.*Ä
Chapter 7
Metal clusters
and nanomaterials
Jacqueline BELLONI and Hynd REMITA
Introduction
In the past, tiny particles of matter were widely used as pigments in paints, inks,
cosmetics and stained glass. In the 19th century, the properties of colloids attracted
increasing interest and the specific colours of metal colloids were explained by Mie (1908).
Fogging of photographic plates, that is reduction of silver ions into atoms and finally into
tiny metal clusters in the emulsion of silver halides, enabled W.C. Roentgen to discover X-rays
in 1895 and H. Becquerel uranium rays in 1896. This effect is still used in some dosimeters
to detect ionizing radiation and more generally in radiography. Due to the ionization of
liquids and the production of solvated electrons, es−, and of Ht atoms (Chapter 1), metal ions
in solution are likewise reduced by these strong reducing agents into lower valency ions,
and then into atoms. Indeed, Marie Curie used to compare the radiolytic processes to an
"electrolysis without electrodes", since the primary pair electron-solvent cation is formed in
the bulk solution instead of at the surfaces of separated electrodes. After coalescence, the
atoms give rise to clusters containing an increasing number of atoms.
However, such atoms and the smallest clusters (or oligomers), observed after the
radiation-induced generation, appeared to be produced to a lesser extent than expected and
were easily corroded, even if the corresponding bulk metal presented a noble character [1].
In other words, small clusters of, for example, gold, silver, copper, platinum, etc., exhibited
a new reactivity under conditions where the metals have none. We solved this apparent
violation of thermodynamics in 1973 by assigning to both the atoms and the smallest
Ä.,Ä
2!$)!4)/.#(%-)3429
clusters a reduction potential much more negative than for the bulk metal and depending
on the number of atoms they contain [1]. This peculiar behaviour was assumed to be due to
their extremely divided state and to discrete energy levels, different from the band structure
of large metal crystals. This new concept was confirmed in 1977 by A. Henglein’s calculations
of the reduction potential of the silver atom in aqueous solution, found to be 2.6 V more
negative than the bulk metal [2], and by measuring oxidation rate constants of the atom
and the dimer with several electron acceptors. Since then, due to their deep penetration
and the homogeneous production of strong reducing species, ionizing radiation has been
used as a particularly powerful tool, in the steady and the pulse regimes, to observe at room
temperature the early and transient steps of metal atom formation, nucleation and growth
into clusters, in the depth of a solution or of the pores of a support and to determine their
nuclearity-dependent properties [3]. Based on these results, which were obtained under
various conditions, the nucleation and growth mechanisms involving competition between
several processes could be established. Their understanding is a guide to control the final
size, shape and structure of mono- or multi-metallic clusters. The mechanisms are also
inspiring the improvement of some processes for various applications where they play an
important role.
Metal ion reduction and atom coalescence
The atoms are produced in solution by radiation-induced reduction of the metal ion
precursors without any other added electron donor. The species arising from the radiolysis
of water, solvated electrons eaq–, and Ht atoms, are indeed the strongest reducing agents of
this wet process. They easily reduce at room temperature all metal ions, possibly complexed
by a ligand, down to the zero-valent state (Fig. 1). In contrast, sibling radicals which are also
formed in radiolysis, such as tOH radicals in water, are able to oxidize the ions or the atoms
into a higher oxidation state. To prevent this reaction, a scavenger of tOH radicals, such as a
secondary alcohol or formate anion is added to the solution. The secondary radicals formed,
H3C-tCOH-CH3 or CO2t− radicals, respectively, are also strong reducing species (Chapter 1). To
prevent oxidation by di-oxygen or corrosion by protons of the easily oxidisable atoms, the
solutions are de-aerated and, for non noble metals [4], made slightly basic.
Such reduction reactions have been observed directly by pulse radiolysis for
several metal ions. Most of the reduction steps have been observed and their rate constants
determined. Figure 1 presents the example of Ag reduction observed by pulse radiolysis
coupled with time-resolved spectrophotometry. The evolution of the optical absorption
spectrum in the successive fast steps is recorded just before and after the short electron pulse
delivering the irradiation dose, as in a movie filming the fast cascade of reactions initiated
Ä.-Ä
-ETALCLUSTERSANDNANOMATERIALS
by the pulse. Provided the transient species have specific absorption bands (or colours),
they can be identified and the detailed mechanism of their reactions and the corresponding
rate constants is deduced from the variation of their absorbance which is related to their
concentration. The broad and intense band in the near infra-red at the end of the pulse in
Figure 1 is due to hydrated electrons. They react readily here with Ag giving rise to Ag atoms
(Lmax = 360 nm) at the same time as the electron absorbance decays [2]. The binding energy
between two metal atoms is stronger than the atom-solvent or atom-ligand bond energy.
Therefore, the atoms dimerise on encounter. The bonding between atoms or clusters with
unreduced ions is also strong and these association processes are fast, the first one being
that of atoms with excess ions leading to Ag2 (Lmax = 310 nm) (Fig. 1). The atom absorbance
vanishes while that of Ag2 increases. Then, Ag2 dimerise into Ag4 (Lmax = 270 nm)
which progressively coalesce by a cascade of processes into growing oligomers and clusters.
At nuclearity larger than n = 13, the spectrum of yellow clusters (Lmax = 380 nm), due to a
plasmon resonance of electrons confined in the particule, is fully developed [2] (Fig. 1).
Figure 1 : Evolution with time (logarithmic scale) after a 3 ns electron pulse of the optical spectra of transients
formed in an aqueous solution of silver ions as observed by pulse radiolysis.
Because the radiation-induced reducing agents, solvated electrons and radicals are
generated randomly in the bulk of the sample and in the vicinity of metal ions, the atoms
and clusters are formed with a rather homogeneous distribution throughout the solution
Ä..Ä
2!$)!4)/.#(%-)3429
with a remarkable homodispersity. Note that in the pulse regime (at high dose rate), all
reducing species are produced and scavenged within a short time, followed by the steps
of coalescence of atoms separately created (Fig. 2a) [6]. In contrast, in the continuous
irradiation regime (at low dose rate), the association of M ions with atoms, as in the charged
dimers M2 and clusters MO, and the subsequent coalescence processes are faster than
the production rate of the reducing radicals (Fig. 2b). Therefore, the reduction of M ions
occurs mostly in situ on clusters MO already formed. As a consequence, the new atoms
formed are not isolated but contribute to the growth of a smaller number of preformed
nuclei. In fact, the final nuclearity n and the cluster radius, which is proportional to n1/3,
are found to be systematically larger by steady than by pulse irradiation. When a chemical
agent generally chosen as an electron donor D is added, the reduction potential (E°D/D)
is not negative enough to reduce directly isolated metal ions to atoms (see below). Thus, it
essentially reduces ions adsorbed on the nuclei generated by radiolysis and acting as seeds,
a development process which results in still larger clusters (Fig. 2c) [7].
Figure 2. Nucleation and growth of clusters generated by radiolytic radicals at high (a) and low (b) dose
rates, without or with an electron donor D (c). The stabilizing effect of the added polymer prevents exclusively
coalescence beyond a certain limit of nuclearity, but does not prevent successive ion and electron transfers (from
the radicals at low dose rate and from the donor D), which lets the cluster develop up to much larger sizes than
at high dose rate.
Ä&%%Ä
Cluster stabilization
The radiolytic method of cluster synthesis, which is a bottom-up approach, can

be used in various environments (Fig. 3). The control of the final size depends on the
limitation applied to coalescence beyond a certain nuclearity, and to reduction in situ on
nuclei (limitation favoured at high dose rate without chemical donor as shown in Figure 2a).
For free clusters such as nanocolloids in solution, the coalescence may be limited by a
polymeric molecule acting as a cluster stabilizer. Functional groups with high affinity for
the metal, such as carboxyl groups, ensure the anchoring of the molecule at the cluster
surface, while the polymeric chain protects the cluster from coalescing with the next one
through electrostatic repulsion or steric hindrance and thus inhibits at an early stage further
coalescence. Note that the polymer protects the cluster from growth by coalescence but not
from development by metal ion addition and electron transfer in situ from radiolytic radicals
or chemical donors (Figs. 2b and 2c). The polymer should not chemically reduce the ions
fixed on the clusters so as to prevent their growth by the development process (Fig. 2c).
Poly(vinyl alcohol) (PVA) for example generally fulfils this criterion (Fig. 3a), in contrast to
stabilizers also often used as the electron donor in some chemical reductions.
Figure 3 : Radiation-induced metal clusters. (a) Silver nanoclusters stabilized by PVA (10 nm). (b) STM ima-
ging of a single cluster of the blue sol of silver oligomers Ag73+ formed by G-irradiation (n = 4). (c) Clusters of Agn
partially reduced by irradiation and then chemically developed by EDTA. (100 nm large and 15 nm thick).
(d) TEM bright-field image of Nin, PVA clusters (5 nm). (e) Two-dimensional self-assembled array of gold clusters
(PVA) on mica with remarkable homodisperse size (5 nm). (f) Monocrystalline Pt nanotubes with CPCl (10 nm
diameter and a few 100 nm long). (g) Pt nanorods with CTAB ( 3-4 nm thick and 20-40 nm long).
Ä&%&Ä
2!$)!4)/.#(%-)3429
The final size of the clusters depends on the metal ; it decreases as the initial polymer/
ion concentration ratio increases and with PVA lies in the nanometre range (1-10 nm). The
radiolytic method has been used for the synthesis of a great number of noble and non noble
metal nanocolloids in various solvents, water, alcohols, liquid ammonia, etc. Their intense
colours, due to a Mie resonance in light absorption by the electron pool of the particle,
depend on the surrounding medium, on the cluster size and shape, and on the metal (Fig. 4):
in water, yellow for Agn spheres of a few nm (Lmax = 380 nm), purple for Aun (Lmax = 520 nm),
pink for Cun (Lmax = 570 nm), and brown for the othersLmax in the UV). For such small sizes,
no light scattering is observed and the solutions are very clear. Cobalt and nickel clusters are
ferromagnetic [4] (Fig. 3d). As a consequence of the initial homogeneous distribution of atoms,
the size distribution of the clusters is remarkably narrow, as shown in Figures 3a, 3d, and 3e.
Due to their high homodispersity, they can even arrange in a two-dimensional self-assembled
array when the droplet of the nanometre colloidal sol is dried on mica surface (Fig. 3e).
Phase imaging in tapping mode AFM (atomic force microscopy) permits discriminating
between silver or gold clusters and the amorphous polymer stabilizing the nanoparticles
[7]. In the absence of an added radical scavenger, the polymer PVA can be cross-linked under
irradiation during the simultaneous reduction of metal ions (silver, palladium, nickel, etc.).
Finally, after drying, the clusters formed are trapped in a thin polymeric film, which presents
the specific optical absorption band of the metal nanoparticles. In the case of ferromagnetic
nickel nanoparticles, the film can be moved by a magnetic field [6].
Figure 4 : Nanoclusters synthesized by irradiation of water solutions and stabilized by PVA. From left to right:
copper, silver, and gold clusters. The colors of ultradivided metals are known from long. In the Middle Ages, they
were produced in fused glasses under reducing conditions to produce stained glasses.
Another polymer, sodium polyacrylate (PA), displays the property to stop at low dose
the coalescence at only a few atoms, as shown by pulse radiolysis (Fig. 3b)[8]. Such oligomers of
Ä&%'Ä
silver Ag7 or platinum Pt5-7 were observed by STM (Scanning Tunneling Microscopy) (Fig. 3b)
and by their optical absorption spectrum (Lmax = 800 nm for blue silver and 540 nm for
pink nickel oligomers) assigned to a strong metal-acrylate ligandation. The silver and
platinum oligomers stabilized by PA are long-lived even in air, but nickel-PA oligomers are
spontaneously oxidized by water within one day [3].
Some ligands, such as CN− for gold or silver clusters, are also able to stabilize, without
added polymer, small size particles, probably because they are strongly linked to the cluster,
thus protecting them from coalescence by electrostatic repulsion. However, the reduction
potential of these clusters is still more negative than without CN− and they are extremely
2–
sensitive to oxygen [6]. Carbonyl clusters, such as Chini clusters [Pt3(CO)6]m with m = 3-10,
can be easily obtained by radiation-induced reduction of metal ions under CO atmosphere [3].
The synthesis is selective and m is controlled by adjusting the dose (high doses yielding low
m values). The colours (or optical absorption spectra) and redox reactivity also change with
the nuclearity.
Figure 5 : Clusters generated by irradiation in various hard (oxides, semiconductors, metals, zeolites, electrodes,
carbon nanotubes) or soft matrices (liquids, micelles, polymeric membranes, mesophases).
Clusters supported on solids or confined in various structures, hard or soft matrices

(Fig. 5), are produced by letting the ion solution diffuse into the matrix pores or by mixing the
ion and surfactant solutions [3, 6]. The support is often charged superficially with a layer of
counter ions of opposite charge in the liquid phase. Metal ions of the same charge can thus
Ä&%(Ä
2!$)!4)/.#(%-)3429
exchange with these counter ions. Due to the penetrating properties of ionising radiation,
reducing radicals are produced in the vicinity of these ions and yield the metal atoms in
tight interaction with the support. At room temperature, the coalescence is restricted to the
region, surface or intra-pore, where the atoms are fixed, and the sintering is limited by the
strong interaction with the host material.
Another challenge in cluster synthesis is the control of the shape of the nanoparticles
which governs their properties. One-dimensional metal nanostructures such as nanowires
and nanorods are very attractive because of their unique electrical, optical, magnetic
and mechanical properties, and because of their potential applications in nanodevices.
Surfactant molecules self-assemble under certain conditions in water into a large variety of
morphologies, including micelles and liquid crystalline phases such as lamellar or hexagonal
mesophases. Such mesophase structures provide a class of useful and versatile templates
for generating 1D or 2D nanostructures in relatively large quantities. A flexible approach has
been developed for the radiolytic synthesis of nanostructured materials of various shape and
size such as metal nanofibres or lamellae (and also polymer nanowires), using mesophases
with hexagonal, cubic or lamellar symmetry as soft templates [9] (Fig. 5). Metal nanorods of
a few nm diameter and a few Mm length, and lamellae of a few nm thick and tens of nm2 in
area have been obtained.
In the Pt-doped hexagonal mesophase formed from CPCl (cetyl pyridinium chloride),
platinum ions are adsorbed at the surface of the surfactant cylinders. They are reduced
radiolytically into a metal layer as a nanotube of around 10 nm diameter and a few hundred
nm long (Fig. 3f). Extraction of all these nanostructures is achieved by dissolution of the
soft template using alcohol. This possible easy extraction constitutes a marked advantage
over the synthesis in hard templates, such as mesoporous silica or carbon nanotubes, the
dissolution of which is more hazardous for the metal nanostructures.
Because of a preferential binding to a single face of clusters, some surfactants are

known (CTAB, cetylmethyl ammonium bromide, for example) to drive the growth of metal
nanocrystals unidirectionally. Single monocrystalline Pt nanorods, monodisperse in diameter
(3-4 nm) and 20-40 nm long, were recently obtained by orientated coalescence of spherical
seeds of 3-4 nm radiolytically produced in the presence of CTAB surfactant [10] (Fig. 3g).
Note that solvated electrons can also be produced by photodetachment of electrons
from certain anions or photoionization of molecules using UV excitation. Then the mechanism
of metal ion reduction is expected to be quite similar to the radiolytic processes of Figures 2a,
2b and 2c. However, the matrices used as cluster hosts are generally not transparent to light,
and the reduction is often restricted to the surface.
Ä&%)Ä
Nuclearity-dependent reduction potential of clusters
The most important changes in cluster reactivity occur, as for other properties,
at low nuclearity. However, due to the spontaneous coalescence, the small oligomers are
generally short-lived and observable only by pulse radiolysis and time-resolved detection in
the course of their coalescence as in Figure 1. Likewise, the determination of their reactions
is only possible by kinetics methods, studying the influence of an added reactant of known
reduction potential on their usual coalescence processes. Depending on the reactant, metal
clusters MO may behave as electron donors or electron acceptors. In the first case they
are oxidized into the ions (some examples, such as spontaneous oxidation by protons or
oxygen, are given above).
When clusters are electron acceptors, the electron coming from a donor may reduce
the metal ion fixed on the cluster and the cluster nuclearity is incremented by one unit (Fig.
2c). The donor D is produced by the same pulse as the atoms M0 and is observed by a change
in the absorbance at a specific wavelength. The remarkable feature of the kinetics signal at the
donor wavelength is that the absorbance remains constant after the pulse for a rather long
induction time [7]. Nevertheless, during this time, millions of encounters between the donor
and isolated Ag, Ag2 or Ag4 have occurred without reaction. Actually, the electron transfer
requires that E0(Mn/ Mn), which increases with nduring the spontaneous coalescence, becomes
higher than the reference potential E°(D /D). The reduction potential of the reference donor
thus creates a threshold for n, that is a critical nuclearity nc . For n < nc , it is observed indeed
that the coalescence occurs as in the absence of D, and that isolated ions and charged clusters
can not be reduced except by radiolytic species. However, when n ≥ nc , D starts to decay at
the end of the induction time due to the transfer of an electron to Mn. Simultaneously,
the clusters grow autocatalytically by successive reduction of adsorbed ions to atoms and
addition of a supplementary ion. Once formed, a supercritical cluster behaves indeed as a
growth nucleus. The sequence of alternate reactions of ion fixation and reduction makes the
cluster reduction potential more and more favourable for the transfer, so that an autocatalytic
growth is observed up to the total consumption of D or M. Actually, nc is deduced from the
time-dependent donor concentration variation and from a numerical simulation taking into
account the competitive processes of coalescence and electron transfer [7, 11]. By changing
the reference potential in a series of reduction monitors (such as reduced forms of sulfonato-
propyl viologen, methyl viologen or naphthazarin), the dependence of the cluster potential
on the nuclearity has been obtained, for example for silver and copper clusters. The potential
E°(Agn/Agn)aq in water increases with n as shown in Figure 6 [11]. A value of the reduction
potential is reached asymptotically at the nuclearity around n = 500 (radius ≈ 1.25 nm).
More generally, the reduction potential of any metal atom E°(M/M0) is expected to be quite
Ä&%*Ä
2!$)!4)/.#(%-)3429
negative, and the oligomer potential to be lower than that of the bulk metal with a similar
variation as in Figure 6.
Figure 6 : Size-dependence of the reduction potential E° of silver clusters in water (O) and of the ionization
potential IP of silver clusters in the gas phase (Δ). The reduction potentials refer to the normal hydrogen elec-
trode one which is at 4.5 eV above vacuum [11].
The same phenomenon of a required critical nuclearity should exist as well in

chemical reduction by donors having a reduction potential much less negative than
radiolytic species. It is indeed well known that the reduction of a silver salt occurs essentially
on the walls of a vessel where it forms a silver mirror. In contrast to free ions which are not
reducible, the ions adsorbed on the walls present a potential higher than E°(D /D).
The potential E°(Agn/Agn)aq represents the energy difference required to ionize the
cluster in water compared to the ionization energy of the normal hydrogen electrode, which
is 4.5 eV relative to vacuum. Thus, a comparison can also be made between the nuclearity
effects on E°(Agn/Agn)aq and on the ionization potentials IP (Agn) of bare silver clusters in the
gas phase [12] (Fig. 6). It is obvious from this figure that the variations of E° and IP do exhibit
opposite trends vs. n for the solution and the gas phase, respectively. The origin lies indeed
in the solvation free energy of the cation Agn which in water assists the ionization of Agn.
The difference between the ionization potentials of bare and solvated clusters decreases
with increasing n, and corresponds fairly well to the solvation free energy of the cation Agn
deduced from the Born model.
An important consequence of the nuclearity-dependence of the reduction potential

concerns the cluster nucleation/growth process itself. As shown in Figure 2c, the potential
Ä&%+Ä
of most chemical donors D is not negative enough to start a homogeneous reduction

of isolated ions in the bulk with the potential E° (Ag / Ag1) = – 1.8 VNHE [2] (Fig. 6). They
contribute instead by reducing in situ ions adsorbed on supercritical preformed particles (or
on walls) (E° (Agn / Agn) > E° (D / D)) to develop larger ones. In the example of Figure 3c, the
ligand ethylenediamino tetracetate (EDTA) protects the clusters from coalescence during
the first step of irradiation. They are spherical with a size of 10-15 nm. However, if a large
part of the ions are not radiolytically reduced but adsorbed on the former clusters, they are
slowly reduced after days by EDTA in a post-effect as in Figure 2c, whereas solutions of silver
ions complexed by EDTA are stable. The growth is orientated and favours the 111 surface,
giving pellets of 100-150 nm [6]. They are bluish-grey and a second band centred at 1100 nm
appears in their optical absorption spectrum. These pellets are particularly efficient for surface
enhanced Raman scattering (SERS) detection of single molecules adsorbed on Ag. Similarly,
gold clusters formed by partial radiolytic reduction of a solution of gold ions AuCl4− are very
slowly developed by reduction of adsorbed excess ions by PVA to form homodisperse cubic
crystallites which range in size from 10 to 500 nm, depending on the respective parts of
radiolytic and chemical reduction [6].
Bi- and multi-metallic clusters
Clusters containing two or more metals are of high interest for various applications,
particularly catalysis, and their physical and chemical properties can be adjusted according
to the composition [3, 6]. In alloyed clusters, the different metals are all present among the
surface atoms of the cluster and can accelerate the successive steps of the catalyzed reaction.
Radiolytic reducing agents are strong enough to reduce rapidly the ions of any
metal. However, as shown by pulse radiolysis of mixed solutions, further electron transfer
from less noble atoms to the ions of the more noble metal may lead to a privileged reduction
of the latter, then to the reduction of the former in a shell coating the initial noble metal core,
resulting in a core-shell structure (Fig. 7 top). The optical absorption band due to the surface
plasmon is in these cases progressively shifted in mixed solutions irradiated by G-rays (a few
kGy h−1) from that of the noble metal core to that of the coating less noble metal. But, at
high dose rate (delivered for example by pulses of an electron beam with a few kGy s−1),
the reduction of all ions can be achieved rapidly (Fig. 7 bottom) before any inter-metal
electron transfer can occur. In that case, the coalescence occurs between the various atoms
irrespective of their potential but according to their abundance, and an alloyed structure of
the cluster is found having the same composition as the initial ion solution [3, 4]. This structure
is checked by the evolution of the optical absorption spectrum, by X-ray photon spectroscopy
(XPS) and X-ray analysis at increasing dose, and by electron scattering patterns.
Ä&%,Ä
2!$)!4)/.#(%-)3429
Figure 7 : Scheme for the influence of the dose rate on the competition between the inter-metal electron
transfer and the coalescence processes during the radiolytic reduction of mixed metal ion solutions. Top) Low
dose rates favour core-shell segregation of the metals in the cluster due to electron transfer from less noble metal
atoms (in yellow) to noble metal ions (in blue). The noble metal atoms thus constitute the core and are coated
by the other metal. Bottom) High dose rates favour alloying of the metals by sudden reduction of all the ions,
followed by the coalescence of atoms according to their abundance. The cluster size is also smaller (Fig. 2).
Depending on the rate of inter-metal electron transfer competing with the dose
rate-dependent coalescence [6], alloying is spontaneous, even at low reduction rate, as for
Cu3Pd, CuPd, NiPt, CuAu, AgPd and AgPt [3, 4], whereas it is only obtained at high dose rate,
as for AuAg, AuPt and AuPd [3].
Alloyed multi-metallic clusters have been synthesized in the same way [3, 6]. However,
characterization of the alloyed structure is still more difficult than for bimetallic clusters. The
alloyed character is inferred rather from their catalytic properties, for example, which are
enhanced when they are produced at high compared to low dose rate.
Optical limitation
At low light fluence, clusters behave like other molecules and their optical absorbance
is independent of the exciting light fluence because they relax rapidly (Fig. 8) [13]. However,
at very high light fluence (≥ 0.9 J cm−2) delivered by a short laser picosecond pulse (30 ps),
the absorbance of gold clusters increases suddenly, particularly for large clusters (15 nm
Ä&%-Ä
mean radius), indicating a non linear optical behaviour. The metal cluster nanosol behaves
as an ultrafast optical limiter or shutter, which can be used as a shielding against powerful
lasers. From the analysis of the relaxation kinetics, which depends on the energy confinement
(fluence, pulse duration, cluster size), two kinds of mechanisms are found: in the first one,
faster than 1 ns, the gold cluster cumulates the energy absorbed, vaporizes and expands, so
that scattering of light causes a sudden decrease of the direct light transmission [13] (Fig. 8).
Then, in the second mechanism, after energy transfer to the solvent, the cluster relaxes by
condensation, but the new solvent bubble created by the solvent heating around the cluster
now expands, and also behaves as a secondary scattering centre up to the solvent bubble
cooling and the energy dissipation to the bulk. The cluster size relaxation is not strictly
reversible, and the cluster seems to undergo fragmentation during the 20 first pulses. But
the results are then reproducible, and an optical limitation with half efficiency is reached. For
medium fluences, the first mechanism is negligible.
Figure 8 : Non-linear optical effects. Scheme for the formation of two types of scattering centres at increasing
fluences of a laser pulse used to excite gold clusters. High fluence : vaporized and expanded metal clusters limit
light transmission immediately after the pulse and are replaced, after energy transfer to the solvent, by solvent
bubbles acting as secondary scattering centres. Medium fluence : only solvent bubbles are formed. Low fluence :
light transmission depends linearly on the intensity.
Catalytic properties
One of the important applications of metal clusters is catalysis. The clusters are fixed
on a support and their role is to accelerate the reaction between reactants. A catalytic process
generally involves repetitive multistep reactions where reactants are transformed into
Ä&%.Ä
2!$)!4)/.#(%-)3429
products, and where the catalyst, in spite of its transient interactions with the reactants
and products, is steadily regenerated. The catalysts should offer a large specific area in
order to accelerate the access of reactants to the active sites. Ultra-divided metal clusters
are thus particularly efficient in a number of reactions. However, as in the autocatalyzed
growth described above (Fig. 2c), the catalyzed reaction is controlled not only by kinetics
(large area), but also by thermodynamics (size-dependent reduction potential) [14] (Fig. 9).
These aspects are often overlooked, although they concern each elementary step of the
mechanism. Actually, when an overall electron transfer reaction between a donor D and
an acceptor A is catalyzed by metal clusters (Fig. 9), they have the capability to play the
role of an intermediate electron relay due to their reduction properties: in a repetitive cycle,
alternately Mn acts as a donor faced to A, and Mn as an acceptor faced to D [14]. Very small
amounts of Mn/Mn are thus able to transform large concentrations of D/A into D/A−. This
mechanism will reach the optimized efficiency provided E°(D/D) < E°(Mn/Mn) < E°(A/A−) (Fig.
9). This thermodynamic condition also implies that the size of the cluster is small enough to
correspond to a potential value in the appropriate range, between two thresholds imposed
by D and A. Otherwise, for a given system D/A, too a large size cluster, (with E°(A/A−) < E°(Mn/
Mn)), a fortiori a bulk metal, will be unable to transfer the electrons to A, and too a small
one, (with E°(Mn/Mn) < E°(D/D)), will be irreversibly corroded by transferring electrons to
A without receiving electrons from D. Alloying between two or more metals permits one to
adjust the potential according to the requirements of the successive catalyzed steps.
Figure 9 : Scheme of the mechanism of a catalytic electron transfer involving metal clusters as a relay. The
thermodynamic conditions to be fulfilled are that the cluster reduction potential is higher than the donor D and
lower than the acceptor A potential. This implies an optimized range for the cluster size value.
Ä&&%Ä
The mechanism of some catalyzed reactions of electron transfer has been studied
by pulse radiolysis, for instance at the surface of colloidal TiO2 particles [14]. Due to their
very small and homodisperse size and to the strong interaction with the support, radiation-
induced metal clusters are remarkably efficient in various reactions. For example, the radiolytic
method has been used to graft metal nanoaggregates onto anodes or cathodes involved
in the electrochemical chlorine-soda process. In both cases, significant overpotentials are
usually measured on unmodified electrodes. Radiation-induced bimetallic nanoparticles
(such as Pt-Ru and Ni-Ru), once grafted onto bulk metal electrodes (Ti or Ni), display a
remarkable electrocatalytic efficiency. As a synergistic effect, a drastic decrease of the
overpotential is observed when Pt and Ru are alloyed in atomic ratio 2:1 [3]. Similarly, the
electrochemical behaviour of transparent SnO2 counter electrodes grafted with Pt, Ru or Rh
clusters approaches that of bulk Pt electrodes. Platinum Ptx(CO)y and trimetallic Pt-Ru-Sn
alloyed clusters, strongly anchored to the surface of carbon fibres and powders with high
loadings up to 60 Pt wt.%, are very active as electrocatalysts for methanol oxidation in fuel
cells [3] while Pt-Co clusters are very efficient for oxygen reduction.
Radiation-generated platinum and platinum-tin supported on A-Al2O3 or CeO2

appeared as particularly active catalysts in the oxidation of vehicle exhaust gases (deNOx
reactions). The temperature-dependence of the methane conversion in the vapo-cracking
reaction (CH4 )2O $0 )2) has been studied over radiolytic catalysts Ni/A-Al2O3.
The catalytic tests indicate that the reaction starts at a low temperature with a very high
selectivity. When complexed Ni(NH3)6 ions adsorbed on CeO2 are reduced radiolytically at
room temperature, highly dispersed Ni0 and intermetallic phases NiCe and Ni2Ce are detected.
The catalytic activity in the benzene hydrogenation reaction is remarkably high and the total
conversion into cyclohexane is achieved in a markedly low temperature range [15].
Silver Photography
Silver photography is based on the formation, during the exposure, of the latent
image made of silver clusters distributed at the surface of each silver halide crystal embedded
in the gelatine and containing from 0 to 10 atoms.
Although the environment of the clusters is different in solutions and at the surface
of silver halide crystals, we proposed to extend the same growth mechanism, which was
demonstrated by pulse radiolysis for Agn clusters free in solution, to a theoretical explanation
of the development process in photography [7] (Fig. 10). The development occurs in both
cases at the interface between an aqueous solution of an electron donor and a silver cluster
acting as an autocatalytic growing site alternately accepting electrons and silver ions.
Ä&&&Ä
2!$)!4)/.#(%-)3429
Figure 10 : Photographic development mechanism. The reduction potential, E°(Agn+/Agn ), of the latent image
clusters, when in contact with a solution, increases with the number of atoms n. Therefore a nuclearity threshold
for development is created by the redox potential of the developer E°(D+/D). Above the critical nuclearity nc , the
potential E°(Agn+/Agn ) is higher than E°(D+/D), and alternate electron transfer toward Agn+ and Ag+ adsorption
on Agn allows the cluster to grow autocatalytically. On the contrary, when E°(Agn+ / Agn) is lower than E°(D+/D),
corrosion of subcritical clusters takes place by oxidizing molecules, such as D+ or Ox [7].
In fact, the various aspects of the catalytic growth revealed by the kinetic
studies of solutions mimic the characteristics of the development known experimentally
to photographers, namely the existence of a critical nuclearity as a lower limit for the
development, and the dependence of its value on the reduction potential of the developer
[7]. According to the mechanism established in solution, the discrimination induced by
the developer is explained as the consequence of a quantum-size effect on the reduction
potential of the silver nuclei which, at the aqueous interface, does increase with n: the
critical nuclearity nc, as a lower limit, is determined by the threshold imposed by the first
one-electron reduction potential of the developer (nc = 3-5 for usual developers) (Fig. 10).
Sensitization by ions of a noble metal such as gold generates alloyed clusters having a more
positive potential. Only supercritical oligomers with a potential higher than that of the
developer act as growth nuclei. The model is now widely accepted.
Another important application to photography of our understanding of the

cluster formation mechanism in radiation chemistry is the enhancement of the sensitivity
of photographic emulsions (Fig. 11) [16]. The primary effect of the photon absorption by
the silver halide crystals of the emulsion during the exposure is indeed to produce one
Ä&&'Ä
Figure 11 : Photographic latent image formation in undoped and formate-doped and gold-sulfide (S)
sensitized AgBr crystals. Top : undoped crystal with electron-hole recombination. Centre : formate-doped
crystal and hole scavenging step by formate (HCO2–). Bottom : formate-doped crystal and delayed reduction
step of additional silver ions by carboxyl radicals CO2to [16].
Ä&&(Ä
2!$)!4)/.#(%-)3429
electron-hole pair (e–-h) per photon absorbed. However, a fraction only of the holes (20%)
diffuses to the crystal surface and is eliminated by gelatine. Indeed, the effective sensitivity
decreases markedly due to the very fast recombination of part of the initial pairs with
production of fluorescence before electrons reduce silver cations, and to the oxidation by
holes of the newly formed atoms (Fig. 11 top). This analogy with tOH oxidation processes
occurring in irradiated solutions prompted us to use the same scavenging method to inhibit
the fast electron-hole pair recombination and the silver atom oxidation by the holes. When
the silver halide crystal has been doped during the precipitation with formate ions (a low
amount, HCOO–/Ag = 10–6, is sufficient), the holes produced simultaneously with electrons
during the exposure are indeed immediately scavenged by the formate ions HCO2– (Fig. 11
centre) and produce carboxyl radicals CO2to (Fig. 11 bottom). The fluorescence arising
from the recombination is suppressed and the quantum yield is enhanced from 0.20 for
the undoped reference up to 1 atom per photon for the doped emulsion after immediate
developing. This means that the loss of electrons by recombination is totally suppressed.
The yield is still doubled to an effective quantum yield of 2 atoms per photon when the
development is delayed, because a supplementary reduction of Ag by the strong reductant
carboxyl radicals CO2to occurs within 15 minutes (Fig. 11 bottom). This remarkable overall
enhancement by ten-fold is applicable to all processes based on silver imaging, black and
white and colour photography, radiography or holography [16].
Conclusion
Radiation chemistry methods have been shown to be of high potentiality to generate

small- size monodisperse metal clusters, and to study the dynamics of nucleation and growth
of clusters, mono- and multi-metallic, from the monomers to the stable nanoparticle. Their
structure, core-shell or alloy, is governed by the dose rate used. Pulse radiolysis provides
the means to observe directly by time-resolved techniques their reactivity, especially to
determine during the growth their nuclearity-dependent properties, such as the optical
absorption spectrum and the reduction potential. These are of crucial importance for the
understanding of the mechanism of the cluster growth itself, in the radiation-induced as
well as in the chemical or photochemical reduction processes. This knowledge is exploited
to guide the synthesis of the clusters and the control of their size and shape. These specific
properties have important applications in several fields such as electronics, optical limitation,
catalysis, and photographic latent image formation and development.
Ä&&)Ä
References
[1] Delcourt M.-O., Belloni J., Capture de précurseurs de l’hydrazine par les ions Cu au cours de la
radiolyse de l’ammoniac liquide, Radiochem. Radioanal. Lett., 1973, 13, 329-338.
[2] Henglein A., The reactivity of silver atoms in aqueous solutions (a G-radiolysis study), Ber.
Bunsenges. Phys. Chem., 1977, 81, 556-561. Tausch-Treml R., Henglein A., Lilie J., Reactivity of
silver atoms in aqueous solution. II. A pulse radiolysis study, Ber. Bunsenges. Phys. Chem., 1978,
82, 1335-1340.
[3] Belloni J., Mostafavi M., Remita H., Marignier J.-L., Delcourt M.-O., Radiation-induced synthesis of
mono- and multi-metallic clusters and nanocolloids, New J. Chem., 1998, 23, 1239-1258.
[4] Marignier J.-L., Belloni J., Delcourt M.-O., Chevalier J.-P., New microaggregates of non noble
metals and alloys prepared by radiation-induced reduction, Nature, 1985, 317, 344-345. Belloni J.,
Marignier J.-L., Delcourt M.-O., Minana M., Non noble metal catalytic microaggregates, a method
for their preparation and their application in the catalysis of the photoreduction of water., US Pat.
1986, 4,629,709. Belloni J., Marignier J.-L., Delcourt M.-O., Minana M., New mono or multimetal
microaggregates, a method for their preparation and their application in the catalysis of the
photoreduction of water, US Patent 1987, 4745094.
[5] Janata E., Henglein A., Ershov B.G., First Clusters of Ag Ion Reduction in Aqueous Solution, J.
Phys. Chem., 1994, 98, 10888-10890.
[6] Belloni J., Mostafavi M., Metal and Semiconductor Clusters, in: Studies in Physical and Theoretical
Chemistry 87. Radiation Chemistry : Present status and future trends, Jonah C.D., Rao M. (eds),
Elsevier, 2000, 411-452.
[7] Mostafavi M., Marignier J.-L., Amblard J., Belloni J., Nucleation dynamics of silver aggregates.
Simulation of the photographic development process, Radiat. Phys. Chem., 1989, 34, 605-617.
Belloni-Cofler J., Amblard J., Marignier J.-L., Mostafavi M., La photographie révélée, La Recherche,
1990, 212, 48-57. Belloni-Cofler J., Amblard J., Marignier J.-L., Mostafavi M., The principles of
photographic development, Endeavour, 1991, 59, 2-9. Belloni-Cofler J., Marignier J.-L., Amblard
J., Mostafavi M., Guyon S., Roussi G., The Magic of the Image, 26 min film, CNRS, 1989.
[8] Mostafavi M., Keghouche N., Delcourt M.-O., Belloni J., Ultraslow aggregation process for silver
clusters of a few atoms in solution, Chem. Phys. Letters, 1990, 167, 193-197. Remita S., Orts J.M.,
Feliu J.M., Mostafavi M., Delcourt M.-O., Chem. Phys. Letters, 1994, 218, 115-121.
[9] Surendran G., Tokumoto M., Pena dos Santos E., Remita H., Ramos L., Kooyman P.J., Santili C.V.,
Bourgaux C., Dieudonné Ph., Prouzet E., Nanomaterial Synthesis in Highly Swollen Hexagonal
Phase Templates, Chem. Mat., 2005, 17, 1505-1514.
[10] Krishnaswamy R., Remita H., Impéror-Clerc M., Even C., Davidson P., Pansu B., Synthesis of single-
crystalline Platinum nanorods within a “soft” crystalline surfactant-Pt(II) complex, Chem. Phys.
Chem. 2006, 7, 1510-1513.
Ä&&*Ä
2!$)!4)/.#(%-)3429
[11] Khatouri J., Mostafavi M., Amblard J., Belloni J., Ionization potential of clusters in liquids., Z. Phys.
D., 1993, 26, 82-86.
[12] Jackschath C., Rabin I., Schulze W., Electron impact ionization potentials of gold and silver clusters
Me[n] n ≤ 22, Z. Phys. D, 1992, 22, 517-520 ; G. Alameddin, J. Hunter, D. Cameron, Kappes M.M.,
Electronic and geometric structure in silver clusters, Chem. Phys. Letters, 1992, 192, 122-125.
[13] François L., Mostafavi M., Belloni J., Delaire J., Optical limitation of gold clusters. Mechanism and
Efficiency, Phys. Chem. Chem. Phys., 2001, 3, 4965-4971.
[14] Khatouri J., Mostafavi M., Belloni J., Kinetics of electronic transfer in solution catalyzed by metal
clusters, in Photochemistry and Radiation Chemistry : Complementary methods for the study
of electron transfer. Wishart J., Nocera D. (Eds), Adv. Chem. Series, 1998, 54, 293-314. Belloni J.,
Nucleation, growth and properties of nanoclusters studied by radiation chemistry. Application to
catalysis, Catal. Today, 2006, 113, 141-156.
[15] Keghouche N., Chettibi S., Latrèche F., Bettahar M. M., Belloni J., Marignier J. L., Radiation-induced
synthesis of α-Al2O3 supported nickel clusters : Characterization and catalytic properties, Radiat.
Phys. Chem., 2005, 74, 185-200. Chettibi S., Keghouche N., Wojcieszak R., Boudjennad E.H.,
Belloni J., Bettahar M.M., Catalytic properties of CeO2- supported nickel clusters synthesized by
radiolysis, Catal. Today , 2006, 113, 157-165. Redjala T., Remita H., Apotolescu G., Mostafavi M.,
Thomazeau C., Uzio D., Bimetallic Au-Pd and Ag-Pd clusters synthesized by gamma or electron
beam radiolysis and study of the reactivity/structure relationships in the selective hydrogenation
of buta-1,3-diene, Gas and Oil : Science and Technology, 2006, 61, 789-797.
[16] Belloni J., Treguer M., Remita H., De Keyzer R., Enhanced yield of photoinduced electrons in doped
silver halide crystals, Nature, 1999, 402, 865-867. De Keyzer R., Treguer M., Belloni J., Remita H.,
Photosensitive silver halide element with increased photosensitivity, US patent 6,436,625, Aug.
20, 2002.
Ä&&+Ä
Chapter 8
Water radiolysis
in cement-based materials
Pascal BOUNIOL
Introduction
Cement-based materials, that are widely used in nuclear industry, concern two main
fields of applications :
1/ matrices of coating or coverage of radioactive wastes of low and medium level

activities within primary or composite waste-forms (Fig. 1) ;
a b
Figure 1 : Examples of cement-based matrices used in the conditioning of radioactive wastes : a) concrete
container accommodating 4 primary waste forms ; b) mortar for embedding metallic wastes.
Ä&&,Ä
2!$)!4)/.#(%-)3429
2/ concretes for structural purposes or for radiation shielding within installations,

which are able to be exposed to strong irradiation (silos and casemates for reactor
vessel or burned fuels storage and high activity wastes).
The relatively moderate cost, great availability and flexibility in their use, the wide
available composition and the high physico-chemical inertia (e.g. buffer medium, resistance
to fire and irradiation) cause an inevitable interest in the use of cement-based materials. On
the other hand, cement-based materials such as grouts, mortars and concretes constitute
complex mediums of which some ingredients can prove weak with age, when the nature and
the combination of diverse stresses such as drying, leaching, chemical attack, mechanical
effort, irradiation, etc. are considered. Typically in this case, the free water is the major cause
of the encountered problems, when the preservation of the concrete properties and integrity
is the main concern (concept of durability). In the presence of internal- or external-source of
irradiation, the action of ionizing radiation results, in particular, in a decomposition of water
and mainly a production of gaseous dihydrogen. Besides the ageing mechanism, which is
mainly related to the heating and drying (the direct radiation effects are indeed negligible), the
radiolysis phenomena raise a real safety problem regarding the risk of internal pressurization
or gas explosion in ionizing environment. In order to adopt the best countermeasures
in industrial situation (e.g. ventilation of the buildings to provide a maximum of 3 to 4%
H2 in the air) or to try to inhibit the radiolysis while intervening on the process through
formulation of material, addition of specific products, it is always advisable first to evaluate
the level of risk with an adequate accuracy. The aim is, beyond the rough estimation of the
source term H2 alone, to describe correctly the radiation chemistry within the interstitial
liquid in the material, as well as the satellite radiolysis phenomenologies by taking account
of the greatest number of elementary mechanisms. A description as complete as possible
of the radiolysis phenomena in a macroscopic and heterogeneous system (e.g. mineralogy
and porosity resulting from the hydration of cements, transport of gases in porous medium,
transport of the radiation and energy deposition, heterogeneous chemistry gas/solution/
solid phases, etc.) is necessary for the development of an operational model of behaviour.
Though not developed in this study, a high level of coupling exists between these various
phenomena and radiation chemistry. We will mainly focus on this last point in the following
developments.
The cement based matrix: an original medium
Independent of the material implemented, the binder phase ensuring cohesion of

the aggregates and mechanical rigidity comprises :
Ä&&-Ä
2ADIOLYSISINCEMENT BASEDMATERIALS
- anhydrous cement (mixture of mainly calcium silicates and aluminates) ;
- hydration products of these minerals, essentially amorphous silicates and

crystallized Ca(OH)2 (Fig. 2) :
Figure 2 : Hardened paste of hydrated Portland cement observed with MEB : a) tabular crystals of Ca(OH)2
(portlandite) on fissured calcium silicate hydrate substratum ; b) zoom on the intrinsic porosity of calcium sili-
cate hydrate.
- a multimode porosity, including fine pores of just a few nanometers as well as

coarser capillary sections up to hundreds of nanometers, occupied by residual
liquid and air (Fig. 3) :
Figure 3 : Distribution of pore sizes of an hydrated Portland cement paste 4 months old (initial mass ratio
water/cement of 0.4) obtained by mercury intrusion. The pore family related to the calcium silicate hydrate
(some nm) and that of capillary porosity (a few hundreds of nm) can be easily distinguished. The finest porosity
is completely saturated with interstitial liquid while the capillary porosity is partially.
Ä&&.Ä
2!$)!4)/.#(%-)3429
All these make cement matrix a composite, multiphasic and very heterogeneous
material, that is difficult to be studied. In addition, cement matrix has also an evolutive
character due to the progress of hydration, modifying the microstructure in its early ages
(from a few months to several years). From a physical point of view, the amorphous character
of the main hydrated phase containing numerous defects allows the material to adapt to
various conditions of temperature and irradiation. In addition, the presence of a porosity
accessible to gases ensures an effective transport of radiolytic dihydrogen, essentially by
gas diffusion. This means that, when H2 is evacuated, it is replaced by air in which the O2
fraction interferes with the radiolytic processes. From a chemical point of view, the presence
of alkaline elements in cements (K and Na) makes the pore liquid in contact with the solid
phases to be far from being pure water, since it is an aqueous solution of pH higher than 13
with an ionic strength close to 0.2 mol/l. The important proportion of crystallized Ca(OH)2
(portlandite) with an already high balanced pH (12.45 at 25 °C) provides this medium with a
reinforced buffer power, with respect to which the chemical variations induced by radiolysis
remain overall very limited.
In the case of a concrete structure, the presence of steel reinforcements can further
complicate the system by causing a local heterogeneity since they constitute in terms of
dose rate (driving of the radiolysis) or by the influence of iron ionic species on radiolysis
at the interface with the cement paste. Although still not very well-studied, the resulting
radiolysis-corrosion coupling (as a function of temperature) is identified as the only aspect
which could possibly affect the concrete durability under radiation [1].
Effect of cement chemistry on radiolysis
Even if radiolysis does not, or only very weakly, influence the chemistry of the cement
medium, the reverse assertion is not true. The radiolysis of the interstitial liquid only relates
to the solvent (H2O) which gives approximately 18 species, well-known for most of them
(Chapter 1). Due to the very high pH, only the basic forms will play a role. Thus, the equilibria
listed below implying the main radiolytic species are displaced towards the right :
HO2t0) j O2t )2O pKa = 4.80 (1)

Ht0) j eaq pKa = 9.60 (2)
H2O20) j HO2 )2O pKa = 11.68 (3)
t
0)0) j Ot )2O pKa = 11.90 (4)
Before leading to the formation of these various species in the solution at the homogeneous
stage, the high concentration of OH ions interferes with the primary production processes
Ä&'%Ä
Figure 4 : Evolution of primary yields for the main species of the water radiolysis (gamma radiation) accor-
ding to the pH at 25 °C. The pH of the cement based matrices, beyond 13, is in the most basic zone.
within the radiolytic spurs due to a higher encounter probability. Figure 4 shows thus that
between pH 12 and 14, the primary yields of tOH and eaq radicals increase significantly,
whereas the primary yields of H2 and H2O2 decrease more moderately. At the end of
approximately 10 6 s, about sixty secondary reactions between species resulting from the
primary stage initiate the already known mechanisms involved in the radiolysis of water
(Chapter 1) with a specific orientation, due to the very high basicity of the medium :
– Respective attacks of H2 by the oxidizing radical Ot , and of HO2 by eaq are faster
in comparison with the homologous reactions constituting the chain-reaction
highlighted by A.O. Allen [2] at neutral pH. This results in a persistence of that chain
in alkaline medium under an alternative form, which is theoretically more effective
to recycle H2 and inhibit the water decomposition:
initiation :
Ot )2 m eaq k = 1.2 s 108 l mol 1 s 1 (5)
propagation :
eaq )02 m Ot 0) )2O k = 3.5 s 109 l mol 1 s 1 (6)
Ä&'&Ä
2!$)!4)/.#(%-)3429
– The peroxide and superoxide species, which naturally tend to accumulate, are
prone to very slow reactions of disproportionation (reactions (7), (8)) (< 1 l mol 1 s 1).
These reactions are also the cause of the secondary dioxygen formation (always
less compared to H2), that occurs at long time. Among these reactions, specially the
three following can be listed, the last one (mixed type) being a basic equivalent of
the Haber-Weiss reaction :
H2O2)02 m O20) )2O (7)

O2t 02t m O2022 (8)
HO2 02t m O20) 0t (9)
Like in neutral medium, the presence of dioxygen partially interferes with the chain-
reaction. That is to say that in a closed system, safe from air, a regulation of the radiolysis leading
to a steady state is always possible by a mass action mechanism relating to the whole reaction
system. This also means that, in the presence of a strong reductant inducing a ratio of initial
concentrations such as [H2] / [O2] > 2, the chain-reaction is strongly activated, causing O2 to
disappear and resulting in a low and stable concentration for H2. This mechanism is similar
to that implemented in the pressurized water nuclear reactors where a partial pressure of H2
is applied in order to obtain a slack pressure and an anoxic medium (Chapter 4). Under these
conditions, at any time the amount of H2 produced is equivalent to the amount of H2 consumed.
With a very high pH, the second peculiarity of the cement medium lies in the presence
of important amounts of calcium in the form of portlandite Ca(OH)2. Paradoxically, the content
of calcium in solution appears rather moderate, approximately 3 s 10 mol l–1 at 25 °C.
However, this might be of a secondary importance compared to the fact that portlandite
provides a permanently available and quasi-inexhaustible source of calcium. This aspect can
be considered as interesting since a new solid compound, the calcium peroxide octahydrate,
can be formed due to the increase in the peroxide concentration during the radiolysis [3].
With a very low solubility (Ks = (Ca)(O22 )(H2O)8 = 2.5 s 10 11), CaO2.8H2O precipitates indeed
in contact with the portlandite as soon as the solubility product Ks is reached (heterogeneous
precipitation not requiring a supersaturation), and grows gradually with the consumption of
the portlandite. When the two solid phases coexist, the equilibrium relations with the pore
solution already buffered in term of pH result in obtaining a mineralogical super-buffer where
portlandite and CaO2.8H2O respectively control the concentrations of calcium and peroxide
in the solution. The concentration of the whole of the species, and especially of O2 and H2,
becomes constant by propagation to all other equilibria, which results in a second mode of
radiolysis regulation within the interstitial liquid. Precipitation is however conditioned by a
strong production of radiolytic peroxide (i.e. by a high dose rate (Fig. 5)).
Ä&''Ä
Figure 5 : Evolution of the relative under-saturation index σ according to the duration of gamma irradiation
of a Portland cement paste for two dose rates (simulation). The solubility product Ks is reached for the highest
dose rate that causes an earlier regulation of the radiolysis in the pore solution.
The effect of the cement material porosity on radiolysis
Another set of relations of the heterogeneous type is also established between

the interstitial liquid and the gaseous phase present in the material porosity. At the end
of the hydration process of a Portland cement paste having an initial mass ratio of water/
cement < 0.42 (index generally associated to a good quality), the gaseous phase occupies
approximately 25% of the total porosity in a closed system. Taking into account the low
solubility of H2 and O2 initially formed in aqueous phase by radiolysis, the latter passes then
very easily into the gaseous phase of the porous volume, thus having significant effects on
the global behaviour of the system :
- in term of quantity of matter, H2 and O2 happen to be mainly in gaseous phase:

