Synchrotron radiation imaging and diffraction for industrial applications
Synchrotron radiation imaging and diffraction for industrial applications
Jose Baruchel 0
Ake Kvick ESRF 0
0 About the author Massimo Altarelli was trained at the University of Rome as a con densed matter theorist. At the ESRF in Grenoble, he was Research Director (1987-1993), and Head ofthe Theory Group (1994-1998). Since 1999 he has been in Trieste on the staff of the International Centre for Theoretical Physics, and Director of the Elettra Syn chrotron Light Source
ious laboratories around the world. Clearly the technical difficulties connected with the requirements on the quality of the electron bunches (emittance, peak current, etc.) grow dramatically as the required wavelength decreases towards the range of the hard xrays. In the pursuit of the higher brilliance of fourth generation sources there is an alternative road to the Linac FEes. This is the approach based on recirculating Linacs, as exemplified by the 4GLS ("4th Generation Light Source") project being designed at Daresbury Laboratory in the UK. The idea is to combine the "few passes" strategy to lower emittance with the advantage of the ring geometry in which a large number of tangential beam lines and experimental stations can be exploited by users in parallel. It consists of a Linac that feeds electrons into one or several arc-shaped sections on which undulators, wigglers and even FEes can be inserted (Fig. 4). At the end of the (last) arc, electrons are either dumped or fed back into the Linac, where they can return part of their energy to the RF fields in the accelerating cavities (this latter version is called "energy-recovery Linac"). The Linac is preferably of the superconducting type, one of the reasons being that it can support a much higher repetition rate (a much smaller time mterval between successive bunches).
 See http://www-ssrl.slac.stanford.eduljhhorne.htrnlforashort
description of the state of the project
he use of the modern synchrotron radiation (SR) sources for
T X-ray imaging and diffraction provides new possibilities for
applied science and technology. They include an enhanced resolu
tion (spatial and/or temporal), the possible use of the coherence of
the beam, in-situ experiments and accurate quantitative measure
ments of local density or strain. Selected applications showing the
improved capabilities in absorption or phase microtomography,
time-resolved diffraction experiments showing the way a lithium
based battery works, or strain measurements on stir-welding
materials, are presented as examples. The combination of diffrac
tion and imaging (i.e. diffraction with micron spatial resolution)
allows further applications, such as the study of grain growth, or
the X-ray topographic investigation of defects responsible for spu
rious modes in single crystal resonators.
The aim of the present paper is to highlight the contribution of
synchrotron radiation X-ray imaging and diffraction to the inves
tigation of industrial topics. These techniques, which allow the
visualization of the volume of systems opaque for other probes and
the determination of their structure, have been dramatically
renewed by the use of modern, "third generation" , synchrotron
radiation (SR) facilities beams. Real timelhigh-resolution experi
ments are performed for both X-ray imaging and diffraction. The
high energy of the available beams allows the deformation within
bulk materials to be investigated. The high coherence of the beam
allows phase contrast imaging to be exploited, the contrast arising
from phase variations across the transmitted beam. Absorption
and phase microtomography provide three-dimensional informa
tion on features like cracks, porosities or inclusions, which are of
high interest for applied topics such as metallurgy, polymers, or
reservoir rocks containing fuel. The spatial resolution of these tech
niques is now in the 1 flm range. Lastly the combination of
diffraction and imaging, through the new tracking techniques, or
the use of Bragg diffraction imaging (X-ray topography) provides
additional information on industrially interesting materials.
<Ol Fig. 1: 3D image
of the distribution
of holes (in white)
within a copper
2 J.lm voxel size
4 Fig. 2: Reconstruction of an open-cell flexible polyurethane
foam at several levels of compressive strain. (a) 0%, (b) 10%, (c)
23%, these percentages being the ratio ofthe variation of vertical
length of the foam over the original, unstrained, length.The 30
renditions represent volumes of 7mm x 7mm x 1.4 mm.
The original features of third generation synchrotron radiation
facilities such as the APS, ESRF, or Spring 8, for imaging and dif
fraction applications, are:
a) the very high intensity of the X-ray beam (factor 106 with
respect to usual X-ray generators); both white and monochro
matic beams can be used
b) the availability of photons spanning the whole range from the
infrared to hard X-rays (up to 300-400 keV)
c) the design ofbeamlines optimised for a given set of techniques
d) the small size of the electron beam cross-section « 100 Jlm),
which leads to high brilliance, and to a sizeable lateral coherence
of the X-ray beam.
