HIGH-RESOLUTION ELECTRON MICROSCOPY OF QUASICRYSTALS

An Al3Mn-type Al3(Mn, Pd) crystal and an Al–Mn–Pd decagonal quasicrystal (DQC) in an Al70Mn20Pd10 alloy are studied using a spherical aberration (Cs)-corrected scanning transmission electron microscope (STEM) with high-angle annular dark-field (HAADF) and annular bright-field (ABF) techniques, together with atomic-resolution energy dispersive X-ray spectroscopy (EDS). Mn and Pd atomic positions in the Al3(Mn, Pd) structure projected along the b-axis (pseudo-tenfold rotational axis) are represented by separate bright dots in observed HAADF-STEM images. Besides, Al as well as Mn and Pd atomic positions are represented as dark dots in ABF-STEM images. Most Mn and Pd atomic positions in the Al3(Mn, Pd) structure can be observed on atomic-resolution EDS maps. On the basis of the good correlation between the STEM images and the EDS maps, and also considering the structure of the Al3(Mn, Pd) crystal, which was determined by X-ray diffraction using a single crystal, observed HAADF and ABF-STEM images of the Al–Mn–Pd DQC have been interpreted. Pd and Mn atomic positions in the Al–Mn–Pd DQC can be detected on the observed EDS maps. It can be seen that Pd is enriched around the centre of the columnar clusters, having a decagonal section with 2 nm in diameter. It can therefore be concluded that Pd plays an important role in the stabilization of the decagonal clusters, which form the Al–Mn–Pd DQC structure.

Elemental and structural changes in illite/smectite mixed-layer clay minerals during diagenesis in Kimmeridgian- Volgian (- R yazanian) clays in the Central Trough, North Sea and the Norwegian ­Danish Basin

The formation process of the decagonal quasicrystalline phase (D-phase) with 1.2 nm periodicity from a crystalline approximant in the Al70Pd13Mn17 alloy has been identified by high-resolution electron microscopy. At the early formation stage, decagonal atom clusters, which are the most important structural units of the D-phase, were observed to segregate from the crystalline matrix with an atomic correlation determined by the structural characteristics of the crystal. A non-equilibrium structural state (a mixture of the segregated atom clusters and the crystalline matrix phase) was also observed, and its structural characteristics clearly revealed. The equilibrium D-phase, whose structure is characterized by an aperiodic arrangement of decagonal atom clusters with definite linkages and a long-range quasiperiodic correlation, was formed by fully annealing at 800degreeC for a long time. The linear phason strain in the equilibrium D-phase formed from the crystalline approximant may be related to the structur...

ABSTRACTConsolidated powders of nanocrystalline Cu and Pd have been studied by x-ray diffraction (XRD) and high resolution electron microscopy (HREM) as part of an investigation of the mechanical behavior of nanocrystalline pure metals. XRD line broadening measurements were made to estimate grain size, qualitative grain size distribution and average long range strains in a number of samples. Mean grain sizes range from 4-60 nm and have qualitatively narrow grain size distributions. Long range lattice strains are of the order of 0.2-3% in consolidated samples. These strains apparently persist and even increase in Cu samples after annealing at 0.35 Tm (498K) for 2h, accompanied by an apparent increase in grain size of e2x. Grain size, grain size distribution width and internal strains vary somewhat among samples produced under apparently identical processing conditions. HREM studies show that twins, stacking faults and low-index facets are abundant in as-consolidated nanocrystalline Cu samples. Methodology, results and analysis of XRD and HREM experiments are presented.