(about 57 and 93 molecules remain in solution, respectively, for 1000 in the gas
phase at 25 °C) ;
- in the gas phase, H2 is not attacked by the oxidizing radicals while in the solution
the relative enrichment in O2 favours the radiolytic decomposition of water and
subsequent formation of hydrogen ; therefore, H2 accumulates in the gas phase ;
- the accumulation of H2 gas strongly attenuates the efficiency of the regulation

mechanisms of the solution radiolysis previously described ; reaching a global
Ä&'(Ä
2!$)!4)/.#(%-)3429
equilibrium remains still possible but after a longer duration, since the hetero-
geneous equilibrium for gas/solution (Henry’s law) moreover introduces an
additional “resistance”.
The saturation degree of the pores obviously constitutes a control parameter of

the radiolysis. When increased, it leads more directly to equilibria in solution, and becomes
more efficient in the regulation of the phenomenon. In contrast, the regulation mechanisms
based on a high H2 concentration in solution are not very active and that causes production
and accumulation of gaseous dihydrogen in the system. The opening of porosity to the
atmosphere (open system) results in an alternative for this case: the expansion volume of
radiolytic gases becomes infinite and then, the decomposition of water occurs indefinitely. In
this last case, the fast evacuation of H2 by gas diffusion is accompanied with the penetration
of O2 in exchange. This would also interferes with the regulating chain-reaction of the “Allen”
type (reactions (5), (6)). In the macroscopic waste form level, the operational consequences
that arise due to previously mentioned issues, lead to conflicting situations in gamma
radiolysis: in a closed system the total pressure increases up to a moderate plateau value,
while in an open system the total pressure remains lower but ensures a constant H2 release.
From a more fundamental point of view, the porous medium also raises very
interesting questions on the effect of water confinement within the pores in nanoscale.
Thus, whatever the topological characteristics the network may be, it is legitimate to think
that the tight mixture of liquid and solid phases with a considerably developed interface
impacts not only the modes of energy deposition due to the irradiation, but also the
probabilities of meeting of the various species in solution. This last aspect is being studied
through the modification of the primary yields of the water radiolysis in porous media
model (glasses with monodisperse porosity). The first results clearly confirm the effect of
the confinement with a very significant increase in the primary production of the molecular
species H2 and H2O2 [4]. One of the hypotheses about this behaviour proposes that, if the
characteristic pore size is lower than the mean free diffusion path of the primary radicals,
the latter are forced to recombine, giving more molecular products. This effect starts to be
less significant when the pore size increases and then disappears when it becomes much
higher (a few tens to a few hundreds of nanometers) than the size of the radiolysis spurs. In
theory, the interference of porosity with the chemical reactivity should also impact all the
reactions in solution, consecutive to the primary stage, with a more or less important effect
on the kinetic constants, depending on the size and the charge of the species. This field of
interest should be developed in the future. For the moment, it does not warrant significant
corrections to the current model describing the radiolysis in cement based-materials in
which primary yields and kinetic constants still result from determinations in solution, in
Ä&')Ä
the absence of porous environment. In any case, it will be supposed to bring an answer to
certain questions of operational relevance: do primary yields still keep a meaning in porous
environments ? Would the super-buffer character of the cement medium and the presence
of a gas porosity be able to attenuate the range of the elementary mechanisms highlighted
in porous environment ?
Global phenomenological approach and simulation
The combination of the various phenomena involved in radiolysis allows us to

propose a model based on the previously evoked knowledge: primary yields, secondary
reactions, initial concentrations, dose rate, solid-solution and solution-gas exchanges [5].
A more accurate description of the system moreover results in studying it, according to
determining operational parameters such as the pH, the saturation degree of the pores,
the radiation source decay and the temperature. When the relations existing between the
various phenomenological compartments are examined (Fig. 6), it is seen that in terms of
internal pressure or quantity of released H2, the global result of radiolysis in a cement matrix
cannot be simplified to a simple addition of elementary phenomena.
In a system characterized by a strong heterogeneity on various scales and by

a large variety of mechanisms, the global behaviour proceeds, on the contrary, from the
complexity and the numerous couplings. The most obvious demonstration for this results
in the production of responses, at best non linear, indeed with thresholds. Figure 6 shows
thus that, starting from various initial parameters (intensity and type of the irradiation,
composition of the interstitial liquid), the evolution possibilities of the radiolysis are multiple,
with more or less broad loops according to operating conditions. The two solid-liquid and
liquid-gas heterogeneous equilibria, the impact of which is very strong because of quasi-
direct feedbacks on the medium, can be recognized in particular on the upper part of the
diagram (related to the radiolytic chemistry). With an operating mode in an open system,
the dimension of transport is superimposed with a feedback at the same time less direct
and more complex on the reactional medium (smaller H2 solubilization, contamination by
entering O2). In all cases, it is interesting to emphasize that the equilibrium pressure in the
porosity plays the central role in these feedbacks, more than the solution concentration.
This leads to consider that the cement matrix could exhibit a global behaviour more or less
similar, independently of the nature of the mechanisms in solution.
Ä&'*Ä
2!$)!4)/.#(%-)3429
Figure 6 : Physico-chemical parameters and their couplings involved in the radiolysis of cement-based ma-
terials at the level of a waste form. Heterogeneous equilibria (solid solution and solution-gas) are located by a
double red arrow.
An example of simulation of the radiolysis in a closed and partially saturated cement

medium illustrates perfectly the non-linearity of the behaviour for various dose rates (Fig. 7).
Hence, this kind of simulation shows that :
- partial pressure of H2 in the porous network is not proportional to dose rate D’,
except at the beginning of irradiation ;
- an equilibrium can be reached by the action of different mechanisms (by precipitation

of CaO2.8H2O for D’ = 0.2 and 0.4 Gy/s ; by mass action for D’ = 0.1 Gy/s) ;
- a close equilibrium plateau can be obtained under different conditions of irradiation

and mechanisms (note the important delay of setting to equilibrium in the case of
the mass action).
Ä&'+Ä
Figure 7 : Compared evolutions of the H2 partial pressure in a Portland cement paste with various gamma
constant dose rates (simulation for a degree of pore liquid saturation equal to 0.75). Dose rate of 0.1 Gy/s is not
sufficient to cause the precipitation of CaO2.8H2Ο and an earlier regulation of the radiolysis in the medium.
Although simulations happen to be the most economic and fastest way to carry
out the evaluation of the H2 source term in various cement-based materials, it is advisable to
perform periodical experiments, preferably following an important conceptual modifi-
cation, in order to check certain predictions and also to fix certain parameters of the model.
These experiments generally are related with the test-bars of cement material irradiated
in closed vessel or the material poured within air-tight mini-containers (Fig. 8).
Figure 8: Radiolysis experiments on cement pastes leading to know the total pressure and the composition
of gases according to the duration of gamma irradiation : a) sealed glass phial around a test-bar 4 s 4 s 16 cm ;
b) mini-containers out of stainless of 500 cm3 (allowing a quasi-integral filling of cement paste) with outlets for
P(H2) measurement in line.
Ä&',Ä
2!$)!4)/.#(%-)3429
The follow-up of the time dependence of the total pressure and the H2 partial
pressure, supplemented by a complete analysis of the gas space by chromatography, allows
comparison between experiment and simulation reasonably under well-defined operating
conditions. One should keep in mind that a badly characterized porosity or an inaccurate
dose rate constitute the most common causes of divergence, with the strongest impact.
The currently selected model is based on strong assumptions of homogenization
(e.g. dose rate, average concentrations and gradients) and can be described with an
1-D geometry. This model enables us to calculate the total and partial pressures to a satisfying
order of magnitude as well as a minority presence of O2 even safe from the air, and also a
similar evolution of the long-term behaviour (1 year and more).
Besides the fundamental studies which will determine if the cement matrix can
be still regarded as an overall homogeneous medium in spite of the incidence of the
porous environment on the primary phenomena and the reactions between species, the
prospects for the model improvement still remain to be numerous. They concern primarily
the chemistry of elements naturally present in cement (Fe, S, C) and interfering with the
radiolysis, as well as the chemistry of elements deliberately added to the wastes (N, Cl). The
temperature influence studies are of a great extent since this parameter affects practically the
whole of the physical and chemical phenomena proceeding during radiolysis : modification
of the kinetics rate constants, primary yields, equilibrium constants, transfer coefficients, etc.
With this point of view, the temperature-dependence of the solubility product of CaO2.8H2O
appears as an important data to be acquired.
Mainly motivated by the safety problem in relation with the “hydrogen” risk, studies
in progress about radiolysis find their application in the evaluation of gas term-source for
cemented radwastes in storage stage (i.e. at the beginning of the exploitation, when dose
rates are the highest). They also concern the packaging concepts integrating high activity
primary waste forms (nuclear glasses, burned fuel) where the coupling of radiolysis with
temperature or corrosion is particularly involved.
References
[1] Zuili D., Couvreur F., Bouniol P., Gorse D., Electrochemical study of the effects of gamma-ray
irradiation and temperature on the corrosion behaviour of reinforced concrete, séminaire
CEFRACOR Corrosion Acier, January 2001, Paris.
[2] Allen A.O., The radiation chemistry of water and aqueous solutions, Van Nostrand, New York,
1961.
Ä&'-Ä
[3] Bouniol P., Aspart A., Disappearance of oxygen in concrete under irradiation: the role of peroxides
in radiolysis, Cement and Concrete Research, 1998, 28, 1669-1681.
[4] Rotureau P., Renault J.P., Lebeau B., Patarin J., Mialocq J.C., Radiolysis of confined water: molecular
hydrogen formation, Chem Phys Chem, 2005, 6, 1316-1323.
[5] Bouniol P., État des connaissances sur la radiolyse de l’eau dans les colis de déchets cimentés
et son approche par simulation, Report CEA-R-6069, Direction des systèmes d’information CEA/
Saclay, 2004.
Ä&'.Ä
Chapter 9
Obtaining high performance
polymeric materials by irradiation
Xavier COQUERET
Introduction
The interaction of high-energy radiation with organic monomers and polymers

induces various chemical transformations of increasing importance from both scientific
and technological standpoints. Cross-linking and scission in polymers and networks,
polymerization of pure and blended monomers, grafting onto synthetic and natural
polymers as well as chemical activation of organic materials by oxidation can be induced by
irradiation under soft conditions. After more than fifty years of complementary basic and of
application-oriented research, the radiation chemistry of polymers has been found to lead
to many applications of current technological, economical and societal interest [1,2]. Better
understanding of the primary events and subsequent chemical reactions mediated by ionic
or free radical intermediates leads to an increasing number of industrial processes that
include: food, drug and medical device sterilization, cross-linking of temperature-resistant
thermoplastics and elastomers, surface patterning for micro- or nanotechnologies, curing of
coatings and composites.
High energy radiation exists under many forms, but the most important types for the
potential applications they have in polymer science and technology are X-rays and gamma
rays on the one hand, and accelerated electrons on the other. The recent development of
swift ion beam accelerators that induce reactions exhibiting some common features with
widespread radiation treatments cited at first will be also accounted for.
Ä&(&Ä
2!$)!4)/.#(%-)3429
The designation ionizing is sometimes improperly used to qualify high-energy

radiation since it is in fact too restrictive because the process of ionisation refers specifically
to the ejection of an electron from an atom or molecule thereby creating an ion-pair.
Rigorously speaking, high-energy particles or photons interacting with matter may activate
a variety of physical and chemical effects, including heat generation, atomic displacements,
electronic excitation of atoms and molecules, breaking of chemical bonds, and nuclear
reactions (Chapters 1-3). The specific effects depend on the type of radiation, the target, and
the irradiation conditions. In many situations, the chemistry that results is actually mediated
by free radical species produced by a number of pathways that depend strongly on the
chemical constitution, on the purity, and on the physical state of the irradiated matter [3].
The capability of high-energy radiation to bring about chemical changes with a

high degree of control over the spatial and temporal features of the exposure has led to
the development of a technology known as radiation processing. In the particular field of
polymers, but not only in that field, the technique has led to products exhibiting improved
thermal, chemical or wear resistance without the use of additives. For example, radiation
processing has been used to produce rubber latex gloves and other similar items free from
sulfur and from toxic nitrosamines, with less allergenic effects to wearers and with less
emission of sulfur oxides during final elimination by burning [4]. Many other applications will
illustrate the large variety of benefits that have been allowed by the gradual improvement of
basic knowledge and by the availability of well-adapted and reliable radiation sources.
Energy deposition resulting from radiation-matter interaction
The complete description of the interaction of some radiation with a substrate is

not limited to the detailed inventory of the induced chemical effects. The spatial distribution
of the deposited energy, and hence of the resulting chemical effects, is another feature of
prime importance (Chapter 1). The location of effects and of the side-reactions can indeed
be mastered to a large degree by the appropriate selection of the radiation type and of beam
characteristics.The energy loss per unit length of particle track is termed the stopping power
in nuclear physics and Linear Energy Transfer (LET) in radiation chemistry. Heavy charged
particles are referred to as high LET radiation, whereas X- or gamma rays, and fast electrons
are radiation with low and intermediate LET, respectively. LET depends on the nature of
the medium, and of the energy and electric charge of the interacting particle [5]. The data
reported in Table 1, given for water, illustrate this essential feature of polymer processing.
The kinetic energy of incident radiation is dissipated by various cascade processes

involving a myriad of elementary interactions that impart the affected atoms with enough
Ä&('Ä
0OLYMERSANDHIGHENERGYRADIATION
Table 1. Comparison of linear energy transfers (LET) for various elementary particles in water.
Rest Mass LET

Particle Charge Observation
(a.u.) (nJ m−1)
1 MeV (EB or secondary
Electron 0.00055 –e 0.032
to 60Co gamma)
Alpha 4.004 F 14.4 from 210Po
127 n+
I 127 OF 1120 65 MeV
energy to expel outer shell electrons, or at least give rise to electronic excitation. In a simplified
description, heavy charged particles (e.g., proton, deuteron, alpha) lose kinetic energy via small
energy transfers to atomic electrons in the medium. The more energetic interactions eject
electrons from their parent atoms and generate primary ion-pairs. The more energetic secondary
electrons can initiate additional ionizations, whereas less energetic secondary electrons induce
electronic excitations. At the end of the secondary electron tracks, several ion pairs and excited
species are concentrated in a small volume called spur. Only a small fraction of the initial energy
is transferred at each event. The cascade of events along the path of the moving particle thus
produces a track consisting of clusters of ions or spurs (Chapter 1) ( Scheme 1).
Scheme 1 : Cascade of events along the path of a high energy particle (electron) interacting with a covalent
molecule AB.
Ä&((Ä
2!$)!4)/.#(%-)3429
Electrons tracks are less dense than the tracks of heavy charged particles and the
spurs are more widely spaced (Chapter 4). The induced physical processes are comparable in
many regards to the effect of exposure to heavy particles or ions. X- and gamma rays actually
are indirect methods of producing fast electrons in matter. As a consequence, chemical
effects can be considered very similar in nature. Some specific features may nevertheless
arise from the significantly different dose rates (i.e. amount of absorbed energy per unit
time) as a consequence of different LET values.
Versatility of radiation processing applied to polymers
A few selected examples illustrate the appropriate choice of these different radiations.
Highly penetrating gamma or X- rays are well suited, due to their deep penetration, for
radiation processing, such as cross-linking of plastic parts, the items being packaged in
cardboxes (Fig. 1). On the other hand, medical supplies such as polyolefin syringes, tubing,
gloves, blood bags and medical prostheses packaged in sealed films, all made of radiation-
Figure 1 : 10 MeV electron beam accelerator (Rhodotron with its scanned beam circled with the red line) in an
EB/X-ray sterilization facility (courtesy of IBA, Belgium).
Ä&()Ä
resistant polymers, can be sterilized without degradation at the relatively low doses which
are sufficient to kill bacteria and viruses. Thanks to such a treatment under high energy
radiation, biomedical items can be prevented from contamination during their storage until
they are used.
Owing to the low apparent density of the boxes, palletised products can be treated
in a single pass in the irradiation chamber. Heat-shrinkable films that find increasing
applications in automated packaging processes are produced by cross-linking extruded
polyolefin materials with electron beam (EB) accelerators. Treating the thermoplastic wire
cable insulation with low to medium energy beams together with high intensity gives a high
through-put and an adapted penetration depth of some hundreds micrometers (Fig. 2).
Figure 2 : 700 keV EB accelerators for cross-linking cable insulators. By design of the system, the cable is driven
several times through the beam of the electron gun circled by a red line (courtesy of Acome, France).
Swift heavy ion beams with extremely high LETs are used to degrade strongly
polymer chains within tracks over a range of a few tens of micrometers. After irradiation,
the top surface of bulk thermoplastics such as polyethyleneterephtalate, polycarbonate,
polyvinylidene fluoride or polyimide materials as well as thin films made of the same material
are subsequently etched by a wet chemical treatment that reveals pores of which the shape,
surface density and dimensions can be controlled by choosing appropriate conditions. High
aspect ratio cylindrical or conical traces of diameter ranging from a few nanometers to some
Ä&(*Ä
2!$)!4)/.#(%-)3429
micrometers can be produced to form functional porous membranes or to supply templates

for the fabrication of nanowires, nanosensors and many other types of mico- or nanodevices
(Fig. 3).
Figure 3 : Isolated nanometric size pores in a polyvinylide-

nefluoride (PVDF) membrane obtained by Kr ion irradiation
at GANIL and subsequent oxidative etching of a 15 μm-thick
PVDF film (courtesy of LSI/CEA-CNRS).
These few examples illustrate the versatility and the broad applicability of radiation
processing to polymers, from simple performance improvement for commodity plastics to
advanced developments of radiation chemistry for nanotechnologies.
Radiation sources
At present, the two most common high-energy radiation dedicated to industrial

polymer processing are gamma rays and electron beams (EB) (Chapter 2). There are about
200 gamma industrial facilities in operation worldwide, chiefly exploiting the spontaneous
emission of highly penetrating 1.17 and 1.33 MeV gamma rays from 60Co radioelement
[6]. Electron accelerators cover a broad range of particle energy, from 50 keV for machines
dedicated to surface treatments and coating processing, to very high power 10 MeV
accelerators for bulk processing. Some facilities in the high energy range can be equipped
with a tungsten target intercepting the electron beam to produce X-rays by bremsstrahlung,
the resulting radiation being very similar in nature and features to the 60Co gamma rays.
The total number of industrial electron accelerators in the world amounts to

approximately 1 500 units [6]. Compared to gamma sources, such EB machines have a high-
dose rate and therefore short processing times. Because they produce radiation only by
means of physical processes, EB facilities are installed and run with easier public acceptance. A
limited number of large facilities held by national agencies or related institutions implement
swift heavy ions as well as synchrotron radiation lines and develop their applications through
large national and international programs.
In addition to these industrial or large facilities, there exist a large number of

laboratory accelerators producing various types of radiation for research studies and for
Ä&(+Ä
discrete production purposes. As an example, lithographic EB maskers are increasingly used

by micro- and nanoelectronics R&D laboratories and by related production sites.
Chemical effects induced by high-energy irradiation
The chemical and biological effects of ionizing radiation can be referred to as

direct effects when the radiation energy is deposited in molecular targets where chemical
transformations are induced. If energy absorption occurs in the external medium (for
example water in aqueous solution or biological systems), leading to the formation of radical
intermediates which can diffuse to come to react with the molecules, the observed chemical
effect is said indirect (Chapter 1).
Various chemical steps generally take place between the end of purely physical
processes of energy deposition in polymers and the resulting chemistry. When the energy
imparted to a molecular electron is lower than its lowest ionization potential, the resulting
excitation (Scheme 1) proceeds according to optical selection rules.
The subsequent chemical response is expected to be similar to that obtained by

conventional photochemistry. However, ionizing radiation beams have higher penetration.
If the interaction involves larger energy deposition, ionization takes place and produces an
unstable radical cation and a slow electron. Simultaneous or subsequent dissociation of the
radical cation occurs. Thermalized electrons can then induce more conventional chemistry
by combination with a cationic species or by electron capture that may occur also with
dissociation of the molecular assembly. Negative ions produced by electron attachment
can in turn combine with a positive ion by a charge transfer mechanism. Primary and
secondary events thus produce excited molecules that can dissociate into free radicals or
into molecular products. The former homolytic process may induce chain reactions that
can turn into a chemical amplification of the primary effect. Most of the chemical processes
caused by ionizing radiations are interpreted in terms of free radical mechanisms [7].
Long-lived cationic chain reactions involving carbenium or oxonium intermediates are
unlikely because of the high reactivity of active centers that make transfers and deactivation
reactions very fast unless special conditions (purity of involved chemicals, low temperature)
are maintained throughout the process. The efficiency of a chemical pathway, expressed by its
G-value (number of events produced per 100 eV) or by its radiochemical yield in standard units
(moles J–1), quantifies the number of events corresponding to a given chemical transformation
per absorbed energy unit. The current unit of absorbed dose is the gray (Gy), corresponding to
1 J per kg of irradiated material (Chapter 1).
Ä&(,Ä
2!$)!4)/.#(%-)3429
The response of a medium submitted to high-energy irradiation is dependent

on the chemical constitution. When irradiated, molecules with carbon-carbon double or
triple bonds may produce free radicals which are able to add on a second molecule. As
a consequence of the generation of free radicals in these organic compounds, acrylated
monomers undergo rapid polymerisation by a chain reaction (Scheme 2). Charged radicals
(radical anions or radical cations) may as well initiate chain reactions. When the target
molecules have previously formed polymers, various types of chemical reactions can then
take place, as a function of polymer chemical structure and of irradiation conditions (radiation
type and intensity, temperature, ambiance, presence of other chemicals). These reactions
essentially include chain scission, cross-linking, oxidation and grafting (Scheme 3).
Scheme 2 : Sketch representing the chain reaction of polymerization induced by exposing monomers to
high-energy radiation.
Scheme 3 : Sketch representing chemical events that can be induced by high energy radiation in polymers.
Ä&(-Ä
Depending on the ultimate use, the range of radiation dose to apply to materials
covers several orders of magnitude (Fig. 4) [8]. However, at low doses, polymers are generally
not significantly degraded, in contrast with bacteria and viruses which can be killed. This
difference makes it possible to apply widely radiation-induced sterilization processes to
medical supplies, which are often made of and packaged in polymers. In order to avoid
confusion, it is worth mentioning that dose, an energetic parameter of the treatment, is
independent to a certain extent of the energy of the single particle or photon that interacts
with the target material. The latter governs the penetration depth, also affected by the LET in
the target material, whereas the first one corresponds to the cumulated energy deposition.
Figure 4 : Typical radiation dose ranges for the different utilizations of high-energy radiation in polymer proces-
sing. Doses for desinfection, sterilization and sprout prevention (biological effects) are given for comparison.
Chain reactions, essentially polymerizations, can be achieved with medium doses, as a

result of the chemical amplification by purely thermal processes of radiation-induced initiation
(Scheme 2). Processes involving single steps or short kinetic chain length reactions require much
higher doses. This is generally the case for the radiation cross-linking of rubbers and thermoplastics.
Cross-linking
When polymers are submitted to high-energy radiation, contrasting behaviours

are observed with a dominant effect of cross-linking or of chain scission, depending on the
nature of the repetition units (Scheme 3).
Ä&(.Ä
2!$)!4)/.#(%-)3429
Table 2. Radiolytic yield values of cross-linking G(X) and of scission G(S) values
for some common polymers.
G(X) G(S)
Polymera) Chemical structure
(10 mol J ) (10 mol J–1)
–9 –1 –9
b)
polyethylene CH 2 CH 2 n 300 90
b) CH 2 CH
polypropylene n 250 110
CH 3
CH 2 CH CH CH 2 n
CH 2 CH n *
polybutadiene CH 380 -
CH 2
CH 2 CH n
poly(methyl acrylate) 55 18
O
O CH 3
CH 2 CH n
poly(n-butyl acrylate) 63 18
O
O CH 2 CH 2 CH 2 CH 3
CH 2 CH n
CH 3
poly(t-butyl acrylate) O
16 18
O C CH 3
CH 3
CH 3
CH 2 C n
poly(methyl methacrylate) - 120 - 350
O
O CH 3
c)
CH 2 CH
poly(vinyl chloride) n 33 23
Cl
CH 2 CH n
polystyrene 5 <2
O O
poly(ethylene terephtalate) C C O CH2 CH2 O 3 - 20 7 - 20

n
a) Ionization at 25 °C in the absence of oxygen. b) Dependent on microstructure and of crystallinity. c) Unplasticized.
The general trends of the competition between cross-link formation (X) and chain
scission (S) can be discussed for some common polymers on the basis of the values of G(S)
and G(X) gathered in Table 2. Polyethylene (PE) and polyvinylchloride (PVC) form networks,
whereas polymethylmethacrylate (PMMA) only degrades strongly with evolution of low
molecular weight fragments. EB-lithography utilizes this degradative property to engrave
nanometric lines in PMMA films for electronic applications (Fig. 5). Tailoring copolymers
including repeating units with intrinsically opposite behavior allows controlling the overall
behavior [9].
Ä&)%Ä
Figure 5 : Nanometric size transistor

grid obtained by localized degrada-
tion (central engraved line) of a PM-
MA-based multilayer coating with
EB lithographic equipment (courtesy
of IEMN, Villeneuve d’Ascq, France).
Obviously, the degree of substitution of the main chain has a determining influence
on the behavior under irradiation. Polymers including aromatic nuclei as side-groups
(polystyrene-PS) or in the main chain (polyethyleneterephtalate-PETP) give a strongly
attenuated response to high energy radiation. It must be added – to avoid oversimplification –
that morphological and ambiance parameters (microstructure, crystallinity, glassy or rubbery
state, degree of plasticization, presence of oxygen) may strongly alter the reported data.
Cross-linking is the largest commercial application of radiation processing applied to

polymers. Polymers with cross-linked chains can maintain their shape and useful properties
at a higher temperature without flowing (Fig. 6). Generally, the materials exhibit after cross-
linking better mechanical and chemical resistance. Processing is performed at high speed
under the electron beam and neither requires heating to high temperature nor the use of
the added cross-linkers necessary to the non-radiative processes.
Figure 6 : Pictures illustrating the compression resis-

tance (after heating at 250 °C for 1 min. under stress.
a) Electric cable with a radiation cross-linked PE
insulating material. b) Deformation of cables with
uncross-linked PP. c) Thermoplastic PE materials sub-
mitted to the same test (courtesy of Acome, France).
A number of manufactured products comprising thermoplastics (polyethylene,

polyvinylchloride, polyamides, …) or elastomers (natural rubber, nitrile rubber, ethylene
propylene rubber, silicones) are processed by cross-linking to improve their performances in
insulated electric wires and cables, multilayered films for cooking pouches, shape memory
tubes, pressure resistant water pipes, expandable foams, automotive parts exposed to motor
Ä&)&Ä
2!$)!4)/.#(%-)3429
heat, etc. Unsaturated additives are sometimes used to promote radiation cross-linking, in
order to improve productivity or to influence favourably the inherent trend to competing
cross-linking and scissioning.
Irradiation can be very useful to stabilize polymer blends, avoiding physical

phase-separation of the components or making it possible to block a desired but unstable
morphology by sudden cross-linking. Various other applications exploit the shape memory
effect that results from radiation cross-linking of semicrystalline thermoplastics: resettable
fuses for electric devices, heat-shrinkable insulating tubings and food packages including
multilayered metal-polymer complexes are some examples worth mentioning.
Degradation and post-irradiation oxidation
Some important commercial applications take advantage of various chemical

effects induced by radiation and resulting in scissioning, branching and partial oxidation.
Surface oxidation is achieved in nanolithographic processes using X-rays or EB in order
to enhance resolution. If they are well-controlled, the cumulated degradation effects can
improve the processability or the compatibility of the plastic in blends with other materials.
For example, fluorinated polymers can receive a treatment facilitating melt processing.
Radiation-treated polypropylene yields fiber-reinforced composite materials with enhanced
performances. Oxidation appears of interest in this particular use. However, for many other
applications, the very efficient reaction of generated macroradicals with dissolved and
gradually diffusing oxygen forms peroxy derivatives that will induce long-term degradation
of polymer properties. Chain reactions further initiated by thermal or photochemical
peroxide decomposition can ruin the structural and functional properties of the materials.
Anti-oxidants are thus added to minimize the detrimental effect of radio-oxidation. This
generic phenomenon represents a key issue in the development of radiation-sterilizable
materials for biomedical devices and prostheses.
Radiation grafting
Radiation activation offers a rather unique possibility to generate free radical centers
in a reactive or in a latent form within polymer materials for initiating grafting polymerization.
In the pre-irradiation technique, the material is treated first, preferably in the presence of
oxygen, to induce the formation of radicals or peroxy species that are subsequently activated
thermally or chemically in the presence of the monomer to be grafted. In semi-crystalline
polymers, the long-lived free radicals formed in the ordered domains can also be responsible
of the grafting upon appropriate thermal activation and monomer diffusion (In Scheme 4
Ä&)'Ä
step b, the monomer is glycidyl methacrylate GMA) [10]. An alternative process consists in
treating the polymer-reactive additive mixture by simultaneous irradiation grafting.
Scheme 4 : Simplified representation

of the pre-irradiation method (step a)
applied to the grafting reaction of
glycidyl methacrylate (GMA) onto
polypropylene (step b) for the functio-
nalization of the material by reaction
with nucleophiles (step c).
Ä&)(Ä
2!$)!4)/.#(%-)3429
The technique is applicable to polymer substrates under the form of films, fibers,
powders or bulk objects. Free radicals are produced in relatively high concentration during
the ionizing treatment of synthetic (polyolefins, polystyrene, polyamides, fluorinated
polymers) as well as natural polymers (cellulose, starch, chitosan). In principle, there is no
restriction regarding the type of monomer to form the graft polymer, but sufficient affinity
of the monomer for the polymer substrate and for the grafted polymer favours the grafting
yield. Suppression of homopolymer formation (i.e. without grafting) is possible, by using
additives such as Mohr’s salt (FeSO4, (NH4)2SO4, 6H2O), that quench diffusing free radicals,
but convert hydroperoxy species into free radical by a redox reaction. Various industrial
applications have been developed. Textile industry uses this process to improve fiber
dyeability (grafting of polar polymer), to reduce hydrophilicity and soil stainability (grafting
of fluorinated polymers) or to keep sterile medical dressings (grafting of antiseptics). Super-
absorbents have been produced by grafting hydrophilic synthetic polymers onto cellulose
or starch substrates. Plastic-based biomedical devices are modified with hydrophilic layers to
enhance biocompatibility (prostheses), to reduce the risks of haemostasis (catheters, artificial
kidney, extracorporeal circulation devices). Novel chromatographic supports, membranes
with ion-binding properties for decontamination of wastewater as well as for fuel cells are
produced by grafting and are under intense study. Control of graft length and microstructure
is under investigation to improve the performances of the grafted materials.
Grafting of functional side-groups without forming long polymer chains may be

achieved in a similar way by the reaction of activated polymeric materials with low molecular
weight compounds carrying functional groups of appropriate reactivity. The physical
stabilization of unstable blends of amorphized starch with reactive plasticizers has been
achieved by EB-irradiation [11].
Radiation-induced cross-linking polymerization
Radiation curing of inks and coatings by cross-linking polymerization is another

important application that expends rapidly. It represents a most promising alternative to
solvent-based processes that are progressively banned for their detrimental impact on the
environment. This technology offers several advantages in terms of ecoconception (energy
saving, reduction of volatile organic compounds VOCs) [12]. In-situ polymerization of
restorative resins infused into damaged archaeological objects made of wood or into weak
artistic pieces of porous structure is another illustration of the unique in-depth chemical
effects that can be induced under soft conditions by high energy radiation [13].
Ä&))Ä
High-energy radiation treatment is an incomparably more suited alternative to intense

UV light for initiating in depth the cross-linking polymerization of solvent-free compositions
including monomers and reactive prepolymers. The rapid polymerization process is triggered
by exposure to the radiation. Owing to the presence of multifunctional monomers, a
3-dimensional covalent network is formed, causing the drying of the initial formulation and
yielding generally mechanically and chemically resistant materials (Scheme 5).
Scheme 5 : Radiation-induced cross-linking polymerization of monofunctional ( ) and multifunctional

( ) monomer blends.
The liquid or semi-solid formulations can be handled and applied to substrates by

various processes (sprayer, roller, blade coater) as solvent-based coatings. Acrylate derivatives
and some vinylaromatics are well-suited to give fast curing by free radical mechanism.
Epoxy or vinyl ether functional analogues can be cured by a cationic process requiring the
incorporation of onium salts initiators. Pigments and additives are added to the formulation
in order to facilitate the coating application, to enhance curing efficiency and to adjust
performance and properties of the coating. UV radiation is extensively used for labels, cartons
and containers in the packaging industry. High energy radiation curing provides additional
advantages compared to UV light. No initiator is needed for free radical polymerization, and
the penetrating power of electrons is greater than UV light at usual concentration in photo-
initiator and/or in formulations with high pigment or filler concentration. Electrons are more
efficient than UV rays in opaque coatings, and composites and they activate monomers even
after vitrification of the network undergoing gradual curing. This results in lower amounts of
unreacted monomers and other extractables. Grafting reactions can also take place at the
substrate-coating interface, thus improving adhesion. Recent advances in formulation and
in process control give insight into major industrial developments in the food packaging
industry and into industrial coatings (flooring, outdoor building parts).
Ä&)*Ä
2!$)!4)/.#(%-)3429
Curing of fiber-reinforced polymer composites using high-energy electron beams

offers similarly significant advantages for fabricating a variety of aerospace, ship and ground
vehicle components (Fig. 7) [14]. The EB curing process greatly reduces the time required
to cross-link the polymer matrix compared to conventional heat curing. The process allows
curing without external thermal activation for high throughput. It yields materials with
reduced residual stresses for good part fit-up. Curing times are shorter, and the curable
formulations have longer storage stability. The absence of highly volatile monomers renders
the industrial process safer and more environment friendly.
Figure 7 : Glass/poly-
ester 10 meter-long ship
hull cured by EB and X-ray
induced polymerization.
EB gun is circled by red line.
(Courtesy of Intermarine
and Astrium Space Trans-
portation.)
Emulsion Polymerization
Emulsion polymerization is a type of radical polymerization that usually starts with a

dispersion incorporating water, one or some hydrophobic monomers, and a surfactant [15].
Polymerization actually takes place not in the large droplets of dispersed monomers but in the
latex particles that form spontaneously in the first instants of the process. These latex particles
are typically 100 nm in size, and comprise many individual monomer swollen polymer chains
surrounded by the surfactant, forming the so-called emulsion. Their number is essentially
stable along the reaction, but their size increases gradually since monomer continues to
diffuse and to swell the polymer particle. Because initiation takes place at a slow rate in the
continuous aqueous phase, the kinetic chain length inside a single particle is extremely long
(polymer chains of high molecular weight) and, at a given instant, half of the population of
latex particles is undergoing polymerization, whereas the other half has been visited by an
even number of free radicals, thus having been terminated by bimolecular termination. The
inactive state lasts until polymerisation is resumed by a newly entered free radical.
Ä&)+Ä
Emulsion polymerization is used to manufacture several commercially important

polymers. These emulsions increasingly find applications in adhesives, paints, paper coating
and textile coatings, for they contain little or no amount of volatile organic compounds.
Most of them are currently obtained by conventional thermally initiated polymerisation
process, but gamma radiation from 60Co sources can be used to generate at a low and well
controlled rate the hydroxy radicals in the water phase, so as to sustain the alternating
switching on/off of the polymerization in the heterogeneous medium [16]. The use of
radiation makes it possible to decouple the thermal activation of the polymerization process
desired for increasing the production rate from the other effect of temperature influencing
molecular weight. Indeed, when a thermal initiator is used, it also decomposes faster at
higher temperatures, producing more polymers for the same reaction time, but with shorter
polymer chains. This feature of interest associated with radiation-induced initiation is likely
to receive increasing industrial development in the coming years.
Conclusion
The main features and the early achievements of radiation processing in the domain
of polymer materials have been briefly presented, and are summarized in Scheme 6. Efficient
processes for the improvement of thermoplastics and rubbers properties have been made
available for decades and continue to be improved [17]. The current trends show the
enormous potential of radiation-based processes to address environmental concerns by
proposing performing and societally acceptable alternatives to outdated technologies
based on unsuitable chemistries and unreasonable energy consumption. Both commodity
and technological products will increasingly rely on radiation treatment of polymers. Use
of radiation in recycling technologies and for new utilization of biomass is expected to play
an increasing role in a near future. This trend is obvious in speciality fields prone to intense
innovation, as the sector of biomedical materials and devices where nanotechnologies and
the fine use of radiation chemistry have lead to considerable advances (Scheme 6).
Acknowledgements
The author wishes to express his thanks to Dr. C. Lagreve (Acome), Dr. M.C. Clochard (CEA-
DRECAM), Dr. B. Defoort (EADS-Astrium), Mr. G. Massaro (IBA), Dr. M. François (IEMN), for supplying
original illustrations appearing in this article. Sponsoring of research programs on electron-beam
initiated cross-linking polymerization by the EADS Foundation as well as support of the Polynano
project by MENESR (ACI Nanosciences n°120) are grateftully acknowledged.
Ä&),Ä
2!$)!4)/.#(%-)3429
Scheme 6 : Industrial applications of monomer and polymer irradiation.
References
[1] Chapiro A., “Radiation chemistry of polymeric systems”, Wiley, New-York, 1962.
[2] Ivanov V.S., “Radiation chemistry of polymers”, VSP Publishers, Utrecht, 1992.
[3] Turner J., “Atoms, Radiation, and Radiation Protection”, Wiley, New-York,1995.
[4] Sarma K.S.S., Majli A.B., Electron Accelerators and their Applications, IANCAS Bulletin, 1999, 15 (4).
[5] Swallow A.J., “Radiation Chemistry - An introduction”, Longman, London, 1973.
[6] Machi S., Prospects for the application of radiation processing and the activities of the IAEA,
Radiat. Phys. Chem., 1988, 52, 591-597.
[7] Lund A., Shiotani M., “Radical Ionic Systems Properties in Condensed Phases”, Kluwer, Dordrecht,
1991.
Ä&)-Ä
[8] Bly J.H., “Electron Beam processing”, International Information Associates, Yardley, 1988.
[9] Turgis J.D., Vergé C., Coqueret X., Composition effects on the EB-induced cross-linking of some
acrylate and methacrylate copolymers, Radiat. Phys. Chem., 2003, 67, 409-413.
[10] Le Thuaut Ph., Martel B., Crini G., Maschke U., Martel B., Coqueret X., Morcellet M., Grafting of
cyclodextrins onto polypropylene nonwoven fabrics for the manufacture of reactive filters - Part 1 :
Synthesis parameters, J. Appl. Polym. Sci., 2000, 77, 2118-2125.
[11] Olivier A., Cazaux F., Coqueret X., Compatibilization of starch-allylurea blends by electron beam
irradiation: Spectroscopic monitoring and assessment of grafting efficiency, Biomacromolecules,
2001, 1260 -1266.
[12] Drobny J.G., “Radiation Technology for Polymers”, CHIPS, Weimar, 2002.
[13] Cœuré P., Khoi Tran Q., Procédés radiochimiques pour la conservation des biens culturels, Actualité
Chimique, 1999, 225, 141-143.
[14] Defoort B., Lopitaux G., Dupillier J.M., Larnac G., Coqueret X., Electron-beam initiated polymerization
of acrylate compositions, 6: Influence of processing parameters on the curing kinetics of an epoxy
acrylate blend, Macromol. Chem. Phys., 2001, 202, 3149-3156.
[15] Odian G, “Principles of Polymerization”, 3rd edition, Wiley-Interscience, New York, 1991.
[16] Wang X., Zhang Z., Preparation of polystyrene latex particles by gamma-rays-induced emulsifier-
free emulsion polymerization, Radiat. Phys. Chem., 2006, 75, 1001-1005.
[17] Clough R.L., High energy radiation and polymers: a review of commercial processes and emerging
applications, Nucl. Instrum. Methods, B, 2001, 185, 8-33.
Ä&).Ä
Chapter 10
Radiosterilization
of drugs
Bernard TILQUIN
Introduction
In the early 1990s, whilst radio-sterilization of medical devices was flourishing, the
radiolysis mechanisms in irradiated solid-state drugs were not or poorly described [1]. The
existing analytical methods failed actually to detect final radiolytic products since they were
various and present only at the level of traces. Moreover, the European Pharmacopoeia (Ph.
Eur.) tests are not adapted yet for the detection of chemical changes in irradiated drugs.
The improvement of the sensitivity of to-day analytical methods enables us now to study
the behaviour of a drug under ionizing radiations and to determine the conditions of the
radiosterilization treatment.
For a better understanding of the context of this work, it seems fundamental to

remind first some notions of sterility and sterilization. Then we will dwell on radiosterilization
by describing the process and its advantages, without forgetting its limitations. These latter
lead us to a review of the investigations on the radio-induced degradation of drugs. In fine
we propose the latest regulations on radiosterilization.
The effects of irradiation on many different types of substances are now studied at
CHAM (Analytical Chemistry Department at the University of Louvain, School of Pharmacy)
[2] and in other groups [3]. All the most advanced analytical tools are explored to detect the
traces and to study radiosterilization : Electron Spin Resonance (ESR) is the most sensitive
technique and the spectra were found to be specific to free radicals produced in irradiated
Ä&*&Ä
2!$)!4)/.#(%-)3429
solid molecules. HPLC (High Performance Liquid Chromatography) coupled with various
detection systems, such as mass spectrometry (MS) – Diode Array Detector (DAD), is the
most promising technique for analyzing traces of final products.
Notion of sterility
Sterility can be defined as a complete absence of viable organisms. In order to reach

sterility, all the living germs are supposed to be eradicated. This concerns pathogen as well
as non-pathogen micro-organisms. The latter could cause diseases under optimal growth
conditions.
The means of attaining sterility is called sterilization. In fact, this killing process
must be rather considered as an inactivation. The death of a cell is indeed effective when it
cannot reproduce itself. For this reason, the effect of a sterilization process can be followed
by survival curves.
As the notion of absolute sterility, as it was defined, is not applicable in practice, the
notion of probability is used, where the likelihood of the presence of a contaminated unit
or article should be very low. The likelihood of the presence of contaminated unit is used to
define the efficiency of the sterilization (Security Assurance Level, SAL).
It is noteworthy that, at least presently, sterilization is not required for all drugs. It
depends on the type of administration. Hence, it concerns mostly ophthalmic preparations,
sterile topical products and injectable solutions, including intramuscular, intravenous and
sub-cutaneous ways.
Sterilization methods
The techniques which are described in the US and the European Pharmacopoeias
[4,5] in order to reach nowadays sterility are the following.
Steam or dry heating

First of all, thermal sterilization, which includes the use of steam or dry heat, is the
earliest method used. Steam sterilization is up to now the most widely employed sterilization
process. It is carried out in a chamber called autoclave. A typical autoclave cycle corresponds to
a period of 15 minutes at 121 °C. In the case of dry-heat, the process is undertaken in an
oven with heated and filtered air, distributed uniformly. The cycle is of two hours at 160 °C.
Ä&*'Ä
2ADIOSTERILIZATIONOFDRUGS
Both methods enable reaching the recommended SAL value of at least 10−6. They are
considered as the techniques of reference. However, a major limitation is that they are
unapplicable to heat-labile substances (and packaging when used as a terminal process).
The number of drugs, which cannot withstand high temperatures, is increasing especially
with the new generation of treatments based on the use of proteins. Alternatives to thermal
sterilization are therefore necessary.
Filtration
Among them, sterilization by filtration is frequently used. The filter consists in
a membrane, which removes physically the micro-organisms. The fragility of this filter
membrane is one of its most serious disadvantages. Moreover, it is not terminal and the
operations that are carried out after the sterilization by filtration must be done under aseptic
conditions. Finally, a SAL of 10−6 cannot be reached.
Gas sterilization
Gas sterilization is a process in which the active agent is a gas. The agent generally
used is ethylene oxide, which is an alkylator. It has major drawbacks. On one hand, it is highly
flammable and on the other hand, it is carcinogenic, toxic and mutagenic. Thus the residual
ethylene oxide left inside the drug is highly hazardous. Lastly, as for all the sterilizing gases,
its efficiency is restricted by a limited penetration into the innermost system areas.
Aseptic processes
Another option is the aseptic processing, which involves a series of aseptic steps.
Each of them must be highly controlled since the introduction of micro-organisms must
be avoided all along the process. Aseptic filtration or aseptic processes are time-consuming,
expensive, difficult to monitor and the SAL value is limited to 10−3-10−4, so alternative
methods for sterilization of thermosensitive solid drugs are needed.
Radiosterilization
Though Minck reported already in 1916 that X-rays reduced the microbial population,
the application of radio-sterilization started only in 1950. The ability of ionizing radiation to
inactivate microorganisms has been well documented [6]. In fact, until the 1990s, the analytical
techniques used to measure the drug degradation failed to detect final products present
in traces, and no systematic study could thus performed on parameters such as absorbed
Ä&*(Ä
2!$)!4)/.#(%-)3429
dose, dose rate, oxygen content or temperature of irradiation. For that reason, the radio-
sterilization was very restricted in pharmacy, whilst it was flourishing for medical devices,
particularly for thermo-labile polymeric materials (Chapter 9). Then, due to the development
of highly sensitive analytical techniques, the radiation chemistry research could be oriented
towards the study of the radiolysis mechanisms in irradiated solid-state drugs or in aqueous
solutions and the analysis of the final products in terms of their potential toxicity . The aim
of the research is to determine the conditions under which sterilization is achieved at the
required SAL value for micro-organisms eradication with a safe and minimized production
of non-toxic radiolytic products.
Since the inactivation of bacteria follows an exponential decay process with a

limiting value tending towards zero, the absolute sterility can never be obtained. The D10
value is the absorbed dose required to reduce a microbial population to 10% of its initial
value, so that in industrial applications, a SAL value of 10−3 is reached after 3 s D10 value.
For drugs, a dose of 25 kGy (or kJ kg−1) is generally higher than 6 s D10 value and achieves
the minimum SAL required of 10−6. Many reviews demonstrated that for drugs with low
bioburdens (initial contamination by microorganisms), sterilization was achieved with doses
even lower than 15 kGy.
The radio-sterilization process is based on the exposure of a product to ionizing

radiation. In Pharmacopoeias, two types of radiation are considered : gamma rays or electron
beams. Generally, the gamma sterilization is more documented [7]. The irradiation types are
characterized by their dose rate (much higher for electron beams) and their penetrating
power (much higher for gamma rays) (Tab. 1).
Table 1. Properties of irradiation types.
Specificity Gamma rays High energy electrons
Dose rate 1 kGy/h 107 kGy/h
Penetration depth 1-2 m 6-5 cm for 10 MeV
Irradiation Facility Radio-active source Electron accelerator
Advantages
By comparison with earlier methods, numerous advantages are admitted for the
radio-sterilization. The process :
Ä&*)Ä
- can reach a very high efficiency (SAL = 10−6) due to the high living cell sensitivity
to radiation.
- is a cold process applicable to thermo-sensitive molecules and thermosensitive

packaging.
- permits to sterilize the drug inside its packaging (terminal method).
- enables to sterilize homogeneously large containers (due to the penetration of

gamma rays).
- enables to sterilize large batches within a short pulse train (electron beam).
- produces related compounds (radiolytic products) in lower concentration that

initial impurities.
- is easily monitored and controlled, the variable being the absorbed dose.
- is an economical process.
- does not induce – it would be needless to repeat again – any radioactivity.
Disadvantages
- The major disadvantage of radio-sterilization is the peculiarity of the radiation
chemistry induced simultaneously to sterilization, which requires a special study :
the irradiated drug is thus considered as a “new” drug. While micro organisms are
killed, various products of degradation of the main compounds may appear in
traces amounts though they are often the same as in thermal sterilization.
- In the solid state, the drugs are radio-resistant until the Standard Absorbed Dose
(SAD) (25 kGy) but some changes in the organoleptic properties may be reported
at this dose ; by decreasing the doses, some discoloration and/or smell may be
reduced.
The SAD [4] may be lowered if it is validated provided a SAL of 10−6 is achieved . The
doses required to achieve this SAL are available in format when ANSI/AAMI/ISO methodology
is used [8].
Ä&**Ä
2!$)!4)/.#(%-)3429
The chemical changes are much more important in aqueous solution than in solid
state ; as a rule, a solute in diluted aqueous solutions is completely degraded after 1.5 kGy.
In pharmacy, excipients (isotonic solutions) are needed ; new approaches are improved with
these additives and are now able to inhibit the phenomenon
Dose distribution
The dose mapping inside the irradiated object is important. In the dose range
25-32 kGy, regulatory agencies require a ratio of Dmax/Dmin of 1.28 maximum value. The
distribution of absorbed dose depends on the radiation used and is affected by the geometry
of the vials (containers) as well as the nature (electron density) of packaging material.
Penetration of gamma-rays is more homogeneous. For E-beams, the dose depth

distribution is related to the energy levels. If possible, double E-beams are recommended
to flatten the dose distribution curve. Another way is to use scattering foils and reflection
plates. The packaging material (vials and stoppers) must be suitable.
Solid-state Sterilization
In the solid state, the drug is directly ionized and excited by the ionizing radiation.
The reactive species formed may not diffuse (except for a few electrons and mobile hydrogen
atoms) and radical fragments are trapped in the solid matrix, what is called the cage effect.
The cage effect favours recombination reactions of fragments into the initial molecule and
thus the drug transformation yield is usually very low.
The capacity of the European Pharmacopoeia purity tests to detect chemical

changes induced in irradiated drugs was explored. The pharmacopoeia tests failed to detect
products from radiolysis of chloramphenicol even at extremely high doses (> 400 kGy), and
the only noticeable changes detected were in the color of the drug [9].
The techniques of ESR were used [10] to study radicals formed in solid state. ESR
proved to be a very sensitive technique to detect radicals (10−10 mol l–1) trapped for days,
weeks and even years and discriminate between irradiated and non-irradiated drugs. Radical
amounts increased linearly with absorbed doses up to 10 kGy, the radiolytic yield was
constant, and ESR post-dosimetry was proposed as a new method to identify radiosterilized
drugs [11]. Thermoluminescent signals are present prior to any irradiation in all the B-lactam
drugs tested [12]. The irradiated drugs showed a modification in the quality and quantity of
the thermoluminescent temperature peaks.
Ä&*+Ä
Gas chromatography (GC) with an infrared (IR) detector was introduced as a

method to detect volatile radiolytic products, some of which were hypothesized to be
responsible for the bad smells emanating from irradiated drugs. Thiocyanic acid was held
responsible, for example, for the sulfurous smell in irradiated ampicillin. The head-space
(HS) injection technique for GC and the on-line MS detection allowed new approaches to
detect radiosterilization [12]. Many volatile radiolytic products were identified from the mass
spectral libraries. Some of the compounds identified such as aldehydes, esters and sulfides
were quite malodorous. A few of the volatile radiolytic products came from the degradation
of drug molecules by the ionizing radiation, whereas residual solvents played a key role in
the formation of other volatile radiolytic products.
The HPLC-UV-DAD was used for the analysis of final products ; results showed that
many related compounds of cefotaxime were present prior to irradiation. Some of them
had maxima around 320-380 nm and could be responsible for the change in color of the
dissolved irradiated drug. The radiolytic products (25 kGy) were present in traces and were
all below the qualification limit of 0.1%. The HPLC-UV-MS results showed that some radiolytic
products detected were impurities already present in the non-irradiated cefotaxime and
increased after irradiation. The other products were unique to radio-sterilization [14].
For some β-blockers, the radical yields were found to be very low ( <10−9 mol J−1)
and they were considered to be radioresistant [11]. The radiolytic related compounds were
also analysed by HPLC-UV, in order to test for a correlation between radical and product
yields. For the β-blockers, no correlation was established between the yields of radicals and
of final products [15].
Aqueous solution sterilization
In aqueous solutions, the water is ionized and excited by the ionizing radiation,
and the reactive species (eaq−, tH and tOH) (Chapter 1) formed may diffuse and react with
the solute. The amount of reactive species formed by water radiolysis is proportional to the
absorbed dose multiplied by their respective radiation chemical yields.
The solute, which is in rather low concentration is indirectly transformed by the

numerous reactive species of the water radiolysis, so the yield of solute transformation is
much higher than in the pure solid state.
There is a consensus that in aqueous solutions the degradation of the drug is

proportional to the absorbed dose, because the yields are constant in a certain domain
Ä&*,Ä
2!$)!4)/.#(%-)3429
of concentration (solute dilution curve, Chapter 1). The greater the absorbed dose the
greater the solute degradation. To a given absorbed dose corresponds a given amount
of reactive radicals formed by the water radiolysis. Thus, if the concentration of the drug
solute is increased in diluted solution, the relative percentage of degradation of the solute is
decreased [16]. In pharmacy, typical concentration is 10–3 mol l–1.
Moreover, the absorbed doses should be low, since at higher doses, radiolytic
products may interfere [17].
In order to inhibit degradation by the radical scavenging, two approaches were

studied : one is to let an excipient to play the role of a radical scavenger, so protecting the
drug from the radical attack , the second is to freeze the solution into a solid state where the
cage effect favours the recombination of primary radicals of the solvent prior to the reaction
with the drug.
Radioprotection of drugs
The degradation of the drug solute results from the indirect effect through
reactions with products of water radiolysis. Thus, addition of scavengers of these reactive
species, able to react faster than the drug, will protect the drug solute. In order to optimize
the radioprotection of the drug solute, the radical mechanisms of its degradation should
be studied. The choice of the scavenger and its concentration will depend on the solute
concentration and reactivity with the radical(s) responsible for the solute degradation. This
could be further complicated by additional direct effects.
At the radio-sterilization doses, simulation predicts a greater loss than in the

experimental results, possibly because some radiolytic products react with the water
radiolysis radicals thus protecting the drug solute [16]. The simulations also predict similar
solute concentrations without dose rate effects for E-beam or gamma irradiations whereas
the opposite is found in experimental results[17]. The complexity of the radiolysis mechanisms
at sterilization doses appears with the increase of the analytical efficiency.
The use of radio-protecting excipients was shown to be more promising to lower

the loss of potency of drugs in aqueous solution without oxygen than an increase in the dose
rate [16,17]. Type and concentration of the chosen radioprotector are important parameters.
Metoprolol tartrate [18] is commercialized as an injectable drug with 0.9% sodium chloride
as an isotonicity agent, but sodium chloride does not protect metoprolol tartrate from
degradation by both E-beam and gamma irradiations. 1,2-propanediol (2%) significantly
Ä&*-Ä
reduces the loss in metoprolol tartrate, because it reacts rapidly with hydroxyl radicals ;
mannitol (5%) also reacts quite well with hydroxyl radicals and protects the drug. Pyridoxine
or nicotinamide (vitamins) in metoclopramide aqueous solutions give recoveries far above
90% up to a 15 kGy (E-beam or gamma rays), 90% meet the pharmacopoeial specifications
concerning the chemical potency. Both excipients react with the hydrated electrons and the
hydroxyl radicals [19].
Frozen aqueous solutions

Lowering the irradiation temperature in order to freeze the drug aqueous solution
limits the diffusion of the radicals which are the reactive species (indirect effect) and favours
their recombination in the cage.
In a few recent studies, the degradation of the drug solute was lowered by irradiating
frozen aqueous solutions [20,21]. Their results show that :
- Frozen solutions of vitamins and glucose are stable to irradiation.

