Complex materials that exhibit special properties are regularly studied nowadays both in basic and applied research. Examples given in this chapter show the great variety of studies, be it the phase transitions in a quasicrystal, the dislocation microstructure of deep Antarctic ice, the nanometer-sized metal cluster formation in glass, the complex structure of a zeolite in a textured sample or the analysis of stress in a ceramic and a bond layer deposited on a superalloy. The techniques used range from diffraction to grazing-incidence small-angle scattering, via high-energy diffraction, topography, powder diffraction and EXAFS, demonstrating the versatility of ESRF beamlines and the effort made to improve and adapt the techniques to new challenges.


Phason driven phase transition from the icosahedral i-AlPdMn phase to the F2M AlPdMn modulated phase

Quasicrystals are long-range structures with a lack of translational invariance. Just how perfect quasicrystals are and how the long-range quasiperiodic order can propagate are still matters of controversy. Phason fluctuations, which are specific to quasicrystals, play a particular role. In the high-temperature region they give an entropic contribution to the free energy of the quasicrystal.

The relative weight of the energetic or the entropic term leads to two ideal models: the energetically or the entropically stabilized quasicrystals the latter being called a random tiling model. In this context the study of phase transitions is particularly interesting, since it directly addresses the question of stability.

For a small shift in chemical composition the icosahedral phase is only stable above 720°C and transforms reversibly into the F2M-AlPdMn phase for a lower temperature [1]. High-resolution electron microscopy images of the low temperature F2M-AlPdMn phase showed that its structure is a multidomain structure with an overall icosahedral symmetry. Each domain has a size of the order 1 µm, with a cubic symmetry, but remains quasiperiodic [1]. The overall icosahedral symmetry is given by the five different orientations of the cube with respect to the icosahedron.

A complete room-temperature study has been carried out on beamline ID1 together with an in-situ temperature study on the CRG D2AM beamline. Several 2D X-ray diffraction maps of the F2M-AlPdMn phase have been recorded. They present an extremely rich structure as shown in Figure 84. The same area recorded in the i-AlPdMn phase shows only one icosahedral Bragg reflection labelled 19/28 on the Figure. In fact the F2M phase corresponds to a superstructure or a commensurate modulation over the F2 phase 1, [1, 2].

Both the F2 type reflections (labelled SF2) and the i-type reflections are surrounded by satellite reflections located at a position qS1, and their linear combination, in agreement with electron diffraction patterns. The wavevector qS1 lies along a direction parallel to the icosahedral 3-fold axes (black straights in Figure 84), and is commensurate with the i-AlPdMn reciprocal lattice. It corresponds to a wavelength of the modulation equal to 170 Å.

About 150 reflections were measured close to the Pd edge in order to obtain structural information. This confirmed that the S1 satellite intensities can be reproduced by a simple sine phason wave distortion of the high temperature icosahedral phase. The asymmetry of integrated satellite intensities located at +qS1 and ­qS1 is in agreement with a superstructure indexing of these reflections. The intensity distribution of SF2 reflections is much more complicated, but seems to be related to the main Bragg peak intensity.

The phase transition was followed by in-situ diffraction using either a scintillator detector or a 2D CCD camera located at 40 cm from the sample, and an incoming X-ray energy of 17 keV. A series of oscillation pictures (± 0.05°), gave images with an isotropic resolution equal to 0.005 Å-1. With this setting the phase transition could be followed, with a time-scale of the order of 2 mn. As previously shown, the phase transition is reversible. Figure 85 shows a series of images coming from the high-temperature phase. The diffuse scattering intensity, which is present in the high-temperature phase, increases and becomes structured along the 3-fold axes. The final state shows clearly the satellite reflections. The intensity distribution of the diffuse scattering is directly correlated to the phason component (which is called Qper) of the diffraction peaks, i.e. to phason fluctuations [3, 1]. All these results strongly suggest that the phase transition is driven by phason fluctuations. However its exact microscopic description has not yet been achieved.

[1] M. de Boissieu, M. Boudard, T. Ishimasa, E. Elkaim, J.P. Lauriat, A. Létoublon, M. Audier, M. Duneau, A. Davroski, Phil. Mag. A. 1998, 78, 305.
[2] T. Ishimasa and M. Mori, Phil. Mag. B., 1992, 66, 513.
[3] M. de Boissieu, M. Boudard, B. Hennion, R. Bellissant, S. Kycia, A.I. Goldman, C. Janot and M. Audier, Phys. Rev. Lett., 1995, 75, 98.