Nanocrystalline materials are of interest because of their small grain size (less than 100 nm) and unusual properties [1-3]. These properties are primarily the result of the large fraction of atoms that reside in grain boundaries or interfaces. It is recognized that the grain boundaries in nanocrystalline materials may differ from those in systems having larger grain sizes. The "gas-like" grain boundary structure was proposed by Gleiter and co-workers according to the results of X-ray diffraction (XRD) [4], neutron diffraction [5] and extended X-ray absorption fine structure [6]. The most direct information regarding the atomic structure of grain boundaries or interfaces can be obtained from high resolution electron microscopy (HREM) observation. Recently, the results of HREM investigations on nanocrystalline palladium have revealed that the grain boundaries show an "extended" structure not typically observed in conventional systems [7]. However, Siegel and Thomas observed that the atomic structure of grain boundaries is consistent with that typically observed in coarse-grained materials [8]. Somewhat imperfect or disordered regions at the grain boundaries have also been observed by Li et al. [9] . These conclusions on the structure of grain boundaries were based upon the experimental results of the nanocrystalline materials synthesized by the inert gas condensation and in situ compacting technique [10]. Recently, a new method (called the amorphous crystallization method) for synthesizing nanocrystalline materials has been developed by means of crystallization from the amorphous precursor [1113]. This method involves the formation of an amorphous alloy as a raw material, and subsequent heat treatment to form nanometre-sized crystallites within the amorphous matrix. Fully crystallized nanocrystalline materials with an average grain size less than 40 nm and uniformly distributed crystallites can be produced by controlling the annealing temperature and time [14, 15]. To compare the microstructures in nanocrystalline materials prepared by the two different methods, HREM studies were carried out on nanocrystalline palladium synthesized by the inert gas condensation and subsequent in situ compacting technique, and on nanocrystalline Ti70Ni20Sil0 and (Fe0.99Mo0.01)78 Si9B13 alloys produced by the amorphous crystallization method. The nanocrystalline palladium samples were synthesized by the inert gas condensation and in situ compacting technique as described in [10]. Raw palladium with a purity of 99.99 wt% was used.

The nanostructure and chemical distribution in semi-insulating polycrystalline oxygen-doped silicon (SIPOS) deposited on (001) Si and its isothermal transformation behavior at 900 °C were investigated by high resolution electron microscopy (HREM) and electron energy loss nanospectroscopy (EELS). The structure of the as-deposited film, which contained 15 at. % oxygen, was amorphous. No evidence for nanocrystalline second phases was found. It was similar in appearance to amorphous silicon. After annealing for 30 min at 900 °C in an inert environment (N2), a dispersion of small nanocrystals, identified as silicon by imaging, diffraction and EELS, formed in the amorphous SIPOS matrix, with a thin precipitate free zone (PFZ) adjacent to the Si substrate. The SIPOS matrix oxygen concentration increased to 36 at. % and the matrix remained amorphous after annealing. No other phases were observed in annealed specimens. Changes in Si–L near edge fine structure and low loss peaks in EELS spectra from SIPOS with increasing oxygen concentration indicated that it is a solid solution supersaturated with silicon. Microstructures indicated that the Si nanocrystals formed during a homogeneous precipitation reaction.

A detailed characterization of the ceramic-metal interface was carried out for a Si3N4-Ni system bonded in vacuum at 1 273 K. High resolution electron microscopy and analytical electron microscopy were used in the characterization of the interface. Reaction products such as nickel silicides were not present at the interface. It was demonstrated that Si3N4 reacts with Ni in accordance with the solid-state reaction mechanism proposed in this study. The final reaction product is thought to be Ni with Si and N in solid solution.

Microstructures inZrC-dopedWC-Coalloys,whichwerepreparedbyliquid-phase andsolid-phasesintering,wereinvestigated mainlybyhigh-resolution transmission electron microscopy (HRTEM) and energy dispersive X-ray spectrometry (EDS). HRTEM study has revealed thatmulti facet structures, which are often observed in VC-doped alloys, were not formed in ZrC-doped alloys. In addition, the segregation of Zr atWC/Co interfaces could not be detected by EDS analysis even with nano size electron probe. Doped ZrC was found to exist as other carbidegrains without dissolving into -phase, i.e., Co phase by EDS-mapping performed for liquid-phase sintered alloys. In the case of liquid-phasesintering, the inhibition effect for the grain growth caused by ZrC-doping was confirmed to be very small as previously reported. In contrast, itwasfoundthatZrC-dopinglargelyretardsWCgraingrowthinthecaseofsolid-phasesintering.TheretardationofWCgraingrowthobservedinsolid-phase sintered alloys was considered to be closely related to a low sinterbility due to a low wettability between Co and ZrC.[doi:10.2320/matertrans.MC200826](Received January 6, 2009; Accepted March 5, 2009; Published April 25, 2009)Keywords: cemented carbides, grain growth, ZrC, high-resolution transmission electron microscopy (HRTEM), energy dispersive X-rayspectrometry (EDS)

Platelets of the Al—Pd—Mn decagonal phase have been observed to grow in epitaxy on the icosahedral modification of the same system. A resulting strain field in the icosahedral phase has been characterized using high-resolution electron microscopy. It appears that one fivefold axis is preserved, in coincidence with the tenfold axis of the decagonal phase. In the particular case of a thin icosahedral-phase domain squeezed in between two decagonal-phase platelets, the strain field can be qualitatively analysed in terms of combined phason–phonon strain fields.