- The degradation of cyanocobalamine (vitamin B12) is reduced by freezing the
solution.
- Calcium pantothenate solutions (vitamin B5) are successfully sterilized at – 196 °C.
- Insulin may be radiosterilized by freezing the solutions.
- Injectable drugs containing 1 % morphine hydrochloride and 100 UI / ml heparin
are sterile and not degraded when irradiated with 25 kGy at – 80 °C and – 196 °C,
respectively.
Cryo-irradiation is thus a promising method to sterilize protein solutions, as well as

preparations derived from human blood.
At 10 kGy and without excipients, the loss of human insulin in aqueous solutions
is almost complete. Addition of radio-protecting excipients (free radicals scavengers) and
cryo-irradiation allowed to decrease insulin degradation. The best radio-protector used was
ascorbic acid in aqueous solution and oxidized glutathione in the frozen solutions.
Only the combination of these two approaches (addition of scavenger and freezing)
enables the irradiated human insulin in aqueous solution to meet the European Pharmacopoeia
requirements for chemical potency [22].
Ä&*.Ä
2!$)!4)/.#(%-)3429
Guidelines
ICH (standing for International Conference on Harmonization of Technical Requirements

for Registration of Pharmaceuticals for Human Use) is a project that brings together the regulatory
authorities of Europe, Japan and the United States and experts from the pharmaceutical industry
in the three regions to discuss scientific aspects of product registration. For the European region,
the organism that publishes and distributes the guidelines adopted by ICH is the European
Agency for the Evaluation of Medicinal Products (EMEA) [23].
In the Note for guidance and development of pharmaceuticals, the section on the
manufacturing process stipulates clearly that, wherever possible, products intended to be
sterile should be terminally sterilized in their final container. The methods of sterilization
considered as terminal by the European Pharmacopoeia [5] are steam, dry-heat, gas or
gamma-radiation sterilization. If terminal sterilization is not possible, sterilization filtration
or aseptic processing are authorized but the choice must be fully justified.
One last consideration, which is given in the Note for guidance on limitations to the
use of ethylene oxide in the manufacture of medicinal products, must be taken into account.
First of all, as described previously, ethylene oxide is a cytotoxic molecule, a carcinogen and
a mutagen agent. Epidemiological data have clearly shown that the incidence of leukaemia
and other tumours is larger for workers exposed to this substance at their work place. In view
of these toxicological backgrounds, the use of ethylene oxide is only acceptable when two
conditions are fulfilled : it must be “pharmaceutically absolutely necessary” and the residual in
the product cannot exceed a limit of 1 ppm. Obviously workers should be also protected.
All these recommendations are summarized in the Decision trees for the selection
of sterilization methods, edited by the EMEA [23] (Fig. 1). Two cases are considered : on one
hand, the aqueous products and on the other hand, the non-aqueous liquid, semi-solid and
dry powder products. Figure 1 shows the order of preference of the sterilization methods
for the second group. The terminal ones are ranked in the first place. Among them, thermal
sterilization is still referred as the best choice, radio-sterilization ranking right after. Since
gas sterilization is excluded and non-terminal methods are listed as the last choice, radio-
sterilization now precedes all these methods. It is deemed as the recommended alternative
method to thermal sterilization.
Ä&+%Ä
Figure 1 : Decision tree for sterilisation choices for non-aqueous liquid,semi-solid or dry powder products [23].
Ä&+&Ä
2!$)!4)/.#(%-)3429
For aqueous solutions, some delay is necessary for research results to be applied
in industry and so a new selection tree for aqueous solutions taking into account radio-
sterilization is expected.
To conclude, radio-sterilization of drugs appears as a promising terminal process,

which is recommended in Guidelines and, when applied under controlled conditions,
ensures the required Security Assurance Level for eradicating the dangerous micro-
organisms without producing toxic radiolytic products. For a new generation of drugs, to
which heat-sterilization cannot be applied due to their thermo-lability, radio-sterilization is
the only alternative treatment.
Acknowledgements
The author wish to thank all the Ph.D pharmacists of the CHAM laboratory with special
thanks to A. Engalytcheff, A. Maquille and C. Slegers for their contribution to this manuscript. Thanks
to F. Gelders for assistance in the preparation.
References
[1] Jacobs G.P., A review: radiation sterilization of pharmaceuticals, Radiat. Phys. Chem., 1985, 26,
133-142.
[2] Tilquin B., Comprendre la radiostérilisation. J. Pharm. Belg., 2007, 62,10-20.
[3] Radiosterilization and decontamination of pharmaceutical and pharmaceutical raw materials.
IAEA, CRP Report, Vienna (A), 2005.
[4] USP, “Sterilization and sterility assurance of compendial articles”, in The United States
Pharmacopoeia 24, The National Formulary, United States Pharmacopeial Convention INC.,
Rockville, MD, 2000, 19, p. 2143-2148.
[5] Ph. Eur., “Textes Généraux sur la Stérilité”, in European Pharmacopoeia 5th Edition, p. 475-477,
Council of Europe, Strasbourg, 2005.
[6] Nordhauser F.M., Olson W.P. (eds) Sterilization of Drugs and Devices, Technologies for the 21st
Century Interpharm Press Inc., Buffalo Grove, IL, 1998.
[7] Reid D., Gamma processing technology: an alternative technology for terminal sterilization
parenterals, J. Parenteral Sci. Technol., 1995, 49, 83-89.
[8] ISO 11137. Sterilization of health care products - Requirements for validation and routine control-
Radiation sterilization, International Organization for Standardization, Geneva, Switzerland, 1995.
[9] Zeegers F., Gibella M., Tilquin B., Analysis of some products from the radiosterilization of solid
chloramphenicol, Radiat. Phys. Chem., 1997, 50, 149-153.
Ä&+'Ä
[10] Zeegers F., Tilquin B., Étude par R.P.E. de la décroissance radicalaire dans les medicaments irradiés,
ESR applications in organic and biorganic material. Catoire E. (eds), Springer-Verlag Berlin, 1992,
292-305.
[11] Engalytcheff A., Deridder V., Debuyst R., Tilquin B., Determination of radical yields in solid-state
drugs as one technique to identify drugs that will withstand radiosterilization : radioresistance of
beta blockers, Radiat. Res., 2003, 160, 103-109.
[12] Stocker P., Gibella M., Tilquin B., Lesgards G., Thermoluninescence of some ionized pharmaceuticals,
J. Chim. Phys – Chim. B.D., 1999, 96, 174-177.
[13] Barbarin N., Rollmann B., Tilquin B., Role of residual solvents in the formation of volatile compounds
after radiosterilization of cefotaxime, Int. J. Pharm., 1999, 178, 203-212.
[14] Barbarin N., Tilquin B., Study of non-volatile degradation compounds produced by radiosterilization
of cefotaxime, Radiat. Phys. Chem., 2001, 60, 359-367.
[15] Engalytcheff A., Vanhelleputte J.P., Tilquin B., HPLC detection and quantification of radiolytic
products of eight β-blockers irradiated in the solid state and hypotheses on their origins,
Pharmaceut. Res., 2004, 21, 1103-1108.
[16] Slegers C., Tilquin B., Theoretical approach to the destruction or sterilization of drugs in aqueous
solution, Radiat. Phys. Chem., 2005, 72, 363-365.
[17] Slegers C., Tilquin B., Final product analysis in the e-beam and gamma radiolysis of aqueous
solutions of metoprolol tartrate, Radiat. Phys. Chem., 2006, 75, 1006–1017.
[18] Slegers C., Maquille A., Deridder V., Sonveaux E., Jiwan J.L.H., Tilquin B., LC-MS analysis in the
e-beam and gamma radiolysis of metoprolol tartrate in aqueous solution: Structure elucidation
and formation mechanism of radiolytic products, Radiat. Phys. Chem., 2006, 75, 977–989.
[19] Maquille A., Private Communication, 2007.
[20] Juanchi X., Albarran G., Negron-Mendoza A., Radiolysis of cyanocobalamin (vitamin B12), Radiat.
Phys. Chem., 2000, 57, 337-339.
[21] Moyne P., Botella A., Peyrouset A., Rey L., Sterilization of injectable drugs solutions by irradiation,
Radiat. Phys. Chem., 2002, 63, 703-704.
[22] Terryn H., Maquille A., Houée-Levin C., Tilquin B., Irradiation of human insulin in aqueous solution :
first step towards radio-sterilization, Int. J. Pharm., 2007, 343, 4-11.
[23] EMEA (European Agency for the Evaluation of Medicinal Products), Decision trees for the selection
of sterilization methods, CPMP (Committee for Proprietary Medicinal Products) / QWP / 054 / 98
Corr), 2000.
Ä&+(Ä
Chapter 11
Food irradiation:
wholesomeness
and treatment control
Jacques RAFFI et Jacky KISTER
Food irradiation by X- or gamma-rays and by electron beam has been studied for more
than four decades, mainly to reduce spoilage losses and to improve hygienic quality [1-3]. This
treatment is legally accepted in more than 50 countries, but still prohibited in other ones by
lack of public acceptance, despite numerous researches sponsored by the International Food
Irradiation Project (IFIP), the International Atomic Energy Agency (IAEA) (in particular via the
program for Analytical Detection Methods for Irradiation Trade or ADMIT), the Food and Agriculture
Organisation (FAO), the World Health Organisation (WHO) and the European Union (via two
concerted actions and the working group of the European Committee of Normalisation or ECN).
Principles and wholesomeness of food irradiation
Action of ionising treatments

In principle X-rays, as well as electron and gamma rays, can be used ; the main
effect of gamma and X-rays is Compton scattering, i.e. ionisation of atoms giving rise to
“primary electrons” ; these electrons, like those generated in an accelerator, lead to a number
of other “secondary electrons” and ions. This cascade of secondary electrons (about 4 000
to 6 000 ionisations per initial secondary electron) loses energy in ionising the foodstuff
molecules, with consequent production of free radicals and, thereby, of “radiolytic products”
(Chapter 1).
Ä&+*Ä
2!$)!4)/.#(%-)3429
The physical, chemical and biological effects are linked to the irradiation dose
which is the quantity of energy absorbed by the matter, measured in gray: 1 Gy = 1 J/kg.
The positive effects (increase of hygienic quality) are mainly linked to radiolytic effects on
DNA molecules of viable cells from micro-organisms or living foodstuffs. The doses applied
may vary from 0.05 to 50 kGy, but usually between 3 and 10 kGy. The possible applications
are put in Table 1 but the true choice was made, comparing the cost of the process with the
obtained advantage. Thus it depends on the foodstuff, on the irradiation goal, on the chosen
technique (gamma or X-rays, electron beam) and, also, on the country involved.
Table 1. Doses recommended for possible applications of food irradiation.
Dose (kGy) Purpose
0.05 to 0.15 Sprouting inhibition
0.15 to 0.5 Insect desinfestation and parasite disinfection
0.5 to 1 Delay of physiological process (e.g. ripening) (a)
1 to 3 Spoilage of pathogenic micro-organisms - Shelf-life extension
1 to 7 Elimination of pathogenic micro-organisms (b)
7 to 25 Decontamination of certain food additives and ingredients (b)
a) Mainly for developing and non temperate countries. b) The main application in developed and temperate countries such as France.
Irradiation facilities
Up to now, only electron beam accelerators and gamma rays cells (Fig. 1) were used
for commercial applications. The power of these facilities is measured in watts for electron
beam accelerators, in becquerels for gamma-rays irradiators ; in this last case, we generally
speak of activity, instead of power: 1 becquerel (Bq) corresponds to a source where one
disintegration happens per second (the old unit, 1 Ci = 3.7 s 1010 Bq).
National authorities control the security of irradiation facilities ; the maximal energy
of these radiations is fixed, according to WHO recommendations, by construction of the
accelerator or by choice of the radioactive isotope (mainly cobalt 60) in order to avoid any
induction of additional radioactivity in the foodstuff.
Ä&++Ä
&OODIRRADIATION
Figure 1 : Scheme of a gamma irradiation facility (usually 1 to 3 x 10 8 Ci, i.e. about 1017 Bq). When not used, the
60
Co radiactive source is stored down in the pool. During irradiation, the source is raised to the upper position and
parcels are moved in front of and around the source. A thick concrete shielding surrounds the irradiation chamber.
Nutritional effects
The consequences of the radiolysis depend on the dose, on the temperature during
the treatment, on the food constituents [4,5], and on the mobility of the foodstuff. However,
some general rules can be drawn :
O Irradiation has practically no effect upon the proteins.
O The only important radiolytic effect upon carbohydrates is a depolymerization of

polysaccharides, which can lead to positive consequences (better digestibility, decrease
Ä&+,Ä
2!$)!4)/.#(%-)3429
of heating time, irradiated prunes judged as softer by the consumers…) or negative

ones (strawberries irradiated above 3 kGy judged as too soft by the consumers…).
O Irradiation of lipids, especially unsaturated ones, may lead to unwanted flavour,

which is the main technological limit of the treatment. Even if lowering the
irradiation temperature can always reduce the quantities of these undesirable
radiolytic products, this generally increases the treatment cost.
O Like thermal treatment, radiation processing causes some loss of vitamins, but
they are only important at sterilisation doses, level which is only used for special
applications such as diets for immunologically-incompetent patients.
O Consequently, irradiation “will not lead to nutrient losses to an extent that would
have an adverse effect on the nutritional status of individuals or populations” [6-8].
Wholesomeness of irradiated foodstuffs
First at all, let us point out two main ideas :
O To ascertain that a chemical compound is dangerous is a scientific absurdity ;

curare can be use in medicine, i.e. it is not always a poison. On the opposite, sugar
may be dangerous (for every one, not only for diabetic patients)… but surely at
high concentrations. What is dangerous, in fact, is “a couple” linking the chemical
structure of a compound with the quantities ingested.
O The absolute wholesomeness of a product (or of a treatment) is impossible to

prove (if it exists) ; what we can only prove is the toxicity, depending of the use of
the product, using assays on animals fed with irradiated diets.
Extensive animal feeding studies designed to detect the potential presence of toxic
substances in various irradiated foods have been carried out since the 1950s, mostly in USA
and Europe. None of the studies carried out under the auspices of the IFIP (24 involved
countries, 67 technical reports published in 12 years) showed any indication that the
irradiated foods contained radiation-produced carcinogens or other toxic substances. In a
French study, for example, nine chemical compounds that had been identified in irradiated
starch were fed daily to rats in amounts calculated to be 800 times the amounts the animals
might be expected to consume from a normal daily intake of irradiated starch. No toxic effect
was found even at this exaggerated rate of intake (chapter 8 of [1]).
Ä&+-Ä
&OODIRRADIATION
Radiation chemistry studies had shown that the radiolytic products of major food
components are present in low quantities (a maximum of about 10 ppm at 10 kGy) and,
moreover, do not depend on the food from which they were derived. Moreover, for major
food components [4,5], these products had also been identified in foods subjected to other
accepted types of food processing such as appertisation, or under natural auto-oxidation.
This understanding is very important for considering general clearance of irradiated foods:
– when foods of similar composition are similarly irradiated, their chemical responses
are similar and they are, accordingly, toxicologically equivalent (for instance: beef,
chicken and pork meats) ;
– when an irradiated food in a class of similar foods is cleared as safe, then other
members of that class are, correspondingly, wholesome (for instance strawberries
and raspberries).
Many investigations have been devoted to possible health hazards : in 1980, a Joint
FAO/IAEA/WHO Expert Committee meeting [6] concluded that “the irradiation of any food
commodity up to an overall average dose of 10 kGy presents no toxicological hazard ; hence,
toxicological testing of foods so treated is no longer required”. Moreover, in September 1997, an
expert study group from WHO, FAO and IAEA, came to the conclusion that “foods treated with
doses greater than 10 kGy can be considered safe and nutritionally adequate when produced
under established Good Manufacturing Practice” [7, 8].
Since then no scientific result has raised doubts about these conclusions.
Methods of detection of irradiated foods
These methods may be helpful to label the foodstuffs, to control trade and for
information of the consumers. However, the most difficult problem has been that the
chemical changes that occur in irradiated foodstuffs [4,5] are very small and generally similar
to those produced by classic food treatment processes (heating, freezing) or natural spoilage
(auto-oxidation).
Many detection methods have been discussed in the literature [3,9,10], but
considerable progress has been made, due to the IAEA ADMIT program and, mainly, to
the European concerted actions [11-14]. We briefly describe here the published European
protocols leading to a proof of irradiation or a proof of no irradiation (Tab. 2).
Ä&+.Ä
2!$)!4)/.#(%-)3429
Table 2 . Detection of irradiated foodstuffs (DIF) with ECN protocols.
Protocol Date Proof Methods Application scope
crude meat, cheese,

1996
EN 1784 DIF containing fat by GC analysis of hydrocarbons avocado, papaya,
2003 (1)
mango
1996 DIF containing fat by GC/MS analysis of crude meat

EN 1785
2003 2-alkylcyclobutanones whole liquid egg
EN 1786 1996 DIF containing bone by ESR spectroscopy meat and fish
1996 pistachio nuts,

EN 1787 DIF containing cellulose by ESR spectroscopy
2000 paprika powder
1996 aromatic herbs,

EN 1788 DIF containing silicate studied by TL
2001 spices, shrimps
dried figs, mangoes

EN 13708 2001 DIF containing crystalline sugars by ESR
papayas & raisins
Protocol Date Screening Methods Application scope
herbs, spices,
EN 13751 2002 DIF using photostimulated luminescence
seasonings, shellfish
DIF using Direct Epifluorescence Filter Technique/

EN 13783 2001 Poultry meat, spices
Aerobic Plate Count (DEFT/APC)
Meats, spices,
EN 13784 2001 DNA Comet Assay for the DIF
dried fruits
DIF using the limulus amoebocyte lysate test in

EN 14569 2004 conjunction with a gram negative bacterial count Poultry meat
(LAL/GNB)
(1) The second date corresponds to a first revision of the protocol, which must be done, at least, every five years.
GC: gas chromatography, MS: Mass Spectrometry, ESR: electron spin resonance, TL: Thermoluminescence.
Two “chemical” protocols voted by the ECN are relative to food containing lipids ;
the percentage of radiolytic products from lipids such as volatile hydrocarbons, aldehydes
or butanones is directly linked to the chemical composition of lipids. In the case of another
treatment such as heating there is no direct correlation between the different percentages.
For instance, the two main hydrocarbons radio-induced from each fatty acid of
structure CN:M (i.e. having chains with N carbon atoms and M double bonds), have structures
JO$/. XJUIPOFDBSCPOMFTTUIBOUIFQBSFOUGBUUZBDJE
BOE$/. XJUIUXPDBSCPOT
less and one extra double bond more), leading to protocol EN 1784 (Fig. 2).
Ä&,%Ä
&OODIRRADIATION
Figure 2 : When avocado samples have been irradiated, the lipids in CN ; M (N carbon atoms and M double
bonds) lead to two main hydrocarbons in C N–1 : M and C N–2 : M+1: the three pie - charts are quite identical
(in contrast with any of other food treatments). This constitutes a specific test of the irradiation.
The natural amount of aldehydes in foodstuffs is too high, with regard to the radio-
induced one, to allow the establishment of a protocol. But the radio-induced cyclobutanones
lead to the protocol EN 1785 and it must be noted that, up to now, the butanones seem to
be characteristic of the irradiation treatment.
Three protocols used the Electron Spin Resonance (ESR) technique ; ESR allows
the observation of unpaired electrons, especially free radicals induced by irradiation, if
they are stable during commercial storage of the food. This only occurs in the solid and dry
components of the food, where the reactivity of the radicals with each other or with water
is low. The first protocol (EN 1786) is relative to meat and fish bones (Fig. 3), the second one
(EN 1787) to food containing cellulose such as berries, and the third one (EN 13708) to food
containing crystallized sugars.
Figure 3 : ESR spectra of unirradiated (a) and 3 kGy irradiated (b) bones. The magnetic field in the abscissae
is in mT (milli tesla) and the intensity is in a.u. (arbitrary unit).
Ä&,&Ä
2!$)!4)/.#(%-)3429
A last European protocol uses thermoluminescence (TL) : irradiation results in

positive ions and free electrons, which can be relatively stable in solid and dry foodstuffs.
When the food is fastly heated, the recombination of ions and electrons leads to light
emission (luminescence). It was later pointed out that the seat of the phenomenon is not
in the vegetable material itself but in contaminating mineral particles adhering to the
products. On this basis, an European protocol (EN 1788) has been approved for herbs, spices
and shrimps.
Some other methods were also approved but only lead to an irradiation presumption
(i.e., in fact, a proof of treatment) and can be used as screening protocols (Tab. 2) as they are
easy to use in food control laboratories.
A possible future of food irradiation
We must notice that each country has its own law [15], a complexity that hampers
the development of the process. In Europe, the Brussels Commission wrote a first directive
(European “law”) in 1986 but the European Parliament rejected it ; one reason for that was
the impossibility, in 1988, to bring the proof whether or not a foodstuff has been irradiated.
Since the publication of the first European protocols, a new directive was written and then
voted by the European Parliament in 1999 [16]. This directive was now adapted in each law
of European countries but the final positive list (i.e. included authorized irradiated foodstuffs)
is still under discussion. Moreover the European Union is also blocked by :
O the attitude of certain countries (such as Germany, despite the fact this country was
one main contributor to this technology and to the proof of its wholesomeness) ;
O the lacking enthusiasm of other members (a consequence of a bad use of the

“precautionary principle”?) ;
O the lobbying of certain associations of consumers, chemical industrials and food

marketing networks.
But Europe is not alone in the world: the Food and drug administration (FDA)
approved for instance, on December 1997, the use in the US of irradiation to kill harmful
bacteria such as Escherichia coli in beef, lamb and pork. The FDA was acting on a three years
old petition since 1994, but gave the allowance after the recall last August of 11,000 tons of
hamburger meat feared tainted with E. coli. If applied in France, the accident, which occurred
in South-West of France in late 2005, could not have happened.
Ä&,'Ä
&OODIRRADIATION
Some developing countries which export into Europe products disinfested or

debacterised by chemical fumigation well know that, once, this will be completely forbidden
(Toronto agreement, December 1995) ; but a working group of the French Academy of Medicine
found that, even in France, at present time, some industrialists prefer to use this forbidden
treatment (without labelling) instead of irradiation (which requires a labelling). WHO also point
out the interest of irradiation to fight against problems of salmonella, lysteria etc. [17-20].
The next years will probably be a turning point with regard to the development of
food irradiation :
– When the directive will be applied, it will be easy to inform the consumers (who at
present have no choice) with an “official development” of the treatment, including
good labelling of the irradiated foodstuffs, or
– Whether or not there will be official development in Europe, there will be in all
cases introduction in Europe of foreign irradiated foodstuffs, due to the need
to replace some chemical treatments and to fight problems of salmonellae, but
more or less visible for the consumers which means no opportunity for a good
information. This is also why, despite a lot of existing texts on food irradiation (see for
example the Codex Alimentarus, [21]), there is a demand for an ISO protocol for
“Good processing practices for the irradiation of foods for human consumption” [22].
Moreover, last year, the USA asked the World Trade Organization to recognize a
technical rule on processing and handling of irradiated food.
The choice is clear. The issue of consumer protection must not be understood as a
means to prevent the consumer to be informed and to make an informed choice [2] : “Shall
we let Sound Science decide the safety of irradiated food?” [23]. Now the ball for acceptance by
the public is in the court of policy and health authorities but also in the court of consumers
associations…
References
[1] Vasseur J.-P., Ionisation des produits alimentaires, Tec & Doc - Lavoisier - APRIA - Teknéa, Paris,
1991.
[2] Ehlermann D., Four decades in food irradiation, Rad. Phys. Chem., 2005, 73, 346-647.
[3] Raffi J., The state of food irradiation and of detection of irradiated foodstuffs, Res. Adv. Food Sci.,
2002, 3, 11-19.
Ä&,(Ä
2!$)!4)/.#(%-)3429
[4] Elias P., Cohen A., Radiation chemistry of major food components, Elsevier, Amsterdam, 1977.
[5] Elias P., Cohen A., Recent advances in food irradiation, Elsevier, Amsterdam, 1983.
[6] Joint FAO/IAEA/WHO Expert Committee, Wholesomeness of Irradiated Food, WHO, N°659,
Geneva, 1981.
[7] WHO, Food irradiation – Sky’s the limit, Press Release WHO/68, 1997, 19 September.
[8] Joint FAO/IAEA/WHO Study Group, High-dose irradiation: Wholesomeness of food irradiated
with doses above 10 kGy, WHO Techn. Rep. Ser., 1999, 890, Genève.
[9] Delincée H., Analytical detection methods for irradiated foods. A review of the current literature,
IAEA-TECDOC-587, Vienna, IAEA,1992.
[10] Raffi J., Methods of identification of irradiated foodstuffs and relative products, in “Handbook of
food analysis”, Nollet L. (ed), 2nd edition, Marcel Dekker/CRC Press, 2004, 1919-1940.
[11] Raffi J., Stevenson M.H., Kent M., Thiéry J.M., Belliardo J.-J., A European intercomparison on Electron
Spin Resonance identification of irradiated foodstuffs, Int. J. Food Sci. Technol., 1992, 27, 111-124.
[12] Raffi J., Delincée H., Marchioni E., Hasselmann C., Sjöberg A.-M., Léonardi M., Kent M., Bögl K.-W.,
Schreiber G., Stevenson M.H., Meier W., New methods for the detection of irradiated food, CEC,
BCR, Luxembourg, 1993, EUR 15261 EN.
[13] McMurray C., Stewart E., Gray R., Pearce J., Detection methods for irradiated foods, current status,
Royal Chem. Soc, 1996.
[14] Raffi J., Stachowicz W., Migdal W., Barabassy S., Kalman M., Yordanov N., Andrade E., Prost M.,
Callens F., Final report of Copernicus contract CIPA-CT94-0134 on “Establishment of an eastern
network of laboratories for identification of irradiated foodstuffs”, CEC, 1998.
[15] Database on approvals for irradiated foods, Food & Environmental Newsletter, 2006, 9, 21-59.
[16] Directive 1999/2/CE du Parlement Européen et du Conseil du 22/2/99, J. Off. Comm. Eur., 13/3/99.
[17] WHO/FAO, Food Irradiation, A technique for preserving and improving the safety of food, Geneva,
1988.
[18] WHO, Study on the safety and nutritional adequacy of irradiated food, Press Release WHO/68, 19
September 1994.
[19] WHO, Multi-drug resistant salmonella typhimutium, fact sheet N°139, January 1997.
[20] WHO, Food safety – Joint FAO/IAEA/WHO Study Group on high Dose Irradiation, Weekly
Epidemiological record, 1998, 73, 9-11.
[21] Codex Alimentarus, CAC/RCP 19-1979, Rev. 2-2003, recommended international code of practice
for irradiation processing of food, 2003.
[22] ISO/CD 22810, Food irradiation – Good processing practices for the irradiation of foods intended
for human consumption, 2006.
[23] Loaharanu P., Shall we let Sound Science decide the safety of irradiated food? American Council of
Science and Health, Facts & Fears, 15 December 2006, http://www.acsh.org/factsfears/newsID.895/
news_detail.asp
Ä&,)Ä
Part III
Radiation damage
to biomolecules,
radioprotection and
radiotherapy
Chapter 12
Radiation-induced
damage to DNA : from model
compounds to cell
Thierry DOUKI and Jean CADET
Deoxyribonucleic acid (DNA) is a key biomolecule involved in the storage of genetic

information and its transmission from one cell generation to the other at each cellular
division. These biological functions are made possible by the complementarity between
nucleic bases on the two opposite strands: thymine specifically pairs with adenine and
guanine with cytosine. In the DNA molecule, bases are linked to a sugar (2-deoxyribose),
and the resulting nucleosides are bound to a phosphate group. The obtained nucleotides
are the basic monomeric units of DNA and form a chain by linkage between two consecutive
deoxyribose rings through a phosphodiester bond. A wide array of chemical and physical
agents, including ionizing radiations, is able to damage the chemical structure of DNA.
A first result of such an event is the killing of the cell. Another major consequence of the
formation of DNA damage is the modification of the genetic information (a mutation)
at the damaged sites in the newly synthesized strand upon replication of DNA (Inset).
This process not only involves the irradiated cell but all the following generations and may
represent an initiating event in the formation of tumors. In the present review, emphasis will be
placed on the chemical pathways involved in the radical degradation of the DNA components
(Fig. 1). These processes are mostly accounted for by the attack of the highly reactive tOH
radical produced upon radiolysis of the water molecules (Chapter 1) surrounding DNA and by
the direct ionization of nucleic components by the incident radiation. The relative contribution
of the indirect and direct effects to radiation-induced DNA damage has been and still is an
issue of debate and a subject of intense investigations (Chapter 13). Another specific feature
Ä&,,Ä
2!$)!4)/.#(%-)3429
Inset : Induction of mutation by DNA damage

The figure below illustrates how a damaged base (a thymine in this case) leads to a change
in the genetic information, namely a mutation. The two DNA strands are symbolized by the
vertical lines linked by the nucleobases. The modified thymine appears in grey. Following
apparition of the damage, the hydrogen bonding properties are lost or modified. Upon the 1st
replication, a guanine instead of an adenine may be incorporated in front of the damage. In the
following generations of cells, the guanine is normally paired with cytosine, leading to a T:A to
C:G mutation.
of the interaction between ionizing radiation and the genome is the co-localization of several
lesions along the track of the incident particle, resulting in highly deleterious clustered damage.
Comprehensive mechanisms of degradation are available on most of the DNA nucleobases
and the 2-deoxyribose moiety as inferred from studies on model systems and isolated DNA.
In contrast, much less information is available on cellular DNA despite recent development of
biochemical and chemical analytical tools aimed at measuring base lesions.
Radiation-induced degradation pathways of DNA monomers
Extensive studies on monomeric compounds, such as bases and nucleosides, have

led to the isolation and the characterization of the final radiation-induced degradation
products [1]. In addition, electron spin resonance, pulse-radiolysis studies and comparison
Ä&,-Ä
2ADIATIONDAMAGETO$.!
Figure 1 : The different pathways involved in radiation-induced damage to DNA. 1) Indirect effect corresponds
to the degradation of DNA by the reactive species produced upon water radiolysis. The most damaging radical
is tOH that may react with deoxyribose and the four bases. 2) The direct effect is the ionization of DNA compo-
nent by the incident radiation. The resulting positive charges migrate within the double-helix towards guanine.
3) Both mechanisms take place along the track of the incident radiation, leading to the localization of several
damages in a small portion of DNA. (Ade: Adenine, Thy: thymine, Gua: guanine, Cyt: cytosine, dR: 2-deoxyribose.)
with photosensitization experiments have provided information on the structure and the
reactivity of the main radical intermediates produced in the radical reactions [2]. Altogether
this has permitted to propose detailed mechanisms for the main radiation-induced
degradation pathways. The emphasis is placed here on oxidation reactions in the presence
of oxygen that are predominant in cells. It should though be kept in mind that reductive
pathways, that mostly involve initial addition of electron or hydrogen atom to pyrimidine
bases, may also give rise to stable final products under anaerobic conditions.
tOH radical-induced damage to nucleobases

Among the species produced upon radiolysis of water, hydroxyl radical (tOH) is the
most reactive. Indeed, its reaction rate with the four bases and related nucleosides is diffusion-
controlled. The main reactive sites of hydroxyl radicals on nucleobases are the double-bonds
of the heterocycles. Accordingly, addition of tOH at the C8 position of adenine and guanine
yields the corresponding reducing 8-hydroxy-7,8-dihydropurin-7-yl radical (Fig. 2). Oxidation
of this intermediate leads to the formation of related 8-hydroxypurines that are in dynamic
equilibrium with their more stable 8-oxo-7,8-dihydropurine tautomeric form. Competitive
reduction of the latter purine radical gives rise to imidazole ring opened compounds: the
Ä&,.Ä
2!$)!4)/.#(%-)3429
formamidopyrimidine (Fapy) derivatives. Addition of tOH to the C4 position of the moiety, that
is in fact the predominant radical reaction, yields an oxidizing oxyl type radical subsequent to
a fast dehydration process. Addition of superoxide radical to the latter intermediate gives rise
to an imidazolone product that is quantitatively hydrolyzed into its oxazolone derivative.
Figure 2 : Hydroxyl radical-induced degradation of guanine. Addition of tOH to the C8 position leads to
the reducing 8-hydroxy-7,8-dihydroguan-yl radical that is then either oxidized into 8-oxo-7,8-dihydroguanine
(8-oxoGua) or reduced into formamidopyrimidine lesion (FapyGua).
In the case of pyrimidine bases, hydroxyl radical preferentially adds to the C5

and, to a lesser extent, to the C6 positions of thymine and cytosine (Fig. 3). The obtained
hydroxyl radical adducts may undergo a fast reaction with molecular oxygen. Reduction
of the resulting peroxyl radicals yields related hydroperoxides, that then give rise to final
stable products such as 5,6-hydroxy-5,6-dihydropyrimines (glycols ; ThdGly for thymidine)
by reduction, together with ring rearrangement and fragmentation products. tOH radicals may
Figure 3 : tOH radical-induced degradation of thymine. Oxidation involves either tOH addition to the C5-
C6 double bond or tOH -mediated hydrogen abstraction from the methyl group. Molecular oxygen then adds
to the resulting carbon-centered radicals. This reaction yields hydroperoxide that further evolves into stable
products, such as thymidine glycol (ThdGly), 5-hydroxymethyl-uracil (HMdUrd) and 5-formyl-uracil (5-FordUrd).
N- represents the bond to the DNA backbone.
Ä&-%Ä
also abstract a hydrogen atom from the methyl group of thymine. In the presence of oxygen,
the radical produced leads to the formation of a hydroperoxide, 5-(hydroperoxymethyl)
-2’-deoxyuridine. The two final stable products of this degradative pathway are the
corresponding alcohol and the aldehyde, namely 5-(hydroxymethyl)-2’-deoxyuridine
(HMdUrd) and 5-formyl-2’-deoxyuridine (FordUrd), respectively.
One-electron oxidation reactions

Lesions induced by the direct ionization of DNA by ionizing radiation are difficult to
study in aqueous solution because of the predominant effect of the species produced upon
water radiolysis (Chapter 13). However, the chemical pathways triggered upon ionization of
bases and nucleosides have been unraveled by one-electron oxidation reactions provided
by chemical and photochemical systems. It was thus shown that hydration of the purine
radical cations gives rise to the reducing 8-hydroxy-7-yl radical also observed upon tOH
addition at position 8 (Fig. 4). Competitive deprotonation of the radical cations leads to the
formation of related oxidizing neutral radical that gives rise to imidazolone and oxazolone
from guanine and to hypoxanthine from adenine. Study of the one-electron oxidation of
2’-deoxyguanosine has been hampered by the consumption of 8-oxo-7,8-dihydroguanine
as soon as it is formed due to its low ionization potential by the oxidizing guanine radical.
Figure 4 : Reactivity of radical cations of guanine (upper panel) and thymine (lower panel). In both cases,
the radical cations produced by ionization deprotonate or are hydrated to yield the same radicals than those
produced upon reaction with tOH radicals.
One-electron oxidation reactions of pyrimidine bases are also well established, mostly
through the study of the menadione-mediated type I photosensitization of 2’-deoxycytidine and
thymidine [3]. The radical cations once formed may undergo two competitive pathways.
Ä&-&Ä
2!$)!4)/.#(%-)3429
Hydration of the thymidine radical cation leads to the predominant formation of the
oxidizing 6-hydroxy-5,6-dihydrothymid-5-yl radical whereas deprotonation mostly generates
the methyl-centered radical (Fig. 4). As already pointed out, these two pyrimidine radicals
are also produced upon reaction with tOH. However, the methyl-centered radical represented
only 5% of base radicals when produced by tOH while it corresponds to 30% of the thymidine
radicals produced upon one-electron oxidation. Type I photosensitization of 2’-deoxy-
cytidine leads to the formation of the 6-hydroxy-5,6-dihydro-2’-deoxycytidil-5-yl radical
also produced upon reaction with tOH. Additional oxidative pathways involve the formation
of 2-deoxyribonolactone and free cytosine as well as production of 2’-deoxyuridine as the
result of deprotonation of the pyrimidine radical cation at C1’ and NH2 group, respectively.
Radiation-induced damage to oligonucleotides

and double-stranded DNA
Studies on isolated DNA and sugar model compounds have allowed the
determination of the main tOH-mediated degradation pathways to the 2-deoxyribose
moiety. In addition, data are also available on the influence of the double-helix structure on
the fate of purine and pyrimidine base radicals.
Sugar damage
Studies involving gas chromatography measurements, isotopic labeling experiments
and synthesis of photolabile precursors of sugar radicals have provided relevant information
on the tOH induced hydrogen abstraction to 2-deoxyribose that in most cases leads to
DNA strand breaks [4] (Fig. 5). The most reactive site of the sugar ring of DNA is C4’. In the
presence of molecular oxygen, the resulting carbon centered radical gives rise to a peroxyl
intermediate that may decompose predominantly into phosphoglycolate with the release of
base propenal. Hydrogen abstraction at the 5’ position (Fig. 5) is also a common reaction that
yields among other oxidation products the 5’-aldehyde derivative under aerobic conditions.
These reactions, as well as other minor ones involving H2’ and H3’, lead to frank strand breaks.
In contrast, abstraction of H1’ in the presence of oxygen gives rise to 2-deoxyribonolactone
without cleavage of the phosphate-sugar backbone. This oxidized abasic site is unstable
under even mild alkaline conditions leading to the formation of a strand break.
Radiation-induced base damage within isolated DNA

Some of the final radical-degradation products of nucleobases have been detected
and quantified within isolated DNA. Interesting notable changes in the product distribution
Ä&-'Ä
Figure 5 : Examples of DNA single strand breaks produced upon tOH-mediated hydrogen abstraction to the
2-deoxyribose moiety. Attack at the H4’ position induces the release of base propenal and the formation of
phosphoglycolate, while abstraction of H5’ leads to the formation of a 5’-aldehyde.
have been observed within DNA duplex with respect to that obtained with free bases or
nucleosides. The most striking one is the predominant formation of 8-oxo-7,8-dihydroguanine
in DNA under aerated conditions whereas the related nucleoside is barely detected upon
gamma irradiation of aerated aqueous solutions of dGuo. Indeed, 8-oxoGua bond to the
phosphodiester backbone can no longer diffuse and react with oxidizing species like in
solutions of free nucleosides. Moreover, formation of FapyGua and FapyAde at the expense
of the 8-oxo-purine lesions was observed under reducing conditions, as observed with the
free nucleosides. These results confirm the common origin of these two classes of modified
purines that arise from the reducing 8-hydroxy-7,8-dihydropurin-7-yl radicals (Fig. 2).
Data have also been obtained on the formation of oxidized pyrimidines within
isolated DNA exposed to ionizing radiation in aerated aqueous solutions, conditions where
t
OH is the main reactive species. Thymidine glycols and the methyl oxidation products
HMdUrd and FordUrd were found to be produced in significant yields. Interestingly, the ratio
between the yields of these two classes of damage is much lower than for the free nucleoside.
This may be explained by a higher accessibility to tOH of the methyl group than the C5-C6
double bond in double-stranded DNA. The available information on oxidatively generated
damage to the cytosine base within double-stranded DNA is much more limited than for the
other bases. Evidence was however gained for the formation of 5-hydroxy-2’-deoxycytidine
together with low amounts of 5-hydroxy-2’-deoxyuridine.
Ä&-(Ä
2!$)!4)/.#(%-)3429
Tandem lesions
Extension of the study of base damage from free bases or nucleosides to
oligonucleotides and isolated DNA has permitted the identification of a new class of base
lesions that involves the reaction of a DNA radical with an adjacent nucleotide on the
same strand [5]. Thus, exposure to X-rays of dinucleotides carrying thymine and guanine
in aerated aqueous solution gave rise to lesions exhibiting both an 8-oxoGua moiety and
a formylamine residue (dF), produced by the oxidation of the guanine and thymine bases,
respectively. These 8-oxoGua/dF tandem lesions, which are produced by a single initial radical
event, were also found to be generated in a significant yield within isolated DNA (Fig. 6) [6].
Mechanistic studies showed that the initial step in the formation of 8-oxoGua/dF was the
addition of tOH to the C5-C6 double bond of thymine, followed by reaction of molecular
oxygen and addition of the resulting thymine peroxyl radical to the C8 position of guanine.
Other types of tandem lesions arising from a single radical event involve crosslinking of
thymine to an adjacent purine via its methyl group and cyclonucleosides that arise from the
intramolecular addition of the 5’-carbon centered radical to the C8 position of a purine base
of the same nucleotide.
Figure 6 : Detection of the 8-oxoGua/formylamine tandem lesions in isolated DNA exposed to gamma radiation
in aerated aqueous solution. DNA was digested by a mild enzymatic treatment in order to release the lesions without
cleavage of the phosphodiester bond. The sample was then injected onto a reverse phase HPLC column. The detection
was provided by a mass spectrometer monitoring the main fragmentation of the two tandem base lesions.
Charge transfer
The double-stranded structure of DNA not only affects the chemistry of tOH-induced
damage but also may modulate oxidation reactions that take place upon initial direct
ionization of a given nucleotide. Indeed, the positive holes created in the latter process have
Ä&-)Ä
been shown to migrate through DNA double-helix, as extensively discussed in Chapter 14.
They are eventually trapped at sites exhibiting the lowest ionization potential, namely the
guanine bases and especially those located on the 5’-end of GG doublets. This was inferred
from electron spin resonance studies performed at low temperature on DNA samples
exposed to ionizing radiation and from photoionization experiments of oligonucleotides
under more physiological conditions. In the latter works [7-9], damage was mapped at the
sequence level by electrophoretic techniques following conversion of the base damage into
strand breaks by either alkaline treatment or incubation with specific N-glycosylase repair
enzymes. Although the range of charge migration and the underlying mechanism(s) are still
a matter of debate, the predominance of base damage over direct strand breaks and the
overwhelming degradation of guanine are common features reported by all the groups. It
may be added that direct measurement of base damage by a more specific chromatographic
approach confirmed these observations [10].
Radiation-induced damage to cellular DNA
Measurement techniques
A real gap exists between the amounts of mechanistic information gathered in
model systems and isolated DNA and those available at the cellular level. The measurements
of oxidatively generated lesions have been hampered by the low level of modifications
(in the range of 1 per million normal nucleotides) and usually the low amounts of the
available DNA samples. The bulk of the quantitative data on radiation-induced damage to
cellular DNA are related to single (SSB) and double-strand breaks (DSB). Indeed, a number
of electrophoretic methods allow their detection either following DNA extraction such
as pulsed-gel electrophoresis for monitoring DSB or in the whole cell such as the alkaline
comet assay or alkaline elution for measuring both SSB and DSB. The latter methods may
also be used to quantify base damage after conversion into strand breaks by using DNA
repair glycosylases that cleave DNA at the site of the lesion. These biochemical techniques
are sensitive and are not subjected to artifactual oxidation. However, they do not provide
specific information on the chemical nature of the damage since repair enzymes excise
a broad spectrum of lesions. More specific data on modified bases can be obtained by
applying chromatographic methods that require DNA hydrolysis into monomeric units
prior the measurements. Liquid chromatography associated with electrochemical detection
has been used extensively for the quantification of 8-oxo-7,8-dihydro-2’-deoxyguanosine.
A more versatile approach involves tandem mass spectrometry that makes possible the
quantification of several modified nucleosides in one analysis. Gas chromatography coupled
to mass spectrometry has also been proposed. However, flaws in the sample treatment that
Ä&-*Ä
2!$)!4)/.#(%-)3429
involved a silylation step at high temperature have led to the overestimation of the level of
lesions. The limitations of chromatographic methods include a need for rather large amount
of DNA and occurrence of spurious oxidation upon extraction of DNA. Recent efforts have
been devoted to minimize the latter risk. Altogether, less than 10 modified bases can be
accurately quantified within cellular DNA while more than 50 different types of radiation-
induced lesions have been identified in nucleosides.
Radiation-induced DNA damage in cells