Absorption and phase microtomography
The principle of microtomography is very similar to that of the
well-known medical scanner. When applied to materials investi
gation, it consists in recording a series of radiographs (typically of
the order of 1000) for different angular positions of the sample,
which rotates around an axis perpendicular to the beam. Several
laboratory microtomographs have been commercially produced
over the last years
(Riiegsegger et al., 1996, Sasov and Van Dyck,
. But the best images, in terms of spatial resolution, signal
to-noise ratio and quantitative exploitation, are obtained using
synchrotron radiation. This results from the high intensity, practi
cally parallel and monochromatic incoming beam. In this
approach there is no image magnification, and the spatial resolu
tion mainly results from the effective pixel size of the detector. The
range of pixel sizes available at the ESRF goes from 0.3 Jlffi to 30
!lm, and a big effort is being produced to enhance the spatial reso
lution down to the 100 nm range. The total acquisition time is in
the 10-2 ("fast tomography") to1 hour range, and the recorded
data is often several Gigabytes.
An increasing number of applied or industrial research labora
tories require the use of microtomography to solve some of the
problems they encounter. To achieve this requirement they either
buy beamtime and expertise, or have access to the synchrotron
facilities through collaboration with Universities or research
groups and peer-reviewed proposals. The industrial requirements
include, most of the time, confidentiality, rapid access and full ser
vice. Most facilities, and in particular the ESRF, propose such a
service, going from the experiment to the data analysis (volume
.reconstruction, extraction of relevant parameters ... ).
Figure 1 is an example of such application-oriented investigation.
A strong impact on a copper target creates porosity within the bulk
metal. The investigation of the pore sizes and distribution is of high
importance for applications. The only way to image these pores
inside such a sample is X-ray microtomography
et al., 2001)
. The spatial distribution of pores within one of these
copper samples was determined (figure 1), and shown to be less
homogeneous, both from the point of view of size and oflocation,
than what was expected from theoretical models.
Another example of absorption tomography, which implies a
specially designed sample environment cell, is the in-situ investiga
tion of an open-cell polyurethane foam at several levels of
compressive strain (figure 2). This work, performed by a research
group from ICI Polyurethane in collaboration with a group of the
University of Cambridge, correlates the macroscopic behaviour
(stress/strain curve) with the local structure modifications. It
shows that the initial phase of compression, which is associated
with a linear elastic response, corresponds to a bending of the
struts, whereas the plateau in the stress/strain curve is linked to the
collapse a whole band of cells (Elliott et al., 2002). This result was
known from surface observations, but no volume evidence was
The X-ray beams produced at third generation synchrotron
radiation facilities exhibit a high degree of coherence. This results
from the small source size 0' (in the 50 Jlffi range) and the large
source to sample distance L (in the 100 m range). The transverse
coherence length de= 'A-L/2 0', is in the 100 Jlm range, and allows
"phase images" to be recorded by just varying the sample-to-detec
tor distance ("propagation technique", reviewed for instance by
Cloetens et al., 1999a). The great advantage of this new type of
imaging is the increased sensitivity it provides, either for light
materials such as polymers, or for composites made up of materi
als with neighbouring densities (for example AI and SiC). A first
use of the phase images relies on the visualisation of the phase
jumps that occur at the edges of a particle or porosity imbedded
in a matrix having a different index of refraction. Phase microto
mography based on the visualisation of the edges was used, for
instance, to understand the mechanisms of degradation in AI-SiC
composites. It was possible not only easily to visualize the SiC
reinforcing particles, but also to observe the nucleation and prop
agation of cracks when the material is submitted, in situ, to tensile
stress. The cracks appear first in the elongated particles, and their
number is 50% more than suggested by surface investigations (Buf
fiere et al. 1999).
4 Fig. 3: Tomographic images of an AI-Si alloy, quenched from
the"semi-solid" state a) Detector to sample distance D= 0.7 cm b)
0=60 cm c) "holotomographic"image, d) 3D rendering shOWing
the former"liquid" phase (courtesy L. Salvo and P. C1oetens).