Fine clays (<0.1 μm) extracted from an acid soil developed in a granite saprolite from the Massif Central, France, were characterized by X-ray diffraction (XRD) using a curve decomposition program, and high-resolution transmission electron microscopy (HRTEM) associated with a method of impregnation of moist samples. Direct measurement of d-spacings were performed on HRTEM photographs. Decomposition of XRD patterns indicated 5 to 6 different clay phases including chlorite (and/or hydroxy-interlayered vermiculite), vermiculite/smectite, illite/vermiculite and illite/smectite mixed layers. Expandable phases with decreasing layer charge (vermiculite, high- and low-charge smectite) were shown in the clay assemblage. When performed on K-saturated samples subjected to wetting and drying cycles, HRTEM observations were consistent with the XRD results. The major clay mineral phases identified by the decomposition of XRD patterns were also found by direct measurement of d-spacings on HRTEM images. Vermiculite and high-charge smectite appeared to be impregnated with preservation of their initial hydration state, whereas low-charge smectite interlayers were penetrated by the resin molecules during the impregnation procedure. It was concluded that the decomposition of XRD patterns gave a realistic analysis of the clay phases present in a complex soil clay sample, as well as the direct measurement of a limited number (50) of clay crystals on HRTEM images.

Symmetric, asymmetric and overlapped twins were observed in epitaxial NiSi2 thin films. These twin structures were studied by high-resolution electron microscopy. The twin-boundary planes of the observed symmetrical twins were found to be (1―11―)1-(1―11)2, while those of the asymmetrical twins were (1―11)1-(11―5)2 and (001)1-(22―1)2. The structural unit model was applied to explain the observed twin-boundary structures. The summetrical twin boundary is comprised of a uniform array of six-atom ring structural units, while the asymmetrical twins contained a mixture of structural units of the (1―11―)1-(1―11)2 and (1―12)1-(11―2)2 symmetrical twins. The {111} lattice plane containing Si was found to be the terminating lattice plane at the boundary core of the symmetrical twin in which case the bond length between Si atoms of two grains was determined to be the same as that of Si-Si bonds in bulk Si crystal, and the Ni atoms are sevenfold coordinated. High-resolution images of overlapped twins show triple periodicity along the directions of {111} plane normal which is dependent not only on the defocus, the total foil thickness and the thickness ratio of twins, but also on the stacking order of the two grains.

Exchange bias(EB) is an interfacial magnetic phenomenon. After a bilayer film made of a ferromagnet and antiferromagnet goes through a field-cooling process from above the antiferromagnet's Neel temperature, these two materials will couple magnetically at their interface. Thus, a nickel-iron/nickel-oxide bilayer can magnetically couple and enable EB. This phenomenon is related to the interfacial arrangement of atomic magnetic moments. However, it is quite difficult to experimentally observe the direction of magnetic moments for individual atoms, and even harder to resolve the necessary atomic arrangement in polycrystalline films such as NiFe/NiO. Thus, we try to simulate the magnetic moments of each atom coupling to permit EB at the interface of nickel-iron/nickel-oxide system by theoretical simulations and mapping these results onto experiments on the films grown on a crystalline MgO substrate. According to random-field model, we can obtain the exchange bias from the difference of interfacial energies before and after the change of magnetic field direction, then incorporating the thickness effects of the ferromagnetic layer and its magnetization. Afterwards, we compared the simulated results with the experimental data and determined reasonable configurations of atomic magnetic moments by comparison with results from HRTEM (high resolution transmission electron microscopy). The result of VASP simulations for exchange bias without roughness and with two kinds of roughness (2.20A and 4.43A) at the interface is 3.08±4.59(Oe), -24.27±3.13(Oe) and -195.53±10.88(Oe), respectively, and 0.00±3.88(Oe), -4.17±1.96(Oe) and -11.02±3.12(Oe) for different directions of applied field. All simulations revealed that interfacial roughness can strengthen the EB, which is consistent with experimental results. Thus, we can conclude that the configuration of atomic magnetic moments is similar to real case with the interfacial roughness as also shown from HRTEM. Afterwards, we calculated for the effect of interfacial dislocations on EB. When dislocations appear around the interface with 2.20A roughness, EB increased from -24.27±3.13(Oe) to -143.46±4.37(Oe), and from -4.17±1.96(Oe) to -82.30±7.29(Oe) for different directions of applied field. As EB is essentially a spin-spin interaction between ferromagnet and antiferromagnet at the interface, any kind of modification of the interfacial spin configuration (e.g. quantity of the uncompensated interfacial antiferromagnetic spins and the relative orientation between antiferromagnetic and ferromagnetic spins) will affect the final exchange coupling. Thus, it is reasonable to conclude that an interface roughness can strengthen the interfacial exchange coupling by enhancing the defect-generated uncompensated condition. This is in agreement with recent Monte Carlo simulations [J. Spray and U. Nowak, J. Phys. D: Appl. Phys. 39, 4536 (2006)].