Evidence is now available in a wide variety of cell types for the formation of single
and double-strand breaks in a ca. 25:1 ratio within DNA after exposure to X- and G-rays.
Biochemical approaches showed that modified bases are produced in amounts similar to
those of single strand breaks. It was also found, by using both enzymatic and HPLC-MS/MS
assays that purine and pyrimidines bases are damaged to a similar extent (Fig. 5). Thymine
glycols were found to be produced with the highest frequency. Methyl oxidation products
of thymine including HMdUrd and FordUrd were also shown to be efficiently generated.
8-OxoGua was found to be another major lesion but FapyGua was at least equally frequent.
The observation of an efficient radiation-induced formation of FapyGua strongly suggests
that cellular DNA is in a reducing context likely due to a relatively low intracellular oxygen
concentration and the presence of reducing species. Formation of FapyAde and 8-oxodAdo
was also observed but the overall yield of these two adenine lesions was about one order
of magnitude lower than that of related guanine damage, suggesting a much lower
Figure 7 : Yield of formation of oxidized bases in human THP1 monocytes exposed to G-rays or 266 nm high
intensity laser beam. Yields are expressed in lesion/10 6 bases per 10 Gy and per mJ, respectively. Red, blue and
green bars correspond to guanine, adenine and thymine damage, respectively.
Ä&-+Ä
susceptibility of the adenine moiety to radical oxidations in DNA. Altogether, the available
data show the predominance of the hydroxyl radical-induced damage within the DNA of
cells exposed to low linear energy transfer (LET) radiation. Interestingly, ionization of cellular
DNA by exposure to high intensity 266 nm laser pulses leads to the overwhelming formation
of 8-oxodGuo (Fig. 7), in agreement with occurrence of charge transfer, guanine bases being
the sink for this process.
A distribution of oxidized bases similar to that obtained with G-rays is observed upon
exposure of human cells to high LET particles such as heavy ions [11]. However, with both
12
C and 36Ar ions beams, a decrease in the yield of all lesions was observed [12]. This
observation is reminiscent of the effect of increasing LET on the radiolytic yield of tOH that
decreases as the result of efficient radical recombination in the particle track. A same trend
has been observed for both single and double-strand breaks [13]. These results strongly
suggest a major role of the hydroxyl radical chemistry in the formation of radiation-induced
DNA damage in cells.
Clustered damage
The decrease in the yield of lesions upon increasing LET seems contradictory with
the well established increase in cell lethality for a same dose of radiation (relative biological
effectiveness). Another striking result is the relatively low yield of damage induced within
DNA when a lethal dose is applied to cells. Indeed, 4 Gy of G-rays which is a lethal dose is
needed to double the basal level of 8-oxodGuo as inferred from comet assay measurements.
Both observations can be explained by the implication of other types of radiation-induced
damage that are likely to have strong biological impact. Indeed, in contrast to other types of
oxidative stresses which involve the production of oxidizing species homogeneously within
the cell, ionizing radiation triggers induction of chemical reactions mostly in localized area
along the incident track. In the recent years, modelization studies have shown that even
low LET particles are able to induce a significant fraction of damage that consists in two
individual modifications located only a few nucleotides apart on the same strand or on the
opposite strands [14]. Interestingly, the complexity of the clustered damage increases with
increasing LET. Such a process explains why DSBs are considered as specific radiation-induced
DNA lesions. However, DSBs represent only a fraction of the complex, so-called “clustered”,
lesions that may consist in two oxidized bases or a modified base and a strand break, or any
larger combination of these individual lesions (Fig. 1). Observation of an increase in the ratio
between the yields of double- and single-strand breaks with increasing LET strongly supports
this hypothesis [13]. More direct evidence for the formation of clustered damage has been
provided by the use of repair enzymes in order to convert closely located modified bases (or a
Ä&-,Ä
2!$)!4)/.#(%-)3429
modified base and as SSB) into double-strand breaks [15]. Recent studies involving synthetic
oligonucleotides have shown that base damage and/or strand break located within a cluster
are less efficiently repaired by comparison with isolated lesions [16,17]. Therefore, clustered
lesions may persist inside the genome and interfere with DNA transcription and replication,
leading ultimately to mutagenesis and cell death.
Conclusion and perspectives
Radiation-induced damage to DNA arise from either the reaction of highly reactive
t
OH radicals produced by radiolysis of water or, to a lesser extent, of the direct ionization
of DNA components. In addition, the clustering of the lesions along the radiation track
is expected to be highly damaging to the cell. Recent data, mostly based on the use of
microbeam of charged particles, have revealed other levels of complexity. Indeed, deleterious
effects such as mutagenesis are observed in cells that are not directly hit, a process so-called
the “by-stander” effect. Another interesting effect is the induction of mutations in cells that
were hit only in the cytoplasm and not in the nucleus where DNA is located. In addition,
other radiation-induced reactions, such as dissociative capture of low energy electrons and
K-shell ionization events, have also been proposed to be involved in the induction of DNA
lesion. However, evidence for the implication of the two latter processes at the cellular level
is still required. It is clear that major efforts are still needed to gain further insights in the
complex mechanism of formation of radiation-induced damage to DNA and to better assess
the biological role of these modifications.
References
[1] Cadet J., Berger M., Douki T., Ravanat J.-L., Oxidative damage to DNA: Formation, measurement and
biological significance, Reviews Physiol. Biochem. Pharmacol., 1997, 131, 1-87.
[2] Steenken S., Purine bases, nucleosides, and nucleotides: aqueous solution redox chemistry and
transformation reactions of their radical cations and e- and OH adducts, Chem. Rev., 1989, 89,
503-520.
[3] Wagner J.R., van Lier J.E., Johnston L.J., Quinone sensitized electron transfer photooxidation of
nucleic acids: Chemistry of thymine and thymidine radical cation in aqueous solution, Photochem.
Photobiol., 1990, 52, 333-343.
[4] Knapp Pogozelski W.K., Tullius T.D., Oxidative strand scission of nucleic acids: route initiated by
hydrogen abstraction from the sugar moiety, Chem. Rev., 1998, 98, 1089-1107.
[5] Box H.C., Dawidzik J.B., Budzinski E.E., Free radical-induced double lesions in DNA, Free Radic.
Biol. Med., 2001, 31, 856-868.
Ä&--Ä
[6] Bourdat A.-G., Douki T., Frelon S., Gasparutto D., Cadet J., Tandem base lesions are generated by
hydroxyl radicals within isolated DNA in aerated aqueous solution, J. Am. Chem. Soc., 2000, 122,
4549-4556.
[7] Schuster G.B., Long-range charge transfer in DNA: transient structural distortions control the
distance dependence, Acc. Chem. Res., 2000, 33, 253-260.
[8] Treadway C.R., Hill M.G., Barton J.K., Charge transport through a molecular π-stack: double helical
DNA, Chem. Phys., 2002, 281, 409-428.
[9] Giese B., Long-distance charge transport in DNA: the hopping mechanism, Acc. Chem. Res., 2000,
33, 631-636.
[10] Douki T., Angelov D., Cadet J., UV Laser photolysis of DNA: effect of duplex stability on charge-
transfer efficiency, J. Am. Chem. Soc., 2001, 123, 11360-11366.
[11] Pouget J.-P., Frelon S., Ravanat J.-L., Testard I., Odin F., Cadet J., Formation of modified bases in
cells exposed either to gamma radiation or high-LET particles, Radiat. Res., 2002, 157, 589-595.
[12] Douki T., Ravanat J.-L., Pouget J.-P., Testart I., Cadet J., Minor contribution of direct ionization to
DNA base damage induced by heavy ions, Int. J. Radiat. Biol., 2006, 82, 119-127.
[13] Roots R., Holley W., Chatterjee A., Irizarry M., Kraft G., The formation of strand breaks in DNA after
high-LET irradiations: a comparison of data from in vitro and cellular systems, Int. J. Radiat. Biol.,
1990, 58, 55-69.
[14] Nikjoo H., O’Neill P., Terrissol M., Goodhead D.T., Quantitative modelling of DNA damage using
Monte Carlo track structure method, Radiat. Environ. Biophys., 1999, 38, 31-38.
[15] Sutherland B.M., Bennet P.V., Sidorkina O., Laval J., Clustered damage and total lesions induced in
DNA by ionizing radiation: oxidized bases and strand breaks, Biochemistry, 2000, 39, 8026-8031.
[16] Blaisdell J.O., Wallace S.S., Abortive base-excision repair of radiation-induced clustered lesions in
Escherichia coli, Proc. Natl. Acad. Sci. USA, 2001, 98, 7426-7430.
[17] David-Cordonnier M.-H., Boiteux S., O’Neill P., Efficiency of excision of 8-oxo-guanine within DNA
clustered damage by XRS5 nuclear extracts and purified human OGG1 protein, Biochemistry,
2001, 40, 11811-11818.
Ä&-.Ä
Chapter 13
Mechanisms of direct radiation
damage to DNA
Michael D. SEVILLA and William A. BERNHARD
Introduction
It has been estimated that approximately 50% of the DNA damage from G-irradiation
is due to direct-type effects, which encompass two types of events: those stemming from
energy deposited in DNA itself and those starting with ionization of the DNA solvation shell
and rapidly followed by transfer of radical cations and ejected electrons to DNA (quasi-
direct effects) [1-3]. The 50% remaining damage is from water radicals via the indirect effect
(Chapter 12).The large contribution of direct-type effects stems from the fact that the DNA of
higher organisms is highly condensed in the form of chromatin, effectively excluding water
near DNA and, thereby, reducing the importance of the indirect effect [1]. For these reasons
a full understanding of the direct-type effects is required for a comprehensive model of the
effects of radiation on DNA in vivo [3]. Calculations indicated that about 90% of the energy
deposited in liquid water is via ionization ; presumably, a similar fraction would apply in the
case of DNA [4]. Ionization creates sites of electron-loss (a hole or radical cation) and sites
of electron-gain (radical anion). These radical species account for the majority of direct-type
damage observed in DNA. The other portion of the energy is deposited as excitations, which
at least for low LET (linear energy transfer) radiation do not appear to give rise to significant
amounts of chemical damage. Thus, one should be able to account for nearly all the stable
direct-type damage produced in DNA by tracking the fate of the initially formed sites of
electron-loss and electron-gain. Not all ejected electrons upon reaching low energies, <20
eV, terminate by one-electron reduction of a particular DNA site. Some low energy electrons
(LEE) terminate by dissociated electron attachment (DEA) [5], (Chapter 5). It appears,
Ä&.&Ä
2!$)!4)/.#(%-)3429
however, that damage due to DEA for low LET radiation is relatively minor in comparison to
the inelastic process that results in radicals directly formed by one-electron gain and one-
electron loss.
Trapped radicals
Given in Figures 1a and 1b are the structures and labels for free radicals and some
stable end products which are most important to direct radiation damage to DNA and
are discussed here. Because all of these radicals are inherently reactive, it is advantageous
to irradiate DNA samples at low temperatures, 4 K or 77 K, where the radiation produced
radicals are trapped and stable for long time periods and then perform measurements
[6]. The relative yield of each specific primary electron-loss and electron-gain radical
trapped in G-irradiated DNA depends on the temperature and other variables such as DNA
conformation, DNA hydration state, DNA packing, total dose and the LET of the radiation.
Table 1 shows the relative initial yields of the radicals trapped in DNA hydrated to 14 mol
H2O/mol nucleotide [7,8]). The samples were irradiated and measured at 77 K. The first row
shows that for double stranded (ds) DNA exposed to G-rays (low LET radiation) four types of
radicals predominate. Two are due to electron loss, Gua(N1-H)t and dRibt, and two are due to
electron gain, Thyt¦BOE$ZU /)
t. The formation reactions for these radicals are shown in
Figure 2. These results suggest that, at 77 K, the neutral sugar-phosphate radicals originating
from the deoxyribose-phosphate backbone contribute a larger fraction of damage at high
LET than at low LET. Although not shown in the Table 1, tOH radical located in the outer
hydration layer of DNA is also found in small quantities, 0-5%, depending on the degree of
hydration [9]).
Table 1. Approximate initial relative radicala yields in Irradiated double stranded DNA at 77 K [7,8].
Radiation Initial LET Guat+ Adet+ Cytt−

a
Thyt– b
dRib(Cn-H)t RO(O)2Pt–
type (keV/μm) a
Gua(N1-H)t a
Ade(N-H)t Cyt(N3+H)t
G ≈ 0.3 35% < 5% 29% 27% ca. 10% —
36
Ar18+ ≈ 400 22% < 5% 24% 28% ca. 27% ca. 0.1%
a) Note that Guat+ in DNA undergoes reversible deprotonation from N1, Adet+ rapidly deprotonates from its NH2 group (pKa=1) and
Cytt– undergoes rapid protonation at N3. b) May include a small but unknown fraction of neutral base radicals.
Ä&.'Ä
$IRECTRADIATIONDAMAGETO$.!
Figure 1a : Nomenclature used in text for free radical structures and stable end products derived from oxida-
tion and excitation pathways.
Ä&.(Ä
2!$)!4)/.#(%-)3429
Figure 1b : Nomenclature used in text for free radical structures and stable end products derived from oxida-
tion and excitation pathways.
Base radicals
Electrons are selectively captured by the pyrimidines and the holes formed on the
base stack, or transferred to the base stack, are nearly all trapped by Gua, giving Guat. This
species is in a reversible equilibrium with its deprotonated form Gua(N1-H)t (Fig. 2). While
initial distributions between the pyrimidines are nearly equal upon annealing, or irradiation
to high dose, the excess electron transfers to cytosine as the most stable site as a result of
protonation of Cytt–BU/GPSNJOH$ZU /)
t [10]. Molecular orbital calculations show that
the major factor that affects the relative yields of DNA base ion radicals is the difference in both
the ionization potential and electron affinity of the DNA bases [11]. In addition, experimental
and theoretical work shows that variations in relative radical yields are also affected by the
Ä&.)Ä
DNA strandedness. For example, the transfer of both the hole and the excess electron is more
facile through stacked bases in dsDNA than in ssDNA in which the base stacking is disrupted.
Figure 2 : Important reaction pathways stemming from the ionization of DNA (with a model of DNA in the
common B-form configuration). Electron loss pathways are on the left and electron gain pathways on the right.
Notation: LEE (low energy electron), ssb (single strand break), ht (hole transfer), and et (electron transfer).
While the deprotonation reactions from nitrogen sites described above are reversible,
ion radicals of the DNA bases may also react irreversibly in characteristic ways to form neutral
base radicals as shown in Figure 2 [1,3]. DNA base cation radicals may deprotonate from
carbon sites and/or undergo hydroxide ion addition. For the thymine radical cation, Thyt
can form the Thy(Me-H)t radical by loss of a proton from the methyl group before transfer of
the hole to the more stable guanine site [1]. For the guanine radical cation, Guat reaction
XJUIXBUFSZJFMETUIFIZESPYZHVBOJOFSBEJDBM (VB $0)
t on annealing to temperatures
over 200 K. This species undergoes subsequent oxidation to form 8-oxoguanine or reduction
to form FapyGua (2,6-diamino-4-hydroxy-5-formamidopyrimidine) (Fig. 2) [12]. Pyrimidine
anion radicals formed initially protonate irreversibly at the C5-C6 position. Thyt– protonates at
$UPGPSNUIFXFMMLOPXO EJIZESPUIZNJOZMSBEJDBM 5IZ $)
t. Cytto, and its reversibly
QSPUPOBUFE GPSN $ZU /)
t QSPUPOBUF JSSFWFSTJCMZ BU $ PS $ UP GPSN $ZU $)
t and
$ZU /) $)
t (Fig. 2) [10]4VCTFRVFOUPOFFMFDUSPOSFEVDUJPOPG5IZ $)
t followed
by protonation leads to the stable end-product of dihydrothymine DHThy (Fig. 2). In a similar
reaction sequence, DHCyt is formed. In water, deamination of DHCyt gives DHUra, a stable
lesion found in DNA [3,10]. The fate of the radiation-induced hole is then chiefly 8-oxoGua
and FapyGua, whereas, the excess electron results in the 5,6-dihydropyrimidines, DHThy and
Ä&.*Ä
2!$)!4)/.#(%-)3429
DHUra. While these are the major products, if they are spatially isolated from other lesions,
they are readily repaired enzymatically and consequently have no deleterious effect ; but when
formed in clusters, the likelihood of cellular injury increases dramatically [2].
In the past it has been often suggested that neutral DNA base radicals may abstract
a hydrogen atom from the sugar phosphate backbone to form a strand break ; however,
theoretical work clearly indicates that abstraction from the DNA sugar backbone by DNA
base neutral radicals is thermodynamically unfavorable in most cases [11]. It is likely that
strand breaks are largely accounted for by direct ionizations of the sugars with additional
strand fragmentations induced by low energy electrons (LEE) (discussed below).
DNA sugar radicals
There are long-standing questions regarding the identity and the yields of the sugar
radicals in irradiated DNA [13,14]. The DNA sugar-phosphate backbone makes up about
50% of the electron density of the DNA and, therefore, about 50% of all holes and electrons
are initially generated on the DNA backbone. Nearly all ejected electrons are eventually
trapped on the pyrimidine bases (Cyt and Thy) (Tab. 1). However, of the holes generated
in the DNA backbone, both base release studies [15] and ESR (electron spin resonance)
evidence on sugar radicals [14] indicate that there is a partial transfer of radiation-induced
holes to the bases. Without hole transfer from the sugar backbone to the DNA bases, about
25% of all radicals resulting from holes and electrons would then be expected on the sugar
phosphate backbone. However, ESR work suggests about 10% of the total radicals are on
the sugar phosphate backbone [14,16]. In agreement with this later value, product analysis
also suggests that hole transfer to the bases reduces the sugar damage from the calculated
maximum of 25% to ca. 13% of the total [15]. (Note that based on initial ionizations 50% of
the holes correspond to 25% of all radicals.) There is a clear correlation between most sugar
radicals and the subsequent formation of a strand break. However, studies on plasmid DNA
indicate that the yield of trapped sugar radicals fall short of the yield of strand breaks by
about 0.07 μmol/J [16]. This shortfall is yet to be explained and it has been suggested to
involve double oxidations at a single site. One additional known contributor to strand break
formation, which does not involve a radical cation precursor, is the mechanism of DEA by
LEE [19]. The actual yield of strand breaks due to LEE is not known and is presumed to be
relatively small.
Two of the sugar radicals stabilized at 77 K in G-irradiated DNA have been tentatively
identified ; these are the dRib(C1-H)' and dRib(C3-H)' radicals [14]. These are most likely
formed by deprotonation of sugar cation radicals. In recent work [8], with both gamma and
Ä&.+Ä
ion beam irradiated DNA, evidence was reported for two radical species that result from
fragmentation of the DNA sugar phosphate backbone. They are a C3’ dephosphorylation
radical, dRib(C3’–Phos)t, and a phosphoryl radical, RO(O)2Pt–. Both these species are
associated with immediate single strand breaks (SSB). Deprotonation of a deoxyribose-
phosphate radical cation cannot explain these fragmentation reactions. Based on the well
known radiation chemistry of alkyl phosphates, it is likely that dRib(C3’–Phos)t, and RO(O)2Pt–,
are formed by dissociative electron attachment likely via LEE (Fig. 2).
Recent work has also found another mechanism for production of sugar radicals,
i.e. deprotonation from electronic excited states of DNA base cation radicals [17]. This work
suggests that combination of an excited state with a DNA base cation such as Guat results in
hole transfer to the sugar followed by deprotonation to form a sugar radical. This mechanism
is considered to be important in high LET radiation damage in which excitations are copious.
Thus, three mechanisms for sugar radical formation in DNA are now known :
1. The major pathway is hole deprotonation of the directly ionized sugar phosphate
backbone.
2. Secondary pathways likely to be more important in high LET radiation are :

– dissociative electron attachment caused by LEE (low energy electrons) and
– electronic excitation of existing DNA base cation radicals.
Three of the deoxyribose carbon centered radicals, dRib(C3’-H)t, dRib(C4’-H)t, and

dRib(C5’-H)t, give rise to prompt SSB via a phosphate elimination reaction. The strand break
mechanism for the dRib(C1’-H)t radical differs substantially. Like the three radicals above, it will
release a free unaltered base but instead of forming a strand break, a subsequent one-electron
oxidation results in deoxyribonolactone (dRibonoLac) formation (Fig. 1a) [1,18]. This abasic
site is quite stable. Heat and a catalyst, such as spermine, turn this lesion into a strand break
plus 5-methylenefuranone (5MFur) [18]. The later is a signature end product of C1’ sugar
damage.
Low Energy Electrons
As describe above DNA strand breaks can be induced by LEEs. Recent work has
shown LEEs, with energy as low as 1 to 5 eV [19], are effective in this regard. The decay
of localized transient anion states (resonances) within DNA is the principal mechanism
leading to SSB and double strand breaks (DSB) by electrons with energies below 15 eV [19].
Theoretical calculations support the fragility of the DNA backbone to LEEs and suggest
Ä&.,Ä
2!$)!4)/.#(%-)3429
that even 1 eV electrons can cause SSB, perhaps via a “resonant capture mechanism” [20].
From these reports, it is clear that there may be a very low energy threshold for strand
break induction by LEE which act directly on DNA. As mentioned above, ESR studies do find
evidence for specific radical species (dRib(C3’–Phos)t and RO(O)2Pt¦) in low abundance that
likely result from LEE induced strand cleavage.
Transfer of electrons and holes, induced by irradiation,

through duplex DNA
As pointed out above ESR studies show that after low temperature G-irradiation,
electrons are trapped on cytosine and thymine and holes on guanine (Tab. 1). However,
the distances that excess electrons and holes travel before being irreversibly trapped have
been in active dispute in the literature until recently [21]. Earlier results were contradictory
and gave electron migration distances from just a few DNA bases pairs up to hundreds
of base pairs. The structure of DNA with overlapping P-orbitals of the stacked bases even
gave rise to the suggestion that DNA is a “molecular wire.” However, recent work [21] shows
that DNA is in fact a poor semi-conductor, which is by definition an insulator. Since hole
and electron migration within DNA and its hydration layer are processes important to the
radiation chemistry of DNA, a careful and systematic investigation of these processes has
been performed over the last several years in several laboratories. These investigations
included consideration of rate of transfer through the stacked bases and the role of the
hydration sheath, base sequence, temperature, inter duplex as well as intraduplex transfer ;
these efforts have helped to elucidate charge migration processes in DNA [21].
With regard to hole/electron transfer distances, these recent ESR studies clearly
show that after the initial migration and trapping, electron and hole transfer distances at 77 K
are limited to ca. 10 base pairs and that transfer occurs predominantly via tunneling [21].
However, at ambient temperatures, holes can actually range up to 100 base pairs via an
activated hopping mechanism, from Gua to Gua. A point of controversy in the literature
was the value of B(the tunneling decay constant), which governs the reduction in the rate
of electron transfer, k, as the tunneling distance, D, increases [k =koexp(-BD)]. Early work
had suggested values of B as low as 0.1 Å−1 ; these results are now known to be a result
of activated processes, “hopping” (Chapter 14). ESR work has clearly shown that electron
transfer, by tunneling, through DNA, occurs with a B of 0.7-0.9 Å−1 [21]). For inter-duplex
transfer, a higher beta value (near 1.3 Å−1) is suggested.
Ä&.-Ä
High LET irradiation studies
The effect of LET on a variety of radiation endpoints, including mutations, double

strand breaks, single strand breaks, transformations, cell inactivation, cell death, chromosome
aberrations and chromatin breaks have been extensively studied. However, for high LET
radiation, an understanding of the chemical processes that lead to strand breaks after the
direct deposition of energy in DNA (and its solvation shell) is in its infancy. Modern track
structure theory commonly posits two distinct zones of energy deposition for high LET ions.
In one, commonly referred to as a “core,” glancing collisions generate continuous ionizations
and excitations ; knock-on collisions, with small impact parameters and large depositions of
energy, also occur in the core. The second zone of energy deposition results from a cascade of
high energy electrons (D-rays) which are emitted at larger angles from the core ; these result
in a "halo" or "penumbra" consisting of widely spaced spurs (Chapter 4). The processes that
occur in this region are thought to mimic those found in low LET radiation, in which radiation
damage results largely from scattered high-energy electrons. The chemical processes that
occur in the penumbra are expected to be similar to those of low LET irradiation such as
gamma irradiation. However, the processes in the core are clearly of a different character
and one expects different mechanisms. Indeed, ion beam studies show that sugar radical
formation is greatly augmented at high LET (second row of Tab. 1). As mentioned above for
DNA samples subject to ion beam irradiation at 77 K, two immediate strand break radicals,
dRib(C3’–Phos)t and RO(O)2Pto, have been identified in surprisingly high yields compared
to low LET radiation such as G-radiation [8]. The high energy density along the track core of
the ion results in ionizations and excitations in close proximity creating reaction profiles not
found at low LETs. Thus, these radicals have been suggested to result from LEE electrons,
perhaps by interacting with an already vibrationally excited sugar ring. Lower levels of Guat
found in high LET samples (see Tab. 1) with correspondingly higher levels of sugar radicals
are in accord with the now well established mechanism that excitations of Gt directly result
in sugar radicals [17].
Acknowledgments
We appreciate the careful reading and helpful suggestions provided by David Becker. The authors
thank the National Cancer Institute of the NIH for support, grant R01-CA32546 to WAB and grant
R01-CA045424 to MDS.
References
[1] Becker D., Sevilla M.D., The chemical consequences of radiation damage to DNA, in “Advances in
Radiation Biology”, Lett J.T., Sinclair W.K. (eds), Academic Press, San Diego, 1993, p. 121-180.
Ä&..Ä
2!$)!4)/.#(%-)3429
[2] Ward J.F., The Complexity of DNA Damage: Relevance to Biological Consequences, Int. J. Radiat.
Biol., 1994, 66, 427-432.
[3] Bernhard W.A., Close D.M., DNA damage dictates the biological consequence of ionizing
radiation : the chemical pathways, in “Charged Particle and Photon Interactions with Matter”,
Hatano Y., Mozumder A. (eds), Marcel Dekker, New York, 2004, p. 431-470.
[4] Uehara S., Nikjoo H., Goodhead D.T., Comparison and Assessment of Electron Cross Sections for
Monte Carlo Track Structure Codes, Radiat. Res., 1999, 152, 202-213.
[5] Huels M.A., Handorf I., Illenberger D., Sanche L., Resonant dissociation of DNA bases by
subionziation electrons, J. Chem. Phys.,1998, 108, 1309-1312.
[6] Sevilla M.D., Becker D., ESR Studies of Radiation Damage to DNA and Related Biomolecules, Royal
Society of Chemistry Specialist Periodical Report: Electron Spin Resonance, 2004, 19, 243-278.
(See previous reviews referenced within.)
[7] Wang W., Yan M., Becker D., Sevilla M.D., The influence of hydration on the absolute yields of
primary free radicals in gamma-irradiated DNA at 77 K. II. Individual radical yields, Radiat.
Res.,1994, 137, 2-10.
[8] Becker D., Razskazovskii Y., Callaghan C., Sevilla M.D., ESR of DNA Irradiated with a Heavy Ion
Beam (16O): Evidence for Damage to the Deoxyribose Phosphate Backbone, Radiat. Res., 1996,
146, 361-368. (See references within.)
[9] a.-B7FSF5 #FDLFS% 4FWJMMB.% :JFMETPG0)tJO(BNNB*SSBEJBUFE%/"BTBGVODUJPOPG%/")ZESBUJPO
)PMF5SBOTGFSJO$PNQFUJUJPOXJUI0)t'PSNBUJPO 3BEJBU3FT 145, 673-680.
b. Debije M., Strickler M., Bernhard W.A., On the efficiency of hole and electron transfer from the
hydration layer to DNA: An EPR study of crystalline DNA X-irradiated at 4 K, Radiat. Res., 2000,
154, 163-170.
[10] Debije M.G., Bernhard W.A., Thermally stable sites for electron capture in directly ionized DNA:
free radicals produced by the net gain of hydrogen a C5/C6 of cytosine and thymine in crystalline
oligodeoxynucleotides. J. Phys. Chem. A, 2002, 106, 4608-4615.
[11] Colson A-O., Sevilla M., Application of molecular orbital theory to the elucidation of radical
processes induced by radiation damage to DNA, in “Theoretical and Computational Chemistry:
Computational Molecular Biology”, Lesczycyski J. (ed), 1999, 8, 245-277.
[12] Shukla L.I., Adhikarya A., Pazdro R., Becker D., Sevilla M.D., Formation of 8-oxo-7,8-dihydroguanine-
radicals in gamma-irradiated DNA by multiple one-electron oxidations, Nuc. Acid Res., 2004, 32,
6565-6574.
[13] Close D.M., Where are the sugar radicals in irradiated DNA?, Radiat. Res., 1997, 147, 663-673.
[14] Shukla L.I., Pazdro R., Becker D., Sevilla M.D., Sugar Radicals in DNA: Isolation of neutral radicals in
Gamma Irradiated DNA via hole and electron scavenging, Radiat. Res., 2005, 163, 59-602.
[15] Razskazovskiy Y., Debije M.G., Bernhard W.A., Direct Radiation Damage to Crystalline DNA: What
is the Source of Unaltered Base Release?, Radiat. Res., 2000, 153, 436-441.
[16] Purkayastha S., Milligan J.R., Bernhard W.A., Correlation of Free Radical Yields with Strand Break
Ä'%%Ä
Yields Produced in Plasmid DNA by the Direct Effect of Ionizing Radiation, J. Phys. Chem. B, 2005,
109, 16967-16973.
[17] Adhikary A., Malkhasian A.Y. S., Collins S., Koppen J., Becker D., Sevilla M.D., UVA-Visible photo-
excitation of guanine cation radicals produces sugar radicals in DNA and model structures, Nuc.
Acid Res., 2005, 33, 5553-5564.
[18] Roginskaya M., Bernhard W.A., Marion R.T., Razskazovskiy Y., The Release of 5-Methylene-2-
Furanone from Irradiated DNA Catalyzed by Cationic Polyamines and Divalent Metal Cations,
Radiat. Res., 2005, 163, 79-84.
[19] Huels M.A., Boudaiffa B., Cloutier P., Hunting D., Sanche L., Single, Double, and Multiple Double
Strand Breaks Induced in DNA by 3-100 eV Electrons, J. Am. Chem. Soc., 2003, 125, 4467-4477.
[20] Barrios R., Skurski P., Simons J., Mechanism for Damage to DNA by Low-Energy Electrons, J. Phys.
Chem. B, 2002, 106, 7991-7994.
[21] Cai Z., Sevilla M.D., Studies of excess electron and hole transfer in DNA at low temperatures, in
“Topics in Current Chemistry: Long Range Transfer in DNA, Vol II”, Shuster G. (ed), Springer-Verlag,
New York, 2004, 237, 103-128. Also see other Chapters in the volume, especially those by Douki
T., Cadets J., and Berlin Y.A. and Siebbeles L.D.A.
Ä'%&Ä
Chapter 14
Charge motion in DNA
Yuri A. BERLIN and Laurens D. A. SIEBBELES
Introduction
Double stranded deoxyribonucleic acid (DNA) represents a macromolecular

assembly consisting of two intertwined helices with an aromatic P–stack core, where
the bases of the pyrimidine deoxynucleotides (thymine, T ; cytosine, C) and purine
deoxynucleotides (adenine, A ; guanine, G) participate in Watson-Crick base pairing (A :T ;
C : G) (Chapters 12 and 13). This unique structure discovered more than 50 years ago defines
such biological properties of DNA as its capability for coding, storage and propagation
of genetic information. Therefore it is not surprising that for a long time, DNA has almost
exclusively been the province of biologists and life scientists.
During the past fifteen years, however, the situation has gradually changed mainly due
to the development of new methods for direct manipulation of single DNA molecules and for
their labeling with redoxactive probes. Using these experimental techniques, researchers were
able to expand information about molecular characteristics of DNA to include its mechanical,
electronic, and transport properties [1]. These properties and new experimental methodology
for their studies have attracted attention of scientists working in different fields that are often
far apart from biology and genetics. In particular, the ability of DNA to serve as a medium for
long-range charge transfer has stimulated interest in the possibility to exploit this molecule in
nanoscale electronics, molecular computing, and in electrochemical biosensoric devices. The
same property is shown to be important for developing new ways to detect structural changes
due to protein binding and base mismatches. For these potential applications, the elucidation
of mechanisms responsible for charge transport phenomena in DNA turns out to be crucial.
Ä'%(Ä
2!$)!4)/.#(%-)3429
This challenging problem is also vital for current research on oxidative damage of DNA, which
may cause apoptosis, mutations, and cancer.
A key structural element, which determines transport properties of DNA, is the array
of π-stacked base pairs. The striking resemblance of the base pair stack to conductive one-
dimensional aromatic crystals prompted the proposal that the interior of the double helix
can provide a one-dimensional pathway for charge migration due to formation of a P-band
across different stacked bases [2]. The validity of this early mechanistic picture was tested in
experiments on the photo-induced oxidation of DNA, dated back to the early 1990s (for review
see e.g. [3]). In typical experiments designed to study the transfer of an electronic hole (i.e. a
positive charge corresponding to an ionized nucleobase), a donor and an acceptor of these
charge carriers are intercalated in the stack of native base pairs or chemically attached to the
phosphate-sugar backbone. For properly chosen donor and acceptor species, such chemical
modifications enable one to generate holes under irradiation of the sample by light due to
the removal of electron from nucleobases to the photoexcited donor. The subsequent hole
transfer from the donor to the acceptor bridged by the base pair sequence can be probed
by measuring, for instance, the quenching of the fluorescence of the donor for sequences
of different lengths or the damage yield at certain sites along the sequence. Later, a similar
approach was also utilized in experiments on the photo-induced reduction of DNA aimed to
probe transport of excess electrons along the stack of base pairs [4]. In the latter case negative
charge carriers were generated using the photoexcited donor to inject electrons into DNA or to
transfer them directly to the acceptor. Some additional information about motion of electrons
and holes in the interior of the double helix has become available from the series of works on
low-temperature G-radiolysis of crystalline DNA, its ice and glassy aqueous solutions [5].
The experimental studies mentioned above have shown that the ordered π-electron
system of the natural DNA bases in duplex B-form DNA (here simply referred to as DNA) indeed
provides an appropriate pathway for the motion of excess positive and negative charges
once generated on extended and chemically well-defined stacks of base pairs. However the
observed dependence of the charge transfer efficiency on the base pair sequence [6] suggests
that the actual behavior of generated electrons and holes is more complicated than the band-
like picture of charge motion proposed at the initial stage of investigations [2].
Structural disorder in DNA and disordered energy landscape
Qualitatively the plausible scenario of the DNA-mediated charge transfer can be

inferred from the consideration of the relevant energy landscape. The latter shows how the
energy of a charge carrier changes as a hole or an excess electron is consecutively deposited
Ä'%)Ä
#HARGEMOTIONIN$.!
on each nucleobase involved in the formation of the π–pathway between a donor and an
acceptor. To construct the simplest landscape possible, one should take into account the
energetics of individual bases and the structural disorder arising from the choice of A, T, G,
or C at each substitution base site along the backbone of the DNA helix. These two factors
lead to the static energy disorder that determines the compound mechanism of the entire
process of charge transfer as well as mechanisms governing its elementary steps.
In the case of holes the static energy disorder arises due to the differences between
individual nucleobases in the values of the oxidation potential and the ionization energy.
To be more specific, G is known to be the most easily oxidized nucleobase since its in vitro
reduction potential is about 0.4 eV lower than that of A, and significantly lower than the
reduction potentials of C and T. The same conclusion follows from the results obtained
for energies of a positive charge located at different native nucleobases using various
computational methods of quantum chemistry [7-9]. Therefore the energy of a hole when
residing on A, C, or T sites is higher than on G. Hence the G base is a more probable place for
the positive charge to be localized than three others. This hierarchy of hole energies (G < A
< C,T) also holds when stacking interactions between neighboring nucleobases are taken
into account [9], although their ionization potentials become smaller in comparison with
the values found for individual bases. In particular, ionization potentials of GG doublets and
GGG triplets formed by stacking two or three adjacent Gs on the same strand are shown
to be less than the ionization potential of a single G base. As a consequence, the energy of
holes on GG and GGG sites is lower than the energy of G by several tenths of eV.
Thus, due to different energetics of nucleobases, structural disorder in the stack

of A:T and G:C pairs gives rise to static disorder in energies of holes residing on individual
bases. Furthermore, a close examination of the resulting energy landscape reveals three
main groups of states. The first group consists of states with the lowest energy. They serve as
deep hole traps arising when several adjacent Gs are stacked on the same strand and form
a multiple GG…G unit (Fig. 1). Hole states associated with individual G bases belong to the
second group, which is intermediate in energy between deep traps and holes residing on A,
T, and C bases. For this reason, these states will be defined henceforth as “intermediate”. Three
other native nucleobases A, T, and C are responsible for the formation of the third group
comprising of hole states with the highest energies. Since A, T, and C can be considered as
a building block of the bridge connecting two neighboring G bases, all states belonging to
the third group will be referred to as “bridging”.
Three groups of states discussed above are separated by two energy gaps. At
room temperature the width of both gaps exceeds a typical thermal energy Eth given by the
Ä'%*Ä
2!$)!4)/.#(%-)3429
product of the Boltzmann constant kB and temperature T. Therefore, in the simplest case
considered here the groups do not overlap and can be considered as isolated.
Figure 1 : A fragment of the DNA double helical structure (A), and energy landscapes for hole (B), and excess
electron motion (C) along the stack of base pairs.
The main elements of the energy landscape for positive charge transfer mentioned above
are shown in Figure 1B, using the fragment of the DNA duplex schematically depicted in
Figure 1A as an illustration. In this particular case, the doublet G7G8 is an example of the
deep hole trap, sites G1 and G3 exemplify intermediate states, while sites T2, T4, T5, and T6
correspond to “bridging” states.
Certainly the energy landscape of Figure 1B is oversimplified. In particular, energies

of holes on each of single G’s are assumed to be equal. Moreover, the hole energies for T2, T4,
T5 and T6 bridge units are also supposed to be identical for simplicity. Meanwhile, quantum
mechanical calculations [7-9] suggest that flanking bases can affect the hole energies on A,
T, G, and C bases due to stacking interactions, thus further increasing the degree of static
energy disorder within the DNA P-stack.
A similar structure of the energy landscape can be expected for the process of excess
electron transfer on the basis of the redox potential data available in the literature, see e.g. [10].
However, now the traps have to be pyrimidine deoxynucleobases stacked on the same strand
(e.g. TT dimers or TTT triplets), the intermediate states should be associated with the anions
T and/or C , while bridging states will correspond to G and/or A bases (see Fig. 1C). This
conclusion directly follows from the hierarchy of the measured reduction potentials, which
decrease in the order C z T >> A > G [10]. Electron affinities calculated both for individual
nucleobases and for their different trimers exhibit the same trend [11].
Ä'%+Ä
#HARGEMOTIONIN$.!
Mechanisms of charge transfer :

the entire process and elementary steps
The energy landscapes discussed suggest two distinct mechanisms of charge

transfer between the donor D and the acceptor Ac located at the opposite ends of the
nucleobase sequence [12,13].
If bridging (and for certain systems also intermediate) states in the landscape
are much higher in energy compared with D and Ac, a charge will be transferred via
superexchange mediated tunneling. Two characteristic features of this coherent quantum
mechanism should be mentioned in the context of electronic properties of DNA. First,
the two-center superexchange charge transfer occurs in a single step and hence does not
involve genuine chemical intermediates. Second, the rate of the whole process, kCT, rapidly
decreases with the donor-acceptor distance R following the familiar exponential law
kCT = k0 exp(−βR), (1)
where k0 is the pre-exponential factor and β is the falloff parameter. In the case of charge
transfer in DNA the β values are theoretically expected to be of the order of 1Å−1.
Another situation arises if intermediate states are comparable in energy with D. Now
a charge can be injected from the D site to the proximal base (G in the case of holes and T or
C in the case of electrons) with the subsequent temporal localization in the corresponding
intermediate state of the energy landscape. Thereafter, a hole or an electron is able either
to return back to D or to undergo a transition to the adjacent unoccupied intermediate
state through the intervening nucleobases associated with the bridging states (e.g. A or T
for holes and G or C for electrons). The latter transition represents the first step in a series of
consecutive incoherent hopping transitions that allows charge carriers to move along the
stack of base pairs using intermediate states as stepping stones. As a consequence, a charge
is able to reach a remote Ac site separated from D by several hundred angstro̊ms, where
hopping transport is terminated by trapping.
Thus, unlike the coherent single-step superexchange, incoherent hopping in DNA

involves several steps, i.e. injection of charge carrier, their transport along P-pathway due
to successive transitions between nucleobases with appropriate energetics, and trapping.
In addition, there are several other distinctions between hopping and superexchange
mechanisms. In particular, the former mechanism implies the formation of reactive chemical
intermediates (e.g. G t and pyrimidines anions for hole and electron transfer, respectively),
Ä'%,Ä
2!$)!4)/.#(%-)3429
which propagate from the site of their generation to the distant site of the reaction. By contrast,
the single-step superexchange does not include any intermediate active species. The distance
dependencies of the charge transfer rate for two mechanisms turn out to be also distinct. For
unbiased hopping on a long one-dimensional regular lattice with (N+1) sites separated by the
distance a, this dependence can be approximated by the algebraic function [14]
kCT d 1/Nz 1/(aR) (2)
rather than by the exponential law (1) typical for superexchange.
The above-mentioned features of the superexchange and hopping mechanisms

were documented in a number of charge transfer processes observed in DNA. As has been
demonstrated in a number of experiments, the single-step superexchange mechanism
dominates in DNA oligomers with short base pair sequences (R < 20 Å). Representative
examples of such systems and superexchange-driven reactions are given in Table 1. By
contrast, experiments with longer nucleobase sequences (R > 20 Å) reveal that the multi-
step hopping mechanism prevails. According to the current consensus, the latter mechanism
governs a number of processes in different duplexes listed in Table 2.
Table 1. Examples of DNA oligomers with short base pair sequences and superexchange driven elemen-
tary processes observed in these systems a).
Elementary
System Reaction Scheme
Process
Stilbene (S) Hole generation on
capped DNA 1
S*-(AT bridge)-G m S−-(AT bridge)-G
a single G base
hairpins
Charge
S−-(AT bridge)-G m S-(AT bridge)-G
recombination
Photo-induced hole
Acridine (Acr)
generation on Acr*-(AT bridge)-G m Acr−-(AT bridge)-G
modified DNA
a single G base
Hairpins
and double Hole trapping by
G-(short AT bridge)-(GGG) m G-(AT short bridge)-(GGG)
stranded short GGG triplets
oligonucleotides
Stilbenediether
(Sd) capped DNA Electron injection 1
Sd*-(GG bridge)-T m Sd-(GG bridge)-T−
hairpins
Pyrene(Py)- Photo-induced … − U − C − … m … − Ut¦ − C − … m … − U − Ct¦ − …
modified duplexes generation of | | |
containing uracil
bases (U) negative charges Py* Pyt Pyt
a) Asterisk denotes the excited state of the corresponding structural unit.
Ä'%-Ä
Table 2. Examples of DNA oligomers with long base pair sequences and processes govern by multi-step charge hopping in these systems a).
System Process Reaction Scheme

+
Hole migration from the site of G T C A G C T C A G T C T G C A (G G G) m
C A G T C G A G T C A G A C G T C C C
4’-modified oligonucleotides generation to the GGG trap via
+
with the GGG unit intrastrand and interstand GmG G T C A G C T C A G T C T G C A (G G G)
(“zigzagging”) transitions C A G T C G A G T C A G A C G T C C C
Migration of a hole selectively

generated at the G site to the triple
G TT G TT G … TT (GGG)m G TT G TT G … TT (GGG)
G trap along one strand of duplex
with TT bridges of equal length
AQ*-T(GG)T(GG)T(GG)T(GG)T(GG)T(GG)TATA
n
Anthraquinone Photo-induced propagation of AQ−-T(GG)T(GG)T(GG)T(GG)T(GG)T(GG)TATA
(AQ)-linked duplex DNA
radical cations n
Ä'%.Ä
oligomers …
n
AQ−-T(GG)T(GG)T(GG)T(GG)T(GG)T(GG)TATA
1
Stilbenedicarboxamide Sa*−AAGAGA(GGG)mSa−−AAGAGA(GGG)m…
Photo-induced charge separation
(Sa) capped DNA hairpins …mSa−−AAGAGA(GGG)
DNA oligomers with

internally conjugated Photo-induced excess electron
aromatic amine donor X*−(A :T and G:C pairs)−Y m X−( A:T and G:C pairs)−Y−
transfer
(X) and 5-bromo-2’-
deoxyuridine acceptor (Y)
DNA hairpins with flavin- Excess electron-transfer-based F*−(A:T and/or G : C pairs)−TT
capped donor (F) and the
repair of cis-sin TT dimer m F−(A : T and/or G : C pairs)−TT−
TT-dimer acceptor
a) Asterisk denotes the excited state of the corresponding structural unit.
#HARGEMOTIONIN$.!
2!$)!4)/.#(%-)3429
As has been emphasized in earlier publications [15-17], there is no dichotomy

between the two mechanisms of charge transfer in DNA considered above. On the contrary,
each can contribute to the mechanistic picture of the entire process. The superexchange
mediated tunneling controls the rate of short-range (< 20 Å) elementary hops of a
charge carrier between neighboring nucleobases that produce intermediate states in the
corresponding energy landscape, while multi-step hopping is responsible for the long-
range migration of charge along the stack of nucleobases.
Figure 2 : Efficiency of hole transfer from site-selectively generated Gt to the GGG triplet across A:T bridges
of various length RAT . Points correspond to the experimental data of Giese et al. [18]. The length dependence
calculated for the same system [16] is shown by the solid line. The intersection of the dotted line with the hori-
zontal axis gives the length of A:T bridge and the number of A:T pairs, at which the rates of quantum tunneling
and classical thermally-induced transitions become equal.
The superexchange mediated tunneling, however, is not the only plausible

mechanism of short-range steps of hopping motion in DNA: at finite temperatures this
mechanism is in competition with classical thermally-induced transitions of charge carriers
between two neighboring “resting” sites (G for holes, C and/or T for excess electrons). As can
be seen from equation (1) and Figure 2, the tunneling rate exponentially decreases with the
distance separating these two states, while the rate of the thermally-induced transition Wth
is mainly determined by a thermal population of the bridge. Therefore Wth does not vary
with distance, but depends on the energy gap between intermediate and bridging states,
Eib, in accordance with the Arrhenius law
Wth = W0 exp[-Eib/(kBT)], (3)
Ä'&%Ä
#HARGEMOTIONIN$.!
where W0 is the pre-exponential factor. Due to the distinction in distance dependencies, a

changeover from the superexchange mediated tunneling to the thermally activated regime
of the elementary hopping step can be expected as the distance between neighboring
intermediate states becomes equal to a certain critical value [16]. In the case of hole transfer,
the tight-binding model for the elementary step of hopping motion [16] suggests that a
positive charge can be transferred between two neighboring G sites via superexchange
only if the A :T bridge connecting these two sites consists of less than 3-4 base pairs. Since
the mean plane-to-plane distance between base pairs in B-DNA is known to be 3.4 Å, this
corresponds to the situation where the AT bridge has the length RAT b 14 Å. Otherwise (i.e. for
RAT > 14 Å), the elementary hopping step includes thermal activation of holes into the tight-
binding band followed by their ballistic or hopping motion along the A :T bridge. For holes,
the latter process is known in the literature as A-hopping. Recent experiments (for review, see
[18]) provide strong evidence for such thermally-induced transitions through long bridges
with the number of A :T pairs nAT = 4-10 (RAT z 17-37 Å) and nAT = 4-16 (RAT z 17-58 Å). Moreover,
measurements [18] of the hole transfer efficiency as a function of RAT for the process

G ¥ T ´ (G G G) G ¥ T ´ (G G G)
m (nAT = 1-16) (3)
C ¦§ A µ¶ n C C C C ¦§ A µ¶ n C C C
AT AT
and the analogous theoretical dependence [16] were found to be in good agreement (see
Fig. 2), thus supporting theoretical predictions concerning two competing mechanisms of
elementary hopping step.
Thus, a plausible scenario for the entire process of charge transfer from D to Ac
along the sequence of nucleobases involves variable-range hopping between intermediate
states corresponding to the bases with the appropriate oxidation or reduction potentials.
Short steps made by a moving charge in this multi-step transport process occur due to
the coherent superexchange mediated tunneling. By contrast, long steps require thermal
activation of charge carriers needed to overcome the energy gap between intermediate and
bridging states. Once this thermally-induced transition has completed, electrons or holes
can reach the next “resting” site undergoing ballistic motion or hopping along the pathway
provided by bridging states.
As follows from the detailed kinetic analysis of this scenario [13,15], the model of
variable-range hopping allows quite accurate predictions of both sequence and distance
dependencies for the efficiency of charge transfer through stacks with various combinations
of base pairs. This, in turn, provides reasonable estimations of the distance scale for the
Ä'&&Ä
2!$)!4)/.#(%-)3429
propagation of charge in DNA duplexes. Based on these estimations verified by steady-state

experiments (see e.g. [19] and references therein) one can conclude that typically the upper
limit for the distance traveled by charges in DNA is about 200-300 Å. Charge transfer over
such large distances can be accomplished because of the weak distance dependence of the
reaction rate (cf. equation (2)) offered by variable range hopping. In addition, a coexistence
of quantum and classical steps of hopping process is also favorable to the long-range
migration of electrons and holes in the interior of double helix.
How much time a charge needs to move through DNA
Although steady-state experiments have yielded a wealth of knowledge about the

mechanism of charge transfer in DNA, the results obtained with this type of measurements are
insufficient to decide how fast a charge can be transferred over a certain distance. Absolute
values of the charge transfer rate can be obtained from time-resolved pump-probe laser
experiments, such as those performed by Lewis and coworkers [20]. These experiments were
carried out on DNA sequences containing a stilbenedicarboxamide (Sa) electron acceptor.
Figure 3 shows a few examples of the DNA sequences studied.
Figure 3 : Examples of DNA sequences containing a stilbenedicarboxamide electron acceptor Sa. The rates
of charge transfer between the primary G (green) and the distal GG doublet (red) strongly depend on the inter-
vening base sequence as explained in the text. For all sequences, hole transitions between G and GG are shown
by arrows. Numbers next to the arrows are the values of rate constants for the hole transfer between G and GG.
Corresponding average times for this process are given in parentheses.
In the experiments, the Sa molecular unit is photoexcited using a laser pulse.