A. Létoublon (a), M. de Boissieu (a), J.P. Simon (a), J.F. Bérar (b), M. Capitan (c), J.L. Joulaud (c).

(a) LTPCM, ENSEEG, St Martin d'Hères (France)
(b) Laboratoire de Cristallographie, CNRS, Grenoble (France)
(c) ESRF




Presence of a characteristic diffuse scattering line shape in icosahedral quasicrystals

After the discovery of quasicrystals a large effort was made in order to determine the bulk structure of these quasiperiodic systems. Special attention was paid during the last year to the study of the diffuse scattering intensity. Up to now, these studies have led to quite a range of results and interpretations, with some apparent discrepancies between the different conclusions. On the one hand, neutron studies revealed a background diffuse scattering slightly modulated in the reciprocal space that has been attributed to an apparently important chemical and geometrical partial disorder. This disorder was considered as an intrinsic feature of the icosahedral structures. On the other hand, high resolution X-ray diffraction studies have shown that quasicrystals have a densely-filled reciprocal space that should be taken into account in the background intensity of the low-resolution neutron scattering maps. The ESRF facilities (ID1) were used to tackle this problem by taking advantage of a high-resolution configuration with a still-comfortable brilliance for recording scattering diffusion maps in a limited part of the two-fold plane in reciprocal space.

Figure 86 (left side) shows the diffuse scattering map obtained for a thermodynamically stable icosahedral phase of the AlPdMn system by means of X-ray diffraction. This phase exhibits diffuse intensity that can be classified into two kinds:

1- a diffuse scattering around Bragg peaks with specific shapes depending on the point symmetry of the peaks and extending preferentially along the three-fold directions of the quasicrystal;

2- zones of the reciprocal space between Bragg peaks where the background intensity is extremely low, close to the standard background noise. The first type of diffuse intensity has been shown to be extremely dependent on the quasicrystal composition.

The first type of diffuse scattering appears as a very robust feature of the quasicrystals. The question is posed to determine whether they could be characteristic of hypothetical pre-transitional behavior or not. This icoshedral diffuse scattering can be fairly well reproduced qualitatively by theoretical simulations based on the Jaric and Nelson approach which assumes simple phason-phason correlations in the equilibrium quasicrystal. The good agreement between the simulations (right side in Figure 86) and the experiment shows that there is no need to invoke a nearby phase transformation for reproducing the line shape of the Bragg peaks, especially the typical streaks along the three-fold directions. They result from the icosahedral symmetry out of which a generic icosahedral hydrodynamic matrix is constructed. They are, to some extent, a sort of signature of the true icosahedral nature of these quasicrystals.

M.J. Capitan (a), Y. Calvayrac (b), J.L. Joulaud (a), M. Bessiere (c) and D. Gratias (b), to be published.

(a) ESRF
(b) CECM.-CNRS, Vitry sur Seine(France)
(c) LURE, Université Paris Sud, Orsay (France)




Microstructure in deep Antarctic ice: a clue to the climate 4000 centuries ago?

The deformation of glacier ice provides direct information on the flow of polar ice sheets. X-ray diffraction imaging ("topography") in turn provides, through the visualization and characterization of dislocations, information about the deformation processes suffered by the single crystal sample. A 3326 m long ice core recently recovered at Vostok in East Antarctica provides information on the past climate over time scales extending to more than 4000 centuries.

Very large ice single crystals (about 50 cm) were found below 3000 m where the temperature is near the melting point. This large grain size appears to be related to the occurrence of continuous recrystallization during the slow deformation of ice by dislocation glide.

The analysis of the dislocation microstructure within these single crystals by X-ray topography, taking advantage of the good geometrical resolution available at the ESRF, thus appeared highly interesting. The samples were large (~ cm3) ice single crystals from the Vostok core.

Figure 87 shows a 0002 projection topograph obtained on ID19. The image corresponds to a part of a thin tube with axis along the c-axis. This 3D structure is clearly displayed on the section topograph and schematic drawing given in Figure 88. The rocking curves do not exhibit the presence of sub-grain boundaries, but indicate a continuous change in lattice orientation, as produced by a three-dimensional distribution of dislocations.