Phase imaging based on the detection of the edges does not
allow the local phase to be extracted quantitatively, and its spatial
resolution is limited by the occurrence of the fringes used to visu
alize the borders. A more quantitative approach of phase imaging
and tomography was developed. It is based on the combination of
several images recorded at different distances. An algorithm, ini
tially developed for electron microscopy by the Antwerp group,
was successfully adapted to the X-ray case, and allows the "holo
graphic" reconstruction of the local phase, well beyond the images
of edges (Cloetens et al., 1999a). Once the phase maps are obtained
through holographic reconstruction, there is no conceptual diffi
culty in bringing together many maps corresponding to different
orientations of the sample, and in producing the tomographic,
three-dimensional, reconstruction procedure. For each of the
angular positions of the sample, the phase map is retrieved using
images recorded at several (typically four) distances. The highest
accessible spatial frequency is determined by the resolution of the
detector. This combined quantitative phase mapping and tomog
raphy procedure, called holotomography (Cloetens et al.,1999b),
provides a very useful approach to the characterization of materi
als on the micrometer scale.
The "holotomographic" procedure was applied to an AI-Si
alloy; quenched from a temperature where a "mainlyAI" solid phase
is surrounded by an AI-Si liquid one. This is an important indus
trial material because in the "semi-solid" state it is possible to give
a desired shape to the alloy, this shape being retained when cooled
to room temperature. The density difference between the two
phases is in the 1-2% range. For each angular position of the sam
ple 4 images were recorded at different distances, yielding 3D
quantitative images that show, with a voxel edge size of 1 filll' the
distribution of electron density, hence of mass density, in the sam
ple. The highest accessible spatial frequency is determined by the
maximum resolution of the detector (-1-2 flm). Fig 3 compares
the possibilities of absorption microtomography (fig. 3a, where the
phases are indistinguishable, and only iron-rich metallic inclusions
are visible) with edge-enhancement phase microtomography (fig.
3b, where the phase boundaries are outlined by a black-white line)
and with holotomography (fig. 3c, where the two phases are clear
ly observable through their grey level), and shows in 3d) a 3D
rendering of what was the ''liquid'' phase.
Diffraction experiments on industrially related topics
The accessibility of hard penetrating X rays that can be focussed
into sub micron spots while retaining extremely high probe inten
sity has opened up the field of diffraction to industrially interesting
fields. New experiments with millisecond time-resolution, sub
micron spatial resolution with bulk probing capability of several
tens of centimetres in high Z materials can now be performed to
map out stress/strain fields, to follow reactions in real time or to
study in situ processes in complex environments such as inside
reaction vessels or even in working devices like batteries or other
Time-resolved studies tend to fall into three categories: the second
to several minute processes, millisecond to microsecond processes
and extremely fast pico-second reactions. The last category can
only be studied at present by "stroboscopic" experiments were the
process is repeated many times to obtain sufficient counting statis
tics. Most industrial reactions however fall into the first two
categories and are usually irreversible ruling out stroboscopic mea
surements. The present synchrotron facilities provide sufficient
x-ray flux: to perform one-shot experiments well down in time to
- --_... -----
the microsecond regime. The 2 dimensional diffraction data is
recorded by fast read-out detectors such as large CCD cameras.
An example is given by the in-situ study of an electrochemical
reaction experiment on a working LixCo02battery (Morcrette et al.
2002). Strong diffraction spots from higWy crystalline current col
lectors and packaging material have in the past compromised the
weaker diffraction from the material under study. In order to de
emphasise the strong diffraction it is now possible to use micro
beams and combine diffraction rapidly collected from many dif
ferent areas. In the case of the LixCo0 2 battery 16 different
orientations were collected in less than 2 secs and the pixel by pixel
medians of the images collected by the CCD camera could be
used to reduce the influence of the electrodes. One-dimensional
powder patterns from 2 D images can thus be produced with suf
ficient precision to obtain accurate time-resolved behaviour of the
phase fractions during the working cycle of the battery. The struc
tures of the phases that are stable only under applied voltage
12 lJ I~
Ti .. '}+ltC
... Fig. 5: (a) The diffraction pattern during the reaction between A
(anatase),G (graphite) and M (magnesium).The timeresolution is
0.65 msec. (b) The diffraction patterns captured while the reaction
between the metals take place.The time interval is 0.65 ms.