Upon photoexcitation Sa accepts an electron from DNA. This initially leads to the
formation of the anion radical Sato and a hole on the G site nearest to Sa (colored green
in Fig. 3). The hole generated on this proximal G is able either to recombine with Sato or
Ä'&'Ä
#HARGEMOTIONIN$.!
to jump to a distal GG doublet (colored red in Fig. 3). The latter process leads to the
formation of (GG)t, but does not terminate the motion of positive charges, since in the
case under consideration a hole can be back transferred to the primary G site. The decay
kinetics of the transient of the measured transient optical absorption of the anion radical
Sato depends on the rate of recombination of Sato with a hole on the primary G, and on
the rates for reversible charge transfer between the primary G and the distal GG doublet.
The absolute values of these rates could be obtained from a theoretical analysis of the
measured decay kinetics of the transient optical absorption due to Sato anions. The values
obtained for the rate constants for hole transfer from the primary G to the distal GG are
6 s 107 s−1, 0.33 s 107 s−1 and 0.05 s 107 s−1 for sequences a, b and c, respectively. The inverse
values of these rates correspond to the average charge transfer times, which are equal to
17 ns, 300 ns and 2000 ns for sequences a, b and c, respectively. These very different values
show that the process of charge transfer is strongly sequence dependent. Comparison of
the rate constants for hole transfer in sequences a and b demonstrates that hole transfer
between the proximal G site and the distal GG doublet via an intervening adenine (A) is
almost twenty times faster than via thymine (T). The rate constant for hole transfer decreases
more than six fold upon introduction of an additional A :T base pair (sequences b and c).
In order to provide theoretical insight into the factors governing the rate of charge
transfer, quantum chemical calculations have been carried out, which are described in detail
in [8]. As discussed below, a qualitative explanation of the relative rates in the different DNA
sequences can be obtained from a Hückel (tight-binding) type model. For a quantitative
description of the charge transfer rates, it is necessary to include the effects of the charge-induced
structural reorganization of DNA, and the aqueous environment used in the experiments.
In the Hückel type calculations the hole is described quantum mechanically by a

wave function, which is a time-dependent linear combination of orbitals on the individual
nucleobases. Charge transfer is affected by the structural conformation of the nucleobases
in the DNA sequence. In particular, twisting of base pairs around the axis of the DNA helix
strongly affects the rate of charge transfer. Therefore effects of twisting were included in the
calculations. Initially the wave function is localized on the proximal G site. The time evolution
of the wave function is then obtained by solving numerically the Schrödinger equation. In the
calculations the charge is forced to decay at the distal GG doublet. The rate of charge transfer
from the proximal G site to the distal GG doublet can be obtained from the probability for
the hole to survive trapping at the GG doublet.
Figure 4 shows the calculated decay of population of charge on the initial G site,
due to charge transfer from this site to the distal GG doublet, for the three different
Ä'&(Ä
2!$)!4)/.#(%-)3429
DNA sequences in Figure 3. Hole transfer is seen to be fastest for DNA sequence a and
becomes slower on going to sequences b and c. This trends in the kinetics of hole transfer is
consistent with the results deduced from the experiments discussed above. The calculated
energy of a hole localized on the proximal G site in sequence a is 0.3 eV lower than it would
be if it were localized at the A nucleobase between the proximal G and distal GG doublet. This
energetic difference is much larger than thermal energy at room temperature (kBT ~ 0.025 eV).
Hence, the hole migrates from the G site to the GG doublet by quantum mechanical
tunneling through the barrier provided by the intervening A nucleobase. The height of the
barrier due to the T nucleobase intervening the G site and the GG doublet in sequence b
is equal to 0.8 eV. The much higher energetic barrier for tunneling via a T nucleobase as
compared with A causes the charge transfer rate for sequence b to be much smaller than for
sequence a. The reduction of the charge transfer rate on going from a barrier consisting of a
single A nucleobase in sequence a to the barrier of two adjacent adenines in sequence c is a
consequence of the increased length of the barrier in the latter case.
Figure 4 : Probability for the hole to survive trapping by the GG doublet in the DNA sequences a, b, and c
shown in Figure 3.
The Hückel type calculations offer a qualitative explanation of the trend observed
for the experimental rates of hole transfer between the proximal G and the distal GG doublet.
However, as can be seen in Figure 4, the calculated timescales for charge transfer range from
a few picoseconds for sequence a to several hundreds of picoseconds for sequence c. These
times are about three orders of magnitude smaller than the average charge transfer times
obtained from the experiments. The difference between the calculated and experimental
timescales for charge transfer is due to charge-induced structural deformations, which are
not included in the Hückel type calculations. An excess charge in DNA induces an internal
Ä'&)Ä
#HARGEMOTIONIN$.!
structural reorganization of the nucleobases, which reflects the fact that the equilibrium
structure of a cation differs from that of a neutral molecule. An additional effect is due to
the fact that the experiments involve DNA in aqueous solution. The charge in DNA causes
the surrounding dipolar water molecules to reorganize and to adopt an energetically
favorable orientation around the charge. The effects of structural reorganization were taken
into account by using a semiclassical superexchange model based on the Marcus theory
for electron transfer. The absolute values of the experimental rates for charge transfer in
the sequences in Figure 3 could be reproduced by invoking a reorganization energy with a
magnitude about 1 eV, which is close to literature data [21,22].
The information on the kinetics and mechanism of charge transfer obtained from
the combination of the experimental and theoretical studies discussed above can be used
to obtain insight into the mobility of holes in DNA. This information is important in order to
understand the conducting properties of DNA. The results can be utilized to predict oxidative
cleavage patterns in long sequences of DNA with several multiple guanine-containing sites.
The results of experimental studies on the conductance of DNA are still highly controversial,
and a large variety of possible electronic behavior has been suggested, ranging from
DNA as an insulator to a superconductor (see the review in [19] and references therein).
The experimental studies usually focus on current-voltage dependence measurements.
The charge carrier mobility (i.e. the velocity of the charge per unit strength of an external
electric field) cannot be deduced from these measurements since the charge carrier density
is unknown. To the knowledge of the authors, the only study, from which a value of the
mobility for hole hopping along stacks of A:T base pairs can be deduced is the work of
Takada et al. [23], who determined the rate for A-hopping between adjacent nucleobase
pairs to be 2 s 1010 s−1. The mobility of holes undergoing hopping motion between A bases
is then calculated to be 9 s 10−4 cm2 V−1s−1.
The mobility of a charge moving along a stack of either G:C or A:T base pairs with the
Gs or As on the same strand of the double helix is of particular interest, since for these stacks
charge motion can occur by hopping between identical and adjacent bases. In this case,
the mobility will be much higher than for a DNA consisting of both G:C and A:T base pairs,
since in the latter case charge transfer occurs by relatively slow tunneling steps between
G nucleobases that are separated by A:T base pairs. Using the results from the quantum
chemical calculations discussed above the mobility of a hole along a stack of either G:C or
A:T base pairs can be calculated using Marcus type (polaronic) hopping theory for charge
transfer. For a reorganization energy equal to 1 eV the mobility of holes is calculated to be
10−4 cm2 V−1s−1 and 2 s 10−5 cm2 V−1s−1 for stacks of G:C and A :T base pairs, respectively.
A slightly different value of the reorganization energy equal to 0.63 eV must be used to
Ä'&*Ä
2!$)!4)/.#(%-)3429
reproduce the mobility of 9 10−4 cm2 V−1s−1 for holes on a stack of A :T base pairs, as deduced
from the experiments by Takada et al. [23] More accurate calculations of the mobility require
knowledge of the reorganization energy for the specific sequences studied in experiments
on the conductance of DNA. Nevertheless, the estimated mobility values given above are
more than six orders of magnitude smaller than the relative high mobility values of charges
in conventional semi-conductors such as crystalline silicon. However, the estimated mobility
values for stacks of G:C or A:T base pairs are comparable to that found for structurally
disordered polymers in opto-electronic devices. Although the mobility of charges in DNA is
strongly limited by the relatively large reorganization energy, DNA may find application in
future electronic devices based on organic molecules.
Epilogue
Quantum mechanical calculations of the energy landscape, together with

theoretical modeling of charge dynamics, offer the detailed mechanistic picture and
quantitative kinetic information on charge transfer in DNA over distances as long as 50-300 Å.
Biological implications of the obtained results pertain to the generation of DNA damage
induced by radiation and by chemical reactions. In both cases, specific radical reactions
with nucleobases occur, which finally lead to the formation of oxidized guanine followed by
cation (hole) migration. For such situations, the present theoretical approach based on the
multi-step hopping kinetic scheme with the individual hopping rates being inferred from
quantum mechanical calculation is particularly relevant.
Due to the existence of the fundamental relation between conducting and charge
transfer properties of the same nanostructured systems [14], the same kinetic-quantum
approach helps to explain electrical conductance through DNA placed in the nanogap
between two electrodes, and is therefore relevant to DNA nanoelectronics. Similar to other
molecular wires, the current through the DNA molecule connecting two electrodes is
determined by two processes. These include the injection of charge carriers onto the stack
of base pairs and their transport along the stack. Our theoretical studies focused on the
transport aspect of the problem provide information on the key characteristics of charge
carriers in DNA, namely their drift mobility, and enable us to establish main factors affecting
the value of this important quantity. We show that the drift mobility of holes and, hence, the
DNA electrical conductance strongly depend on the reorganization of nucleobases and the
water surroundings, on static and dynamic disorders in the stack of base pairs as well as on
the base pair sequence. Certainly, the application of DNA in molecular electronic devices
must take into consideration these and some other factors that promote or inhibit DNA-
mediated charge transport.
Ä'&+Ä
#HARGEMOTIONIN$.!
The sequence-dependence of DNA conductance predicted by theory and verified

experimentally is probably the most important outcome of our investigations, thus making
them relevant for a number of electrochemical DNA-based technologies. The latter ones
address primarily such fields as biosensors in which DNA is used as a conducting spacer,
molecular recognition and sequencing methods that rely on interfacing of DNA to
macroscopic electrodes. For all these applications, the kinetic-quantum approach serves
as a useful and quite general theoretical background. More importantly, this approach
allows us to demonstrate the high sensitivity of DNA conductance to mistakes in base
pairing (the so-called mismatches). This lays the groundwork for new highly reliable, fast
and accurate method for the detection of mismatches in a single DNA molecule that relies
on the measurement of molecular electric conductance and does not require such time-
consuming procedures as labeling and target amplification.
Acknowledgements
The Netherlands Organization for Scientific Research (NWO) is acknowledged for financial
support. Yuri Berlin is grateful to the Chemistry Division of the ONR, the NASA URETI Program, to DoD/
MURI and DURINT Programs for their support of the research. We also wish to thank many colleagues,
particularly A.L. Burin, E.M. Conwell, and J. Jortner for useful discussions.
References
[1] See collection of articles published in special section “Double helix” of Nature, 2003, 421, 396-453.
[2] Eley D.D., Spivey D.I., Semiconductivity of Organic Substances. Part. 9 – Nucleic Acid in the Dry
State, Trans. Faraday Soc., 1962, 58, 411-415.
[3] Grinstaff M.W., How do charges travel through DNA? – An update on a current debate, Angew.
Chem. Int. Ed., 1999, 38, 3629-3635.
[4] Wagenknecht H.-A., Reductive electron transfer and transport of excess electrons in DNA, Angew.
Chem. Int. Ed., 2003, 42, 2454-2460.
[5] Cai Z., Li X., Sevilla M.D., Excess electron transfer in DNA: Effect of base sequence and proton
transfer, J. Phys. Chem. B, 2002, 106, 2755-2762.
[6] Meggers E., Michel-Beyerle M.E., Giese B., Sequence dependent long range hole transport in
DNA, J. Am. Chem. Soc., 1998, 120, 12950-12955.
[7] Senthilkumar K., Grozema F.C., Guerra C.F., Bickelhaupt F.M., Siebbeles L.D.A., Mapping the sites
for selective oxidation of guanines in DNA, J. Am. Chem. Soc., 2003, 125, 13658-13659.
[8] Senthilkumar K., Grozema F.C., Guerra C.F., Bickelhaupt F.M., Lewis F.D., Berlin Y.A., Ratner M.A.,
Siebbeles L.D.A., Absolute rates of hole transfer in DNA, J. Am. Chem. Soc., 2005, 127, 14894-
14903.
Ä'&,Ä
2!$)!4)/.#(%-)3429
[9] Voityuk A.A., Jortner J., Bixon M., Rösch N., Energetics of hole transfer in DNA, Chem. Phys. Lett.,
2000, 324, 430-434.
[10] Steenken S., Telo J.P., Novais H.M., Candeias L.P., One-electron-reduction potentials of pyrimidine
bases, nucleosides, and nucleotides in aqueous solutions – consequences for DNA redox
chemistry, J. Am. Chem. Soc., 1992, 114, 4701-4709.
[11] Voityuk A.A., Michel-Beyerle M.E., Rösch N., Energetics of excess electron transfer in DNA, Chem.
Phys. Lett., 2001, 342, 231-238.
[12] Bixon M., Giese B., Wessely S., Langenbacher T., Michel-Beyerle M.E., Jortner J., Long-range charge
hopping in DNA, Proc. Natl. Acad. Sci. USA, 1999, 96, 11713-11716.
[13] Berlin Y.A., Burin A.L., Ratner M.A., On the long-range transfer in DNA, J. Phys. Chem. A, 2000, 104,
443-445.
[14] Berlin Y.A., Ratner M.A., Intra-Molecular electron transfer and electric conductance via sequential
hopping: Unified theoretical description., Radiat. Phys. Chem., 2005, 74, 124-131.
[15] Berlin Y.A., Burin A.L., Ratner M.A., Charge hopping in DNA, J. Am. Chem. Soc., 2001, 123, 260-
268.
[16] Berlin Y.A., Burin A.L., Ratner M.A., Elementary steps for charge transport in DNA: Thermal
activation vs. tunneling, Chem. Phys., 2002, 275, 61-74.
[17] Berlin Y.A., Beratan D., Kurnikov I.V., Ratner M.A., Burin A.L, DNA electron transfer processes: Some
theoretical notions, Top Curr. Chem., 2004, 237, 1-36.
[18] Giese B., Long-distance electron transfer through DNA, Annu. Rev. Biochem., 2002, 71, 51-70.
[19] Treadway C.R., Hill M.G., Barton J.K., Charge transport through a molecular π-stack: double helical
DNA, Chem. Phys., 2002, 281, 409-428.
[20] Lewis F.D., Letsinger R.L., Wasielewski M.R., Dynamics of photoinduced charge transfer and hole
transport in synthetic DNA hairpins, Acc. Chem. Res., 2001, 34, 159-170.
[21] Siriwong K., Voityuk A.A., Newton M.D., Rösch N., Estimate of the reorganization energy for charge
transfer in DNA, J. Phys. Chem. B, 2003, 107, 2595-2601.
[22] LeBard D.N., Lilichenko M., Matyushov D.V., Berlin Y.A., Ratner M.A., Solvent reorganization energy
of charge transfer in DNA hairpins, J. Phys. Chem. B, 2003, 107, 14509-14520.
[23] Takada T., Kawai K., Cai X., Sugimoto A., Fujitsuka M., Majima T., Charge separation in DNA via
consecutive adenine hopping, J. Am. Chem. Soc., 2004, 126, 1125-1129.
Ä'&-Ä
Chapter 15
Genome maintenance mechanisms
in response to radiation-induced
DNA damage
Evelyne SAGE and Bertrand CASTAING
Introduction
Maintenance of DNA integrity (the molecular support of genome) is essential to

minimize heritable mutations and to promote healthy survival of organisms. Indeed, loss
of genome integrity has long been implicated in genetic disorders, ageing and cancer. The
genome is under constant threat arising from environmental agents (such as solar ultraviolet
radiation, ionizing radiation and chemicals), from endogenous metabolic byproducts (such
as reactive oxygen species) or from replication errors and arrests, that can alter its physical
and chemical structure and corrupt its encoded message. A plethora of damages arises
in DNA and their consequences are diverse and generally adverse. Many lesions block
transcription, thus inactivating genes. They also interfere with DNA replication, triggering
cell division and cell death. When left unrepaired or repaired inaccurately, DNA lesions result
in irreversible mutations or genetic rearrangements, contributing to carcinogenesis as long-
term effects in higher organisms. So, DNA repair systems appeared early in life (they exist in
Archae bacteria) and have been maintained during evolution. In addition, stability of DNA is
so important that cells evolved a whole strategy in response to DNA damage, as summarized
in Figure 1 [1]. Damage on DNA constitutes a signal which is transduced by a cascade of
events (protein phosphorylations) to protein effectors that mediate cell cycle arrests in order
to give time to the cells for repairing their DNA. Indeed, the activation of the thus-called
DNA integrity checkpoints prevents cell cycle progression when DNA is damaged, and limits
Ä'&.Ä
2!$)!4)/.#(%-)3429
the fixation of mutations and their transmission to the offspring. Inability to properly face
a genotoxic threat leads to genomic instability, and eventually to tumoral transformation.
In multicellular organisms, a cell which is too heavily damaged induces its own death by
apoptosis, thus ensuring a protection for the organism.
Figure 1 : Cellular response to DNA damage. Sensor proteins are able to detect the presence of DNA damage
and alert the cellular machinery so that the cell can face the stress and protects itself. The sensor then activates
a transducer and through a series of proteins activation by phosphorylations, DNA integrity checkpoints (red ar-
row), in charge of the cell cycle control are activated and play their role which is to stop the cell cycle progression
in order to allow repair to fully operate. Indeed, DNA replication should not occur on damaged DNA because of
possible mistakes when copying modified bases, so the G1/S interphase checkpoint prevents the cell to enter into
S-phase and to start replication. The G2 checkpoint verifies DNA for the presence of double strand breaks and
incomplete replication, so that an altered DNA copy is not transmitted to the daughter cell.
A wide diversity of lesions arises in DNA. For instance, over hundred oxidative
modifications produced by reactive oxygen species (ROS) have been identified. Also,
ionizing radiations produce over hundred of different DNA modifications (Chapters 12, 13,
14). Meanwhile, DNA lesions have been classified in two main groups, base modifications
and strand breaks. Base modifications comprise addition of or substitution by bulky
residues forming bulky adducts (such as cyclobutane pyrimidine dimers produced by
ultraviolet radiation, or benzo(a)pyren adducts), as well as base alkylation, or base oxidation
and reduction produced by ROS or ionizing radiation. Strand breaks comprise single and
double strand breaks, typically produced by ionizing radiation and radiomimetics. Most
of sugar modifications generate single strand breaks. Inter- and intra-strand crosslinks
implicate covalent bounds between the crosslinking agent and DNA bases ; they enter into
the category of base damage. DNA-proteins crosslinks induced, for example, by ionizing
radiation are yet poorly understood. Notably, one agent does produce neither a unique DNA
Ä''%Ä
$.!DAMAGEREPAIR
lesion, nor a unique class of lesions. Typically, ionizing radiation induces predominantly base
damage, i.e. oxidized and reduced bases and abasic sites, single and double strand breaks, as
well as clustered lesions (Fig. 2). Clustered lesions, also called Multiply Damaged Sites, result
from heterogeneous energy deposition by ionizing radiation in DNA. They consist of two or
more closely spaced oxidized bases, abasic sites and single strand breaks distributed on both
strands within 1-2 helical turns (Chapter 12) [2-3]. They are thought to largely contribute to
the deleterious effect of ionizing radiation, since they may be difficult to repair [4-5].
Figure 2 : DNA damage induced by ionizing radiation. A) DNA damage and repair. All the constitutive elements
of DNA (sugar-phosphate backbone and bases) are possibly modified by ionizing radiation. Single strand breaks (SSB),
oxidized bases and abasic site are processed by base excision repair (BER), double strand breaks (DSB) by homologous
recombination and non homologous end joining (HR and NHEJ) and DNA-protein crosslinks by nucleotide excision
repair (NER). B) Quantitative measurement of radiation-induced and spontaneous DNA damage.
DNA repair processes have to face quite a diversity of chemical and physical
alterations of DNA. No single repair process can cope with all kinds of damage, but several
DNA repair systems have evolved, adapted to the class of lesions to be eliminated. Excision
repair takes care of base damage (Fig. 2) ; it is divided into several sub-pathways, including
nucleotide excision repair (NER) and base excision repair (BER), depending on the type of base
modification. Double strand breaks are mainly eliminated by two repair modes, homologous
recombination (HR) and non-homologous end-joining (NHEJ). We will further focus on the
description of pathways involved in the processing of radiation-induced DNA damage.
Ä''&Ä
2!$)!4)/.#(%-)3429
Excision repair pathways for removing base damage
Most base modifications in the cellular genome are cleared by base excision repair
(BER) or nucleotide excision repair (NER) pathways [6-7]. The first step in these processes is the
recognition of the damage by proteins which have to detect one base damage among 104 to
106 normal base pairs. Then, the excision step allows the removal of a short single stranded
DNA fragment that contains the damage. Then DNA repair synthesis by a DNA polymerase and
its accessory proteins which fills the gap is carried out, and then, ligation of the repaired strand
which restores DNA continuity (Fig. 3). Those are error-free repair processes, meaning that an
original and intact DNA, on the chemical, structural and functional point of view, is restored.
The type of base modification will determine which protein will come first and
which system, BER or NER, will be used to cure the damaged DNA. NER is responsible for
the removal of bulky adducts such as dimerized pyrimidine bases produced by UV radiation
from sunlight, or benzo(a)pyren or psoralen adducts and others (Tab. 1). Bulky adducts are
Table 1. Radiation-induced DNA damage recognition proteins in human cells.
Repair protein Function/Substrates*
Base excision repair

- Uracil DNA glycosylase Ung - glycosylase for U
- Uracil DNA glycosylase Smug1 - glycosylase for U, 5-hmU, 5-fU
- 8-oxoguanine glycosylase Ogg1 - glycosylase for 8-oxoG and FapyG opposite C
- Myh - glycosylase for A opposite 8-oxoG
- Nth1 - glycosylase for thymine glycol, FapyG, 5-OHU, 5- OHC
- glycosylase for 5-OHU, 5-OHC, DHT, DHU, thymine glycol, 8-ox
- Neil1
oG, FapyG, FapyA
- Neil2 - glycosylase for 5-OHU, 5-OHC
- Ape1/Hap1 - hydrolytic AP endonuclease for abasic sites
- PARP1 - Poly(ADP-ribose)polymerase recognizes single strand breaks
Nucleotide excision repair
- damage induced by UV, benzo(a)pyren, cisplatin, psoralen,
- XPC-HR23B
acetylaminofluoren
- XPE - damage induced by UV, cisplatin, MNNG, abasic sites
Double strand break repair
- Ku70/80 - double strand break ends
*U, uracil; 5-hmU, 5-hydroxymethyluracil; 5-fU, 5-formyluracil; 5-OHU, 5-hydroxyuracile; 5-OHC, 5-hydroxycytosine; DHT, dihydrothymine;
DHU, dihydrouracil; 8-oxoG, 8-oxoguanine; FapyG, FapyA, formamidopyrimidine from guanine or adenine; AP or abasic site, apurinic and
apyrimidinic site; MNNG, N-methy, N’-nitro, N-nitrosoguanidine.
Ä'''Ä
$.!DAMAGEREPAIR
Figure 3 : Excision repair pathways for removing base damage. The first step in base excision repair (BER) and
nucleotide excision repair (NER) is the recognition of the damage (red triangle) by two protein complexes (NER)
or a damage specific N-glycosylase (BER). The second step is the incision and excision of damage. In NER, the
incision of the DNA phosphodiester backbone on both sides of the bulky adduct is followed by the removal of a
short single stranded DNA fragment, 29-32 nucleotides long, that contains the damage. In BER, the N-glycosy-
lase excises the damaged base, the DNA phosphodiester backbone is then incised and the ends are processed,
leaving a gap of 1-7 nucleotides. The gap has to be filled by DNA polymerases in conjunction with accessory pro-
teins: repair patch synthesis. The last step performed by a ligase, restores the DNA continuity. Different proteins
are involved in the two last steps, repair replication and ligation, regarding BER and NER.
Ä''(Ä
2!$)!4)/.#(%-)3429
recognized by two protein complexes, which sense the distortion of the double helix in-
duced by the adduct, rather than the adduct itself. In contrast, relatively minor modifications
to the bases of DNA are removed by BER. They are alkylated bases, oxidized and reduced
bases, abasic sites, which do not drastically disturb the DNA double helix structure (Tab. 1).
Interestingly, NER pathway may serve as back-up when BER is not available.
In regard with NER, at least 32 proteins, forming several complexes, are necessary
and sufficient in a successful repair of a bulky lesion. In the case of BER, 5-6 proteins are
necessary to reconstitute the entire process on a single base damage [7]. For both systems,
the action of the proteins or complexes is coordinated and protein interactions allow passing
from one step to the next one. Thus, free repair intermediates which could be entry sites for
nucleases that would degrade DNA, are avoided.
Base excision repair
Ionizing radiation induces the formation of base damage, i.e. oxidized and reduced
bases and abasic sites (Chapter 12), which are repaired by BER. Single strand breaks (SSB)
which are produced by ionizing radiation but also as repair intermediates at the early steps
of BER (Fig. 3), are processed by BER.
As seen from Table 1 and in contrast to the DNA damage proteins involved in NER
which recognize a broad variety of substrates, in BER the damage recognition proteins have
a rather narrow substrate specificity. In consequence, a defined number of different proteins
exist to deal with the numerous small modifications to bases. Those proteins are DNA
glycosylases, which selectively bind with a high affinity to an altered base and hydrolyse
the N-glycosylic bond to release the altered base from the sugar-phosphate backbone. The
resultant abasic deoxyribose (apurinic, and apyrimidinic (AP) site) is processed further by an
AP lyase or an AP endonuclease, which cleaves the phosphodiester backbone. The single
base gap, single strand nick or flap generated as a repair intermediate, has to carry a 3’OH
terminus in order to allow a DNA polymerase to fill the patch (1 up to 7 nucleotides long).
The DNA polymerase adds nucleotides to this extremity using the undamaged strand as
template and a ligase reseals the DNA molecule.
SSB induced by ionizing radiation typically possess “damaged” termini, in other

words they lack the conventional 5’-phosphate and 3’-hydroxyl moieties that could easily be
religated by a ligase. Instead, the 3’ termini possess monophosphate or phosphoglycolate
end groups, and 5’ termini may carry a hydroxyl group. The unligatable (dirty) 3’ termini are
processed by a second activity of the AP endonuclease APE1. The polynucleotide kinase
Ä'')Ä
$.!DAMAGEREPAIR
PNK phosphorylates the 5’-OH ends. The specialized BER polymerase and the ligase can
then proceed. Anyhow, an early step in the repair of SSB appears to be the rapid binding
by poly(ADP-ribose) polymerase-1 (PARP1), a nick sensor that is activated at SSB. PARP1
synthesizes negatively charged polyADP-ribose that may serve as scaffold for the binding of
the repair proteins involved later in the process [8].
BER has been poorly investigated in radiobiology, although it is a major repair

process for ionizing radiation-induced DNA damage. It is also the main guardian against
DNA damage due to cellular metabolism that results in ROS, methylation, deamination,
hydroxylation.
Double strand breaks repair by homologous recombination and

non-homologous end joining
For simplicity, we will describe double strand breaks (DSB) repair in mammalian cells
[7]. When a DSB occurs, if the ends are ligatable, it will quickly be repaired at low cost by non-
homologous end joining pathway (Fig. 4). The first step is the recognition of the DSB by the
proteins Ku70/80 which bind at the ends and thus protect DNA from degradation by nucleases.
Next, the cell has to maintain the two ends of the broken molecule in a physical proximity to
avoid the loss of genetic information. This is realized by protein-protein interactions, DNA-PKcs
is recruited and interacts with the complex Ku70/80. Then, ligation may occur by the complex
XRCC4/Ligase4. Most of DSB created by ionizing radiation have unligatable (dirty) ends. The
cell must then "clean up" the end to generate 3'-hydroxyl groups and 5'-phosphate groups.
The repair process is more complicated, requires other proteins and it will take more time.
So, after recognition of the DSB by Ku70/80, a signal is transduced to the cell-cycle apparatus
to delay the progression through the cell cycle and up-regulate the transcription of the DNA
repair genes (see Introduction). The complex rad50/Mre11/Nbs1 is recruited as well as Artemis
protein which cleans and resects the ends. The core end joining proteins proceed to end
joining. Unfortunately, this process may be associated with gain or loss of a few nucleotides
and this is why non-homologous end joining pathway is considered as an error-prone repair
process. Non-homologous end joining is the major DSB repair pathway used by mammalian
cells, and the only one operating when cells are in G1 phase.
If a DSB is produced when DNA is replicating (S-phase) or in G2 phase when cells carry
two gene copies on a chromosome (Fig. 1), it will be repaired by homologous recombination.
This repair mechanism involves a strand exchange between the damaged DNA duplex and
the same sequence from the intact homologous double stranded DNA molecule. Firstly, the
complex rad50/Mre11/Nbs1 is recruited and resects the ends. The 3’ single stranded end
Ä''*Ä
2!$)!4)/.#(%-)3429
Figure 4 : Double strand breaks repair by non-homologous end joining (NHEJ) and homologous recombi-
nation (HR).
Ä''+Ä
$.!DAMAGEREPAIR
thus created are then coated with rad51 protein forming a filament with the help of other
proteins. A correct positioning of the duplexes allow invasion of the homologous sequence
by the coated single stranded end. The intact double stranded copy is used as template
to properly heal the broken ends by DNA synthesis. The cross-junctions are resolved and
ligation occurs to restore DNA integrity. The whole transaction is orchestrated by a number
of known proteins (Fig. 4) leading to an accurately repaired DNA.
Mutagenesis
In the absence of repair, a transient DNA damage event may be transformed into
irreversible, heritable mutational event. Left unrepaired, base damage may lead to mutation
during DNA replication when the replicative DNA polymerase inserts the incorrect base
opposite the lesion. For base damage which blocks the replication machinery, specialized
DNA polymerases evolved to bypass the blocking lesion and performed the so-called
translesion synthesis (TLS), thus overcome the replication blockage, allowing cell to pursue
its cycle [7]. Unfortunately, TLS DNA polymerases are often error-prone and insert the wrong
nucleotide opposite the lesion and the price to pay for a cell to survive is a mutation. In
most cases, those mutations are base substitutions or gain or loss of 1-2 nucleotides. The
best known case is the change of a GC pair into a TA pair at 8-oxoG, which is the mutational
signature of 8-oxoG. Most of repair processes are error-free, while, as mentioned earlier, DSB
repair by end joining is often error-prone and lead to loss of a few nucleotides. Unrepaired
DSB lead to a more drastic loss of genetic information, so that ionizing radiation may induce
100 to 30 kilo base-pair deletions and loss of parts of chromosomes. Mutations lead to
inactivation or absence of proteins. Depending on which gene is mutated and so which
protein is inactivated, in which organ, the biological consequences may drastically differ:
cancer upon inactivation of a tumor suppressor gene, genetic disease in the offspring.
Chemistry and structural biology of DNA glycosylases :

the case of the formamidopyrimidine-DNA glycosylase (Fpg)
The prokaryote Fpg protein (the functional homologue of the eukaryotic Ogg1),
(Tab. 1) catalyses the removal of oxidized purines such as imidazole-ring opened purines
(Fapy) and the major oxidized product of purines, the 8-oxoguanine (8-oxoG) [9]. Both
lesions can be generated in DNA by irradiation (UV and gamma) and are associated with
replication arrests (Fapy residues) and G to T transversions (8-oxoG). Fpg cleaves the
N-glycosidic bond between the damaged base and its associated sugar (DNA glycosylase)
thus generating an AP site in DNA. The enzyme is also associated with an AP lyase activity
which eliminates the resulting AP site by successive cleavages at 3’ and 5’ sides by a
Ä'',Ä
2!$)!4)/.#(%-)3429
BD-elimination reaction. Then, the one nucleotide gap is filled in by DNA polymerase
and finally, the DNA strand is rejoined by DNA ligase (Fig. 3). Fpg achieves its catalytic
process using its P1 N-terminal amino group to perform a nucleophilic attack at the C1’
of the damaged nucleoside. During this process, the enzyme forms an imino enzyme-
DNA intermediate (Fig. 5). Fpg is probably one of the BER enzymes the more documented at
the atomic level (for review, see [10]).
Figure 5 : Reaction mechanism catalised by of the Fpg protein-DNA glycosylase and AP lyase process. The
deprotonated N-terminal proline P1 of Fpg (in blue) is exquisitely positioned inside the active site for nucleophile
attack (1) at the C1’ anomeric position of the damaged nucleoside (in black). The initial condensation step led to
the formation of a transient N,N-disubstituted hemiaminal (2). This unstable intermediate looses the damaged
base (B*) by the hydrolysis of the N-glycosidic bond and forms a protonated Schiff base intermediate (3) between
P1 and the C1’ of the resulting abasic site (DNA glycosylase activity). From the Schiff base, β-elimination (clea-
vage at the 3’ side of the sugar) produced the covalent intermediate (4). After δ-elimination and the hydrolysis of
the imino enzyme-DNA intermediate, the sugar is released as the unsaturated 4-oxo-2-pentenal (5, 6). (AP lyase
activity). Repair is achieved by DNA polymerase and DNA ligase.
Nucleoside chemistry to obtain stable complexes

of Fpg bound to DNA
Two general approaches have been used so far to generate stable Fpg/DNA
complexes enabling to study at the molecular level the interaction of this enzyme with
its substrates, reaction intermediates and products. One of these relies on site-directed
mutagenesis of the active site to abolish catalysis, independently from the Fpg specific DNA
binding [11]. An alternative approach relies on the design and synthesis of uncleavable
substrate analogues or inhibitors (Fig. 6). One class of inhibitors recently used consists of
Ä''-Ä
$.!DAMAGEREPAIR
nucleobase analogues with a stabilized N-glycosydic bond, that are not processed by Fpg
[12,13]. The second class of inhibitors more extensively used contains cyclic or non cyclic AP
site analogues that mimic the AP site structure in the catalytic transition state, preventing
the Fpg end products synthesis (Fig. 6) [14,15].
Figure 6 : Substrate analogues of the Fpg protein. The synthetic derivatives of oxidized guanine (7, 8, 9, 10)
and abasic site analogues (11, 12, 13, 14) are high affinity ligands for Fpg. Chemical “mutations” of natural
damages are indicated by red atoms.
DNA binding and nucleobase flipping by Fpg
The challenge of Fpg is to detect and excise one lesion among million of normal
bases, all the more this lesion does not disturbed the local DNA structure. Recently, the
crystal structure of Fpg interrogating undamaged DNA was solved using disulfide cross-
linking technology [16]. The protein searches its substrate by sliding along DNA. During this
process Fpg protein inserts a probe residue into the DNA helical stack and severely buckles
the normal target base pair. Crystal structures of Fpg bound to analogue-containing DNA
have revealed how Fpg, after recognizing the damage in DNA, exposed the C1’ anomeric
centre of the damaged nucleobase to the nucleophilic attack of N-terminal proline of the
enzyme (Fig. 7). Fpg solved this problem by flipping the target nucleobase out of the DNA
helix and stabilizing it in an extrahelical conformation into the active site of the protein [17,18,
19]. This extrusion of the damage results firstly from a strong curvature of the DNA helix
centred on the target site induced by the Fpg binding to the damaged strand, and secondly
from the intercalation in the minor groove of three strictly conserved residues which fill in
the hole resulting from the nucleobase extrusion (Figs. 7A and 7B). The amino acids (Met75,
Arg109 and Phe111) inserted into DNA prevent the double helix to collapse locally and allow
specific contacts with the orphan pyrimidine on the opposite strand, maintaining it in an
Ä''.Ä
2!$)!4)/.#(%-)3429
intrahelical conformation (Fig. 7B). In addition to elucidate the recognition of the damaged
base by the enzyme, structural studies indicate that Fpg specifies a pyrimidine opposite the
lesion. This structural observation is also supported by biochemical experiments indicating
that Fpg negatively discriminates a purine opposite the lesion [20].
Figure 7 : Crystal structure of Fpg bound to cFapydG-containing 14-mer DNA duplex A) Overview of the
complex showing the strong DNA torsion induced by Fpg binding. DNA double helix is in grey and the peptide
main chain fold of Fpg is represented by a ribbon multicolour. The side chain of the active site Pro1 is indicated in
magenta and the zinc binding domain of the protein in green. B) Enlargement of the structure at the DNA target
site showing the intercalation of the protein by the minor groove. The phophodiester backbone of DNA strands
are represented by white cylinders. Fpg is represented by its solvent accessible surfaces (blue, red and white areas in-
dicate basic, acidic and neutral regions, respectively). In the enzyme binding pocket, the damaged nucleoside (its C,
N and O atoms are coloured in yellow, blue and red, respectively) is stabilized in an extrahelical conformation while
the estranged cytosine on the strand opposite the lesion in an intrahelical conformation. C) Network of hydrogen
bonds made by the enzyme with the FapydG lesion inside the catalytic site. The peptide main chain of Fpg is
shown in yellow ball-and-sticks, the side chains of amino acids residues involved in the recognition of the damage
in pink and carbon atoms of DNA in grey balls. N, O and P atoms are indicated by blue, red and purple balls, respec-
tively. Covalent links and hydrogen bonds are indicated by sticks and red dashed lines, respectively. A red “w” letter
indicates a structural water molecule.
Fpg recognition mode of FapyG residue and mechanism

of catalyzing N-glycosidic bond cleavage
The extruded FapydG residue in an anti-conformation is inserted into the binding

pocket of the enzyme active site [19]. This pocket consists in aromatic and hydrophobic
residues that can modulate the size and the shape of the pocket, and in polar residues that
Ä'(%Ä
$.!DAMAGEREPAIR
make specific hydrogen bonds with the damaged guanine (Figs. 7B and 7C). Fpg specifies an
oxidized purine, establishing hydrogen bonds with all the hydrogen donors and acceptors
of the Watson-Crick face of the FapydG residue. In addition, Fpg recognizes, through water
molecule-mediated interactions, all the structural determinants of the imidazole ring-
opened moiety of FapydG (especially, the protonated N7 and the C8O groups of the damage
which distinguish clearly the normal guanine from FapyG ; Fig. 7C). All these interactions
constrain the nucleobase in an optimal conformation inside the active site of the enzyme
for catalysis. In this conformation, the C1’ anomeric centre of the extrahelical damaged
nucleoside is exposed to the nucleophilic attack of the N-terminal proline which results in
an imino enzyme-DNA complex (Figs. 5 and 7C). This intermediate can be easily trapped by
its irreversible reduction [17].
Conclusion
During the past decade, our understanding of the molecular mechanism by which
Fpg (and other DNA glycosylases) recognizes and excises base lesions has advanced greatly
by combining organic chemistry, biochemistry and structural biology approaches. Despite
our understanding of Fpg process, several aspects must be now best documented, especially
substrate specificity and catalysis, and how this enzyme recognizes its substrate in DNA
before the damaged nucleoside flipping.
Acknowledgements
We are very grateful to all people who participated to our respective laboratory work. Our
work is supported by CNRS, Institut Curie, EDF, CNES, ARC, the Région Centre and the Ligue nationale
contre le cancer (Indre).
References
[1] Iliakis G., Wang Y., Guan J., Wang H., DNA damage checkpoint control in cells exposed to ionizing
radiation, Oncogene, 2003, 22, 5834-5847.
[2] Ward J., DNA damage produced by ionizing radiation in mammalian cells: identities, mechanisms
of formation and reparability, Prog. Nucleic Acid Res. Mol. Biol., 1988, 35, 95-125.
[3] Nikjoo H., O'Neill P., Terrissol M., Goodhead D., Modelling of radiation-induced DNA damage: the
early physical and chemical event, Int. J. Radiat. Biol., 1994, 66, 453-457.
[4] Blaisdell J., Harrison L., Wallace S., Base excision repair processing of radiation-induced clustered
DNA lesions, Radiat. Prot. Dosimetry, 2001, 97, 25-31.
Ä'(&Ä
2!$)!4)/.#(%-)3429
[5] Eot-Houllier G., Eon-Marchais S., Gasparutto D., Sage E., Processing of a complex multiply
damaged DNA site by human cell extracts and purified repair proteins, Nucleic Acids Res., 2005,
33, 260-271.
[6] Cline S., Hanawalt P., Who's first in the cellular response to DNA damage?, Nature reviews, 2003,
4, 361-372.
[7] Hoeijmakers J., Genome maintenance mechanisms for preventing cancer, Nature, 2001, 411,
366-374.
[8] Caldecott K., Mammalian DNA single-strand break repair: an X-ra(y)ted affair, BioEssays, 2001, 23,
447-455.
[9] Zaika EI, Perlow R.A., Matz E., Broyde S., Gilboa R., Grollman A.P., Zarkov D.O., Substrate
discrimination by formamidopyrimidine-DNA glycosylase, J. Biol. Chem., 2004, 279, 4849-4861.
[10] Huffman J.L., Sundheim O., Tainer J.A., DNA base damage recognition and removal: new twists
and grooves, Mutation Res., 2005, 577, 55-76.
[11] Fromme JC, Verdine G.L., DNA lesion recognition by bacterial repair enzyme MutM, J. Biol. Chem.,
2003, 278, 51543-51548.
[12] Wiederholt C.J., Delaney M.O., Greenberg M.M., Interaction of DNA containing FapydA or its
C-nucleoside analogues with base excision repair enzymes. Implications for mutagenesis and
enzyme inhibition, Biochemistry, 2002, 41, 15838-15844.
[13] Ober M., Linne U., Gierlich J., Carell T., The two main DANN lesions 8-oxo-7,8-dihydroguanine and
2,6-Diamino-5-formamido-4-hydroxypyrimidine exhibit strongly different pairing properties,
Angew. Chem. Int. Ed., 2003, 42, 4947-4951.
[14] Castaing B., Boiteux S., Zelwer C., DNA containing a chemically reduced apurinic site is a high affinity
ligand for the E. coli formamidopyrimidine-DNA glycosylase, Nucleic Acids Res., 1992, 20, 389-394.
[15] Castaing B., Fourrey J.L., Hervouet N., Thomas M., Boiteux S., Zelwer C., AP site structural
determinants for Fpg specific recognition, Nucleic Acids Res., 1999, 27, 608-615.
[16] Banerjee A., Santos W.L., Verdine G.L., Structure of DNA glycosylase searching for lesions, Science,
2006, 311, 1153-1157.
[17] Serre L., Pereira de Jésus K., Boiteux S., Zelwer C., Castaing B., Crystal structure of the Lactococcus
lactis formamidopyrimidine-DNA glycosylase bound to an abasic site analogue-containing DNA,
EMBO J., 2002, 21, 2854-2865.
[18] Fromme J.C., Verdine G.L., DNA lesion recognition by the bacterial repair enzyme MutM, J. Biol.
Chem., 2003, 278, 51543-51548.
[19] Coste F., Ober M., Carell T., Boiteux S., Zelwer C., Castaing B., Structural basis for the recognition of
the FapydG lesion (2,6-diamino-4-hydroxy-5-formamidopyrimidine) by the Fpg DNA glycosylase,
J. Biol. Chem., 2004, 279, 44074-44083.
[20] Castaing B., Geiger A., Seliger H., Nehls P., Laval J., Zelwer C., Boiteux S., Repair and binding of a
DNA fragment containing a single 8-oxoguanine by wild type and mutant Fpg protein, Nucleic
Acids Res., 1993, 21, 2899-2905.
Ä'('Ä
Chapter 16
Pulse radiolysis studies
of free radical processes
in peptides and proteins
Chantal HOUÉE-LEVIN and Krzysztof BOBROWSKI
Introduction
The story of free radical reactions in biological systems began many years ago,
when Harman postulated that they played a prominent role in ageing [1]. Since the sixties,
the relevance to biological as well as industrial processes became more and more clear
every year. It is now current to invoke free radicals in ordinary life, in cooking, in prevention
of ageing processes, etc. and the scientific knowledge gave a basis to these assertions [2].
It is now beyond doubt that free radicals processes in proteins are involved in all steps of
life, going from conception to death induction. They are believed to be part of the cellular
defence against oxidative stress and at the same time responsible of severe damage like
atherosclerosis [3] Protein free radicals are also enzyme active sites [4].
The studies by the methods of radiolysis provided a wealth of knowledge about

the kinetic and thermodynamic controls of radical reactions, the importance of which
is no more questioned. Indeed, it is known that the chemical events initiated by ionizing
radiation, are the same as those that take place in normal and deleterious events of every
day’s life. In this review we focus on some of the major knowledge that was acquired by the
use of pulse radiolysis and steady-state gamma radiolysis of aqueous solutions of amino
acids, peptides and proteins (Inset). The potential role of pulse radiolysis (Chapter 2) for
studying biomolecules has been acknowledged rather early. In most cases, pulse radiolysis
Ä'((Ä
2!$)!4)/.#(%-)3429
method has been very valuable in identification of radicals, establishing their structures and
exploring their reactivity (Chapter 1). We would like also to show how the reactions that were
evidenced took their place in biological events, thus how a synergy between pure chemistry
and biology was established.
Inset : Protein structure
A protein is a polymeric chain. (A) Monomers are amino acids. (B) In polymerization, residues
are linked by the peptidic bond, which is an amide function.
A/ An amino acid. R1 will be the residue 1 in scheme B.
B/ The protein
However this polymeric chain is folded. The folding has a tremendous importance in the action
of the protein and in its recognition by biological partners and/or by the degradation pathways.
The different levels of structure are presented in Figure 1. (For more details, see for instance [i]).
[i] Voet D., Voet J.G., Biochemistry 2nd edition, John Wiley New York.
The simplest amino-acid, glycine
One of the more recent results concerns tOH-induced oxidation of glycine, the
simplest amino acid [5]. Two main radical products H2Nt-CH2-CO2 and aminyl radicals
HNt-CH2-CO2 have been identified and their subsequent reaction pathways including
decarboxylation with parallel formation of tCH2NH2 and B-fragmentation into the respective
Ä'()Ä
/NE ELECTRONPROCESSESINPEPTIDESANDPROTEINS
Figure 1 : The several levels of structure of a protein.
imine and carboxyl radical CO2t , respectively. The possible initiation of amino acid
decarboxylation by C-centred radicals are considered to be of general significance and
interest in chemical and biological systems.
Ä'(*Ä
2!$)!4)/.#(%-)3429
From amino acids to proteins
It was, however, soon recognized that in aqueous solutions peptides and proteins
(Inset) and free radicals derived from them behaved differently from amino acids and free
radicals derived from them. These differences arise from the following :
O The folding (Fig. 1) renders some amino acids not available to free radicals. Indeed,
since the reactions are fast, their control is kinetic and not thermodynamic. One
can imagine that the amino acids that are on the surface are more reactive than
those in the interior.
O Some amino acids are charged; thus the protein creates an electrostatic field
around itself, which orients the reactivity of the charged radicals.
O The reduction potentials of amino acids are sensitive to the environment.
O Functional groups present in side chains of neighbouring amino acids can be involved
in reaction mechanism and thus may affect reaction pathway of radicals formed.
Thus kinetics as well as thermodynamics aspects of the free radical reactions are
modified. Some examples are detailed in what follows.
Thiyl radicals
Thiyl radicals are important reactants in several enzymes and form in vivo during
conditions of oxidative stress [6]. They have been considered for a long time as rather
unreactive species. However, recently several reactions of thiyl radicals with biomolecules
have been described (catalysis of cis-trans isomerization of unsaturated fatty acids, addition to
the pyrimidine bases C5-C6 double bonds, and hydrogen abstraction from polyunsaturated
fatty acid, thymine and peptide CA-H and side chain C-H bonds) [7]. More recently, the
intramolecular addition of peptide cysteine thiyl radicals (CysSt) to phenylalanine (Phe)
yielding alkylothio-substituted cyclohexadienyl radicals was demonstrated in the peptides
Phe-Cys and Phe-Gly-Cys-Gly (Fig. 2) [8].
This addition reaction might be of great biological significance since the

intramolecular addition of CysSt to aromatic ring of Phe can compete with addition of O2 to
CysSt, assuming biologically relevant tissue concentration of O2. It presents a possible free
radical pathway to thioether-containing peptide and protein cross-links.
Ä'(+Ä
Figure 2 : In the peptide Phe-Gly-Cys-Gly (formed by Phenylalanine, Glycine, Cysteine and Glycine amino acids
linked by peptidic bonds), the sulphur atom is very close to the phenyl ring. An intramolecular addition of the
CysStradical on the aromatic ring of Phe (marked with an arrow) leads easily to the alkylthio-substituted cyclohexadie-
nyl radicals. Atom colors: Cyan: carbon; blue: nitrogen; red: oxygen; yellow: sulphur. H atoms are not shown.
Disulfide radicals
Disulfide radical anions might play an important role in oxidative stress. In cellular
media, they can be formed by oxidation of protein thiol functions (by tOH radicals, for
instance, Chapter 1), followed by dimerization :
1SPUFJO4)tOH m Protein/St)2O (1)

