Search

We investigate the effect of direct electric current (DC) on grain growth in 100 nm thick iron-carbon films with carbon concentrations between 0.7 to 4.4 at%. The application of DC-current during annealing at 550 °C confirms the expected transport of carbon in the direction of the electric current and the unexpected formation of elongated, abnormally large carbide and ferrite grains along the current direction in the carbon-rich regions. The formation of elongated grains is explained by electromigration-induced carbon flux divergences that result from the carbide precipitates. This presents a possible scenario for controlling microstructure evolution in iron by using DC electric currents. Changes for alternating current (AC) pulses had been observed before.

Brittle materials, such as oxide glasses, are usually very sensitive to flaws, giving rise to a macroscopic fracture strength that is much lower than that predicted by theory. The same applies to bulk-metallic glasses (BMGs), with the important difference that these glasses can exhibit certain plastic strain prior to the catastrophic failure. Here we consider the strongest metallic alloy known, a ternary Co55Ta10B35 BMG. We show that this macroscopically brittle glass is flaw-insensitive at the micrometer scale. This discovery emerges when the testing pre-cracked specimens with self-similar geometries, where the fracture stress does not decrease with increasing pre-crack size. The fracture toughness of this ultra-strong glassy alloy is further shown to increase with increasing sample size. Both these findings deviate from our classical understanding of fracture mechanics, and are attributed to a transition from toughness-controlled to strength-controlled fracture below a critical sample size.

Taking advantage of in-situ fracture testing method inside a transmission electron microscope, crack evolution in a low alloy steel under low triaxiality conditions is studied and the interaction between cementite particles and the crack is revealed. It is found that the ferrite matrix is the major void initiation site due to the low stress triaxiality in the TEM sample, which contrasts the behavior under plane strain conditions in bulk specimens, where voids are typically found to initiate by decohesion at the particle/matrix interface. This work reveals that fracture behavior proceeds differently under low triaxiality conditions, such as in thin films, and demonstrates the possibility to avoid interface decohesion and thereby to enhance ductility in steels.

Taking advantage of in-situ fracture testing method inside a transmission electron microscope (TEM), crack evolution in a low alloy steel under low triaxiality conditions is studied and the interaction between cementite particles and the crack is revealed. It is found that the ferrite matrix is the major void initiation site due to the low stress triaxiality in the thin TEM sample (plane stress condition), which contrasts the behavior under plane strain conditions in bulk specimens, where voids are typically found to initiate by decohesion at the particle/matrix interface. This work reveals that fracture behavior proceeds differently under low triaxiality conditions, such as the shear lip region of fractured bulk sample, and demonstrates the possibility to avoid interface decohesion and thereby to enhance ductility in steels.

We investigate the accuracy of a MEMS-based nanocalorimeter for determining specific heat in the temperature range from 300 K to 800 K. By comparing DC differential scanning measurements of materials of known specific heat capacities with detailed finite element studies, the dominant sources for heat losses and inaccuracies have been identified. The main source of error comes from radiative losses at elevated temperatures, but even near room temperature thermal gradients and thermal conductivity in the specimen lead to moderate discrepancies. Furthermore, studies on de-wetted Au films and Bi particles show that specimen emissivities differ strongly from literature values, presumably due to effects associated with the specimen dimensions being smaller than the thermal wavelength. By including the emissivity as a fitting parameter in the simulations, specimen heat capacities can be obtained to an accuracy of better than 7% over the entire temperature range.

We investigate the multiaxial behavior of nanoporous gold (np-Au) using finite element simulations on tomographic reconstructions, and its inferences of the hardness-to-yield strength ratio from nanoindentation and microcompression experiments on np-Au. Plane stress and axisymmetric loading simulations were carried out on real np-Au structures and simulated spinodal structures. The predicted initial yield response of spinodal and np-Au is nearly identical, while disagreeing with the Deshpande-Fleck foam yield criterion. A qualitative comparison of np-Au simulations with Deshpande-Fleck criterion suggests a plastic Poisson ratio of 0.23, corresponding to a hardness-to-yield stress ratio of 2.7 which is in remarkable agreement with experiments.

A melt pocket in the bonding zone of medium carbon steel bonded on low-carbon steel by explosive welding was investigated by laser assisted atom probe tomography. It was found that the structure is nanocrystalline and (sub-) grain boundaries are enriched with carbon. High hardness values in the melt pocket are attributed to a combination of (sub-) grain size and carbon distribution.