Figure 89 shows the image obtained using Bragg reflection from prismatic lattice planes. It corresponds to the projection of tubes with different diameters, similar to those observed on the basal plane topographs, and not to sub-grain boundaries as was deduced from previous measurements with other X-ray sources.

We can conclude that there is no significant distortion of the lattice around the c-axis. Such a structure, which has never been observed in crystals grown in the laboratory, should be produced by the bending of the ice crystal through isolated basal dislocations, with dislocation density ~ 106 cm/cm3. This value is consistent with that deduced from the modelling of the deformation of ice along this deep ice core.

These promising results on the characterization of the dislocation microstructure during the "high temperature" deformation of ice are being complemented by the observation of very weak lattice distortions and individual dislocations on several large crystals recovered below 3600 m. The dislocation density is estimated to be less than 104 cm/cm3 , i.e. comparable with the best quality ice single crystals grown in the laboratory.

L. Arnaud (a), J. Baruchel (b), O. Brissaud (a), S. de La Chapelle (c), P. Duval (a), J. Härtwig (b), T. Hondoh (d), V. Lipenkov (e), J. Ocampo (a), E. Pernot (b), to be published.

(a) LGGE-CNRS, Grenoble (France)
(b) ESRF
(c) Ecole des Mines, St. Etienne (France)
(d) ILTS, Sapporo (Japan)
(e) AARI, St. Petersburg (Russia)




First evidence of Cu-Ni alloy nanoclusters formation induced by double ion implantation in silica glass

Metal doped glasses are promising candidates for application in non-linear integrated optics and photonics, i.e. in all optical switching device technology. In fact glasses containing nanometer-sized metal clusters exhibit in the visible range an enhanced intensity dependent refractive index n2 (n = n0 + n2I, where n0 is the linear index and I is the light intensity), roughly seven orders of magnitude larger than silica glass! The interest in these composite materials is also due to the general interest in the behavior of strongly confined systems.

Ion implantation is currently used to obtain this kind of composite. In this study it is shown that by using double ion implantation, the preparation of colloidal structures containing alloy clusters of different metals is possible, with the related possibility of tailoring their optical performances.

Cu, Ni and Cu + Ni implanted silica samples have been investigated, at implantation energies lower than 100 KeV and fluences lower than 1017 atoms/cm2: this induces nanocluster concentration below the percolation limit (dispersed clusters). To determine the local order around Cu and Ni atoms, X-ray absorption spectroscopy (XAS) measurements were performed at the CRG beamline GILDA (BM8) in fluorescence mode at liquid nitrogen temperature, using a seven-element high-purity germanium detector. A complex picture of the clusters structure emerges, consisting generally of partially oxidized metal.

In the doubly implanted sample, the data analysis performed both at the Ni and at the Cu K edges have shown the same first metal-metal distance value of 2.52 Å. This value is in between the Cu-Cu (2.5561 Å) and the Ni-Ni (2.4918 Å) nearest neighbor distance in metals, showing the formation of Cu-Ni metallic alloy nanoclusters. To determine the alloy composition, literature reported lattice parameter of Cu1-xNix bulk alloy were compared with the value deduced from the first shell-metal distance, assuming negligible lattice distorsion with respect to an ideal lattice (Figure 90); the alloy composition found is around Cu60Ni40. This result is in agreement with the absence in the visible absorption spectra of the Cu + Ni sample of the characteristic surface plasmon resonance of Cu clusters (Figure 91): the absorption band centered around 2.3 eV, clearly visible in the Cu implanted sample, is no longer present after Ni implantation. Theoretical absorption spectra of Cu-Ni alloy clusters, embedded in silica, calculated in the framework of the Mie theory, show a good agreement with the experimental data (Figure 91 inset).

The formation of Cu-Ni alloy nanoclusters in silica also induces variations in the non-linear optical properties: the Cu + Ni sample shows n2 values four times larger than the single-metal implanted ones.

F. D'Acapito (a), F. Cattaruzza (b), F. Gonella (b), C. Maurizio (b), P. Mazzoldi (b), F. Zontone (c) and S. Mobilio (d), to be published.