Protein/St1SPUFJO4)m Protein /S—St¦1SPUFJO) (2)
These disulfide radical anions are believed to be strong reductants. Thus they might
counterbalance the action of oxidizing free radicals.
To study their properties, the easiest way is to form them by one-electron reduction
of disulfide bonds. Among radicals from water radiolysis, hydrated electron is the most
powerful reductant. It reacts with almost all amino acids and especially with the disulfide
groups. Using less powerful reductants such as COOt¦ radicals, some selectivity in the attack
appears. An example is displayed in Figure 3.
Ä'(,Ä
2!$)!4)/.#(%-)3429
Figure 3 : Hen egg white lysozyme has 4 disulfide bridges. However, only one of them is easily reduced
A) marked with an arrow. The detail of its structure is shown in B) The disulfide radical anion is stabilized by inte-
raction with the charged end of an Arginine residue (in green). In white: the distances (in Å) between the central
carbon atom and the sulphur atoms. In red: the polypeptidic chain.
In hen egg white lysozyme, out of the four disulfide bridges, one is much more easily
reduced than the others (Fig. 3 A and B) [9]. A study by quantum chemistry could justify this
selectivity by the stabilization of the resulting disulfide free radical by the positively charged
end of arginine. The distances between sulphur atoms and the central carbon of arginine
confirm the strong interaction between these groups (Fig. 3).
Pulse radiolysis studies demonstrated that all disulfide radicals do not have the same
chemical properties. An example is given by the study of protonation equilibrium (Scheme 1).
Scheme 1: Protonation-deprotonation equilibrium of disulfide radical anion.
The resulting protonated disulfide radical undergoes cleavage of the S-S bond.
Since the thiyl radical is a strong oxidant, the protonation equilibrium appears like a switch
Ä'(-Ä
between oxidant/reductant radical. The pKa values of some of these radicals, measured by
pulse radiolysis, are between 5 and 6 (Fig. 4). However in thioredoxin, it is below 3 [10].
Figure 4 : Titration curves for hen egg white lysozyme and thioredoxin disulfide radical. Red : lysozyme; blue :
thioredoxin.
It means that sulphur free radical from this protein are mostly reductant, which
enlightens the role of thioredoxin in restoring sites which were oxidized and thus as
protection against oxidative stress.
One-electron oxidation of methionine

Methionine (Met) is one of the sensitive sulphur-containing amino-acids toward
one-electron oxidation. However its ease of oxidation is also modulated by the structure.
The one-electron oxidation of Met in peptides yields sulfide methionine radical cations
(MetSt) which convert into intermediates that obtain catalytic support from neighbouring
groups containing electron rich heteroatoms (S, N, O) and thus stabilize electron deficient
sulphur centres in S<S, S<N, and S<O-three-electron bonded complexes (Fig. 5) [11].
Interaction with particular peptide or protein domains would likely involve N- and
C-terminal nucleophilic functionalities (NH3, COO ), and nucleophilic functionalities in the
side chains of amino acid residues (Asp, Glu, Lys, Val, Thr). However, very often heteroatoms
in peptide bonds are the only nucleophiles present in the vicinity of the MetSt. It was
Ä'(.Ä
2!$)!4)/.#(%-)3429
Figure 5 : Electronic energetic diagram of the molecular orbitals of three–electron bonds in sulphur radicals
and optical transition. X : S, N, or O atom.
recently shown that such interactions play an important role in oligopeptides of the form
N-Ac-Gly-Met-Gly and N-Ac-Gly-Gly-Gly-Met-Gly-Gly-Gly [12]. Pulse radiolysis studies with
UV/Vis spectrophotometrical and conductometric detection showed for the first time that
MetStin peptides can be stabilized through bond formation with either the oxygen or the
nitrogen atoms of adjacent peptide bonds (Fig. 6).
Figure 6 : Structure of N-Ac-Gly-Met-Gly peptide showing interactions of the sulfur atom (S) in Met with the
oxygen atom (O) (A) and the nitrogen atom (N) (B) (marked with arrows) located in the N-terminal adjacent
peptide bond that lead to the S<O and S<N-bonded radicals, respectively. Atom colors : Cyan : carbon ; blue :
nitrogen ; red : oxygen ; yellow : sulphur. H atoms are not shown.
Ä')%Ä
Moreover, formation of radical transients with S<O bonds is kinetically preferred,

but on longer time scale they convert into transients with S<N bonds in a pH dependent
manner. Ultimately transients with S<N bonds transform intramolecularly into C-centred
radicals located on the AC moiety of the peptide backbone. Another type of C-centred
radicals located in the side chain of Met-residue, A-(alkylthio)alkyl radicals, are formed via
deprotonation of MetSt. C-centred radicals are precursors for peroxyl radicals (ROOt) that
might be involved in chain reactions of peptide and/or protein oxidation. Stabilization of
MetStthrough formation of S<O- and S<N-bonded radicals might potentially accelerate
oxidation and autooxidation processes of Met in peptides and proteins. Considering
that methionine sulfoxide, which is the final product coming from all radicals centred on
sulphur, is restored by the enzyme methionine sulfoxide reductase into MetS, stabilization
of MetStappears as a protection against an eventual peroxidation chain that would develop
from a carbon centred radical.
The amyloid B peptide (BAP), a 39- to 43-amino acid-long peptide is the major
constituent of the neuritic plaques and neurofibrillary tangles in brain. Their progressive
formation is characteristic for the development of the Alzheimer disease. The integrity of
Met35 is very important for the constitution of the senile plaque. In the structure of BAP,
Met35-Stformation can be facilitated by a pre-existing close sulphur-oxygen (S-O) interaction
between the Met35 sulphur and the carbonyl oxygen of the peptide bond C-terminal to
isoleucine (Ile31) that might lower one-electron reduction potential of MetSt/Met couple.
The first experimental evidence that Met is more easily oxidized than in other peptides and
proteins comes from one-electron oxidation of BAP1-40 using azide radicals (N3t) produced
by pulse radiolysis [13]. One-electron oxidation of the natural peptide was compared to that
of the non-natural peptide of reverse sequence (BAP40-1). Circular dichroism showed that
whereas AB1-40 is highly structured, AB40-1 has no regular structure. It appeared that the
oxidation patterns of these two compounds are different: in BAP (1-40) Met35 is mostly the
target of free radicals, whereas in BAP40-1, Tyr10 gets dimerized. This observation underlines
the role of structure in driving the free radical reactions and seems to be indeed relevant to
explain specificity of the AB1-40 in the development of Alzheimer disease.
The pathogenesis of another well-known neurodegenerative disease (Jacob

Creutzfeld disease) seems to be strongly linked to the presence of prion proteins in the brain.
These macromolecules contain multiple Met residues, some of them in close vicinity. Such
structure should favour stabilization of MetStas intramolecular (S<S)t complexes. Since
weak intramolecular non-bonded SvvvO and SvvvN interactions have been recently suggested
in proteins [14], stabilization of MetSt through formation of S<N- and/or S<O-complexes
might potentially accelerate oxidation processes in proteins. The first experimental evidence
Ä')&Ä
2!$)!4)/.#(%-)3429
for the stabilization of MetStas intramolecular complex with N-atom in adjacent amide
group was obtained during one-electron oxidation of calmodulin (CaM-Ca4, wild type),
studied on the microsecond time domain by pulse radiolysis [15]. Calmodulin is a regulatory
“calcium sensor” protein that contains nine Met residues. Stabilization of MetSt by peptide
bonds might be a general phenomenon in proteins.
There are different reaction pathways of decay of the S<N-complex depending on

the structure of the peptide. In model peptides containing N-terminal glutamic acid residue
(G-Glu-Met, G-Glu-Gly-Met-Gly), an intramolecular electron transfer from the carboxylate
group to the electron-deficient center at the nitrogen within S<N-bond followed by
homolytic bond-breakage of the carbon-carboxylate bond leads to formation of A-amino
radicals and CO2 [16]. On the other hand, in model peptides Thr-(Gly)n-Met, n = 0-4, the
intermediary S<N-bonded radicals do not decarboxylate but undergo homolytic cleavage
of the CA-CB bond of Thr, yielding the highly toxic acetaldehyde [17]. Once again, the fate
and the consequences of peptide oxidation depend strongly on the stabilization of MetSt.
Intramolecular long range electron transfer

in peptides and proteins
Intramolecular long-range electron transfer (LRET) plays a major role in many

biological processes, including fundamental energy storage processes such as photosynthesis
and respiration, redox-mediated enzyme catalysis, and in various pathologic processes such
as radiation damage, oxygen toxicity and cellular aging. Most of these processes occur
on ground-state potential energy surfaces, making pulse radiolysis an effective and truly
unique tool for these studies. The typical LRET pulse radiolysis experiment begins with the
rapid selective oxidation or reduction of one-redox site on a macromolecule (formation of
the donor-acceptor complex) followed by the intramolecular LRET. The advantage of pulse
radiolysis is that using either an oxidizing or a reducing radical can generate the donor-
acceptor complex. These radicals have redox potentials covering the approximate range
GSPNoUP7 vs. SCE) [18]. The entire subject of LRET is much too wide to be covered in
detail. Numerous comprehensive reviews are available which present and discuss this topic
in a more detailed manner (see for instance [19] and references therein). Therefore, a few
selected important achievements emerging in this field will be highlighted.
Most LRET processes in biological systems are nonadiabatic. In quantum-mechanical

electron-transfer theory, the rate constant for nonadiabatic ET from a donor to acceptor can
be expressed as the product of the square of an electronic coupling matrix element (HDA)
and a nuclear Franck-Condon factor (FC): kET = (2P/h)[HDA]2(FC). The [HDA] is a measure of the
Ä')'Ä
coupling or the interaction between the orbitals of the donor (D) and the acceptor (A) and is
influenced by the distance and structure of the medium separating the electron donor and
acceptor. The role of the distance, standard free energy change, and reorganization energy
was examined. Influence of secondary structural features of the peptide bridge has been
probed by applying flexible oligoglycine bridges, conformationally more rigid oligoproline
bridges and helical bridges (for review see [20] and references therein) Pulse radiolysis
studies of simple model synthetic peptides have demonstrated intramolecular ET involving
radicals located on the side chains of aromatic, histidine and sulfur containing (methionine,
cysteine) amino acid residues.
Elaboration of LRET mechanism by resolving the parameters that determine specific

rates of LRET has stimulated pulse radiolysis studies in proteins. Examples include generation
of metastable electron donor and acceptor complexes in (1) native and mutant proteins,
(2) proteins with the directed single-site specific mutations, (3) native and mutant multi-
site redox proteins, (4) proteins with the site specific modification with transition metal
complexes covalently attached to a specific surface amino acid residues.
Pulse radiolysis investigations related to oxidative stress
The methods of gamma and pulse radiolysis provided the basis for the understanding
of oxidative stress. Indeed, one cannot imagine how the biological importance of short-lived
transients could have been demonstrated otherwise and how their chemical properties
could have been studied.
Control of superoxide radical anion steady state

A striking example is control of the steady state level of superoxide radical anion (O2t ).
Although superoxide anions are not very reactive, they play a prominent role in oxidative stress
by triggering formation of peroxynitrite. The control of their steady state level is thus vital for
cells. In most living organisms the well-known metalloenzyme superoxide dismutase (SOD),
present in almost all aerobic cells catalyzes the disproportionation of O2t into H2O2 and O2:
O2t 02t ) m H2O202 (3)
SODs are differentiated mainly by the redox-active metal in the active site and by the
localization: copper-zinc (in the cytosol), manganese (in mitochondria) and iron. The iron and
manganese SODs are structurally similar. However, some bacteria can live without the Cu-Zn
SOD, whereas no organisms can live without Mn-SOD. It stresses the vital importance of these
Ä')(Ä
2!$)!4)/.#(%-)3429
enzymes close to the place where O2t are produced. The mechanisms by which Mn-SOD reacts
with O2t are of interest as knowledge of the kinetic parameters, and the reaction pathways may
allow the synthesis of model compounds with specific chemical features. The kinetics of the
dismutation of O2t by Escherichia coli MnSOD were measured by pulse radiolysis, and shown
to fit a mechanism involving the rapid reduction of MnSOD by O2t followed by both the
direct reoxidation of MnSOD by O2t and the formation of a MnIISOD(O2t ) complex [21].
The second system was discovered in prokaryotic cells a few years ago [22]. It is
the recently characterized non-heme iron superoxide reductase (SOR) that catalyzes the
reduction of O2t into H2O2 by an intracellular reductant :
O2t F−) m H2O2 (4)
For studies about the mechanisms of action of this enzyme, pulse radiolysis also played
a prominent role. Current research on improving knowledge about these enzymes, creating
more performant ones by site directed mutagenesis or testing the ability of destruction of
superoxide by metal complexes also uses pulse radiolysis as a method of screening.
Haem peroxidases are globular proteins with an iron-porphyrin complex as a

prosthetic group. These enzymes are widespread among prokaryotes and eukaryotes. They
catalyze the oxidation of substrates by organic peroxides or hydrogen peroxide. During the
past decades, considerable scientific effort has been put into elucidation of the mechanisms
of reactions catalyzed by these enzymes. Pulse radiolysis technique has made an important
contribution by providing information on the redox states of the enzymes and their
interconversion, as well as on the properties of the free radical intermediates involved [23].
Pulse radiolysis investigations of the reaction of superoxide

radical anions (O2t−) with radicals derived from various amino acids
The following results were all obtained using pulse radiolysis.
Methionine. The reaction of superoxide radical anions (O2t ) with sulfide radical
cation-nucleophile complexes might represent an efficient sulfoxide-forming process in
peptides and proteins containing methionine under conditions where significant amounts
of sulfide radical cation complexes and superoxide are formed simultaneously. The rate
constant for the reaction of O2t with the (S<N) complex was found to be ca. 3-fold slower
as compared to that of the reaction with the (S<S) complex. This drop in reactivity may,
in part, reflect the lower probability of O2t to encounter S-atom in the (S<N) complex as
Ä'))Ä
compared to the symmetrical (S<S)t complex. It is important to note that the reactions of O2t
with the sulfide radical cation complexes proceed 2.5 to 8-fold faster than the reaction of
O2t with superoxide reductases SOD [24]. From a biological point view, it means that sulfide
radical cation-O2t reactions might represent a potential source for sulfoxide formation when
the system is exposed to high concentrations of reactive oxygen species (ROS).
Tryptophan. The reaction of O2t , with the semi-oxidized tryptophan neutral radical
(Trpt) generated from tryptophan (Trp) by pulse radiolysis has been studied in a variety
of functionalized Trp derivatives and in lysozyme [25]. These studies demonstrate that
the reactivity of O2t with the Trpt radical is significantly dependent on the charge in the
vicinity of the Trp residue from which it originates, by the electrostatic field generated
by neighbouring groups and/or other charged cell constituents (e.g. DNA and lipids).
The kinetics of O2t reaction with the semi-oxidized tryptophan neutral radicals (TrptLyz)
has been investigated at various pHs and conformational states. It was found that at pH
lower than 6.2, the apparent bimolecular rate constant is about 2 s 108 M 1s 1 but drops to
8 s 107 M 1s 1 or less above pH 6.3 and in CTAC micelles. Interestingly, at all pHs the rate
constants of the reaction of O2t with Trpt radicals are more than an order of magnitude
greater than rate constants characterizing the intermolecular recombination of Trpt radicals.
The well-established LRET from Tyr residues to Trp radicals-leading to the repair of the TrptLyz
radicals is inhibited by the Trpt-Z[02t reaction. In conclusion, the reaction of O2t with
reactive radicals derived from amino acids in proteins may have important consequences
for the turnover of proteins involved in metabolic reactions.
Conclusion
The studies summarized here brought important knowledge at several levels. In

fundamental science, the evidence of phenomena such as Long range Electron Transfer,
LRET, has stimulated experiments and discussions all over the world. It was discovered that
LRET concerns all polymers. In DNA, it may explain the localization of base lesions induced
by oxidative stress. As for biology, it helped to understand the chemical basis of initiation
and development of the inflammation processes that take place in all diseases. Moreover
this fundamental research had consequences on everyday life: a search for evidences of the
importance of antioxidants in food in relation to quality and duration of life was undertaken.
The results were so positive that the content of antioxidants is now a criterion of quality
of food. The redox reactions of the sulphur amino acids have triggered a research on new
anti-inflammatory drugs, and some of them (such as N-acetyl cysteine) are currently used
in many pathologies. There is no doubt that in the close future other comprehension and
applications of the radiolysis results will appear.
Ä')*Ä
2!$)!4)/.#(%-)3429
References
[1] Harman D., Aging : a theory based on free radical and radiation chemistry, J. Gerontol., 1956,
298-300.
[2] Houée-Levin C., Sicard-Roselli C., Bergès J., Chimie et Biochimie radicalaires, Belin coll. Ellipses, 2005.
[3] Aruoma O. I., Halliwell B., Molecular biology of free radicals in human diseases, OICA international,
London, 1998.
[4] Stubbe J. A., Van der Donck W. A., Protein radicals in enzyme catalysis, Chem. Rev., 1998, 98, 705-762.
[5] Bonifačić M., Armstrong D. A., Carmichael I., Asmus K.-D., B-Fragmentation and other reactions
involving aminyl radicals from amino acids, J. Phys. Chem. B., 2000, 104, 643-649.
[6] Jourd’heuil D., Jourd’heuil F. L., Feelisch M., Oxidation and nitrosation of thiols at low micromolar
exposure to nitric oxide. Evidence for a free radical mechanism, J. Biol. Chem., 2003, 278, 15720-
15726.
[7] Ferreri C., Kratzsch S., Landi L., Brede O., Thiyl radicals in biosystems: effects on lipid structures
and metabolisms, Cell. Mol. Life Sci., 2005, 62, 834-847.
[8] Nauser T., Casi G., Koppenol W. H., Schöneich C., Intramolecular addition of cysteine thiyl radicals to
phenylalanine in peptides : formation of cyclohexadienyl type radicals, Chem. Commun., 2005, 3400-
3402.
[9] Bergès J., Kassab E., Conte D., Adjadj E., Houée-Levin C., Ab initio calculations on arginine-disulfide
complexes modelling the one-electron reduction of lysozyme. Comparison to an experimental
reinvestigation, J. Phys. Chem., 1997, 101, 7809-7817.
[10] El Hanine, Lmoumène C., Conte D., Jacquot J.-P., Houée-Levin C., Redox properties of protein
disulfide bond in oxidized thioredoxin and lysozyme. A pulse radiolysis study, Biochemistry,
2000, 39, 9295-9301.
[11] Bobrowski K, Holcman J., Formation and stability of intramolecular three-electron S<N, S<S,
and S<O bonds in one-electron-oxidized simple methionine peptides. Pulse radiolysis study,
J. Phys. Chem., 1990, 93, 6381-6387.
[12] Schöneich C., Pogocki D., Hug G. L., Bobrowski K., Free radical reactions of methionine in peptides :
mechanisms relevant to B-amyloid oxidation and Alzheimer’s disease, J. Am. Chem. Soc., 2003,
125, 13700-13713.
[13] Kadlčik V., Sicard-Roselli C., Kodicek M., Houée-Levin C., One-electron oxidation of beta-amyloid
peptide: sequence modulation of reactivity, Free Rad. Biol. Med., 2004, 37, 881-891.
[14] Iwaoka M, Takemoto S., Tomoda S., Statistical and theoretical investigations of directionality
of nonbonded SvvvO interactions. Implications for molecular design and protein engineering,
J. Am. Chem. Soc., 2003, 124, 10613-10620.
[15] Nauser T., Jacoby M., Koppenol W.H., Squier T.C., Schöneich C., Calmodulin methionine residues
are targets for one-electron oxidation by hydroxyl radicals : formation of S<N three-electron
bonded radical complexes, Chem. Commun., 2005, 587-589.
Ä')+Ä
[16] Bobrowski K, Schöneich C., Decarboxylation mechanism of the N-terminal glutamyl moiety in
Gglutamic acid and methionine containing peptides, Radiat. Phys. Chem., 1996, 47, 507-510.
[17] Pogocki D., Ghezzo- Schöneich E., Schöneich C., Conformational flexibility controls proton transfer
between methionine hydroxy sulfuranyl radical and the N-terminal amino group in Thr-(X)n-Met
peptides, J. Phys. Chem. B, 2001, 105, 1250-1259.
[18] Wardman P., Reduction potentials of one-electron couples involving free radicals in aqueous
solution, J. Phys. Chem. Ref. Data, 1989, 18, 1637-1755.
[19] Pecht I., Farver O., Pulse radiolysis : A tool for investigating long-range-electron transfer in
proteins, in “Photochemistry and radiation chemistry. Complementary methods for the study of
electron transfer”, Wishart J. F., Nocera D. G. (eds.), Adv. Chem. Ser., 1998, 254, ACS, Washington,
D.C., 65-69.
[20] Houée-Levin C., Sicard-Roselli C., Radiation chemistry of proteins, in “Radiation chemistry: present
status and future prospects”, Jonah C., Rao B.M. (eds.), Elsevier, 2001, 53-84.
[21] Hsu J.L., Hsieh Y., Tu C.K., O'Connor D., Nick H.S., Silverman D.N., Catalytic properties of human
manganese superoxide dismutase, J. Biol. Chem., 1996, 271, 17687-17691.
[22] Jenney F.E. Jr., Verhagen M.F.J. M., Cui X., Adams, M.W.W., Anaerobic microbes: oxygen
,
detoxification without superoxide dismutase Science, 1999, 286, 306-309.
[23] Candeias L.P., Gebicka L,. Investigation of mechanisms of peroxidase-catalyzed reactions by
radiation-chemical techniques, J. Radioanal. Nucl. Chem., 1998, 232, 29-34.
[24] Bonifacic M., Hug G.L., Schoeneich C., Kinetics of the reaction between sulfide raducal cation
DPNQMFYFT <44> BOE <4/> BOE TVQFSPYJEF PS DBSCPO EJPYJEF SBEJDBM BOJPO + 1IZT $IFN "
2000, 104, 1240-1245.
[25] Santus R., Patterson L.K., Hug G.L., Bazin M., Mazière J.C., Morlière P., Interactions of superoxide
anion with enzyme radicals : kinetics of reaction with lysozyme tryptophan radicals and
corresponding effects on tyrosine electron transfer, Free Rad. Res., 2000, 33, 383-391.
Ä'),Ä
Chapter 17
Radiation-induced damage
of membrane lipids
and lipoproteins
Monique GARDES-ALBERT
Introduction
If radiation-induced damages to DNA have long been considered as the only critical
events for the cell machinery, it is now admitted that cell membranes are also significant
biological targets of ionizing radiation [1]. Lipids are the major constituents of biomembranes
and of lipoproteins. Their amphiphilic structure, namely hydrophilic polar heads and
hydrophobic hydrocarbon tails, allowing the build up of dense networks of intermolecular
bonds, favours the packing of membrane lipids into bilayers. In addition, proteins are inserted
into the lipidic matrix where they are associated with lipid domains. The cell membrane is
not a simple barrier defining the boundaries of the cell, since it modulates signals from the
extracellular medium into the cell and it also controls intercellular communications [2].
Ionizing radiations such as G-rays, are able to directly ionize the lipid/protein network
of biomembranes, but also to ionize the water molecules surrounding all the cell constituents.
Molecules of H2O which are present in both intra- and extracellular compartments, undergo
radiation-induced decomposition into free radical species. The proportion of directly ionized
biological targets versus indirectly-mediated damages, via free radicals from water radiolysis,
remain unclear and controversial, often depending on the studied systems. In this chapter,
it will be focused on the indirect effects coming from the reactions of lipids and lipoproteins
Ä').Ä
2!$)!4)/.#(%-)3429
with oxygenated free radicals produced by radiation-induced decomposition of water. These

effects are related to numerous works devoted to free radical-induced damages on aqueous
lipidic model systems (micelles, liposomes, lipoproteins) submitted to ionizing radiation.
Moreover, the mechanism of formation of free radicals from water radiolysis being well
known, together with the radiolytic yields of production of each radical species (Chapter 1),
it is relatively easy to obtain from radiation dose effects, kinetic results on the chemical
reactions of these free radicals with biological targets (such as lipids) dissolved in water
[3]. For these reasons, water radiolysis can be seen as a powerful tool for studying radical
mechanisms on defined biological systems.
Lipid peroxidation has been found as the main damage of membrane lipids and
lipoproteins, initiated by oxygenated free radicals. A key feature of this phenomenon is its
propagating capacity into the lipidic network by a chain reaction, due to the tight packing
of the lipid molecules, leading to lipid hydroperoxides as major initial products. In addition,
radiation-induced oxidative fragmentation of lipids has been described [4], together
with changes in the physical state (fluidity/rigidity, permeability, …) and as a function of
biomembranes [1,2]. In the present review, recent knowledge about the chemical nature and
mechanism of free radical-induced lipidic damages will be examined, from simple models of
lipid aggregates submitted to G-rays. The consequences of the presence of new peroxidized
lipidic products and short-chain fragments into the cell membranes will be discussed.
Oxidative stress phenomena can be indeed responsible for the in vivo peroxidation of
lipoproteins leading to atherosclerosis [5].
Composition of membrane lipids
Fatty acids, glycerophospholipids and sphingolipids, are the three major classes
of lipids entering in the composition of biological membranes. Fatty acids are carboxylic
acids (RCOOH) with long-chain hydrocarbon side groups, usually ranged between 14 and 20
carbon atoms. In Table 1 are listed some common biological fatty acids. In higher plants and
animals, the predominant fatty acid residues are those of the C16 and C18 species [6] named
palmitic, oleic, linoleic and stearic acids (Tab. 1). Over half of the fatty acid residues of plant
and animal lipids are often polyunsaturated (containing two or more double bonds). They
are symbolized by PUFAs (Poly Unsaturated Fatty Acids). For example, 18:2 means that the
hydrocarbon chain contains 18 carbon atoms and 2 carbon-carbon double bonds (see linoleic
acid in Tab. 1 and Fig. 1A). Almost all fatty acid double bonds have the cis-configuration as it
can be seen for linoleic acid as an example, in the Figure 1A.
Ä'*%Ä
2ADIATION INDUCEDLIPIDDAMAGE
Table 1. Names and formulas of fatty acids.
Name1 Symbol2 Structure
Unsaturated fatty acids
Palmitoleic acid 16:1 CH3(CH2)5CH=CH(CH2)7COOH
(9-Hexadecenoic acid)
Oleic acid 18:1 CH3(CH2)7CH=CH(CH2)7COOH
(9-Octadecenoic acid)
Linoleic acid 18:2 CH3(CH2)4(CH=CHCH2)2(CH2)6COOH
(9,12-Octadecadienoic acid)
G-Linolenic acid 18:3 CH3(CH2)4(CH=CHCH2)3(CH2)3COOH
(6,9,12-Octadecatrienoic acid)
Arachidonic acid 20:4 CH3(CH2)4(CH=CHCH2)4(CH2)2COOH
(5,8,11,14-Eicosatetraenoic acid)
Saturated fatty acids

Palmitic acid 16:0 CH3(CH2)14COOH
(Hexadecanoic acid)
Stearic acid 18:0 CH3(CH2)16COOH
(Octadecanoic acid)
1) Common name (systematic name in brackets). 2) Number of carbon atoms : number of carbon-carbon double bonds. The systematic
name of fatty acids derives from the total number of carbon atoms of the chain (for example, “hexadeca” for 16 atoms, “octadeca” for
18 atoms,…), the other numbers designing the carbon atoms bearing the carbon-carbon double bonds. For example, linoleic acid is the
9,12-Octadecadienoic acid which has 2 double bonds beared on the 9th and 12th carbon atoms, the 1st being attributed to the carboxylic
acid function (Fig. 1A).
Figure 1 : Structures of some

membrane lipids. A) Linoleic
acid as an example of PUFA.
B) General structure of glyce-
rophospholipids, R1, R2 being
long chain fatty acid residues
and X a polar group ; the
glycerol backbone is inside
the dashed line. C) Formula
of sphingosine, the simplest
member of the sphingolipids
family. D) Cholesterol which
can be esterified to long
chain fatty acid on its C3 OHt
group.
Ä'*&Ä
2!$)!4)/.#(%-)3429
Glycerophospholipids are, with sphingolipids, the major lipid components of

biological membranes. Glycerophospholipids (general formula in Figure 1B) derive from
glycerol (CH2OH-CHOH-CH2OH) by esterification to two long-chain fatty acid residues (R1
and R2 in Figure 1B), the third alcohol function being transformed into a phosphoryl-X group.
X is a polar group such as an alcohol or an amine. Sphingolipids are derivatives of the amino
alcohol sphingosine (Fig. 1C). In this latter class of lipids, there are several different families
such as ceramides, sphingomyelins and sphingoglycolipids with complex head groups of up
to four sugar residues.
Cholesterol (Fig. 1D), which is the metabolic precursor of steroid hormones, is a

major component of plasma membranes. Its fused ring system brings a greater rigidity than
other membrane lipids. Cholesterol can be esterified on the C3 OH-group to long-chain fatty
acids to form cholesteryl esters which are major components of lipoproteins.
Lipid aggregates and model systems
The common feature of lipids is their amphiphilic structure, characterized by

a hydrophobic part (the non-polar aliphatic acid residue(s)) and a hydrophilic part (for
example, the polar carboxylic acid function in fatty acids or the phosphoryl head in
glycerophospholipids, …). When lipids are dissolved in water, they spontaneously aggregate,
the non-polar tails being associated by weak Van der Waals interactions. Single tailed lipids
as PUFAs tend to form micelles (Fig. 2A) which are spheroidal aggregates where the polar
carboxylate heads are in contact with water. Micelles appear in aqueous solution when the
fatty acid concentration surpasses the critical micelle concentration (symbolized by cmc),
namely the upper limit concentration above which monomers aggregate. The cmc depends
on the nature of the fatty acid and on the solution conditions (pH, temperature, …).
For example, at room temperature, the value of linoleate cmc is close to 2 s 10−3 mol l−1 in
aqueous solution at pH = 10.5 [7]. This means that at concentration below 2 s 10−3 mol l−1,
linoleate anions cannot form micelles but only very small aggregates (oligomers) and
monomers, whereas above 2 s 10−3 mol l−1 they are mainly associated into micelles. The basic
pH (10.5) of the aqueous medium is necessary in order to deprotonate the carboxylic acid
function of PUFAs giving carboxylate groups COO−, namely polar charged heads (surrounded
by polar water molecules). PUFAs micelles in aqueous solution are the simplest lipidic models
allowing a mechanistic approach of lipid peroxidation under ionizing radiations (see section
“Quantitative determination of hydroperoxides”).
Glycerophospholipids and sphingolipids are biological lipids which have two large
hydrophobic tails, one of them being often a PUFA chain. They exhibit a very low cmc,
Ä'*'Ä
Figure 2 : Lipid aggregates. A) Micelle, spheroïdal aggregate of single tailed fatty acid, the polar heads being
in contact outside with water molecules. B) Bilayer, two dimensional fluid, composed of glycerophospholipids or
sphingolipids, each molecule possessing two hydrophobic tails. C) Liposome, spheroïdal vesicle filled with water,
bounded by a single bilayer. D) LDL, lipids/protein/antioxidants aggregate, carrier of cholesterol in blood plasma.
generally lower than 10−6 mol l−1, and tend to form bilayers (Fig. 2B). Lipid bilayers are two
dimensional fluids in which lipid molecules can diffuse lateraly but not across the bilayer [6].
Liposomes are vesicles filled with water, constituted of several bilayers (multilamellar vesicles,
MLV), which can rearrange to form unilamellar vesicles bounded by a single bilayer (Fig. 2C).
According to the size distribution, one considers the small unilamellar vesicles (SUV) with a
diameter of about 50 nm and the large unilamellar vesicles (LUV) whose diameter is superior
to 100 nm. The liposome size distribution depends largely on the method of production of
the vesicles, the sonication (agitation by ultrasonic vibrations) giving mainly SUVs, whereas
the extrusion (submission to high pressures) generates LUVs. SUVs being highly curved, the
unsaturated fatty acyl chains have a better exposure to water and consequently SUVs are
highly oxidizable via radiation-induced water free radicals. By contrast, LUVs (high diameters)
characterized by a tight packing of phospholipid molecules are less penetrable by water
Ä'*(Ä
2!$)!4)/.#(%-)3429
molecules and consequently less oxidizable than SUVs [8]. Liposomes can be considered as rather
good model systems since they reflect the arrangement of lipid bilayers of cell membranes.
However, to get still closer the living membranes, it is necessary to add proteins
into the lipid matrix. Lipoproteins are suitable candidates to study lipid peroxidation and
protein oxidation under ionizing radiation, allowing to specify lipid/protein interactions. Low
Density Lipoproteins (LDLs) (Fig. 2D) isolated from human plasma, are composed of a single
big protein (apolipoprotein B-100) surrounded by lipids (choline phospholipids, cholesteryl
esters and triacylglycerols). In addition, antioxidants (mainly A-tocopherol and B-caroten)
are located in the lipidic part. LDLs are now considered as very good membrane models
allowing to determine the reciprocal influence of the oxidation of both lipid and protein
moieties at the molecular level, and to specify the role of antioxidants under oxidative stress
conditions, specially those of water radiolysis [5, 9]. It can be noticed that other membrane
models such as erythrocyte membranes have been investigated under radiation exposure,
providing interesting results on the post-irradiation effects to the lipid and protein moieties,
together with structural changes in the membrane arrangement (see for example [1]).
Water radiolysis
When liposome or micelle suspensions in diluted aqueous medium, are irradiated

by G-rays (from a radioactive source of 137Cs or 60Co), they are attacked by the free radicals
generated by water radiolysis. Indeed, there is no direct ionisation of the lipids by the G-rays
if their concentration remains lower than 10−2 mol l−1, whereas the molecules of water (in
high concentration of 55 mol l−1 in aqueous solutions) are the only molecular targets to be
ionized (Chapter 1).
The radiolysis of water leads to the formation of the three radical species tOH
(hydroxyl radical), Ht (hydrogen atom) and eaq− (hydrated electron), within the nanosecond
time scale [3]. In aerated medium (that is in the presence of dioxygen, O2 concentration
dissolved in water being 2 s 10−4 mol l−1), the free radicals Ht and eaq− are replaced by HO2t
and O2t¦ radical species, respectively, which are related by an acid/base equilibrium (pKa
(HO2t/O2t¦) = 4.8). The radiolytic yields (G-values expressed in moles per Joule) of each radical
species are well known: 2.8 s 10−7 mol J−1 and 3.4 s 10−7 mol J−1, respectively for tOH and O2t¦
free radicals at pH = 7 (Chapter 1) [3].
Hydroxyl (tOH) and superoxide (O2t¦) radical species are thus the protagonists of
the oxidative stress initiated by water radiolysis to diluted biological targets [10]. But they
exhibit very different properties against lipids. Hydroxyl free radical is the strongest oxidant
Ä'*)Ä
species because its one-electron reduction potential is very high (E° tOH/H2O = 2.34 VNHE at
pH = 7, versus the potential of normal hydrogen electrode (NHE) as reference, under standard
conditions) and its second order rate constants are diffusion controlled, namely close to the
upper limit of 1010-1011 l mol–1 s–1 (Chapter 1). In contrast, superoxide free radical is not an
efficient initiator of lipid oxidation. Indeed, even if its one-electron reduction potential is
high enough (E° O2t¦/H2O2 = 0.93 VNHE at pH = 7), its second order rate constants against
lipids are very low, generally lower than 102 l mol–1 s–1 [11]. In other words, in aqueous
medium, superoxide free radicals do not attack lipids. However, it has been shown that their
protonated form, HO2t, can react more rapidely with PUFAs (103 l mol–1 s–1 [12]) than does O2t¦.
Note that at pH = 7, HO2t radicals represent less than 1% of the total amount of O2to radicals
(pKa (HO2t/O2t¦) = 4.8) in the irradiated medium. Hence, it can be assumed that at pH = 7,
the predominant initiating species are hydroxyl radicals.
Radical-chain mechanism of lipid peroxidation
Hydroxyl radicals are able to initiate the one-electron oxidation of unsaturated acyl
chains symbolized here by LH (as linoleic acid (18:2), for example), by abstracting a single H-atom
from the aliphatic chain leading to a carbon-centred radical Lt (alkyl radical) (reaction (1)).
-)tOH mLt)2O (1)
The removed hydrogen atom is preferentially located on a methylene group (CH2)

between two adjacent C=C double bonds (bis-allylic position) (see Fig. 3 for linoleic acid).
Indeed, such C-H bonds are weaker than others along the saturated carbon chain and they
represent critical target sites [13,14].
The fate of the carbon-centred radical Lt (which is a resonant pentadienyl radical in the case
of linoleic acid, Fig. 3) is to lead, in a very rapid reaction with dioxygen, to the formation of a
peroxyl radical, LO2t (reaction (2)).
Lt 02 mLO2t (2)
In the case of linoleic acid, two isomeric peroxyl radicals are formed (Fig. 3). Such
peroxyl radical species are expected to abstract H-atoms from other fatty acid molecules
(reaction (3)) giving lipid hydroperoxides LO2H (9 and 13-hydroperoxyoctadecadienoic acid
in the case of linoleic acid (Fig. 3 )), which are the initial non-radical products formed during
lipid peroxidation.
LO2t-)mLO2)-t (3)
Ä'**Ä
2!$)!4)/.#(%-)3429
Figure 3: Mechanism of peroxidation of linoleic a cid (18:2) initiated by tOH free radicals from water radiolysis.
LH symbolizes another molecule of linoleic acid in micelles. 9- and 13-HPODE are the main hydroperoxides
formed, possessing a conjugated dienic structure.
The resulting alkyl free radical Lt can in turn react with dioxygen (reaction (2)),
propagating the radical-chain mechanism (reactions (2) and (3)) inside the lipid aggregates.
These radical-chain reactions amplify the peroxidation process since one single tOH radical
can lead to several alkyl radicals (Lt), peroxyl radicals (LO2t) and hydroperoxide molecules
(LO2H). To stop the radical-chain mechanism, two radical species have to react together
(for example reactions (4) and (5), reaction (4) being the most probable in an oxygenated
medium) leading to non-radical products.
Ä'*+Ä
LO2t-02t m non-radical products (4)

LO2t-t mnon-radical products (5)
From this free radical-chain mechanism it appears that hydroperoxides (LO2H) are
the major early reaction products of lipid peroxidation. Their structure are very diversified
since it depends on the number of allylic hydrogen atom(s) to be abstracted along the acyl
chain of the PUFA (see next section). However, the lipid degradation process does not stop
after the formation of hydroperoxides. Indeed, hydroperoxides are not very stable products
(apart from irradiation) and they undergo spontaneous decomposition (oxidative cleavage)
that generates numerous breakdown products (see section “Oxidative fragmentation
of lipids”).
Quantitative determination of hydroperoxides,

initial markers of lipid peroxidation
In micelles
Hydroperoxides (ROOH) are the next non radical products to be formed in aerated
medium, after tOH-attack on the unsaturated carbon chain. Hence their quantification
as a function of radiation doses is particularly of interest. They can be specifically titrated
by reverse-phase HPLC (High Pressure Liquid Chromatography) using detection by
chemiluminescence [15,16]. As it can be shown on a simple lipidic model such as linolenate
(18:3) micelles, increasing radiation doses lead to increasing hydroperoxide concentrations
(Fig. 4). This ROOH formation is linolenate concentration-dependent, i.e. at a given radiation
dose, hydroperoxide concentration is higher in the presence of 25 mM than of 7.5 mM
linolenate (Fig. 4). The highest part of hydroperoxides comes generally from the H-abstraction
on the bis-allylic positions of PUFAs (see section above). Thus conjugated dienes are
simultaneously formed on the unsaturated carbon chain. Hence conjugated dienes (CD) are
also initial markers of lipid peroxidation. They can be determined by spectrophotometric
absorption at 234 nm (maximum absorption wavelength), with a molar extinction coefficient
of 28000 l mol–1 cm–1 [17]. As it can be seen in Figure 4, the concentration of conjugated
dienes formed in linolenate micelles are increasing with the radiation dose, and they are,
like hydroperoxides, linolenate concentration-dependent. However, for a given linolenate
concentration, the CD concentration is lower than that of hydroperoxides whatever the
radiation dose, meaning that some hydroperoxides do not possess a dienic structure.
Ä'*,Ä
2!$)!4)/.#(%-)3429
Figure 4: Formation of hydroperoxides and conjugated dienes (CD) as a function of the radiation dose (dose
rate = 9.8 Gy .min−1) in linolenate (18:3) aqueous micellar solutions at pH = 10.5. Influence of the concentration
of linolenate (25 mM and 7.5 mM) on the peroxidation process [16].
The radiolytic yields of formation of hydroperoxides (G(hydro)) are determined from

the slope of the initial tangent of the curves (concentration versus radiation dose, as in Fig. 4).
An important lipid concentration dependency is observed in micellar medium where, above
the cmc, the yields of oxidized products formation increase steeply with PUFA concentration.
This means that a chain-mechanism is involved in the formation of hydroperoxides and
conjugated dienes. As an example, in the case of a 25 mM linolenate concentration (Fig. 4),
the G-values are 20.0 s 10−7 mol J−1 and 9.3 s 10−7 mol J−1 for hydroperoxides and conjugated
dienes, respectively. These yields are considerably larger than GtOH = 2.8 s 10−7 mol J−1, since
the hydroxyl radicals are initiators of the oxidation chain mechanism.
It has been demonstrated, not only with linolenate (18:3) aqueous solutions but
also with linoleate (18:2), and more recently with arachidonate (20:4) solutions [18], that
the hydroperoxides were composed of two different kinds of molecules according to the
concentration of PUFA in the irradiated medium. Indeed, for a given PUFA, above the critical
micellar concentration, namely when micelles are formed, one type of hydroperoxide
predominates, whereas when monomers (or very small aggregates) are dispersed in solution
(below the cmc), other types of hydroperoxide are formed. This phenomenon is illustrated in
Figure 5 for arachidonate. We can see that monohydroperoxides are produced in micelles whereas
in monomers, either a cyclic hydroperoxide or an aliphatic dihydroperoxide is obtained [18].
Ä'*-Ä
Such differences seem to arise from the mechanism of their formation which privilegiates a
self-rearrangement in monomers, instead of an intermolecular reaction in aggregates. The
ratio of “micellar” hydroperoxide concentration over “monomer” hydroperoxide concentration
appears to be a signature of the balance micelles/monomers in aqueous medium [18].
Figure 5: Proposed structures of hydroperoxides formed in arachidonate (20:4) aqueous solutions

(pH = 10.5) submitted to G-rays. Monohydroperoxides (6 isomers in positions 5, 8, 9, 11, 12, 15) are formed in
micellar medium (high arachidonate concentration) whereas monocyclic and bicyclic hydroperoxides, together
with dihydroperoxides are produced in dispersed monomers (low arachidonate concentration) [18].
In liposomes
In the case of liposomes of PLPC (1-palmitoyl-2-linoleoyl-sn-glycero-3-phosphocholine)

submitted to free hydroxyl radicals attack (generated by G-radiolysis of water), phospha-
tidylcholine hydroperoxides are the initial markers of lipid peroxidation. Since the molecule of
PLPC contains only one bis-allylic position located on the linoleyl acyl chain, the tOH-induced
hydroperoxides are characterized by the OOH groups at positions 9 and 13 on the linoleyl chain,
as in the case of linoleate micelles (Fig. 3). Moreover, these hydroperoxides exhibit a conjugated
dienic structure. A G-value of total hydroperoxide formation has been determined to be equal
to 6 × 10-7 mol J−1, in G-irradiated small unilamellar vesicles (SUVs) of PLPC 250 μmol l−1 [19].
Such a G-value which is 2 fold the G-value of tOH-formation, also indicates a chain-mechanism
of hydroperoxide production. In addition, other products such as hydroxides, epoxides and
Ä'*.Ä
2!$)!4)/.#(%-)3429
fragments have been identified using HPLC/MS analysis. In LUVs (high diameters), because of
the tight packing of PLPC molecules, tOH radicals have to react first with the polar head before
reaching the linoleyl chains, and for this reason kinetic results are very different between LUVs
and SUVs peroxidation, LUVs being less oxidizable than SUVs [8].
Oxidative fragmentation of lipids
Lipid hydroperoxides are not very stable products and their decomposition can
lead to the formation of a lot of breakdown products such as short-chain aldehydes (Fig. 6),
epoxides, ketoaldehydes, and alkanes (pentane, hexane, for example). These hydroperoxide
decomposition may result from the action of traces of reducing metal cations (Fe, Cu) which
catalyses the cleavage of hydroperoxides according to “Fenton like reactions” (reaction (6)),
in which LO2H acts like H2O2. The alkoxyl radicals LOt so formed are able to abstract H-atom
leading to hydroxylated products (LOH) or to generate epoxy-allyl radical by a self-addition
on a double bond or still to undergo a scission giving aldehydic products [13] :
LO2)'F mLOt−0)'F (6)
Consequently, in addition to hydroperoxides, a lot of secondary oxidized lipidic

compounds, mainly short-chain aldehydes, may appear and represent late markers
of lipid peroxidation. As an example, malondialdehyde (MDA, Fig. 6) is known to be the
most abundant lipid peroxidation aldehyde whose determination by 2-thiobarbituric acid
(TBA) is one of the most common assays in lipid peroxidation studies [20]. However, it can
be noticed that the TBA assay method [21] is not specific of MDA titration since it also can
detect a variety of peroxides and secondary degradation products of lipid peroxidation called
Figure 6: Short-chain aldehydic products resulting from the oxidative fragmentation of lipid hydroperoxides.
Ä'+%Ä
2-thiobarbituric acid-reactive substances (TBARS). Hence, the determination of TBARS must

be used as a global index of lipid peroxidation. Aldehydic molecules are involved in cytotoxic
processes because they can easily react with DNA, protein, and phospholipids leading to
covalent adducts. In the case of the 4-hydroxy-2-nonenal (HNE, Fig. 6) which is produced during
peroxidation of PUFAs such as linoleic and arachidonic acids, HNE is known to react with amino
acids (cystein, histidine, lysine) and with cellular target proteins, inactivating enzymes [20].
However, hydroperoxides decomposition is not the only process undergoing short-

chain lipid fragments. Indeed, G-radiolysis of multilamellar liposomes of glycerophospholipids
such as cardiolipin, phosphatidylglycerol and phosphatidylinositol, has been characterized
by the formation of phosphatidic acid resulting from the cleavage of the phosphatidylglycerol
moiety, as the main product [4]. Hence the polar heads of glycerophospholipids appear to be
targeted by tOH-attacks. The determination of phosphatidic acid has been performed using
matrix-assisted laser desorption/ionisation time-of-flight (MALDI-TOF) mass spectrometry
[4]. The formation of carbon-centred free radicals, in B-position with respect to the
phosphoester bond and containing a hydroxyl group, seems to be a key step of this type of
fragmentation process. A similar mechanism has been proposed to explain the formation
of lysophosphatidylcholine in G-irradiated liposomes of PLPC (LUVs) [8]. Radiation-induced
free radical fragmentation of lipids in their polar moiety seems to play an important role as
signaling molecules in living systems [4].
Peroxidation of low density lipoproteins
It has well been established that PUFAs contained in phospholipids and cholesteryl
esters (see composition of LDLs in section “Lipid aggregates and model systems”, and in Fig. 2D)
are targeted sites of radiation-induced oxygen free radicals attack in LDL particles.
Mechanisms of lipid peroxidation of LDLs, initiated by oxygen free radicals such as tOH, tOH/
O2t¦, O2t¦/HO2t have been proposed, based on the determination of the radiolytic yields of
consumption of A-tocopherol and B-carotene (endogenous antioxidants of LDL), and of
formation of conjugated dienes and TBARS (for a review see [5]). However, as we have already
seen, TBARS are not specific markers of lipid peroxidation (section above). Recently, the
determination of defined hydroperoxides such as hydroperoxides of phosphatidylcholines
(linoleate, arachidonate and docosohexaenoate) and of cholesteryl esters (linoleate and
arachidonate) have been developed using HPLC coupled with a chemiluminescence detection
[9]. These hydroperoxide formations are kinetically characterized by three steps, namely (1)
at low radiation doses, the inhibition step resulting from the presence of A-tocopherol and
B-carotene, able to protect LDL by scavenging radiolytic-induced oxygenated free radicals,
(2) at higher radiation doses, the propagation step resulting from the lipid peroxyl radicals
Ä'+&Ä
2!$)!4)/.#(%-)3429
formation (LOOt) by a chain-mechanism, leading to hydroperoxides of phosphatidylcholines

and of cholesteryl esters, and (3) the chain-termination and decomposition phase. The kinetic
behaviour of conjugated dienes is similar to those of hydroperoxides, the propagation step
beginning at the end of the inhibition step, when there is no more antioxidants. As an example
of quantification of lipid markers for LDLs 3 g l−1, the radiolytic yield of CD formation (3.4 s
10−7 mol J−1) is found slightly higher than GtOH (2.8 s 10−7 mol J−1) due to the presence of a
short chain reaction, and the G-value of hydroperoxide formation is 10 times lower than the
yield of CD production, meaning that conjugated dienes other than hydroperoxides (such as
alcohols, epoxides and endoperoxides) are taken into account [9]. Besides, it has been shown
that the phosphatidylcholines located on the outer surface of LDLs were more accessible than
the cholesteryl ester molecules (in the lipidic core), to the free radicals attack.
In addition, the protein moiety contained in LDLs (apoB-100) is also targeted by tOH
radicals, since western blot immunoassays allow to detect carbonyl groups (C=O) into protein
side chains [9]. These assays involve the derivatization to 2,4-dinitrophenylhydrazone of the
carbonyl groups by reaction with 2,4-dinitrophenylhydrazine (2,4-DNPH). It has been shown
that carbonylated apoB-100 appeared at low radiation doses, during the inhibition phase of
CD and hydroperoxides, meaning that the protein was not protected by the endogeneous
antioxidants which are localized in the lipid part of LDLs and that the polypeptidic chain was
attacked by tOH free radicals. This apoB-100 carbonylation was radiation dose-dependent, its
radiolytic yield being relatively low (1.8 × 10–8 mol J–1) because tOH free radicals simultaneously
react on all other molecular components of LDLs (lipids, antioxidants, …). Moreover, the
fragmentation of the carbonylated apoB-100 may occur, but this process which only begins
during the lipid peroxidation propagation phase, seems to be related to the interactions
between the oxidized parts of lipids and protein. Hence, whereas radiolytic-induced
apoB-100 carbonylation appears to be independent on lipid peroxidation, in contrast oxidative
fragmentation seems to be dependent on it. All these results obtained from radiolysis
conditions represent an accurate and rigourous approach of the oxidative stress phenomena
which can be responsible of the in vivo peroxidation of LDLs leading to atherosclerosis [5].
Conclusion
Effects of ionizing radiation on lipid molecules have been understood by studying

model systems which are simpler than the real biological membranes, such as PUFA micelles
and liposomes. The formation of lipid oxidative modifications of PUFAs appears as a dynamic
process initiated by hydroxyl free radicals generated by water radiolysis, amplified by a
propagating-chain mechanism involving alkyl and peroxyl free radicals, and leading not only
to hydroperoxides but also to a lot of other lipidic oxidized end-products. Kinetic data, such
Ä'+'Ä
as radiolytic yields of production of hydroperoxides and conjugated dienes which are early
products of lipid peroxidation, have allowed us to establish the reaction schemes of their
formation. Knowing that hydroperoxides and other lipid oxygenated products are more hydro-
philic than the initial hydrocarbon chains of PUFAs, their appearance can lead to morphological
changes in the bilayers, together with disturbances in membrane functions. Moreover, hydroxyl
free radicals may initiate the oxydative processes not only within the hydrophobic moiety of
lipids but also in the hydrophilic polar heads leading to fragments such as phosphatidic acid
and lysophosphatidylcholine residues. Such fragments seem to regulate a variety of cellular
functions and to stimulate various signalling pathways. Hence, membrane constituents which
are molecular targets of ionizing radiations have to be considered as important factors the
damages of which are susceptible to largely control the viability of the cells.
Acknowledgments
I am especially indebted to Fabrice Collin for his talented designs of lipid aggregates and
molecular mechanisms and for his judicious comments. I would like to acknowledge Prof. Dominique
Bonnefont-Rousselot for her precious collaboration all along the story of oxidized lipoproteins
(EA 3617, Faculté de Pharmacie, Université Paris 5).
References
[1] Benderitter M., Vincent-Genod L., Pouget J.P., Voisin P., The cell membrane as a biosensor of
oxidative stress induced by radiation exposure: a multiparameter investigation, Radiat. Res.,
2003, 159, 471-483.
[2] Albanese J., Dainiak N., Modulation of intercellular communication mediated at the cell surface
and on extracellular, plasma membrane–derived vesicles by ionizing radiation, Experimental
Hematology, 2003, 31, 455-464.
[3] Spinks J.W.T., Woods R.J., Water and inorganic aqueous systems, in “Introduction to Radiation
Chemistry”, 3rd ed., Wiley, New York, 1990, p. 243-313.
[4] Shadyro O.I., Yurkova I.L., Kisel M., Brede O., Arnhold J., Formation of phosphatidic acid, ceramide,
and diglyceride on radiolysis of lipids: identification by MALDI-TOF mass spectrometry, Free
Radical Biology and Medicine, 2004, 36, 1612-1624.
[5] Bonnefont-Rousselot D., Gamma radiolysis as a tool to study lipoprotein oxidation mechanisms,
Biochimie, 2004, 86, 903-911.
[6] Voet D., Voet J.G., Lipids and membranes (Chapter 11), in “Biochemistry”, 2nd ed., Wiley & Sons,
INC., New York, 1995, p. 277-329.
[7] Hauville C., Rémita S., Thérond P., Rouscilles A., Couturier M., Jore D., Gardès-Albert M.,
Determination of the yield of radiation-induced peroxidation of sodium linoleate in aqueous
monomeric and micellar solutions, Radiat. Res., 1998, 150, 600-608.
Ä'+(Ä
2!$)!4)/.#(%-)3429
[8] Vitrac H., Liposomes de phosphatidylcholine comme modèles de la peroxydation membranaire.