Application of the foam scaling laws to the mechanical properties of nanoporous gold (np-Au) implicitly assumes geometrical similarity of structures of different relative density and synthesis methods. While the very few studies addressing the issue of geometric similarity during thermal coarsening show contradicting results, there are no studies covering np-Au from different synthesis routes. This paper performs quantitative morphological and topological comparison of np-Au of ca. 30% relative density synthesized by free and electrochemical corrosion methods, and thermally coarsened samples. Nanoindentation tests were performed to obtain elastic modulus and hardness. Uniaxial compression simulations of three-dimensional structures from FIB-nanotomography have been conducted using finite element analysis, and the results are compared with experiments. The three-dimensional np-Au structure has been quantified in terms of the ligament diameter and length distributions, ligament tortuosity, surface curvature distributions, structural anisotropy, nodal connectivity, and genus. Our results suggest that neither the synthesis route nor the thermal coarsening, to change length scale by up to a factor of 3, significantly altered the morphology and topology. The modulus and strength from both experiment and simulations exhibit a linear dependency on the scaled genus density, modifying the geometry-dependent prefactors in the Gibson and Ashby scaling relations.

Upon reducing the sample size into micrometer scale, an obvious brittle-to-ductile transition accompanied by a drastic change of failure mode from shattering to shear-banding was observed when compressing the brittle but strong Co55Ta10B35 bulk metallic glass (BMG). The shattering failure under macroscopic compression is dominated by splitting cracking, which completely differs from shear-banding and originates from extrinsic defects like inclusions. To reveal the critical conditions for shear-banding and splitting cracking, various micropillar specimens with intentionally introduced holes as extrinsic defects were tested, and the stress distributions at the failure moment were analyzed with finite element simulation. The shear plane criterion was found to be quite effective to estimate the nominal stress required for the failure dominated by shear-banding. However, brittle splitting cracking does not occur although the maximum tensile stress reaches the critical value, which is different from traditional brittle solids. To initiate splitting cracking, a high-tensile-stress region over a critical distance, which depends on defect size and fracture toughness of the BMG, is required. The critical conditions for shear failure and splitting cracking demonstrated in this approach can be used to estimate the failure conditions of various BMG components with complex geometries in a wide range of length scales, and to design tough composites based on brittle BMGs. As an example, a design criterion to avoid brittle splitting fracture of porous BMG materials is proposed.

Nanoporous gold (np-Au) differs from its macroporous counterparts through its ligament and pore length scales, its high relative density, and its very distinct mesoscale cellular architecture. When reexamining the applicability of conventional macroporous foam scaling laws to np-Au, difficulties persist not only in determining the solid properties of nanoscale ligaments, but also because np-Au structure is not self-similar as the relative density changes. Thus, a clear distinction of the effects of relative density and structure is required. This paper aims to capture the role of topology and morphology into the scaling laws by comparing the overall mechanical response of real np-Au structures with the behavior of spinodal and gyroid structures. Quantitative morphological and topological characterization of these structures has been carried out and their role on the macroscopic elastoplastic response of np-Au has been studied using finite element (FE) simulations. The predicted elastic modulus of real np-Au structures from FE simulations is in remarkable agreement with the nanoindentation measurements, and validates the numerical simulations. Quantitative structural analysis reveals that np-Au and spinodal structures are topologically very distinct, but similar in their morphology. On the other hand, gyroids are both morphologically and topologically very distinct from np-Au. The results suggest that the macroscopic stiffness and strength are highly sensitive to the topology, while being relatively much less sensitive to the morphology. The effects of structural topology are captured into modified scaling laws where the geometric pre-factors for the stiffness and strength are found to vary linearly with the scaled genus.

In-situ transmission electron microscopy is used to investigate crack propagation parallel to the interfaces of a Ti/ZrO 2 multilayer. The cracks propagate along the middle of the 100 nm thick polycrystalline Ti layers, causing extensive dislocation activity, void coalescence, and crack bridging. The plastic zone size has been determined from the range of dislocation activity and agrees well with estimates of the toughness obtained from the measured crack tip opening displacement. The toughness is much smaller than in bulk Ti, which we attribute to the constraint on dislocation activity and cropping of the plastic zone by the small Ti layer dimensions.