(a) Consiglio Nazionale delle Ricerche, c/o GILDA CRG, ESRF
(b) Physics Department University of Padova ( Italy)
(c) ESRF
(d) Physics Department University of Roma Tre, Roma (Italy)




Morphological study with grazing incidence small angle X-ray scattering (gisaxs) of Cu-Ni alloy clusters obtained by ion implantation in silicate glasses

Owing to dielectric and quantum confinement effects, glasses containing nanometer-sized clusters enhance their intensity-dependent refractive index values by up to 7 orders of magnitude in visible light. This is of fundamental importance in all optical switching device technology. These composite materials can be easily prepared by ion implantation, which determines the formation of nano-sized clusters usually localized in the first 1000 Å of the implanted sample. The optical properties are strongly dependent on the cluster sizes and the size distribution, as well as on the mutual distance among nanoparticles.

Small-angle X-ray scattering is an experimental technique that allows morphological investigations on nanometer-sized particles. Small-angle scattering investigates only the low q-exchanged region (q 0.6 Å-1) of the diffraction pattern: here we obtain information on the electronic density modulation in a length range from tens to hundreds of Å. Grazing incidence geometry has been used to maximize the signal coming from the thin implanted layer with respect to that originated by the substrate. The results of GISAXS analysis on two samples, prepared by Cu + Ni double ion irradiation in silica (CuNiS) and in soda-lime glass (CuNiL) are reported. Implantation energies were lower than 100 keV, with fluences of about 1017 atoms cm-2. Double Cu + Ni implantation has caused the formation of Cu-Ni alloy nanoclusters in both samples.

The formation of nanoparticles in glasses by ion implantation is influenced by the physical and the chemical properties of the glassy matrix. Soda-lime glass is a more reactive substrate than silica, owing to the presence of non-bridging oxygen atoms. Also the atomic structures are very different: in soda-lime, the glassy network is not as continuous as in the silica. Hence, the characteristics of the cluster formation may be different.

GISAXS measurements have been carried out at the ESRF on ID1. Data were recorded with a 2D gas-filled detector, using a photon beam energy of 8.327 keV. The signal integration provided the intensity function, I(q). Experimental data were analyzed using the so-called "decoupling approximation": cluster position and size were assumed to be independent. The assumption of a spherical shape for the clusters and a hard-sphere interaction among them was used. Consequently, the diffracted intensity is I(q) P(q) * S(q), where P(q) and S(q) are the particle form factors and the particle ensemble structure factor, respectively. Figure 92 shows that the experimental data and the calculated pattern are in good agreement. The mean radius R of nanoparticles is about 25 Å for both samples. The FWHM of the radius distribution is 14 Å for CuNiL and 4 Å for CuNiS. In both samples, there is an evident correlation among cluster positions: the mean inter-particle correlation distance is 85 ± 10 Å. The main morphological difference between the two samples is that CuNiL shows a polydisperse cluster distribution, whereas CuNiS is a monodisperse system. The explanation of this behavior is under investigation.

D. Thiaudière (a), S. Leguiess (a), C. Maurizio (b), A. Longo (c), F. d'Acapito (d), E. Cattaruzza (b), F. Gonella (b), P. Mazzoldi (b), A. Martorana (c), to be published.

(a) ESRF
(b) INFM and Dipartimento di Fisica, Padova (Italy)
(c) ICTPN­CNR and Dipartimento di Chimica Inorganica, Palermo, (Italy)




Crystal structure solution using textured powder samples

Powder diffractionists generally go to great lengths to ensure that the crystallites in their samples are randomly oriented. In principle, however, more information about the relative intensities of individual reflections can be obtained from a sample with preferred orientation [1]. Such a sample can be considered to be something between a single crystal and a randomly oriented powder. By measuring the diffraction pattern with the sample in different orientations, the overlap of reflections in 2 can be reduced, and the solution (or refinement) of more complex structures from powder data made possible.

This approach has not been used in the past, because the tilting of a sample on a conventional laboratory diffractometer results in a violation of the para-focussing condition and results in severe line broadening. The parallel nature of synchrotron radiation combined with a pre-detector analyzer crystal, however, allows high-resolution data to be collected at all tilt-angles.