Action des radicaux hydroxyles et mécanismes réactionnels, Thesis, Université René Descartes
(Paris 5), 2004.
[9] Jedidi I., Thérond P., Zarev S., Cosson C., Massot C., Jore D., Gardès-Albert M., Legrand A., Bonnefont-
Rousselot D., Paradoxical protective effect of aminoguanidine toward low-density lipoprotein
oxidation: inhibition of apolipoprotein B fragmentation without preventing its carbonylation.
Mechanism of action of aminoguanidine, Biochemistry, 2003, 42, 11356-11365.
[10] Gardès-Albert M., Jore D., Aspects physicochimiques des radicaux libres centrés sur l’oxygène,
in “Radicaux libres et stress oxydant”, Delattre J., Beaudeux J-L., Bonnefont-Rousselot D., (eds),
Lavoisier, 2005, 1-23.
[11] Bielski B.H.J., Cabelli D.E., Arudi R.L., Ross A.B., Reactivity of HO2/O2– radicals in aqueous solution,
J. Phys. Chem. Ref. Data, 1985, 14, 1041-1051.
[12] Bielski B.H.J, Arudi R.L., Sutherland M.W., A study of the reactivity of HO2/O2– with unsaturated
fatty acids, J. Biol. Chem., 1983, 258, 4759-4761.
[13] Bors W., Erben-Russ M., Michel C., Saran M., Radical Mechanisms in fatty acid and lipid peroxidation,
in “Free Radicals, Lipoproteins, and Membrane Lipids”, Crastes de Paulet A (eds), Plenum Press
New York, 1990, p. 1-16.
[14] Wagner B.A., Buettner G.R., Burns C.P., Free-radical mediated lipid peroxidation in cells: oxidizability
is a function of cell lipid bis-allylic hydrogen content, Biochemistry, 1994, 33, 4449-4453.
[15] Therond P., Couturier M., Demelier J.F., Lemonnier F., Simultaneous determination of the main
molecular species of soybean phosphatidylcholine or phosphatidyletanolamine and their
corresponding hydroperoxides obtained by lipoxygenase treatment, Lipids, 1993, 28, 245-249.
[16] Hindo J., Hauville C., Rémita S., Thérond P., Couturier M., Jore D., Gardès-Albert M., Evidence of the
formation of different hydroperoxides in irradiated gamma-linolenate solutions: effect of micelle
formation, Radiat. Res., 2000, 153, 201-207.
[17] Pryor W.A., Castle L., Chemical methods for the detection of lipid hydroperoxides, Meth. Enzymol.,
1984, 105, 293-299.
[18] Vitrac H., Hauville C., Collin F., Couturier M., Thérond P., Delaforge M., Rémita S., Jore D., Gardès-
Albert M., Hydroperoxide characterisation as a signature of the micelle/monomer balance in
radiation-induced peroxidation of arachidonate, Free Radical Research, 2005, 39, 519-528.
[19] Vitrac H., Courrègelongue M., Couturier M., Collin F., Thérond P., Rémita S., Peretti P., Jore D., Gardès-
Albert M., Radiation-induced peroxidation of small unilamellar vesicles of phosphatidylcholine
generated by sonication, Can. J. Physiol. Pharmacol., 2004, 82, 153-160.
[20] Uchida K., 4-Hydroxy-2-nonenal: a product and mediator of oxidative stress, Progress in Lipid
Research, 2003, 42, 318-343.
[21] Patton S., Kurtz G.W., 2-Thiobarbituric acid as a reagent for detecting milk fat oxidation, J. Dairy
Sci., 1951, 34, 669-674.
Ä'+)Ä
Chapter 18
Predicting radiation damage
distribution in biomolecules
Marie DAVIDKOVA and Melanie SPOTHEIM-MAURIZOT
Introduction
Because of the high proportion of water in living matter, the effect of low LET ionising
radiation on biological systems occurs, as discussed for DNA in Chapter 12, at a large extent
via the reactive species generated by water radiolysis. Among those, the hydroxyl radical tOH
(Chapter 1) attacks all biomolecules mainly by abstracting a H-atom from the biomolecule
(from the sugar moiety of DNA or from the peptide chain of a protein) or by addition to the
double bonds of aromatic moieties (DNA bases or aromatic moieties of protein side chains).
Consequently, DNA strand breaks can occur on one strand (called single strand breaks, SSB)
or on the two strands in spatially close (a few base pairs away) locations (called double strand
breaks, DSB). All strand breaks that can be revealed at neutral pH by classical electrophoresis
methods (agarose gel electrophoresis, sequencing gel electrophoresis) belong to the
category of frank strand breaks (FSB), mainly resulting from the damage of deoxyriboses.
Several base alterations can lead to strand breaks after a subsequent treatment with alkali
and thus can also be observed by the classical electrophoresis methods. They are called alkali
revealed breaks (ARB). In the case of proteins, the peptide chain breaks can also be revealed
by classical electrophoresis methods (denaturing sodium dodecylsulfate-polyacrylamide
gel electrophoresis), whereas the modification of amino-acids side chain can be observed
by spectroscopic methods and mass spectrometry.
Ä'+*Ä
2!$)!4)/.#(%-)3429
The lesions are non-homogeneously distributed along the biomolecules. The

probability to get a lesion at a given nucleotide or amino-acid is determined by the chemical
reactivity of the attack site toward the tOH radical and by the accessibility of the reactive
sites to this radical. The reactivity is different for the different entities of the biomolecules.
For instance, a cysteine is more reactive than a tyrosine which is less reactive than an
alanine. Moreover, the accessibility strongly depends on the tridimensional structure of
the macromolecule. A tyrosine located in the interior of a folded protein is less reactive
than one located on the surface of the protein (see Chapter 16). In the case of the classical
double stranded DNA, the tridimensional structure of the molecules is modulated by the
sequence of nucleotides. For instance, the probability that an adenosine localised in an AATT
sequence reacts with an tOH radical is lower than that of one located in a GCAT sequence,
because the accessibility of the deoxyribose of that adenosine is lower in the narrow minor
groove of the AATT sequence. The accessibility of the tOH radicals to the reactive sites can
also be modulated by the presence of covalently or non-covalently bound molecules. These
molecules can physically protect their binding sites (masking effect) or can modify the
structure of the “host” molecule (conformational effect). For instance, an electrostatically
bound positively charged polyamine, such as spermine, that can compact DNA, locally and
globally radioprotects the negatively charged DNA. The only commercial radioprotector of
normal tissues in the case of the radiotherapy of human tumors, Ethyol® (Amifostine), once
metabolised, gives rise to a positively charged ligand that binds and protects DNA by several
effects : scavenging of free radicals at the binding site, chemical repair of radiation-induced
DNA radicals and DNA structure modification.
But the most important DNA ligands that naturally radioprotect DNA are the proteins.
Chromatin, the constituent of chromosomes in eukaryotes, is in fact a complex of DNA and
of several structuring proteins (histones). The regulation of gene expression in prokaryotes
involves the tight interaction of proteins (repressors) with specific DNA sequences. The repair
of endogenous DNA damages is realised by a set of repair proteins that, in a first step, bind
to the region of DNA bearing the lesion. In the DNA-protein complex protection is mutual :
the protein radioprotects DNA and DNA radioprotects the protein.
Below here we will focus on an original computational method that accounts for
the experimentally observed radioprotection of the partners in DNA-protein complexes.
Validated by comparison between experiment and calculation, it can be used to predict
the damage extent and distribution in any biomolecule or complex of biomolecules whose
tridimensional structure is known.
Ä'++Ä
2ADIATIONDAMAGEDISTRIBUTION
RADACK, an original model of radiolytic attack to biomolecules
To account for the extent and location of lesions in the case of free DNA, several
models of radiation damage were proposed. The main difference between these models
concerns the representation of the DNA molecule. In the early models DNA was represented
as a cylinder [1] or as a structured quasi-cylindrical volume in which bases and sugar-
phosphate domains were delimited [2]. At the beginning of 90s, the first models taking into
account the atomic structure of DNA emerged [3-5], shortly followed by higher order DNA
structures (nucleosome, chromatin) [6-8]. All these models take into account the radiation
track structure (spatial distributions of energy deposition events), which is obtained by
using Monte Carlo method. Although these models take into account the DNA structure,
they investigate only the canonical B-form of DNA.
The RADACK (contraction of RADiation-induced attACK) model, that we have

developed [9,10], accounts for the experimentally determined probabilities of radiolytic
damages caused by the OHt radical attack in all forms of DNA (B [11], Z [12], triplex [13],
quadruplex [14]), in DNA-protein complexes [15] and has the potential to predict radiolytic
attack probabilities in other molecules or assemblies. Direct ionisation effects are not taken
into account. The determination of relative probabilities of reaction of the target with the
t
OH radicals takes into account two factors : 1) the accessibility of the reactive sites of the
target since it uses the exact tridimensional structure of the macromolecule or assembly
as determined by NMR, crystallography or as built up by molecular modelling, and 2) the
chemical reactivity of the residues (nucleotides or amino-acids).
The atoms of the target are represented by spheres with van der Waals radii, or, if
the atom is reactive toward the tOH radical, by a sphere with a larger radius (proportional to
reactivity parameters) called a Smoluchowski’s sphere. tOH radicals, represented by spheres
of 1.2 Å radius, are generated randomly in a volume of interest (a water volume surrounding
the studied structure). The radical diffusion in random directions is simulated using Monte
Carlo method. The radical is displaced by steps of 1 Å. When the radical encounters a non-
reactive atom of the target, it restarts its movement. When it encounters a reactive atom, a
reaction is counted. When the radical does not encounter any atom, it escapes the volume
of interest and no reaction is counted (Fig. 1). A large number of radicals is generated for a
good precision of the simulation procedure. The counting can be done separately for each
reactive site of the molecule and one can thus obtain the relative probabilities of attack
at each reactive site. In order to pass from such a reaction probability to the probability to
get an observable lesion, one applies experimentally obtained coefficients representing the
efficiencies of such transformation. They include all types of processes (e.g. charge transfer
Ä'+,Ä
2!$)!4)/.#(%-)3429
to preferential sites). Here comes into play also the chemical fate of the species resulting from
the reaction between the target and the tOH radical. Their evolution toward the observable
lesions depends on the nature of these species (e.g. base radicals) and on their environment
(radical transfer to a spatially proximal site can occur). Therefore the coefficients of
transformation can be different for the one and same base radical in two different structures
(guanine radicals in a B- or in a Z-form DNA do not have the same coefficient).
Figure 1 : Principle of RADACK. DNA as considered by RADACK : DNA reactive atoms are represented by their
Smoluchowski’s sphere whereas the non-reactive ones, by van der Waals spheres. Colours : H atoms in cyan,
t
C atoms in grey, N atoms in blue, P atoms in orange, O atoms in red. The OH radicals (in green) move by steps
of 1 Å, react or escape from the volume of interest surrounding the biomolecule. They are presented with
increasing intensity of green colour in the direction of diffusion.
Sequence and structure-dependent damage distribution

in free DNA
Sequencing gel electrophoresis of irradiated DNA has shown that the probability of
FSB or ARB production is not identical at all nucleotides of the right handed helical double
stranded B-form of DNA [11,15]. The variations are even stronger in the left handed helical
double stranded Z-form [12]. In the triple stranded DNA [13] or along the DNA quadruplex
[14] , variations of FSB and ARB production are also observed. We have applied RADACK to
all these different forms of DNA, the exact structures of which are available in the database
(Protein Data Bank PDB) or were obtained by molecular modelling. In all cases we have obtained
a very satisfactory agreement between the relative probabilities of FSB and ARB determined
by sequencing gel electrophoresis and the values calculated with RADACK. The modulation of
FSB distribution is mainly due to the variation of the accessibility of H atoms of deoxyriboses,
whereas the distribution of ARB is due both to variations of accessibility of the bases and to
Ä'+-Ä
the higher reactivity of guanine as compared to adenine, tymine and cytosine. Moreover, the
efficiency of the transformation of a DNA radical into a lesion depends on the chemical nature
of the radical and also on the environment in which the radical is situated.
RADACK may also account for the experimentally observed reduced probability of
getting a damage at the site of binding of a minor groove ligand that has a preference for a
specific sequence (a natural polyamine [16] or a therapeutic radioprotector, Ethyol® [17]) as it
does for the experimentally observed effect of an intercalator (ethidium bromide) [18].
Protection of DNA by a specifically bound protein

(radiolytic footprinting)
Sequencing gel electrophoresis patterns of DNA fragments irradiated in presence

of bound proteins reveal regions in which no FSB are occurring. They are called protein
“radiolytic footprints” on DNA [15]. Such protected regions were observed when irradiating
the operator-lactose repressor complexes (see Inset 1).
Inset 1 : Lactose repressor
The E. coli lactose operon (Fig. 2) is one of the most studied and best understood
systems of gene expression and regulation. The lac repressor is a tetrameric protein (4 x
360 amino acids), a dimer of dimers, that binds tightly the lac operator, a DNA segment of
about 25 base pairs, with a quasi-palindromic sequence. This binding prevents the expression
of the structural genes of the lac operon, coding for enzymes involved in lactose metabolism.
Thus, the repressor negatively regulates lactose metabolism. The repressor can also bind non-
operator DNA, but with a strongly reduced (105 times) affinity. This protein is organised in
domains: the tetrameric core (formed by the C-terminal parts of the four protomers), and the
four headpieces (N-terminal peptides of about 60 amino-acids) connected by hinge regions.
Its interaction with each operator sequence occurs by the binding of two headpieces (Fig. 2).
Due to the mutual interaction in the complex, the conformation of the two partners (DNA
and protein) changes with respect to their free state. DNA bends and the unstructured hinge
region of the repressor monomers becomes helical.
The structure (complete or partial) of the free repressor and of the DNA-repressor complexes
was solved by NMR or X-ray crystallography. They are available in the structural database PDB
(Internet address http://www.rcsb.org/pdb/home/home.do) as for instance the entries 1LQC,
1LCC, 1L1M, 1CJG, 1LBG, 1EFA, 1LBI, 1JWL.
Ä'+.Ä
2!$)!4)/.#(%-)3429
RADACK calculations were performed in which structural parameters of the target

extracted from PDB entries were used. For the complex between the lactose repressor
headpiece and a DNA fragment bearing the specific sequence for the repressor one
observes the very good fit between the probabilities of experimental FSB production and
the probability of attack to deoxyriboses (the sites involved in FSB production) (Fig. 2).
Figure 2 : Footprints of lactose repressor on DNA in the irradiated DNA-repressor complex. A) Sequencing
gel electrophoresis of the DNA fragment irradiated alone (lane 1) or in the complex with the repressor (lane 2)
and the structure of the complex extracted from 1LBG entry of PDB Databank). In lane 3, the Maxam-Gilbert
sequencing of purines (position of A and G). The four monomers of the tetrameric repressor are in blue and green
and the DNA segments are in red and yellow. The arrows point toward the regions were no strand breaks are
observed which are the footprints of the protein along DNA. B) Comparison between the relative probabilities of
FSB production deduced from the experiment and the relative probabilities of attack to deoxyribose (leading to
FSB) calculated with RADACK.
Calculations were performed also for two other “objects” : the same fragment but in a
linear form and for the same fragment bent as in the complex (“stripped” fragment). From the
superposition of the three curves (Fig. 3) one can conclude that the “footprint” of the protein on
DNA is almost entirely due to the masking by the protein of its binding site on DNA [19,20].
Ä',%Ä
Figure 3 : Explanation of the repressor footprint on DNA. Pattern of relative probability of attack to deoxyribose
(leading to FSB) along the irradiated linear free DNA (green), along the fragment irradiated in complex with the protein
(structure 1JWL from PDB) (red) and along the fragment bent as in the complex but “stripped” of the protein (blue).
This result validates the experimental method of radiolytic footprinting as a very

interesting tool for revealing the sites of “contact” between the protein and DNA when they
are not known.
The same procedure was applied to another biologically very important DNA-protein
complex, the nucleosome core particle (Inset 2). It is composed by a DNA fragment wrapped
around a globular core constituted of several small proteins called histones (Fig. 4A).
Inset 2 : Nucleosome
The nucleosome core particle NCP (Fig. 4) is the first organisation level of chromatin,
constituent of chromosomes. It is the “bead” in the “beads on a string” model of chromatine.
A nucleosome core particle is formed by a 146 base pairs DNA fragment wrapped 1.65 times
around the histone octamer (formed by two copies of H2A, H2B, H3 and H4 histones). NCP are
connected by regions of naked DNA called linkers. The nucleotide sequences of the 146 bp
DNA fragments involved in nucleosomes are polymorphic. The regions of contact between
DNA and the core of histones are distributed along DNA with a periodicity of around 10.4 bp,
which is the number of base pairs per helical turn of DNA in the nucleosome.
The structure (complete or partial) of the nucleosome core particle NCP was solved by X-ray
crystallography. They are available in the structural database PDB (Internet address http://www.
rcsb.org/pdb/home/home.do) as for instance the entries 1ID3, 1F66, 1AOI, 1KX3.
The sequencing gel electrophoresis of the DNA fragment irradiated in the

nucleosome core particle NCP (DNA bound to the histones) show a regular alternation of
regions (of around 10.4 base pairs, which is also the number of base pairs on one helical
Ä',&Ä
2!$)!4)/.#(%-)3429
turn of B form-DNA) with high and with low FSB production [15]. RADACK calculations using
structural parameters from a PDB entry revealed exactly the same periodicity of variation in
the pattern of reaction probabilities (Fig. 4).
Figure 4 : Footprint of histones on the DNA of irradiated core nucleosome. A) Nucleosome core particle struc-
ture extracted from 1AOI entry of PDB. The 146 base pairs DNA fragment is in green and yellow, and the histone
octamer is in blue and magenta tones. B) Experimental relative probabilities of FSB deduced from the radioac-
tive profile (electrophoresisi) and relative probability of attack to deoxyriboses leading to FSB (calculated with
RADACK). C) The Fourier transform (power) of the probabilities showing that probability variation has a period of
10.4 base pairs both for the experiment and for the calculated values.
Ä','Ä
Thus the good fit between calculations and experiments validates RADACK once
more [21]. Here again calculations were performed for the straight DNA fragment and for
the fragment bent as in the nucleosome. They show that the histones protect DNA from the
attack by OHt radicals at the contact points between DNA and the core of histones (located
every 10.4 base pairs) by the masking effect. DNA bending does not bring an important
contribution to the effect.
Protection of the protein by the specifically bound DNA
When the complexes between the lactose operator and operator DNA [22] or that
between the repair protein and abasic site-bearing DNA [23] are irradiated, the complexes
are destroyed mainly due to the damage to the protein. When irradiated alone the lactose
repressor or the repair protein Fpg lose their ability to bind DNA at a dose that is much
lower than that necessary for destroying the complexes. The explanation lies in the reverse
radioprotection provided to the protein by the DNA. RADACK calculations using the
structural parameters for a free and a complexed repressor headpiece (Fig. 5) reveals the
most probable sites of protein lesions.
Figure 5 : RADACK applied to a DNA-binding protein. Calculated relative probabilities of reaction of OH t radi-
cals with the amino-acids of lactose repressor headpiece along the free irradiated headpiece (structure extracted
from 1LQC entry from PDB) (black) and along the headpiece irradiated in complex with DNA (structure extracted
from 1CJG entry from PDB) (red).
One can observe that for the same dose of radiation the probability at some reactive
sites (the closest to DNA) is lower when the protein is irradiated in presence of DNA [24]. The
same effect is obtained in the case of free and complexed Fpg [23]. Thus applying RADACK
to free and complexed proteins allows the protection of the protein by the bound DNA to
be revealed and quantified. It also shows that this protection is exerted at the contact points
between the protein and DNA. To validate the applicability of RADACK to proteins, a good fit
Ä',(Ä
2!$)!4)/.#(%-)3429
with experiments is necessary. Mass spectrometry experiments that will allow identification
of the damaged and the protected amino-acids are in progress in our laboratory.
Conclusion
The potential of computational methods to account for radiolysis phenomena in

biomolecule is proven by the good agreement between the calculated and experimental
values of relative probabilities of damage to each entity (nucleotide, amino-acid) of free or
assembled DNA and proteins irradiated in aerated solution. When the structure of the target
of radiation is known, a model like RADACK can reveal the sites and the extent of damage.
Thus, in silico methods can be used to predict the behaviour of a biomolecule or assembly
of molecules exposed to ionising radiation with the condition that their tridimensional
structure is known. The described good fit between the calculated and experimental results
when taking into account only the reactions of tOH radicals with DNA supports the idea that
these radicals are the main damaging agents in the case of low LET radiation exposure of
DNA or DNA complexes in aerated solutions.
Acknowledgements
All people having contributed to the research of our two teams on radiation damage to
biomolecules (M. Charlier, former leader of the French team, V. Michalik, former leader of the Czech
team, S. Eon, J. Franchet-Beuzit, N. Gillard, S. Goffinont, V. Isabelle, S. Ruiz, C. Saint-Marc, C. Savoye,
L. Tartier) as well as our coworkers (B. Castaing, F. Culard, D. Sy, C. Prevost, S. Hugot, C. Flouzat,
C. Durand, J. Gras, S. Giliberto) are warmly thanked. The research of our teams got the financial support
of CNRS, INSERM, Czech Academy of Sciences, Électricité de France, Ligue Nationale et Départementale
contre le Cancer, Association pour la Recherche contre le Cancer, Grants Agency of the Czech Academy
of Sciences, bilateral French-Czech programs (CNRS-CRAS, Barrande-EGIDE) and the European action
COST-P9.
References
[1] Charlton D.E., Goodhead D.T., Wilson W.E., Paretzke H.G., The Deposition of Energy in Small
Cylindrical Targets by High LET Radiations, Radiat. Prot. Dosim., 1985, 13, 123-125.
[2] Charlton D.E., Nikjoo H., Humm J.L., Calculation of initial yields of single- and double-strand breaks
in cell nuclei from electrons, protons and alpha particles, Int. J. Radiat. Biol., 1989, 56, 1-19.
[3] Holley W.R., Chatterjee A., Magee J.L., Production of DNA strand breaks by direct effects of heavy
charged particles, Radiat. Res., 1990, 121, 161-168.
Ä',)Ä
[4] Pomplun E., A new DNA target model for track structure calculations and its first application to
I-125 Auger electrons, Int. J. Radiat. Biol., 1991, 59, 625-642.
[5] Zaider M., Bardash M., Fung A., Molecular damage induced directly and indirectly by ionizing
radiation in DNA, Int. J. Radiat. Biol., 1994, 66, 459-465.
[6] Tomita H., Kai M., Kusama T., Aoki Y., Ito A., Monte Carlo simulation of DNA strand breaks induced
by monoenergetic electrons using higher-order structure models of DNA, Int. J. Radiat. Biol., 1994,
66, 669-682.
[7] Andreev S.G., Khvostunov I.K., Spitkovskii D.M., Talyzina T.A., Biophysical modeling of radiation
damage in DNA and chromatin induced by radiation of varying quality, Radiat. Biol. Radioecol.,
1997, 37, 533-538.
[8] Friedland W., Jacob P., Paretzke H.G., Stork T., Monte Carlo simulation of the production of short
DNA fragments by low-linear energy transfer radiation using higher-order DNA models, Radiat.
Res., 1998, 150, 170-182.
[9] Michalik V., Spotheim-Maurizot M., Charlier M., Calculation of hydroxyl radical attack of different
forms of DNA, J. Biomol. Struc. Dyn, 1995, 13, 1-11.
[10] Begusova M., Spotheim-Maurizot M., Sy D., Michalik M., Charlier M., RADACK, a stochastic simulation
of hydroxyl radical attack to DNA, J. Biomol. Struct. Dyn., 2001, 19, 141-158.
[11] Sy D., Savoye C., Begusova M., Michalik V., Charlier M., Spotheim-Maurizot M., Sequence-dependent
variations of DNA structure modulate radiation-induced strand breakage, Int. J. Radiat. Biol. , 1997,
72, 147-155.
[12] Tartier L., Michalik V., Spotheim-Maurizot M., Rahmouni A.R., Sabattier R., Charlier M., Radiolytic
signature of Z-DNA, Nucleic Acids Res., 1994, 22, 5565-5570.
[13] Barone F., Begusova M., La Nave E., Matzeu M., Mazzei F., Sy D., Radiation damage to triplex DNA
induced by gamma-rays: a footprinting study and Monte Carlo simulation, Int. J. Radiat. Biol.,
2000, 76, 731-740.
[14] Begusova M., Tartier L., Sy D., Michalik V., Spotheim-Maurizot M., Charlier M., Monte Carlo
simulation of radiolytic attack to 5’-d[T4G4]4 sequence in a unimolecular quadruplex, Int. J. Radiat.
Biol., 1999, 75, 913-917.
[15] Franchet-Beuzit J., Spotheim-Maurizot M., Sabattier R., Blazy-Baudras B., Charlier M., Radiolytic
footprinting. B-rays, G photons and fast neutrons probe DNA-protein interactions, Biochemistry,
1993, 32, 2104-2110.
[16] Sy D., Hugot S., Savoye C., Ruiz S., Charlier M., Spotheim-Maurizot M., Radioprotection of DNA by
spermine : a molecular modelling approach, Int. J. Rad. Biol., 1999, 75, 953-961.
[17] Sy D., Durand C., Hugot S., Savoye C., Swenberg C., Charlier M, Spotheim-Maurizot M., Sequence
dependence of DNA radioprotection by the thiols WR-1065 and WR-151326, Theor. Chem. Acc.,
1999, 101, 114-120.
[18] Begusova M., Spotheim-Maurizot M., Michalik V., Charlier M., Effect of ethidium bromide
intercalation on DNA radiosensitivity, Int. J. Radiat. Biol., 2000, 76, 1-9.
Ä',*Ä
2!$)!4)/.#(%-)3429
[19] Begusova M., Eon S., Sy D., Culard F., Charlier M., Spotheim-Maurizot M., Radiosensitivity of DNA in
a specific protein-DNA complex: the lac repressor-lac operator complex, Int. J. Rad. Biol., 2001, 77,
645-654.
[20] Begusova M., Giliberto S., Gras J., Sy D., Charlier M., Spotheim-Maurizot M., DNA radiolysis in DNA-
protein complexes: a stochastic simulation of attack by hydroxyl radicals, Int. J. Rad. Biol., 2003,
79, 385-391.
[21] Begusova M., Sy D., Charlier M., Spotheim-Maurizot M., Radiolysis of nucleosome core DNA. A
modelling approach, Int. J. Radiat. Biol., 2000, 76, 1063-1073.
[22] Eon S., Culard F., Sy D., Charlier M., Spotheim-Maurizot M., Radiation disrupts protein-DNA
complexes through damage to the protein. The lac repressor-operator system, Radiat. Res. , 2001,
156, 110-117.
[23] Gillard N., Begusova M., Castaing B., Spotheim-Maurizot M., Radiation affects binding of Fpg
repair protein to an abasic site containing DNA, Radiat. Res., 2004, 162, 566-571.
[24] Begusova M., Gillard D., Sy D., Castaing B., Charlier M., Spotheim-Maurizot M., Radiolysis of DNA-
protein complexes, Radiat. Phys. Chem., 2005, 72, 265-270.
Ä',+Ä
Chapter 19
Chemical protection
against ionizing radiation
Caroline PROUILLAC, Christine AMOURETTE and Ghassoub RIMA
Introduction
Since the discovery of radioactivity, applications in various domains have been

developed. Among them, the study of the effects of radioactivity on living matter gave birth
to a new science, radiobiology. In 1950s, when the memories of Hiroshima and Nagasaki were
still in all minds, a new subject of research appeared concerning the development of chemical
radioprotectors. Indeed, the protective action of some organic compounds against damages
inflicted on biologic organisms by gamma radiation was discovered in 1942 by Dale [1]. He
showed that addition of several substances, such as cysteine NH2CH(COOH)CH2SH or thiourea
S=C(NH2)2, to aqueous solutions of enzymes limits their inactivation by X-rays. Patt, in 1949,
showed for the first time the radioprotective activity of cysteine in vivo. Administered into
bone narrow or intravenously at 175 to 575 mg/kg, cysteine allowed 75 to 89% survival in rats
exposed at a dose of 8 Gy [2]. Two years later, Bacq showed that cysteamine, the decarboxylated
derivative of cysteine, has greater radioprotective activity than cysteine [3].
After these works, the intensive research in chemical radioprotection started by the
synthesis of new compounds which will be efficient with lower toxicity. Several chemical
products showed a fair radioprotective activity, especially sulphur compounds such as
aminothiols, disulfides, phosphorothioates, thiazolidines, aminoethylisothiourea and
dithiolanes. However, whereas these compounds have an interesting radioprotective activity,
Ä',,Ä
2!$)!4)/.#(%-)3429
they also have much toxicity which does not allow clinical use, except in the case of WR 2721
[N-(3-aminopropyl)]- 2-aminoethylphosphorothioate, (H2NCH2CH2CH2NHCH2CH2SPO3H2)].
Therefore WR 2721 was the subject of numerous studies [4-5]. It is marketed for a clinical
use (Ethyol®) in radio- and chimio-therapy. It is currently considered as the most efficient
radioprotector although it causes some adverse effects.
Therefore, several organosilicon, organogermanium and organophosphorus derivatives

were recently developed in our laboratory. They showed interesting radioprotective activity
compared to the starting organic compounds, owing to the presence of organometallic
and organophosphorus groups which modify the electronic structure and the chemical and
biological behaviour of these compounds.
Why a chemical protection against radiation?
Since the use of radioisotopes in nuclear power stations, in anti-cancer radiotherapy

or in nuclear weapons, noxious effects of ionizing radiation on human cells are better known.
According to the amount and distribution of exposure, ionizing radiation can locally eliminate
tumors, but can also damage normal tissues. The biological effects of such radiation result
from chemical processes as ionization or excitation of the biological macromolecules, such
as DNA, either in a direct way or in an indirect way via water molecule radiolysis (Chapter 12).
In both cases, many radical species appear then, which have various consequences on
cellular scales such as mutations of DNA, cellular death, or cancer.
At present, only one way may constitute a real protection against radiation, that
is the physical protection. This method uses screens in order to efficiently attenuate the
radiation. The various types of radiation (alpha, beta, gamma- and X-rays) having different
penetration capacity in the matter, then various screens are thus necessary (Fig. 1).
This method remains the most effective, and adequate shieldings are systematically
used around research, industrial, and medical radiation sources or in space (concrete, water
pools, lead, stainless steel, …) (Chapters 2, 6, 11). However, the method is difficult to implement
to avoid any risk under spread hostile irradiation conditions, for example in the event of a nuclear
accident. This is the reason why the objective of different laboratories was for a few years to
set up a personal chemical protection against radiation. The aim of this method is to protect
the system in a way, as effective as possible, by administering a molecule before exposure,
which could limit the biological consequences of the ionizing radiation. Radioprotective
agents act either by scavenging free radicals produced by irradiation or by restoring damaged
biomolecules. In the event of a nuclear accident, this simple and practical method could thus
Ä',-Ä
$E VELOPMENTOFRADIOPROTECTIONAGENTS
be applied to the greatest number of people. Various types of radioprotectors have been
developed in our laboratory. We present here only some of them.
Figure 1: Penetration of the different types of ionizing radiation.
What are the criteria for an efficient radioprotector?
It is essential to establish certain criteria to evaluate the synthesized compounds,

in particular their radioprotective effectiveness and their toxicity. For each of the molecules
we synthesize, first we evaluate its acute (48 hours) and medium-term (2-30 days) toxicity in
male Swiss CD1 mouse. The molecule is injected intraperitonally at different concentrations
to determine the LD50tox/48h, defined as the concentration which kills 50% of the animals at
48 hours. Then, radioprotective effectiveness is determined in the same animal model. In a
preliminary study, the product is administered at the maximum tolerated dose (defined as
½ LD50tox/48h) 15 or 90 minutes before radiation exposure at doses LD100irr/30d and LD100irr/30d
2 Gy (LD100irr/30d is defined as the irradiation dose which kills 100% of the animals 30 days after
exposure). Survival is observed for 30 days. In a second experiment we determine the Dose
Reduction Factor (DRF) for the most radioprotective compounds. DRF is the ratio between
the LD50irr/30d of treated mice and that of non treated mice.
LD50irr/30d protected animals

DRF =
LD50irr/30d non protected animals
It must be higher than 1.3 for clinical use and ideally higher than 2 for management
of a nuclear accident. Of course, an ideal radioprotector will have to demonstrate other
characteristics : compatibility with other drugs, effectiveness for various types of radiation,
low cost, possible synthesis in large quantity, stability… The aim of this work was to study
the incorporation of potentially radioprotective organic compounds in organometallic
Ä',.Ä
2!$)!4)/.#(%-)3429
structures such as the metallathiazolidines, metalladithioacetals and phosphorothioates in

order to decrease their toxicity and to increase their radioprotective activity. It was then to
vectorise the basic organic substances to their target in biological organism.
Organometallic radioprotectors
Metallathiazolidines
Many chemical structures have been studied in the field of chemical protection
against radiation. Among them, cysteamine and methylcysteamine represent simple organic
compounds having an interesting radioprotective activity due to their radical scavenging
and chemical repair ability [3,6]. Nevertheless, these substances present a relative toxicity
(LD50tox/48h ranging between 150 and 250 mg/kg). Thiazolidine structure is one of the most
studied models [6,7], especially because of the prolonged or delayed action it confers to the
radioprotector effect which is however not hardly better than that of cysteamine. The general
structure of these modified cysteamine and methylcysteamine molecules is presented
in Figure 2.
Figure 2 : Structures of thiazolidines derived from cysteamine and methylcysteamine.
In addition, previous works by Voronkov [8,9] have highlighted that the

incorporation of silicon into certain biologically active organic molecules was of great
interest by an exaltation of the required properties. Thus, for a few years, we have studied
thiazolidines with radioprotective activity, of which we modified the structure, by replacing
carbon in position 2 of the cycle by a metal of the group 14 (silicon or germanium). We drew
aside tin and lead because of their intrinsic toxicity. In this spirit, we thus synthesized and
studied sila- and germathiazolidines derived from cysteamine or methylcysteamine, and
germaselenazolidines derived from selenocysteamine or methylselenocysteamine, with the
aim of comparing them with their purely organic equivalents. The general structure of these
compounds is presented in Figure 3.
Ä'-%Ä
Figure 3 : Structure of metallathia-

zolidines [NEI = 2-(1-naphthylethyl)-
2-imidazoline].
The metallathiazolidines are likely, by hydrolysis or cleavage, to release cysteamine,

methylcysteamine or naphthylethylimidazoline. In Table 1, results of toxicological and radio-
pharmacological studies are represented.
Table 1 : Toxicity and radioprotective activity of metallathiazolidines, germaselenazolidines and their

parental organic derivatives [NEI= 2-(1-naphthylethyl)-2-imidazoline].
Metallathiazolidines and Germaselenazolidines

Irr.
LD50tox/48h Survival at 30 days
M X R1 R2 R3 R4 dose
(mg/kg)a (Gy) (%) [t, min]b
40 [15]
1 Si S i-C5H11 i-C5H11 CH3 -CH2CH2NEI 150 8.1
30 [90]
30 [15]
2 Si S n-C6H13 n-C6H13 H -CH2CH2NEI 300 8.1
70 [90]
40 [15]
3 Ge Se i-C5H11 i-C5H11 H H 255 8.1
60 [90]
30 [15]
4 Ge Se i-C5H11 i-C5H11 CH3 H 236 8.1
50 [90]
30 [15]
5 Ge Se n-C6H13 n-C6H13 H H 224 8.1
50 [90]
30 [15]
6 Ge Se n-C6H13 n-C6H13 CH3 H 200 8.1
40 [90]
Parental organic derivatives
Irr.
LD50tox/48h Survival at 30 days
Compounds dose
(mg/kg)a (%) [t, min]b
(Gy)
7 NEICH2CH2NHCH2CH2SH 300 8.1 0 [15]
8 NEICH2CH2NHCH2CH(CH3)SH 212 8.1 0 [15]
9 H2NCH2CH2SeH, HCl 17 8.1 0 [15]
10 H2NCH2CH(CH3)SeH, HCl 10 8.1 0 [15]
a) the administered dose is ½ LD50tox/48h. b) (t) = time between administration of the compound and irradiation.
Ä'-&Ä
2!$)!4)/.#(%-)3429
From this study, if we compare homologous sila- and germathiazolidines, we observe

that silathiazolidines are generally more protective against radiation (higher survival for the
same delay), and have toxicities close to those of the germathiazolidines. These results are
in agreement with the concept of vectorization, but the active substance seems not to act
in the same way in the case of sila- and germathiazolidines. The sila- and germathiazolidines
could thus have different modes of action in the biological environment.
Certain characteristic features can be attributed to sila- and germathiazolidines :
– silathiazolidine structure allows administration of the basic purely organic product,

by exalting its radioprotective properties, and by decreasing its molar toxicity ; i.e.
the LD50tox/48h of silathiazolidines allows injection of more organic compound than
for isolated basic derivative. The parental organic compound is thus potentiated.
– in the case of the silathiazolidines, it appears that the brittleness of the Si-N bond
probably supports the vectorization of the organic radioprotector. In a previous
work [10], it has been established that the purely organic thiazolidines are generally
less active in protection against radiation than their silicon or germanium isologs.
Moreover, previous research has highlighted the mechanism of action of the
organic thiazolidines, a specific isotopic marking having allowed a pharmacokinetic
follow-up of the products after their administration : thiazolidines are hydrolyzed
in ketone and cysteamine which is partly responsible of the protection against
radiation. This observation could, according to the concept of biodisponibility,
explain the classification of the activities of the various thiazolidines, by the ease
of cleavage of the chemical bonds : Si-L > Ge-L > C-L.
Because selenium plays a fundamental role as a biological cofactor of glutathion

peroxidase which protects cellular membranes, nucleic acids and proteins against the
degradation by free radicals, we have also synthesized and studied germaselenazolidines
and germadiselenoacetals derived from selenocysteamine or methylselenocysteamine,
Metalladithioacetals
It is now well-known that the linear compounds having the sequence of
atoms as S-C-C-S have an interesting radioprotective activity [10]. We thus synthesized
metalladithioacetals for the same reasons as those which led us to consider the study of
the metallathiazolidines (modification of a structure with interesting biological properties,
to increase the activity and to decrease toxicity, incorporation of a metal of the group 14 (Si or Ge),
Ä'-'Ä
to exalt the required effects). We were interested in the synthesis and the study of sila- and
germadithioacetals derived from the cysteamine or the N-substituted methylcysteamine.
We thus synthesized metalladithioacetals derived from the naphthylethylimidazoline
(NEI) [11-12] and p-aminobenzophenone. The structure of these compounds is presented
in Figure 4 :
Figure 4 : Structure of metalladithioacetals.
The same interest which led us to study metallathiazolidines can be considered for
the metalladithioacetals :
1) in vivo release of one or more substances known for their radioprotective

properties, a delayed or prolonged release of active substances,
2) a fundamental modification of the electronic structure of these new derivatives

by the incorporation of a metal which can be at the origin of new specific chemical
or biological properties to these organometallic compounds.
A specific interest for the dithioacetals must be underlined ; these molecules make it
possible to deliver, as potentially available, a double quantity of the active substance than that
contained in their thiazolidinic equivalents. Some toxicological and radiopharmacological
results are presented in Table 2.
Ä'-(Ä
2!$)!4)/.#(%-)3429
Table 2 : Toxicity and radioprotective activity of metalladithioacetals, germaselenoacetals and their
parental organic derivatives [NEI= 2-(1-naphthylethyl)-2-imidazoline].
Metalladithioacetals and germaselenoacetals
Irr. Survival at
LD50tox/48h
M X R1 R2 R3 R4 dose 30 days
(mg/kg)a
(Gy) (%) [t, min]b
11 Ge S p-CH3C6H4 CH3 CH3 -NHCH2CH2NEI 200 8.1 40 [15]
12 Ge S p-CH3C6H4 CH3 H -NHCH2CH2NEI 184 8.1 50 [15]
13 Ge S n-C6H13 n-C6H13 CH3 -NHCH2CH2NEI 300 8.1 40 [15]
14 Ge S i-C5H11 n-C6H13 CH3 -NHC6H4(C=O)C6H5 600 8.1 60 [15]
15 Ge S n-C6H13 n-C6H13 H -NHC6H4(C=O)C6H5 600 8.1 40 [15]
16 Ge Se n-C6H13 n-C6H13 H NH2 260 8.1 20 [15]
Parental organic derivatives

Survival at
Irr. dose
Compound LD50tox/48h (mg/kg)a 30 days
(Gy)
(%) [t, min]b
17 NEICH2CH2NHCH2CH2SH 300 8.1 0 [15]
18 NEICH2CH2NHCH2CH(CH3)SH 212 8.1 0 [15]
19 H2NCH2CH2SeH, HCl 17 8.1 0 [15]
20 Ph-(C=O)C6H4NHCH2CH2SH > 300 8.1 30 [15]
21 Ph-(C=O)C6H4NHCH2CH(CH3)SH > 300 8.1 0 [15]
a) The administered dose is ½ LD50tox/48h b) [t] = time between administration of the compound and irradiation.
This study highlighted an overall behaviour equivalent between the sila- and
germadithioacetals in the field of toxicity and protection against radiation, even if some
germadithioacetals are characterized by a radioprotective activity, definitely more marked
than their silicon isologs. For the majority of the metalladithioacetals which we studied, this
structure gets a reduction in the toxicity of the purely basic organic derivatives as well as an
exaltation, sometimes by a large factor, of their radioprotective properties. This observation is
in agreement with the idea of a potentiation of the basic organic substances in mice, owing to
the metalladithioacetal structure. We showed that in vivo vectorization of metalladithioacetals
was taking place in lipophilic medium. Thus, the greater the lipophilic character of the various
vectors related to the metal derivative is, the more important the radioprotective activity.
Ä'-)Ä
Phosphorous radioprotectors
The introduction of an hydrosoluble phosphorous group increases the hydrosolubility

and the activity of these molecules by supporting their transport in vivo in comparison with
their parental aminothiols. Moreover, these kinds of phosphorous derivatives are often
more active because they have an intrinsic radioprotective activity and their metabolites
have radioprotective activity too. Phosphorothioates offer a slow and prolonged release of
radioprotective substances.
WR-2721 (H2NCH2CH2CH2HNCH2CH2SPO3H2) is one of the best currently known

radioprotectors. It was the subject of many studies because of its very high radioprotective
effect [11-12]. In spite of some side effects (vomiting in 25% of the cases, hypotension, etc.),
WR-2721 is currently used for a clinical use, because it preferentially protects normal cells.
The structure of these phosphorated derivatives is presented in Figure 5 :
Figure 5 : Structures of phosphorothioates.
Toxicity and radioprotective activity of phosphorothioates and their parental

organic derivatives are presented in Table 3. Addition of a phosphorothioate moiety to the
organic structures makes it possible to obtain, for these molecules, similar properties to those
observed with the incorporation of silicon or germanium atom in the case of organometallic
compounds (metallathiazolidines or metalladithioacetals) developed by our laboratory
[13-19], for example : delayed or prolonged in vivo release of one or more organic substances
having radioprotective properties (aminothiols), fundamental modification of the chemical
or biological properties.
Ä'-*Ä
2!$)!4)/.#(%-)3429
Table 3. Toxicity and radioprotective activity of phosphorothioates and their parental organic derivatives
[NEI= 2-(1-naphthylethyl)-2-imidazoline; NMI=2-(1-naphtylmethyl)-2-imidazoline].
LD50tox/48h Irr. dose Survival at 30 days

R
(mg/kg)a (Gy) (%) [t, min]b
O
R CH2CH2S P OH
OH
8.1 100 [15]
22 -NMI 168
10.1 70 [15]
8.1 100 [15]
23 -NEI 184
10.1 80 [15]
8.1 10 [15]
24 -HNCH2CH2NMI 280
10.1 0 [15]
8.1 90 [15]
25 -HNCH2CH2NEI 400
10.1 60 [15]
8.1 100 [15]
26 -HNCH2CH2CH2NMI 150
10.1 50 [15]
8.1 100 [15]
27 -HNCH2CH2CH2NEI 260
10.1 70 [15]
CH3 O
R CH2CHS P OH
OH
8.1 100 [15]
28 -NMI 184
10.1 70 [15]
8,1 40 [15]
29 -NEI 260
10,1 0 [15]
8.1 80 [15]
30 -HNCH2CH2NMI 150
10.1 70 [15]
8.1 100 [15]
31 -HNCH2CH2NEI 360
10.1 70 [15]
a) the administered dose is ½ LD50tox/48h. b) [t] = time between administration of the compound and irradiation.
The presence of phosphorus atom does not modify the toxicity. Indeed,
phosphorothioate and corresponding aminothiols have similar LD50tox/48h. On the other hand,
phosphorated pro-drugs, in comparison of aminothiols, have an excellent radioprotective
activity. These molecules are characterized by properties intermediate between lipophilic
and hydrophilic.
Ä'-+Ä
Most of these phosphorated derivatives present still an interesting radioprotective

activity even when mice have been exposed to LD100irr/30d(Z$PNQPVOETEFSJWFEGSPN
naphthylethylimidazoline are less toxic and sometimes more active. Compounds (30, 31)
are less toxic than compounds 28 and 29. However, in derivative 23, the most active of
the series, the substitution of an hydrogen atom by a methyl group, giving compound 29,
induces a great decrease in the activity, and even a total decrease of radioprotective activity
at z 10 Gy. In contrast, in the case of phosphorothioate 30, the presence of a methyl group
greatly improves radioprotective properties in comparison with compound 24 which is
the less active derivative. The Dose Reduction Factor has been determined as previously
described for compounds 27, 28 and 31. Their values are z 1.6, 1.8 and 1.6, respectively
(Tab. 4).
Table 4. Dose Reduction Factor (DRF) of phosphorothioates.
Compound LD50irr/30days (Gy) r2 DRF