.... Fig.6: Schematic of the experimental setup with spiral sUt and
online image plate scanner on 1015B.
showed the existence of intermediate phases as well as the termi
nallayered Co02 structure. Figure 4 gives the diffraction phases
during in situ reduction of the LixCo02 battery with the different
rhombohedral, monoclinic and triclinic phases indicated.
The highest time-resolution for irreversible reactions so far
(l0-30ms) has been obtained in the study of self-propagating high
temperature synthesis (SHS). SHS is based on the characteristic of
highly exothermic reactions to sustain themselves following an ini
tiation by relatively mild conditions. These reactions typically
proceed as reaction fronts with speeds of up to 25cm/sec under
temperatures up to 5000e. An example is given by Contreras et al.
2004 where the production of TiC and TiB2 was studied via two
routes: from elements and from oxides. The reactions were moni
tored by 0.2601 A X rays in situ. The starting powders were
compacted to cylindrical pellets and the reaction was started by
external ignition at the bottom of the pellets. The X-ray beam was
focused on the central part of the pellet and the reaction process
was recorded as the reaction front passed the probing X-ray beam.
The bulk process was followed by recording the diffraction in
transmission by a CCD camera. The detailed process could be
recorded and showed that the entire process took place during less
than 0.1 second. The different synthesis routes produced different
final microstructures that depended on the starting conditions.
The particle size when pure elements were used was 10 flIll where
as the reaction using oxides and Mg had sub-flIll microstructure.
Fig. 5 gives the diffraction patterns during the critical phases of
Combined diffraction and imaging
High spatial resolution
The mapping of residual stresses in components is essential in pro
viding a roadmap for future failure modes in materials and a
detailed knowledge of their distribution can indicate potential alle
viating processes, which may lengthen component lifetime and
avert catastrophic failures such as fractures in railroad tracks or air
plane turbine blades. Until recently only destructive methods or
europhysics news MARCHI APRIL 2004
methods with large measuring gauges have been employed.
Recent deVelopments at the high energy synchrotron facilities have
now produced highly penetrating micro-focus beams with mea
suring gauges down to -5*5*50 flm3• The experimental set-ups
with CCD cameras or narrow receiving slits and micro-precision
translation tables now give spatial resolution on the flm scale.
High precision lattice parameters with 6.d/d -10'6 can thus be
determined with the small measuring gauge and excellent spatial
resolution. A recent example of a novel strain and phase-mapping
technique is given by
Martins et al. 2003
in their depth resolved
investigation of friction stir welding. Friction stir welding is a new
method for solid welding using a spinning tool forced along a joint
line. This new method can overcome problems such as weld poros
ity; use of filler materials and cracking in the heat-affected zone. In
this study the bulk investigation of several mm thick AI plates was
performed, and the depth resolution was obtained by a spiral slit
The experimental set up is illustrated in Fig. 6 and the resulting
distribution of the Si phase and the residual micro strain is given in
Fig. 7 for this weld of a single phase AI alloy and an AISilOMg phase.
The excellent spatial resolution now available opens up entirely
new opportunities for studying dynamical phenomena at the
mesoscopic level. Surprisingly theoretical understanding of many
fundamental materials processes, such as recrystallisation and
deformation, relies on models that have built in assumptions such
as sample homogeneity and absence of grain-grain interactions.
The new experimental facilities can now test these assumptions. An
example is given by Offerman et al. 2002 in their study of grain
growth following phase transformations in a steel sample. Using
the spatial resolving power of the synchrotron beam it was shown
long~udinal strain map
'"C';:~~:::;:~...,-=;;'::::;;;::::;;;::::~ _ ·'000
b) through thickness longitudinal reSidual stress
q~2Olf'','~''---'--''-;'~., ,y. , , :~ " S}'SI<,!r-..c e.m>.r. i l
c) r....··..e~i~;~:d view for Ph~;';;;~P'''''''l
1 d - [
J_ , 1.'- , ,~
. ~o _tro·5m_ete~._r mSml 10 "IS
... Fig. 1: (a) Depth resolved residual macrostrains in
longitudinal direction (parallel to weld). (b) through
thickness long.itudInal residual stress component,
calculated from the strain measurements in three
dimensions. (c) enlarged view of the distribution of the plate.
material (AA60B2 and AA2024) in the stirring zone.
that four different growth mechanisms were present in ferrite and
these were correlated with the local environment of the grains indi
cating that the present models are far from adequate. Fig. 8 shows
the integrated intensity of 5 grains near the cube orientation shown
as a function of annealing time. At present several hundred differ
ent grains can be monitored simultaneously. This method of
simultaneous determination of orientation within a bulk sample
(Lauridsen et al. 2000)
gives interesting access to the mesoscale. It
is now possible, via tracking techniques, to map out grain bound
aries with resolutions down to 5 flm and to follow kinetics and .
dynamics of groups of individual grains during processes such as
heating, torsion etc.