To test this approach, a texture attachment was built and installed on the Swiss-Norwegian CRG beamline (BM1) at the ESRF. A textured sample of the very complex zeolite ZSM-5 was prepared. Pole figures of eight single reflections were measured to determine the texture (the orientation distribution function, ODF), and then four complete diffraction patterns were collected with the sample in different orientations. One of the pole figures and small sections of two of the diffraction patterns are shown in Figure 93. A single set of reflection intensities was extracted from the four patterns using the information from the ODF. As could be shown later, the individual intensities of overlapping reflections could be extracted quite accurately. These quasi single crystal data were used as input to a direct methods program running in default mode, and all 12 of the Si positions and 19 of the 26 O positions were found in the first 40 peaks in the E-map. Similar results were obtained for truly single crystal data.

The method was then applied to the aluminophosphate molecular sieve Mu-9 (framework formula: Al66P72O288 per unit cell, space group: R-3c, unit cell: a = 14.0612 Å and c = 42.2885 Å) which had an unknown structure. The plate-like morphology of this phase easily lent itself to the preparation of a textured sample. Again, the crystal structure dropped out immediately using direct methods. The framework structure shown in Figure 94 is fairly complex with 13 atoms in the asymmetric unit. It has an unexpected 6-connected Al which would have made it difficult to solve the structure using a direct space structure solution algorithm. The latter assume a 4-connected framework.

It has been demonstrated that by using synchrotron diffraction data from a textured powder sample, very complex crystal structures can be solved without making any assumptions about the structure. The structures can be of a complexity which hitherto could only be determined from single crystal data.

[1] R. Hedel, H.J. Bunge & G. Reck, (1997). Textures and Microstructures, 29, 103-126.

Thomas Wessels (a), Christian Baerlocher (a) and Lynne B. McCusker (a), to be published.

(a) Lab. für Kristallographie, ETH Zuerich (Switzerland)




High-energy synchrotron radiation for the analysis of stresses in the bond layer of a thermal barrier coating system

Thermal barrier coating systems are well known today, to improve the thermal resistance of materials to a large extent and, therefore, to increase the efficiency of technical devices like the combustion chamber of turbines. However, due to the extreme temperature gradients between the free surface and the bulk material, high thermal stress gradients arise within the coating system, which may lead to failure by crack initiation and propagation. For this reason, detailed knowledge of the stress distribution in both the ceramic layer and the bond coat, before as well as under, load conditions are of practical importance for reliable life time predictions.

The non-destructive diffraction techniques, applied so far to stress analysis in such coating systems, are either limited to the near surface region of the material (X-ray methods), or they require a large gauge volume (neutron diffraction) which reduces the spatial resolution. Therefore, both conventional X-ray and neutron diffraction are not suited for the stress evaluation in these small buried bond layers of about 150 µm thickness.

Here, the use of high-energy synchrotron radiation offers some fundamental advantages. Due to the high intensity and brilliance of the beam, the gauge volume may be adapted to the layer thickness, which permits a depth-resolved stress analysis even within small buried layers. The first experiments were carried out at the beamline ID15A. Figure 95 shows the experimental set-up.

The thermal barrier coating system, consisting of a 500 µm plasma sprayed ZrO2 - 7wt%Y2O3 ceramic layer and a 150 µm NiCoCrAlY bond layer both deposited on a Ni superalloy In718 substrate with a thickness of 2 mm, was moved through the gauge volume defined by two primary and two secondary slits with a width of 0.8 mm. The diffracted beam was recorded by means of an energy-dispersive detector at a fixed position 2 = 10°. The lattice strains were calculated from the shift of the energy spectra using Bragg's law and de Broglie's relation. For stress evaluation, use was made of Hooke's law. By means of a four-point bending device, different defined loads were applied to the sample to compare the variation in the stress state of both the ceramic layer and the bond coat with the theoretical values. Figure 96 shows the excellent agreement between the theoretical and the experimental results.

The measurements show that high-energy synchrotron radiation diffraction is so far the only non-destructive technique that enables the analysis of the triaxial stress state in the bond coat of a thermal barrier coating system.

Further measurements are planned with respect to an in-situ stress analysis under defined thermal or mechanical loads. This may result in a more detailed understanding of the fatigue behavior of the thermal barrier coating systems and therefore together, with an improved modeling of the stresses, in more reliable life-time predictions.