No compound 6.2 0.970 -
27 10.1 0.982 1.6
28 11.4 0.999 1.8
31 9.9 0.999 1.6
Conclusion
In this work, several silicon, germanium, selenious or phosphorated derivatives have

been synthesized : a great majority presents an interesting radioprotective effect. In many
cases, when the structure was cyclic, we observed a delayed effect. Let us underline the more
important radioprotective activity and the generally lower toxicity of the silicon or germanium
derivatives compared to basic organic materials. This corresponds to a clear potentiation of
these molecules by the organometallic substituents. The results show clearly the important
contribution of silicon, germanium and phosphorus in the origin of the radioprotective
properties of these structures. Phosphorated derivatives from NEI and NMI are the most
efficient. However, they also are more toxic than silicon, germanium or selenious derivatives.
The development of efficient and non-toxic radioprotective compounds should

present an interest not only for military applications or nuclear energy accidents as Chernobyl,
but also in the chemo- and radiotherapy treatments.
Ä'-,Ä
2!$)!4)/.#(%-)3429
Acknowledgements
The authors want to thank the Délégation Générale pour l’Armement (DGA/STTC/DT/SH),
Ministère de la Défense Nationale, France, for its financial support and interest in this research.
References :
[1] Dale W.M., Gray L.H., Meredith W.J., The inactivation of an enzyme (carboxypeptidase) by X- and
A-radiation, Phil. Trans. Roy. Soc., 1949, 242A, 33-62.
[2] Patt H.M., Tyree E.B., Straube R.L., Smith D.E., Cysteine protection against X Irradiation, Science,
1949, 110, 213.
[3] Bacq Z.M., Hervé A., Lecomte J., Fischer P., Blavier J., Dechamps G., Le Bihan H., Rayet P., Protection
contre le rayonnement X par la B-mercaptoethylamine, Arch. Int. Physiol., 1951, 59, 442-446.
[4] Kligerman M.M., Glover D.J., Turrisi A.T., Norfleet A.L., Yuhas J.M., Coia L.R., Simone C., Glick J.H.,
Goodman R.L., Toxicity of WR-2721 administered in single and multiple doses, Int. J. Radiat. Oncol.
Phys., 1984, 10, 1773-1776.
[5] Piper J.R., Stringfellow C.R., Elliot R.D., Johnston T.P., S-2-(omega-aminoalkylamino)ethyl
dihydrogen phosphorothioates and related compounds as potential antiradiation agents, J.
Med. Chem., 1969, 12, 236-243.
[6] Terol A., Fernandez J.P., Robbe Y., Chapat J.P., Granger R., Fatôme M., Study on 2-phenylthiazolidine
derivatives as radioprotectant agents, Eur. J. Med. Chem., 1978, 13, 149-151.
[7] Fatôme M., Poutrain P., Granger R., Orzalesi H., Robbe Y., Randon M., Fernandez J.P., Radioprotective
effects of thiazolidines, Chim. Ther., 1970, 5, 312-317.
[8] Voronkov M.G., Biologically active silicon compounds, Pure Appl. Chem., 1969, 19, 399-416.
[9] Voronkov M.G., Zelchan G., Lukevics E., Silicon and Life, Akademii Verlag, Berlin, 1975, p. 370.
[10] Fatôme M., Poutrain P., Granger R., Orzalesi H., Robbe Y., Randon M., Valentin M., Chevallet P.,
Femandez J.P., Radiation protective agents. Dithiolanes, Chim. Ther., 1970, 5, 327.
[11] Piper J.R., Stringfellow C.R., Johnston T.P., Terminally substituted S-2-(omega-aminoalkylamino)
ethyl dihydrogen phosphorothioates and related compounds as potential antiradiation agents,
J. Med. Chem., 1969, 12, 244-253.
[12] Davidson D.E., Grenan M.M., Sweeney T.R., Biological characteristics of some improved
radioprotectors in “Radiation Sensitizers. Their use in the clinical management of cancer”, Brady
L.W. (ed), Masson, NY, 1980, 309-320.
[13] Célariès B., Amourette C., Lion C., Rima G., Synthesis and radioprotective study of new
siladithioacetals and germadithioacetals, Appl. Organomet. Chem., 2003, 17, 561-569.
[14] Rima G., Satgé J., Dagiral R., Lion C., Fatôme M., Roman V., Laval J.D., Synthesis and radioprotective
activity of new organosilicon and germanium compounds, Metal-Based Drugs, 1998, 5,
139-146.
Ä'--Ä
[15] Rima G., Satgé J., Dagiral R., Lion C., Fatôme M., Roman V., Laval J.D., Synthesis and application of
new organometallic compounds of silicon and germanium in chemical radioprotection, Appl.
Organomet. Chem., 1999, 13, 583-594.
[16] Satgé J., Rima G., Fatôme M., Sentenac-Roumanou H., Lion C., Synthesis and radioprotective
activities of germatranes, silatranes, germylated oxides and sulfides with cysteamine,
methylcysteamine and N(2-thioethyl)-1,3-diaminopropane hydrochloride ligands, Eur. J. Med.
Chem., 1989, 24, 48-54.
[17] Yuhas J.M., Storer J.B., Differential chemoprotection of normal and malignant tissues, J. Natl-
Cancer Inst., 1969, 42, 331-335.
[18] Célariès B., Amourette C., Lion C., Rima G., New phosphorothioates derived from naphthyl-
methylimidazoline and naphthylethylimidazoline : application in chemical radioprotection,
Radioprotection, 2005, 40, 57-71.
[19] Célariès B., “Étude et applications de nouveaux dérivés organosiliciés, -germaniés, -phosphorés
et -séléniés en radioprotection chimique”, Ph. D. Thesis, Université de Toulouse, 2003, p. 292.
Ä'-.Ä
Chapter 20
Advances in radiotherapy :
new principles
Nicolas FORAY and Jacques BALOSSO
Introduction
Since their discovery, ionizing radiations have been recognized to induce deleterous
effects on living cells and locally used to cure tumours. The powerful penetration of high-
energy photon beams has made easier the treatment of a number of deep-seated tumours,
while kilovoltage X-rays are more appropriate for radiotherapy of superficial cancers.
Nevertheless, high-energy photon radiotherapy does not produce optimal radiobiological
effects within targeted tissue region and its development has not fully eliminated the
central problem of irradiated surrounding normal tissues. Moreover, lesions induced in DNA
of tumours (Chapters 12-14, 18) may be repaired (Chapter 15), a process which decreases
the treatment efficiency. Particularly, it must be reminded that proliferating cells (S phase)
(Chapter 15) are more radioresistant than quiescent cells (G0/G1 phase), whatever the type
of radiation used, high-energy photons or X-rays. Hence, two major innovating anti-cancer
strategies involving radiotherapy are proposed to date:
1) to apply powerful radiation to kill a larger yield of tumour cells (notably

proliferating ones) and to use multiple radiation beams whose geometry fits
to the tumour shape by using computerised data obtained from imaging. These
requirements are notably reached by 3D conformal radiotherapy and the intensity
modulated approach.
Ä'.&Ä
2!$)!4)/.#(%-)3429
2) to use drugs, the activities of which are exacerbated by radiation. Radio-

chemotherapy does not systematically lead to strong synergetic effects and a low
number of chemotherapy drugs are known to be overactivated by radiation. The use
of radiolabelled drugs and antibodies is an interesting strategy and has provided
encouraging results. However, again, the success depends on the specific penetrability of
agents into the tumour cells and the cell cycle dependence of their action. By contrast,
the X-rays-induced photoactivation of high Z elements that are contained in some specific
drugs used routinely in diagnostics or in anti-cancer therapy is an alternative to overcome
the difficulties exposed above (as far as photoactivation does not destroy the drug itself).
New advances in radiotherapy
So far, despite of the considerable progress in our understanding of the biology of

tumours, the most efficient way to cure cancers still consists in destroying as much malignant
cells as possible. To this aim, a plethora of physical agents have been proposed. This was
notably the case of powerful ultra-sounds, micro-waves, lasers, etc. However, none of them
is more efficient than ionising radiation. Nowadays, most of cancers are treated with radio-
therapy and about half of them with chemo-radiotherapy.
Anatomical problems
As with surgery, the first difficulty encountered by radiotherapists is anatomy.
Radiotherapy and surgery show the same limits: the infiltrating activity of tumours and their
intimate interplay with normal tissues frequently involve vital organs and render difficult
a precise targeting of the tumour. Surgery as well as any means leading to the physical
destruction of the tumour are applied only in limited cases, notably when tumours are
situated in some anatomical scaffolds that can be crossed without danger. In any other case
(i.e. the great majority), radiotherapy is applied but raises the problem of damage formation
into surrounding tissues : a significant differential effect between healthy tissues and tumour is
therefore required. To date, technological advances in radiation production have progressively
solved the geometrical problem raised by the tumour shapes. However, the differential effect
is mainly obtained by the fractionation of the treatment in multiple sessions and/or by the
use of concomitant chemotherapy to impede the tumour re-growth during the therapeutic
course. Four major approaches are now possible to enhance the success of the treatment :
– to better use computer data generated by tumour imaging ;
– to perform the most precise dosimetry ;
Ä'.'Ä
!DVANCESINRADIOTHERAPY
– to take into account the natural movements of patients, notably breathing,
– and, if possible, to use charged particle beams, such as protons, when therapy
with extreme accuracy in depth is needed.
3D conformal radiotherapy
The radiation dose to be delivered and actually deposited in the different volumes
to be treated unavoidably differs for several reasons, notably :
– because tumours are very heterogeneous tissues and they may elicit very
proliferating and/or very dense regions in which a higher dose is required whereas
some quiescent and/or weakly invading regions require lower doses ;
– some radiosensitive and/or vital organs situated in the vicinity of the tumours
must not be submitted to radiation. This is notably the case with eyes and optical
nerves, brain stem, lungs and kidneys, ...
The tumour shape can be therefore complex, rendering quite difficult a selective
exposure to radiation. One of the major recent advances in external radiotherapy consists in
using numerised tumour imaging from CTscan and/or from nuclear magnetic resonance (NMR)
technology in order to compute the anatomical extensions of the tumour and to modelise
it in 3 dimensions at the scale of the radiation treatment conditions: this is the 3D conformal
radiotherapy. Such an approach allows, as has never been done before, both the tumour
volume to be targeted and sensitive organs to be materialised and protected (Fig. 1) [1].
Figure 1 : A representative example of multiple

isocentric radiation beams for the 3D conformal
treatment of prostate cancer.
Ä'.(Ä
2!$)!4)/.#(%-)3429
Intensity-modulated radiotherapy
According to the tumour shape, data acquired from imaging in the frame of the
3D conformal radiotherapy, the geometry and multiple radiation beams must be set up.
To date, the standard approach has consisted in using interactive softwares that provide
the calculation of the dose distribution to virtual model of the patient reconstructed from
computed imaging. These data are finely adapted to each individual case by iterative manual
correction. Unfortunately, such a standard approach cannot manage a large number of
beams, differing by their size, intensity and time of exposure. In order to avoid an increasing
amount of data to be acquired, an innovating approach is to generate a system leading
by automatic corrections. This is the so-called principle of inversed treatment planning
system: the radiation doses prescribed by the radiotherapist determine the radiation beam
parameters. Obviously, powerful computers are necessary to reach such requirements,
notably for managing complex geometries. In fact, the real technological advance lies in
the development of sophisticated computed operated linear accelerators that deliver
X-rays of high energy through a multi-leaf collimator. By automatically generating a full set
of complementary beams with differential intensity according to the different part of their
sections, the complex requirements of highly conformal radiotherapy can be reached: this
is the intensity modulated radiotherapy (IMRT). This new technology represents the ultimate
physical frontier of X-rays radiotherapy (Fig. 2) [1].
Figure 2 : A representative example of radiation beams modulated by collimators in an IMRT rectum

treatment. PTV : planning target volume.
Ä'.)Ä
Pharmaco-modulation
The chemo-radiotherapy concept gained the status of a global standard for curative
radiotherapy across the 1990s. Basically, it consists in impeding the tumour regrowth during
the therapeutic course. This effect is tremendously important and can explain most of the
effects of regular chemo-radiotherapy. Together with this largely applied concept, we propose
to present innovative approaches that are still in their preclinical state of development.
First clinical trials with photoactivated drugs

Mainly based on the enhancement of photoelectric effect, the concept of
photoactivation of high Z elements has long been developed by physicists but its application
to anti-cancer therapy is recent and can legitimately be attributed to the researchers of the
Norman’s group [2, 3]. In the 1970s, Norman observed chromosomal aberrations in circulating
lymphocytes of patients submitted to urography. Since such a radiodiagnostic treatment
involves iodinated contrast agents, he hypothesised that such aberrations resulted from a
local excess of radiation dose, due to a maximal energy absorption in iodine atoms contained
in contrast agents. This physical effect is notably obtained after irradiating iodine atoms in
water at 33 keV [2-4]. Norman proposed to “exploit” these chromosome damaging effects by
applying it to tumours in an innovating therapy approach consisting in:
– loading the tumour with iodinated contrast agent ;
– imaging the tumour with a modified X-rays scanner system and computing the
tumour position ;
– treating the tumour with the same irradiation X-ray set up by using the computing
data obtained during imaging and by producing photoactivation of iodine
contained in contrast agents.
This technique had the considerable advantage of minimizing patient displacement

since the beam is mobile. Although such a strategy did not overcome the problem of
chromosomal aberrations in normal tissues (observations on which this technique was
based), it opened the wide field of photoactivation of pharmacological compounds
containing high Z elements. This technique has been applied to animals with a limited
success and to humans in a single clinical trial combined with a standard treatment [2,3]. To
our knowledge, since 1990, no other clinical trial has been performed with this technique,
maybe due to its relative toxicity for normal tissues [4].
Ä'.*Ä
2!$)!4)/.#(%-)3429
Synchrotron radiation rescues

the medical application of photoactivation
In addition to the crucial choice of the photoactivable drugs, the main problem
for triggering photoactivation is to produce stable and sufficiently intense monochromatic
X-rays, tunable to the appropriate energy. As mentioned above, Norman’s group had modified
a standard scanner machine to perform experiments with iodinated contrast agents but
the energy was not tuned [3]. In the 1990s, the development of synchrotron in the medical
research area solved this problem. Briefly, synchrotrons are electrons accelerators that
provide extremely high-fluenced X-rays resulting from changing direction the GeV electrons
by bending magnets. The synchrotron X-rays fluence is about 106 times higher than that
of standard medical irradiators and renders possible the production of monochromatic
X-rays by refraction on appropriate crystals. According to the type of inserted devices, the
produced synchrotron X-rays range from a few to hundred keV permitting a theoretical
photoactivation of a number of high-Z elements (at least heavier than iron). However, once
solved the problem of production of high-fluence monochromatic X-rays, physicists and
physicians faced another difficulty : while an X-ray scanner source turns around patients
during diagnostic exams, the synchrotron X-ray beam is necessarily horizontal and fixed.
Consequently, in order to ensure a homogeneous distribution of radiation dose, engineers
developed a turning set-up allowing circular irradiation of samples and patients [5,6].
With regard to the choice of photoactivable drugs, platinum-containing drugs such

as cisplatinum, carboplatinum and oxaliplatinum appeared early to be the best candidates
for photoactivation of chemotherapeutic agents since :
– they contain platinum atoms making possible their photoactivation at 78.4 keV,
corresponding to the K-edge of platinum ;
– they are extensively used with success in various chemotherapy and chemo-
radiotherapy treatment ;
– they target more specifically the proliferating cells with a powerful penetration
capacity into cells ;
– they induce severe DNA adducts preventing successful repair in tumour cells
[5,6].
Recently, photoactivation of cis-platinum (called PAT-Plat) provided by synchrotron

X-rays was applied to rats bearing radioresistant gliomas [6]. After a cisplatinum intratumoral
Ä'.+Ä
injection, a dose of 15 Gy X-rays was delivered by synchrotron radiation into tumour at a

tuned energy corresponding to the platinum K-edge (78.4 keV). This treatment resulted in
the cure of 33% rats and provides still to date the most protracted survival of rats bearing
gliomas [6]. New developments of photoactivation of heavy elements are in progress and
clinical trials in humans may be seriously envisaged and scheduled in the coming years. In
parallel, intensive in vitro experiments still are necessary to propose a molecular model of
the mechanisms involved in the PAT-Plat approach in order to secure its clinical transfer.
Why PAT-plat approach is more efficient to kill tumours?

DNA double-strand break (DSB) is a significant lesion that, if unrepaired, can result in
loss of genetic material (cell lethality), or if misrepaired, can cause genomic rearrangements,
(potential cancer onset) [7]. Mammalian cells possess two major mechanisms for repairing
DSB (Chapter 15) : homologous recombination (HR), mediated by the RAD51 protein and
non-homologous end-joining (NHEJ), dependent upon the DNA-PK protein trimeric complex
[7]. NHEJ acts predominantly during G1 phase and its impairment is likely to be associated
to marked radiosensitivity and immunodeficiency since it is essential for the production of
immunoglobulins [7]. In parallel, HR is a minor DSB repair pathway, alternative to NHEJ. HR
functions to repair breaks that arise during the S/G2 phase (Chapter 15). Defects in HR have
been identified to be more likely in cancer-proned diseases and in cells sensitive to alkylating
agents [7].
Briefly, NHEJ consists in the translocation of the Ku heterodimer all along DNA
(Chapter 15). At a break site, the Ku dimeric protein, component of the DNA-PK complex,
clamps and recruits the catalytic subunit of DNA-PK (DNA-PKcs). Altogether, Ku and DNA-
PKcs forms the serine-threonine kinase DNA-PK that phosphorylates a number of substrates,
leading to the recruitment of ligases that are able to join DNA-ends. Interestingly, whereas
NHEJ is the major DSB repair pathway in mammalians, the translocation of Ku all along DNA
makes it sensitive to any sterical blockage and to large gaps in DNA [7] :
– the first situation (DNA-binding molecules) is notably encountered with cisplatinum-

induced DNA adducts that are known to inhibit the NHEJ process [8]. Consequently,
association between ionising radiation producing DSB and cisplatinum may render
the radiation-induced breaks irreparable (Fig. 3A) as far as :
Othe concentration of DNA adducts is sufficient but not toxic for normal
tissues ;
Ä'.,Ä
2!$)!4)/.#(%-)3429
O the succession of the radiotherapy session and the cisplatinum treatment

are kinetically close [5,6] ;
– the second situation (large gaps) can be encountered with hadrontherapy that
allows very dense energy clusters responsible for large holes in DNA.
In the case of PAT-Plat conditions (Fig. 3B), the presence of cisplatinum DNA-adducts
in tumours prevents the repair of the radiation-induced DSBs. Furthermore, PAT-Plat results
in the emission of low-energy electrons (LEE ; Chapter 13) from cisplatinum molecules
(for 3% of the atoms), and the range of such electrons is limited to DNA surrounding cisplatinum
molecules. Consequently, additional DSBs directly due to the photoactivation are induced
at the close vicinity of DNA adducts, making them irreparable by NHEJ. These specific PAT-
Plat-induced DSBs are likely to enhance the therapeutic index of the PAT-Plat approach [5,6].
However, as mentioned above, although playing a minor role in the DSB repair of mammalians,
HR may serve as an alternative to NHEJ and partly reduce the efficiency of the PAT-Plat strategy.
The relevance of this hypothesis is in progress by testing different models of tumours. One
can however consider to date that HR-deficient tumours are preferentially targeted by
PAT-Plat approach whereas therapeutic index is lower for HR-proficient ones (Fig. 3).
Toward anti-cancer treatments specific to each individual case
The PAT-Plat strategy is a representative example of innovating anti-cancer strategy

that endeavours to associate benefit of both radiation and chemotherapy. However, such
association raises the problem of co-toxicities with concomitant induction of different DNA
damage types that are repaired by different and sometimes interplaying repair pathways.
Hence, pharmaco-modulation of DNA repair necessarily implies a differential tumour
targeting according to their genetic status. In parallel, the tumour shape may be also strongly
dependent upon each individual case and these two physical and chemical approaches raise
an actual strategic problem : the systematic repair gene mutations screening in tumours
is required in order to ensure that each individual case receives the most specific and
appropriate anti-cancer treatment. Personalised anti-cancer treatments will undoubtedly
be one of the major challenges of the radiotherapy of the future.
Ä'.-Ä
Figure 3 : DNA damage in radio-chemotherapy. A) A molecular explanation for the synergetic effects of a radio-
chemotherapy with cisplatinum. The expected result of any radio-chemotherapy is to induce unrepairable DNA
damage into tumours while sparing normal tissues. Radiotherapy induces DNA double-strand breaks (DSB) that
are mainly managed via the NHEJ repair pathway which is active in G0 / G1 cells (Chapter 15). Chemotherapy
with cisplatinum induces large volume DNA adducts that preferentially target proliferative cells. Interestingly,
when the DNA adduct is sufficiently close to a DNA break, the DNA-PK protein, that is essential for repairing
DSBs by NHEJ, is sterically blocked. Consequently, the radiation-induced DSBs become unrepairable. B) The X-ray
photoactivation of cisplatinum additionally enhances the synergy of radio-chemotherapy. Synchrotron X-rays
induce also DSBs at the same rate as standard radiotherapy. However, by using X-rays synchrotron at an energy
tuned at the K-edge of platinum (78.4 keV), low energy electrons (LEE) are emitted locally from platinum atoms,
which play the role of a radiosensitizer, and induce additional DSBs in the close vicinity of DNA adduct sites.
Consequently, such unrepairable DSB are more numerous and localized than in a standard radio-chemothe-
rapy. An extra-synergy is therefore expected into tumour treatment.
Ä'..Ä
2!$)!4)/.#(%-)3429
Acknowledgements
We would like to thank all our Ph.D. students whose contributions to the development of the
PAT-Plat approach was crucial: S. Corde, A. Joubert, M.-C. Biston, Z. Bencokova and J. Gastaldo. This work
was also supported by the Association Pour la Recherche sur l’Ataxie-Telangiectasie (APRAT), the Association
pour la Recherche contre le Cancer (ARC), the Électricité de France (Comité de Radioprotection), the ETOILE
Project (Région Rhône-Alpes and Lyon University), the Région Rhône-Alpes and ROCHE-France.
References
[1] Mazeron J.-J., Maugis A., Barret C., Mornex F., Technique d'irradiation des cancers à La radiothérapie
conventionnelle. Maloine, Paris, 2005.
[2] Norman A., Adams F., Riley R., Cytogenetic effects of contrast media and triiodobenzoic acid
derivatives in human lymphocytes. Radiol., 1978, 129, 199-203.
[3] Rose J.H., Norman A., Ingram M., Aoki C., Solberg T., Mesa A., First radiotherapy of human
metastatic brain tumors delivered by a computerized tomography scanner (CTRx), Int. J. Radiat.
Oncol. Biol. Phys., 1999, 45, 1127-1132.
[4] Joubert A., Biston M.C., Boudou C., Ravanat J.L., Brochard T., Charvet A.M., Esteve F., Balosso J.,
Foray N., The concomitant use of iodinated contrast media and ionizing radiation may result in
severe DNA breaks and radiosensitization of endothelial cells, Int. J. Radiat. Oncol. Biol. Phys.,
2005, 62, 1486-1496.
[5] Corde S., Balosso J., Elleaume H., Renier M., Joubert A., Biston M.C., Adam J.F., Charvet A.M.,
Brochard T., Le Bas J.F., Estève F., Foray N., Synchrotron photoactivation of cis-platin (PAT-Plat)
elicits an extra-number of DNA breaks that stimulate RAD51-mediated repair pathways., Cancer
Res., 2003, 63, 3221-3227.
[6] Biston M.C., Joubert A., Adam J.F., Elleaume H., Bohic S., Charvet A.M., Estève F., Foray N., Balosso
J., Cure of Fisher rats bearing radioresistant F98 glioma treated with cis-platinum and irradiated
with monochromatic synchrotron X-rays, Cancer Res., 2004, 64, 2317-2323.
[7] Chu G., Double-strand break repair, J. Biol. Chem., 1997, 272, 24097-24100.
[8] Turchi J.J., Henkels K.M., and Zhou Y., Cisplatin-DNA adducts inhibit translocation of the Ku
subunits of DNA-PK, Nucleic Acids Res., 2000, 28, 4634-4641.
Ä(%%Ä
Index
A alkali . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46, 265

accelerator . . . . . . . . . . . . . . 3, 5, 17, 22, 26, 30, 135 alkaline comet assay . . . . . . . . . . . . . . . . . . . . . . 185
electron accelerator . . . 22, 26, 62, 89, 90, 134, 136, alkaline earth metal ions . . . . . . . . . . . . . . . . . . . . 46
......................................................................154, 166 alkanes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
electron beam . . . . . . . . . . . . . . . . . . . 18, 80, 158 alpha rays ( -rays) . . . . . . . . . . . . . . . . . . . . 53, 57, 60
Febetron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 amines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
heavy ion beam . . . . . . . . . . . . . . . . . . 29, 54, 187 amino acids . . . . . . . . . . . . . . . . . . . . 70, 76, 77, 234
ion accelerator . . . . . . . . . . . . . . . . . . . . . . . . 30, 75 aminyl radical. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
laser wake-field accelerator . . . . . . . . . . . . 19, 26 ammonia . . . . . . . . . . . . . . . . . . . . . . . . . 35, 41, 44, 69
linear accelerator . . . . . . . . . . . . . . . . . . . . . . 23, 24 apollofix-red . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
Marx-bank impulse generator . . . . . . . . . . . . . 22 apoptosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
particle accelerator . . . . . . . . . . . . . . . . . . . . . . . 19 aqueous solutions . . . . . . . . . . . . . . . . . . . . . . . 7, 36
photocathode electron gun . . . . . . . . . . . . . . . 25 aromatic rings . . . . . . . . . . . . . . . . . . . . . . . . . . 9, 236
single electron bunch . . . . . . . . . . . . . . . . . 24, 25 astrobiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
3
table-top terawatt (T ) . . . . . . . . . . . . . . . . . . . . 26
Van de Graaff accelerator . . . . . . . . . . . . . . 20, 75
acetonitrile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 B
adhesives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 benzene . . . . . . . . . . . . . . . . . . . . . . . . . . 7, 14, 71, 88
advanced oxidation process (AOP) . . . . . . . . . . . 79 beta-rays ( -rays) . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
alcohols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12, 40 biomembranes . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
Ä(%&Ä
2!$)!4)/.#(%-)3429
biomolecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265 concretes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118

Bragg peak . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 conductivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
bremsstrahlung . . . . . . . . . . . . . . . . . . . . . . . . 18, 136 copper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
bridging states. . . . . . . . . . . . . . . . . . . . . . . . . . . . 210 cosmic rays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
by-stander effect . . . . . . . . . . . . . . . . . . . . . . . . . . 188 cost evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
cryo-irradiation . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
C cyclohexane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7, 14
cable insulators . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 cyclotron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
cage effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
cancer . . . . . . . . . . . . . . . . . . . . . . . . . . 291, 297, 298 D
carbofuran . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84, 85 delta rays (D-rays) . . . . . . . . . . . . . . . . . . . . . . . . . . 55
carbon nanotubes . . . . . . . . . . . . . . . . . . . . . . . . 103 desinfection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
carbonyl compounds . . . . . . . . . . . . . . . . . . . . . . . . 8 direct effects . . . . . . . . . . . . . . . . . . . . . . . . 177, 191
carboxyl radical . . . . . . . . . . . . . . . . . . . 11, 114, 235 disulfides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
catalysis . . . . . . . . . . . . . . . . . . . . . . . . . . 85, 109, 236 disulfide radicals . . . . . . . . . . 237, 238, 241, 244
cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177, 186, 220 DNA . . . . . . . . . . . . . . . . . 68, 177, 191, 203, 219, 233
cement-based materials . . . . . . . . . . . . . . 117, 123 cellular DNA . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
Čerenkov light . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 double-strand DNA. . . . . . . . . . . . . . . 195, 182
chain reaction . . . . . . . . . . . . . 57, 60, 121, 138, 255 isolated DNA . . . . . . . . . . . . . . . . . . . . . . . . . . 183
chain scission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 plasmid DNA . . . . . . . . . . . . . . . . . . . . . . . . . . 196
charge . . . . . . . . . . . . . . . . . . . . . . . 39, 184, 207, 267 single-strand DNA . . . . . . . . . . . . . . . . . . . . . 195
charge motion . . . . . . . . . . . . . . . . . . . . . . . . 203 DNA damage . . . . . . . . . . . . . . . . 177, 191, 219, 299
charge transfer . . . . . . . 184, 207, 210, 213, 267 alkali revealed breaks (ARB) . . . . 182, 185, 265
charge tunneling . . . . . . . . . . . . . . . . . . . . . . 207 double-strand breaks (DSB) . . . . 185, 203, 220,
Chini clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 ...............................................................225, 265, 297
chromatography . . . . . . . . . . . . . . . . . . . . . . . . . . 185 frank strand breaks (FSB) . . . . . . . . . . . . . . . 265
cholesterol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252 lesions . . . . . . . . . . . . . . . . . . . . . . . 178, 219, 291
clustered damage . . . . . . . . . . . . . . . . . . . . 187, 221 single strand breaks (SSB) . . . . . 185, 197, 220,
clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 ........................................................................224, 265
alloyed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 tandem lesions . . . . . . . . . . . . . . . . . . . . . . . . 184
core-shell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 DNA repair . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 base excision repair (BER) . . . . . . . . . . . . . . . 221
multi-metallic . . . . . . . . . . . . . . . . . . . . . . . . . 107 homologous recombination (HR) . . . 221, 297
properties. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 non-homologous end-joining (NHEJ). . 221, 297
stabilization . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 nucleotide excision repair (NER) . . . . 221, 223
coalescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 dose . . . . . . . . . . . . . . . . . . . . . . . . 3, 18, 55, 156, 166
coating processing . . . . . . . . . . . . . . . 136, 145, 147 letal dose (LD) . . . . . . . . . . . . . . . . . . . . . . . . 279
comets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 dose rate . . . . . . . . . . . . . . . . 3, 60, 61, 100, 107, 125
Ä(%'Ä
)NDEX
dose reduction factor (DRF) . . . . . . . . . . . . . . . . 279 treatment control . . . . . . . . . . . . . . . . . . . . . . 169

double bonds. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8, 9 wholesomeness . . . . . . . . . . . . . . . . . . . . . . . 168
drugs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 formate ions . . . . . . . . . . . . . . . . . . . . . . . 11, 98, 114
radio-sterilization . . . . . . . . . . . . . . . . . . . . . . 151 Fourier transform infrared spectroscopy (FTIR) . . 75
dust . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 frequency-domain single shot (FDSS) . . . . . . . . 28
dyes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
dye decoloration . . . . . . . . . . . . . . . . . . . . . . . 86 G
gamma rays (G-rays) . . . . 3, 18, 38, 53, 72, 132, 136,
E ...................................................... 154, 158, 165, 167
electron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7, 135 genome . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
electron acceptor . . . . . . . . . . . . . . . . . 105, 204 glycerol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
electron attachment . . . . . . . . . . . . . . 191, 197 gold . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
electron donor . . . . . . . . . . . . . . . . . . . 105, 204 G values . . . . . . . . . . . . . . . . . . . . . . . . 6, 60, 140, 254
fast electron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 free ions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
hydrated electron. . . . . . . 7, 36, 56, 81, 99, 254 water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
solvated electron . . . . . . . . . . . . 3, 35, 37, 48, 97
strongly bound electron . . . . . . . . . . . . . . . . . 48 H
thermalisation . . . . . . . . . . . . . . . . . . . . . . 46, 137 halides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
electron transfer . . 82, 105, 107, 110, 195, 198, 242 high performance liquid
weakly bound electron . . . . . . . . . . . . . . . . . . 48 chromatography (HPLC) . . . . . . . . . . . . . . 157, 186
electron beam treatment . . . . . . . . . . . . . . . . 92, 79 histones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266
electron paramagnetic resonance (EPR) . . . . . . 21 holes. . . . . . . . . . . . . . . . . . . . . . . . . . . . 114, 184, 204
electron pulse-probe analysis . . . . . . . . . 28, 46, 63 hole transfer . . . . . . . . . . . . . . . . . 195, 196, 197
electron spin resonance (ESR) . . . . . . . . . . . . . . 151 hydrocarbon. . . . . . . . . . . . . . . . . . . . 7, 13, 14, 41, 82
electrophile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 hydrogen . . . . . . . . . . . . . . . . . . . 53, 56 , 69, 72, 118
electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . 185 hydrogen atom . . . . . . . . . . . . . . . 7, 10, 81, 98, 254
emulsion polymerization . . . . . . . . . . . . . . . . . . 146 hydrogen peroxide . . . . . . . . . . . . . . . . 6, 53, 82, 244
ethane-diol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 hydronium cation . . . . . . . . . . . . . . . . . . . . . . . . 6, 43
ethanol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 hydroperoxide . . . . . . . . . . . . . . . . . . . . . . . . . 83, 257
ethers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 hydroperoxide radical . . . . . . 6, 58, 80, 84, 122, 254
excitations . . . . . . . . . . . . . . . . . . . . . . . . . . 5, 56, 133 hydroxyl radical . . . . . . . 7, 43, 81, 98, 179, 234, 237,
........................................................................254, 266
F
fast neutrons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 I
fluorescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204 ice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74, 76
fluoride . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 indirect effects . . . . . . . . . . . . . . . . . . . . . . . 177, 191
folding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234, 235 inks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
food irradiation . . . . . . . . . . . . . . . . . . . . . . . . . . . 165 instrumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Ä303Ä
2!$)!4)/.#(%-)3429
intensity modulated . . . . . . . . . . . . . . . . . . . . . . . 291 medical supplies . . . . . . . . . . . . . . . . . . . . . 134, 139

interstellar medium (ISM) . . . . . . . . . . . . . . . . . . . 71 membranes . . . . . . . . . . . . . . . . . . . . . . . . . 136, 144
iodine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295 mesophases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
ionic liquids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 metal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8, 97
ionization potential . . . . . . . . . . . . . . . . . . . . . . . 106 bulk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
ionization . . . . . . . 3, 56, 97, 133, 165, 177, 191, 205 hyper-reduced states . . . . . . . . . . . . . . . . . . . 12
ion-molecule reactions . . . . . . . . . . . . . . . . . . . 5, 72 metal clusters . . . . . . . . . . . . . . . . . . . . . . . . . . 97
ion pair . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 metal ions . . . . . . . . . . . . . . . . . . . 8, 9, 11, 46, 98
isotopic labelling. . . . . . . . . . . . . . . . . . . . . . . . . . 182 metal oligomers . . . . . . . . . . . . . . . . 97, 99, 103
meteorites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69, 77
K methane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
kinetics . . . . . . . . . . . . . . . . . . . 10, 99, 213, 214, 215 methanol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12, 40
competition kinetics . . . . . . . . . . . . . . . . . . . . 18 methionine . . . . . . . . . . . . . . . . . . . . . . . . . . 239, 244
micelles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103, 252
L microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
latent image . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 Mie resonance . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
laser . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 molecular products . . . . . . . . . . . . . . . . . . . . . . . . . 5
light . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 multiply damaged sites (MDS) . . . . . . . . . . . . . . 221
linear energy transfer (LET) . . . 4, 18, 53, 132, 187, mutagenesis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191, 192, 199 mutation . . . . . . . . . . . . . . . . . . . . . . . . 177, 178, 243
lipids. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171, 249
lipoproteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249 N
liposomes . . . . . . . . . . . . . . . . . . . . . . . . . . . 253, 259 nanocolloids. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
liquids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3, 56 nanomaterials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
liquid sodium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 nanometric pores . . . . . . . . . . . . . . . . . . . . . . . . . 136
lithography . . . . . . . . . . . . . . . . . . . . . . . . . . 140, 141 nanorods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
long-range charge transfer . . . . . . . . . . . . . . . . . 203 nanotubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
long-range electron transfer (LRET) . . . . . . . . . 242 nanowires . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
low energy electrons (LEE) . . . . 188, 191, 196, 197, neopentane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199, 298 neutrons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60, 61
nickel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
M nitro-compounds . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
magnetic field effects (MARY) . . . . . . . . . . . . . . . 19 non-homogeneous distribution . . . . . . . . . . . . . . 4
mass spectrometry . . . . . . . . . . . . . . . . . . . 152, 185 non-linear optical behaviour . . . . . . . . . . . . . . . 109
mechanism . . . . . . . . . . 10, 48, 78, 99, 203, 215, 255 nuclear energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
hopping mechanism . . . . . . . . . . 198, 207, 210 nuclear reactors. . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
radical chain mechanism . . . . . . . . . . . . . . . 255 international thermonuclear
superexchange mechanism. . . . . . . . . . . . . 207 experimental reactor (ITER) . . . . . . . . . . . 54, 61
Ä(%)Ä
)NDEX
pressurized water reactor (PWR). . . . . . . . . . 60 platinum-containing drugs . . . . . . . . . . . . . 296

supercritical water reactor (SCWR) . . . . . . 54,62 polarity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
nuclearity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 pollutants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81, 89
nucleation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 polymer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100, 131
nuclei . . . . . . . . . . . . . . . . . . . . . . . . . . . 100, 105, 112 cross-linking . . . . . 102, 131, 138, 140, 144, 220
nucleobase . . . . . . . . . . . . . . . . . . . . . . 178, 179, 205 curing . . . . . . . . . . . . . . . . . . . . . . . . . . . 144, 145
nucleophile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 degradation . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
nucleoside . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 grafting . . . . . . . . . . . . . . . . . . . . . . . . . . 131, 138
nucleosome . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271 oxidation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
scission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
O polymerization . . . . . . . . . . . . . . . . . . . . . . 131, 138
oligonucleotide . . . . . . . . . . . . . . . . . . . . . . . . . . . 184 porosity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
one-electron transfer . . . . . . . . . . . . . . . . . . . . 8, 112 pores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
Onsager radius . . . . . . . . . . . . . . . . . . . . . . . . . . 7, 44 prebiotic molecules . . . . . . . . . . . . . . . . . . . . . 68, 71
optical absorption . . . . . . . . . . . 22, 36, 40, 99, 103 pressure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36, 53, 62
optical limitation . . . . . . . . . . . . . . . . . . . . . . . . . . 108 propane-diol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
organic liquids . . . . . . . . . . . . . . . . . . . . . . . . . . . 6, 12 propanol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
organic molecules . . . . . . . . . . . . . . . . . . . . . . . . . . 8 prostheses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
organometallic compounds . . . . . . . . . . . . . . . . 280 proteins . . . . . . . . . . . . . . . . . . . . . 233, 234, 236, 266
origins of life. . . . . . . . . . . . . . . . . . . . . . . . . . . . 67, 68 protocols . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170, 172
oxidation . . . . . . . . . . . . . . . . . . . . . . . . . 9, 79, 83, 138 pulse radiolysis . . . . . . 3, 28, 36, 46, 63, 83, 98, 233,
oxidative stress . . . . . . . . . . . . . . . . . . 233, 243, 254 . . . . . . . . . . . . . . . . . . . . . . . . . 234, 238, 240, 243
P R
paints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 radiation grafting . . . . . . . . . . . . . . . . . . . . . . . . . 142
pellets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 radiation processing. . . . . . . . . . . . . . . . . . 132, 134
peptides . . . . . . . . . . . . . . . . . . . . 233, 236, 240, 242 radiation-resistant. . . . . . . . . . . . . . . . . . . . . . . . . 134
peroxidation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257 radicals . . . . . . . . . . 80, 137, 144, 151, 192, 194, 233
peroxide radical . . . . . . . . . . . . . . . . . . . . . . . . . . . 255 primary radicals . . . . . . . . . . . . . . . . . . . . . . . . . 7
pesticides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 radical scavenger . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
photoactivation. . . . . . . . . . . . . . . . . . 295, 296, 299 radioactive isotopes . . . . . . . . . . . . . . . . . . . . . . . . 18
photography . . . . . . . . . . . . . . . . . . . . . . . . 111, 114 radioactive wastes . . . . . . . . . . . . . . . . . . . . . . . . 117
photographic development . . . . . . . . . . . . 111 radiosterilization . . . . . . . . . . . . . . . . . . . . . 139, 153
photographic sensitivity. . . . . . . . . . . . . . . . 114 radiotherapy . . . . . . . . . . . . . . . . . . 30, 59, 291, 298
photo-ionisation . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 3D conformal radiotherapy . . . . . . . . 291, 293
picosecond pulse radiolysis . . . . . . . 26, 29, 44, 45, hadrontherapy . . . . . . . . . . . . . . . . . . . . . 59, 298
.............................................................................. 48, 62 intensity modulated (IMRT) . . . . . . . . . . . . 294
platinum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 proton radiotherapy . . . . . . . . . . . . . . . . . . . . 30
Ä305Ä
2!$)!4)/.#(%-)3429
radio-chemotherapy . . . . . . . . . . . . . . 292, 299 spurs. . . . . . . . . . . . . . . . . . . . 4, 43, 57, 121, 124, 199

radiolytic footprints . . . . . . . . . . . . . . . . . . . . . . . 269 sterilization . . . . . . . . . . . . . . . . . . 139, 151, 153, 160
radioprotection . . . . . . . . . . . . . . . . . . . . . . 158, 277 streak camera. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
screens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278 sugars . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
radioresistant . . . . . . . . . . . . . . . . . . . . . . . . 157, 291 sugar backbone . . . . . . . . . . . . . . . . . . . . . . . 196
rate constants . . . . . . . . . . 39, 58, 99, 121, 128, 255 DNA sugar radicals . . . . . . . . . . . . 182, 196, 197
reactivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35, 43 sugar damage . . . . . . . . . . . . . . . . . . . . . . . . . 182
recoil nuclei . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 supercritical water . . . . . . . . . . . . . . . . . . . . . . . . . 54
recombination . . . . . . . . . . . 7, 14, 43, 114, 208, 213 superoxide dismutase (SOD) . . . . . . . . . . . . . . . 243
geminate recombination . . . . . . . . . . . . . . 7, 43 superoxide radical anion . . . . 58, 84, 180, 243, 254
reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 survival . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279
reduction potentials . . . . . 8, 80, 105, 112, 205, 255 synchrotron radiation . . . . . . . . . . 18, 75, 296, 299
relaxation process . . . . . . . . . . . . . . . . . . . . . . . . . . 47
repair . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219, 221 T
temperature . . . . . . . . . . . . . . . . . . . . 36, 42, 53, 125
S tert-butanol . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11, 84
scavengers . . . . . . . . 5, 30, 45, 81, 83, 114, 158, 159 tetrahydrofuran . . . . . . . . . . . . . . . . . . . . . . . . . 41, 44
scavenging factor . . . . . . . . . . . . . . . . . . . . . . . . . . 45 textile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
secondary electron . . . . . . . . . . . . . . . . . . . . . . . . . . 3 thermalisation distances . . . . . . . . . . . . . . . . . . . . . 7
secondary radicals . . . . . . . . . . . . . . . . . . . . . . . . . . 5 thermoluminescence . . . . . . . . . . . . . . . . . 156, 172
security assurance level (SAL) . . . . . . . . . . . . . . 152 thiols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
shape memory. . . . . . . . . . . . . . . . . . . . . . . . . . . . 142 thiyl radicals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
silver . . . . . . . . . . . . . . . . . . . . . . . . . . 45, 99, 101, 111 time-resolved microwave conductivity
silver halide . . . . . . . . . . . . . . . . . . . . . . . . 97, 112 (TRMC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
simulation . . . . . . . . . . . . . . . 41, 50, 54, 55, 125, 267 TiO2 catalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
computational methods . . . . . . . . . . . . . . . . 205 toxicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
molecular dynamics simulation . . . . . . . . . . 41 tracks . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55, 134, 199
Monte Carlo simulation . . . . . . . . . . . . . 59, 267 core . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55, 199
quantum simulation . . . . . . . . . . . . . . . . . 41, 50 penumbra . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
RADACK simulation model . . . . . . . . . . . . . 267 tryptophan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
solvated electron . . . . . . . . . . . . . . . . . . . . . 3, 35, 37 tumours . . . . . . . . . . . . . . . . . . . . . . . . . 177, 291, 299
charge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 tumour imaging . . . . . . . . . . . . . . . . . . 292, 293
mobility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 two-photon ionisation. . . . . . . . . . . . . . . . . . . . . 50
optical absorption . . . . . . . . . . . . . . . . . . . . . . 40
reactivity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 V
solvation cavity . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 viscosity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
solvation dynamics . . . . . . . . . . . . . . . . . . . . . . 36, 48 vulcanization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
spatial distribution of energy . . . . . . . . . . . . . 4, 132
Ä(%+Ä
)NDEX
W
wastewater . . . . . . . . . . . . . . . . . . . 79, 82, 89, 90, 93
water radiolysis . . . . . . . . . . . . . . . 4, 40, 56, 81, 254
water remediation . . . . . . . . . . . . . . . . . . . . . . 79, 89
X
X-rays . . . . . . . . . . . . . . . . . . . 17, 18, 38, 72, 132, 165
pulses of X-rays . . . . . . . . . . . . . . . . . . . . . . . . . 19
Y
yields . . . . . . . . . . . . . . . . . . . . . . . . . . 6, 59, 192, 254
initial yields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
primary yields . . . . . . . . . . . . . . . . 124, 128, 192
Z
zeolites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Ä307Ä

Radiationt Chemistry

Transféré par

Droits d'auteur :

Formats disponibles

Radiationt Chemistry

Transféré par

Informations du document

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Radiationt Chemistry

Transféré par

Droits d'auteur :

Formats disponibles

L I V R E S

#/,,%#4)/.$)2%#4%$"9 Paul RIGNY

Tous droits de traduction, d’adaptation et de reproduction par tous procédés,

© EDP Sciences 2008

Part I / Primary radiation-induced phenomena . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Chapter 2 Tools for radiolysis studies .................................................................. 17

Chapter 3 The solvated electron : a singular chemical species ................................. 35

Chapter 4 Water radiolysis under extreme conditions.

Part II / Radiation chemistry mechanisms and applications . . . . . . . . . . . . . . . . . . . . . . . . 65

Chapter 5 Molecular formation in the interstellar medium ..................................... 67

Chapter 6 Water remediation by the electron beam treatment ............................... 79

Chapter 7 Metal clusters and nanomaterials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97

Chapter 8 Water radiolysis in cement-based materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117

Chapter 9 Obtaining high performance polymeric materials by irradiation ............ 131

Chapter 10 Radiosterilization of drugs ............................................................... 151

Chapter 11 Food irradiation: wholesomeness and treatment control . . . . . . . . . . . . . . . . . . . . . . 165

III / Radiation damage to biomolecules,

Chapter 12 Radiation-induced damage to DNA : from model compounds to cell . . . . . . 177

Chapter 13 Mechanisms of direct radiation damage to DNA ................................. 191

Chapter 14 Charge motion in DNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203

Chapter 15 Genome maintenance mechanisms in response

Chapter 17 Radiation-induced damage of membrane lipids and lipoproteins ......... 249

Chapter 18 Predicting radiation damage distribution in biomolecules .................. 265

Chapter 19 Chemical protection against ionizing radiation .................................. 277

Chapter 20 Advances in radiotherapy : new principles ........................................ 291

Index ............................................................................................................... 301

“L’Actualité Chimique” is a monthly scientific journal meant to convey information

In the present volume, our purpose is to familiarise the larger communities of

Therefore we have called upon internationally recognized experts who kindly

With no pretence of exhaustively covering in detail all the topics of radiation

EMMI Salvatore Silvano

Radiation chemistry is the chemistry initiated by the interaction of high-energy

Radiation chemistry of water

H2O H2Ot F− and H2O* 10−16 s (2)

1 molecule/100 eV = 1.036 × 10−7 mol J−1 (or 0.1036 μmol J−1)

Radiation chemistry of organic liquids

The largest contribution of radiation chemical techniques to general free-radical

Properties of the primary radicals from water radiolysis

The hydrated electron may be visualised as an excess electron surrounded by

eaq− 4n Sn−1 (8)

eaq− 39 (RXt¦) Rt 9− (9)

HOClt¦ ) Clt )2O (12)

Similarly, pulse radiolysis studies of the oxidation of simple aquo-complexes of

Generation of secondary radicals

eaq− /2O N2 0t¦ (19)

eaq− 42O82− SO4t− 4042− (20)

A similar strategy can be used to generate a host of organic radicals as exemplified

eaq− 39 Rt 9− (25)

eaq− .n M(n¦

One-electron reduction and oxidation in organic solvents

CH3OH es−, Ht, tCH2OH, CH3Ot (27)

Radiolysis of solvents such as dichloromethane (CH2Cl2), tetrachloromethane (CCl4),

es− $$M4 CCl3t $M− (31)

The one-electron oxidation of a wide range of compounds has been investigated

In these nonpolar liquids, geminate ion recombination is the dominant process as

Types of ionizing radiation and their methods of generation

X- and gamma rays

radioactive source permanently fixed in shielding. A mechanism is used to transport samples

Particle accelerators can be used as continuous or pulsed X-ray sources by stopping

Particle accelerators for radiolysis

H2O H2OtF− and H2O* 10−16 s (2)

eaq−4n Sn−1 (8)

eaq−39 (RXt¦) Rt9− (9)

HOClt¦) Clt)2O (12)

eaq−/2O N20t¦ (19)

eaq−42O82− SO4t−4042− (20)

eaq−39 Rt9− (25)

eaq−.n M(n¦

es−$$M4 CCl3t$M− (31)

Na metal/)3 liq m NaNH FNH −

4 NH3 – 2 e– m 2 NH2t/)4 m N2H4/)4 (2)

NH3F− m eNH −

OHt)2 ¬ H20)t (1)

Ht02 ¬ HO2t (2)

Ht)2O2 ¬ OHt)2O (4)

OHt)2O2 ¬ Ht)02t (5)

O2)2)t ¬ 2H2O (6)