Precipitation and age hardening processes in A1 alloy 6061 and Comral-85™ metal matrix composite have been investigated and compared. A number of sophisticated techniques were used to characterise fine-scale precipitation in these materials. These included atom-probe field ion microscopy (APFIM), high resolution transmission electron microscopy (HRTEM), conventional transmission electron microscopy (TEM) and differential scanning calorimetry (DSC). A dark-field high resolution technique was developed that enabled imaging of the crystal structures of fine-scale precipitates. The precipitation sequence in alloy 6061 was found to be: independent clusters of Mg and Si atoms >> co-clusters that contained both Mg and Si atoms >> small precipitates of unknown structure precipitates >> mostly B' and some s' precipitates It was found that the Mg:Si ratio in the small precipitates, precipitates and B' precipitates was close to 1. A new structure was proposed for the precipitate: base-centred monoclinic, a = 15.34Â, b = 4.05A, c = 6.83A, s = 106°. The development of hardness during ageing of alloy 6061 was related to different types of precipitates that formed in the precipitation sequence. The clusters of atoms and also the small precipitates had a negligible effect on hardness. Themajority of age hardening was provided by needle-shaped precipitates. The formation of B' precipitates resulted in substantial overageing of the alloy.

Decagonal quasicrystals of Al70Ni20Rh10 and Al70NixIr30-x (x= 15, 17 and 20) alloys prepared by the melt-quenching method have been examined by convergent-beam electron diffraction (CBED) and high-resolution electron microscopy (HREM). All these alloys were found to have a regular periodicity of about 0 4nm along the decagonal axis and belong to the centrosymmetric space group of P105/mmc, which is the same as the stable decagonal quasicrystals of Al-Ni-Co and Al-Cu-Co alloys. However, the HREM study has revealed that the present alloys have specific decagonal atom-cluster columns of about 2nm diameter, while the stable Al-Ni-Co and Al-Cu-Co alloys have pentagonal atom-cluster columns according to the current work of Saitoh et al . and Tsuda et al. Al70Ni20Rh10 produces diffraction spots with the longest lattice spacing of 4 2nm among the quasicrystals examined by us, which indicates very goodlong-range order in the alloy. The present quasicrystals donot show diffuse streaks perpendicular to the decagonal axis in selected-area diffraction patterns, while the stable Al-Ni-Co and Al-Cu-Co alloys do show them.

The use of metal particles in heterogenous catalysis demands a continuous development of particle preparation methods and of methods for studying these particles. Two ways of preparing silver particles for high-resolution electron microscopy (HREM) studies have been used in this investigation; in situ reduction of silver iodide (AgI) in the electron beam and reduction of silver acetate suspended in acetic acid. The reduction of AgI in the electron beam is a straightforward method. Using an intense electron beam the method produces finite silver particles spread uniformly over the amorphous carbon support film. The preparation of silver particles by reduction of silver acetate is similar to particle preparation methods which have previously been used for other metals. The particle sample is air stable and water soluble which makes the specimen preparation for HREM easy. The two preparation methods give particle specimens well suited to HREM. Small silver particles are shown to undergo the same structural rearrangements in the electron beam as previously shown in other metals (e.g. gold, platinum, and ruthenium). Several kinds of surface defects and rearrangements on silver surfaces have been observed and recorded in real time; such as the gradual disappearance of steps on {111}f.c.c. surfaces and atomic column hopping. It is also shown that disordered surfaces become ordered when additional atomic layers are added on top of them. All these dynamic events are presented in surface profile imaging sequences recorded on video with a low-light television camera mounted on a 400kV microscope (JEM-4000EX) with a structural resolution of about 0·16nm.