Industrial application of Bragg diffraction imaging (X-ray topography)
Figure 9 shows the image recorded using the 444 Bragg diffracted
beam from a flux-grown platelet-shaped gallium-substituted yttri
um iron garnet crystal (Y3GaxFes-xO12, with x-I, Ga-YIG). These
Ga-YIG crystals are grown to produce hyper-frequency resonators.
It was observed that some of these crystals exhibit spurious modes,
which are very detrimental to their performance. This X-ray dif
fraction topographic investigation was therefore performed to
identify the crystal defects present on some of the resonators, pro
duced during growth, which are responsible for these spurious
modes. Many types of defects are observable on the topographs,
and one of them, the dissolution bands, appeared to be the origin,
through the occurrence of "closure-like" magnetic domains, of
The above examples give only a few insights of possible studies of
industrial phenomena using diffraction and imaging at the syn
X-raymicrotomography is an invaluable tool to obtain 3D data
on a large variety of materials. The use of modern synchrotron radi
ation sources opens up new possibilities, with "fast tomography" to
investigate the evolution of a system when varying an external para
meter (time, temperature, stress, etc.) and/or improved spatial
resolution, going down to the 100 nm range. In addition phase
images reveal phenomena hardly visible by other means.
Other areas include the rapidly expanding field ofhigh-pressure
diffraction where micro- focussed beams and laser-heating capa
bilities open up new experimental routes to studies of synthesis
.. Fig. 8:The integrated intensity of five grains near the cube in
steel as a function of annealing time
'" ~Topograph of a flux grown Ga-YIG crystal.
and equation of states under non-ambient conditions such as
megabars of pressure and several thousands of degrees of temper
High-resolution powder diffraction is advancing rapidly and
ab initio structure determinations ofvery large structures can now
be performed on materials, which cannot be obtained as single
The current trend in instrumentation development will in the
near future give access to nanometre size beams. This leads to
another type of image through the use of X-ray microbeam-based
scanning imaging approaches, probing for instance the fluores
cence or the absorption near absorption edges. The diffraction
experiments will more and more be combined with complemen
tary techniques such as imaging or Raman spectroscopy, by
performing simultaneous observations. Development in detector
technology will further advance the time- and space-resolving
aspects of the experiments.
About the authors
Ake Kvick graduated from Uppsala University in 1974. Work on
hydrogen bonded systems, zeolites and time-resolved studies with
neutrons and X-rays. Scientist at Brookhaven National Lab. 1980
1989. From 1989 work at the ESRF on synchrotron radiation
diffraction; presently Head of Materials Science Group and editor
of J. Synch.Rad.
Jose Baruchel graduated in Physics at the University of Greno
ble. From 1971 to 1991, work on magnetic domains and nearly
perfect crystals, with X-rays and neutrons. From 1991, work at the
ESRF on synchrotron radiation imaging (topography and tomog
raphy); presently Head of the X-ray Imaging Group.
EDP Sciences' journals
All the information on EDP Sciences'
journals for authors and institutions and
access to the ~Iectronic versions for
readers is available at:
http://www. e dpsciene.:e$~org
europhysics news MARCH!APRIL 2004
What is your strategy for
Sean McCarthy, Managing Director ofHyperion Ltd.
ow many research organisations have a formal written strat
H egy for Framework 6? How many individual researchers
have a personal strategy for Framework 6? This article presents a
checklist for the writing of a simple but practical strategy for
An organisation's strategy for Framework 6
The following are examples of statements of objectives that could
be used in a research organisation's strategy for Framework 6.
They are based on actual discussions with research organisations
on their Framework 6 strategies. (The comments in brackets are
actual quotations from research managers)
To access European Union Funding for research activities. One
of the problems with this simple objective is that the researchers
may concentrate on obtaining any contracts rather than
focussing on the research priorities of the organisation.