[1] W. Reimers (a), M. Broda (a), G. Brusch (a), D. Dantz (a), K.D. Liss (b), A. Pyzalla (a), T.Schmackers (a), T. Tschentscher (c), accepted in Journal of Non-Destructive Evaluation.
[2] W. Reimers (a), A. Pyzalla (a), M. Broda (a), G. Brusch (a), D. Dantz (a), K.D. Liss (b), T. Schmackers (a), T. Tschentscher (c), accepted in Journal of Materials Science Letters.

(a) Hahn-Meitner-Institut, Berlin (Germany)
(b) ESRF
(c) HASYLAB, Hamburg (Germany)




EXAFS studies at the core hole lifetime limit?

The development of third generation synchrotron sources has opened the door to a wide range of new science, in particular making possible studies using very hard X-ray radiation. Extended X-ray absorption fine structure (EXAFS) spectroscopy as a dedicated tool to determine short range interatomic structure, has in general previously been confined to absorption processes below 30 keV. Largely, this was because 30 keV is the typical upper value for the useful photon energy available at second generation synchrotron sources, but perhaps more importantly there was and still is, a degree of opinion that EXAFS is unable to give significant structural information at K edges of higher Z elements - due to the broadening of the signal caused by the short lifetime of the excited state of the photoabsorbing atomic species.

The limitation in available photon energy, although unfortunate, has to some extent been circumventable since elements with K absorption edges beyond 30 keV could be probed via their LIII edges. This solution however has the disadvantage that the LIII absorption edge, which typically occurs at a few hundred electron volts higher in energy, restricts the range of spectral signal that may be used in an EXAFS analysis.

In contrast, the K absorption edge does not suffer from this restriction and consequently analysable data can be obtained over any desired range. This is particularly important for investigations of interatomic structure, as the desired information is contained within the frequency and amplitude of the oscillatory features following an absorption edge - the wider the range of data that can be sampled, the higher the corresponding resolution for interatomic distances and relative coordination numbers. Figures 97 and 98 show the EXAFS signal of europium oxide measured at the LIII and K absorption edges and the corresponding Fourier transforms - the enhancement in structural resolution gained by measuring at the K edge is clearly apparent.

The availability of hard X-rays now enables the range of useful K absorption edges to be pushed to cover the majority of the periodic table. The recently commissioned Si(511) crystals for the BM29 monochromator set a practical upper limit of 74 keV. This new capability comprehensively covers the entire lanthanide series of elements where the problem of the LIII absorption edge limiting structural studies is most acute, and also allows K edge absorption studies to be pursued up to and including the elements tungsten and rhenium (Z = 75). It has been demonstrated that data of extremely high quality can be collected over this energy range and over typical analysable EXAFS spectral k-ranges of 30 Å-1.

In a recent series of measurements to establish the feasibility of EXAFS studies at these energies, data were collected for various lanthanide series oxides and an analysis performed to establish the effect of thermal Debye-Waller signal damping and very short core-hole lifetimes. The result of this work indicates that the detrimental effect of the combination of these two parameters that suppress the EXAFS signal can, to a large extent be removed by working at low temperatures. The core-hole lifetime broadening then becomes the dominant damping term, but its effect becomes manageable when data are collected with excellent signal to noise ratio up to high k-values (25 - 30 Å-1).

The advantages of working at such energies are manifold, not least in the area of sample environment. At energies beyond 30 keV, many materials that were otherwise too absorbing for sample supporting matrices or cells, e.g. aluminium or the more dense organic polymers, become acceptable. This offers great potential for the construction of environments such as pressure cells or chambers for reactive chemistry.

In conclusion, it has been demonstrated that high-quality, high-resolution structural studies can be performed using high-energy absorption edges. Such a capability opens the possibility for significant development in the science of suitable systems under extreme conditions, taking advantage of the penetrating power of hard X-rays. In addition the availability of very high-quality EXAFS spectra at these energies and over large ranges, typically 3 keV-4 keV, allows for tests of conventional EXAFS theory to the limit where relativistic corrections can become appreciable.

M. Borowski (a), D.T. Bowron (a) and S. De Panfilis (a), to be published.

(a) ESRF