To establish the research group as the European (or Internation
al) scientific leader in a scientific area ('to be the best in our
field'). It is important that the scientific areas are well defined (a
niche within an niche). For example, , to be the world leader in
the simulation and design of photovoltaic (solar) systems'.
To access new technologies relevant to the organisation's areas of
excellence ('to avoid missing the train').
To work with the best research partners in the European Union.
('to be the preferred partner of the best scientists in our field')
To provide better education and training to graduates and
post-graduates. ('We want our graduates to excel in any inter
view, anywhere in the world')
• To promote the organisation's scientific and technical excellence
to the scientific/technical community ('more conferences, more
publications, more website hits.. :)
To ensure the research results are used by enterprises and soci
ety ('getting value from our research efforts')
To provide relevant support to researchers ('To streamline the
process of proposal writing, contract negotiation and contract
Bontaz-Carion J. , Nicollet M. , Manczur P. , Pellegrini Y.-P. , Boiler E. , Baruchel J . Dynamic damage in metal: porosity as a testfor damage models, in A . Chiba, S. Tanimura and K. Okamoto eds., Impact Engineering and Application, Proc. 4th Int. Symp. on Impact Engineering , pp. 633 - 638 , Elsevier, Oxford ( 2001 ).
Buffiere J.Y. , Maire E. , Cloetens P. , Lormand G. and Fougeres R. , Characterization ofinternal damage in a MMCp using synchrotron phase contrast microtomography , Acta Mater . 47 , 1613 - 1625 ( 1999 ) Cloetens P. , Ludwig w., Baruchel J. , Guigay J.P. , Rejmankova-Pernot P. , Salome-Pateyron M. , Schlenker M. , Buffiere J.Y. , Maire E. , Peix G. , Hard X-ray phase imaging using simple propagation ofa coherent synchrotron radiation beam . J. Phys. D: Appl. Phys . 32 , A145-A151 (1999a) Cloetens E , Ludwig W , Baruchel J" Van Dyck D. , Van Landuyt J. , Guigay J.P. , Schlenker M. , Holotomography: quantitative phase tomography with micrometer resolution using hard synchrotron radiation X-rays , Appl. Phys.
Lett. 75 , 2912 - 2914 ( 1999b ) Contreras 1 ., Turillas X. , Vaughan G.B.M. , Kvick A. , Rodriguez M.A.
Time-resolved XRD study ofTiC- TiBl composites obtained by SHS . To be published ( 2004 ) Elliott J.A. , Windle A.H. , Hobdel J.R. , Eeckhaut G. , Oldman R.J. , Ludwig W , Boiler E. , Cloetens E , Baruchel J. , Foam compression by microtomography , J. Mat. Science , 37 , 1547 - 1555 ( 2002 ).
Lauridsen E.M , Juul Jensen D. , Poulsen H.P. Lienert U , Kinetics of individual grains a during crystallisation ., Scripta Materialia , 43 , 5561 - 566 ( 2000 ).
Martins R.V , Honkirnaki v., Depth resolved investigation offriction stir welds using a novel strain and phase mapping technique , Texture and Microstructure , 35 , 145 - 152 ( 2003 ).
Morcrette M. , Chahre Y. , Vaughan G.B.M. , Amatucci G. , Leriche J.B. , Patoux 5 ., Masquelier c., Tarascon J.M. In situ x-ray diffraction techniques as a powerful tool to study battery electrode materials .
ElectrochirnicaActa , 47 , 3137 - 3149 ( 2002 ) Offerman S.E, van Dijk N.H , Sietsma J. , Grigull 5 ., Lauridsen E.M, Margulies 1 ., Poulsen H.P. , Rekveldt M. Th ., van der Zwaag S. Grain nucleation and growth during phase transformations . Science , 298 , 1003 - 1005 ( 2002 ).
Riiegsegger E , Koller Rand Miiller R ,. A microtomographic system for the non destructive evaluation of bone architecture, Calcif . Tiss.lnt., 58 , 24 - 29 ( 1996 ).
Sasov A. and Van Dyck D. , Desk-top microtomography: gateway to the 3D world , European Microscopy and Analysis , pp. 17 - 19 ( 1998 ).