We present a study on the plastic deformation of single crystalline stochastic honeycombs under in-plane compression using a crystal plasticity constitutive description for face-centered cubic (fcc) materials, focusing on the very early stage of plastic deformation, and identifying the interplay between the crystallographic orientation and the cellular structure during plastic deformation. We observe that despite the stochastic structure, surprisingly, the slip system activations in the honeycombs are almost identical to their corresponding bulk single crystals at the early stage of the plastic deformation. On the other hand, however, the yield stresses of the honeycombs are nearly independent of their crystallographic orientations. Similar mechanical response is found in compression testing of nanoporous gold micro-pillars aligned with various crystallographic orientations. The macroscopic stress tensors of the honeycombs show the same anisotropy as their respective bulk single crystals. Locally, however, there is an appreciable fluctuation in the local stresses, which are even larger than for polycrystals. This explains why the Taylor/Schmid factor associated with the crystallographic orientation is less useful to estimate the yield stresses of the honeycombs than the bulk single crystals and polycrystals, and why the plastic deformation occurs at smaller strains in the honeycombs than their corresponding bulk single crystals. Besides these findings, the observations of the crystallographic reorientation suggest that conventional orientation analysis tools, such as inverse pole figure and related tools, would in general fail to study the plastic deformation mechanism of monocrystalline cellular materials.

Nanomechanical properties along a single shear band in a Zr-based metallic glass were studied. Spatial mapping of both indentation hardness and modulus reveal complex long-range softening patterns that are indicative of internal stress fields along the shear band. These internal stresses reach values of the order of the yield strength of the tested metallic glass. Time dependent stress relaxation along the shear band is observed, and shear-band cavitation at the micron scale is found. Both the cavitation and the internal stresses are attributed to the non-planar shear plane that during shear-band propagation leads to the development of off-axis stress components relative to the shear direction. The cavities are a signature of a shear-band-to-crack transition, which is supported by stress fields known to develop ahead of mixed mode I and II crack tips.

Based on the statistics of both the stress and strain of a plastic event, the well known size-effect in strength can be linked to a crystal’s critical stress distribution and the universal scaling exponent of intermittent plasticity. We successfully test these hypotheses with small-scale deformation experiments as a function of diameter and aspect ratio, and find that the latter affects the material’s strength in a way that gives direct insight into the underlying critical stress distribution of the deforming volume.

Fatigue behavior of 100 nm and 1.0 µm thick Cu films with 10 nm Ta passivation layers has been studied using cyclic tensile testing. The results show that Ta capping-layer has influences on fatigue damage by suppression of extrusion formation and, thereby, improved the fatigue life dramatically in the 1.0 µm thick Cu film, but does not change the fatigue life of 100 nm thick Cu film for which crack formation is the dominant damage mechanism.

The basic dislocation processes responsible for plasticity of fcc and bcc metals are fundamentally different. Dislocation based deformation in fcc metals is not very sensitive to temperature whereas in bcc metals, deformation is strongly temperature dependent and controlled by the low mobility of screw dislocations. In bcc metals it has been observed that the mobility of screw dislocations can be enhanced in the proximity of the sample surface. In order to investigate this effect, a comparative study of small scaled samples with high surface to volume ratios and selected geometries was performed. The systems that were tested were made of the bcc metals Ta, Mo and Fe. Cu served as a reference material for fcc metals. Microcompression experiments were carried out on focused ion beam (FIB) machined samples. In the FIB machining process, the surfaces of cuboidal pillars were oriented relative to the expected active slip systems. For Ta and Mo the flow stress depended on the orientation of the pillar. This observation is interpreted in terms of the surface enhanced screw dislocation mobility of bcc metals.

The deformation behavior of micron-sized, single crystal specimens of the MAX phases Ti 2 InC and Ti 4 AlN 3 has been investigated under uniaxial compression. Deformation took place by slip on single planes or sets of parallel planes. For highly inclined slip planes, where slip was hindered by the substrate or punch, deformation was accompanied by micro-cracking, layer bending, and the formation of kink bands. This study confirms the importance of obstacles in the deformation behavior and strength of MAX phases.

The effect of dissolved hydrogen in a (1 0 0) vanadium single crystal was studied using compression tests of micropillars. It is observed that the shape of the deformed pillars changes with hydrogen concentration. At low concentrations the pillars deform on a few discrete slip planes and at high hydrogen concentrations the pillars deform to a barrel-like shape. Furthermore, the flow stress increases with hydrogen concentration. Both observations can be attributed to an elevated dislocation activity due to hydrogen.

The size dependence of deformation of Ta was studied using compression tests of focused ion beam (FIB) machined microcolumns. Columns with diameters between 0.5 and 8 μm with 〈1 1 1〉 and 〈1 0 0〉 orientations along the column axis were tested. By comparing results of bcc Ta columns with results from previous experiments on fcc metals it was found that Ta shows significantly higher normalized yield stresses in combination with a weaker sample size dependence. The differences between bcc and fcc metals can be attributed to the different dislocation behaviour of bcc metals, especially to the lower mobility of screw dislocations.

The effect of hydrogen on dislocation nucleation in a vanadium (1 0 0) single crystal has been examined by means of nanoindentation. For electrochemically doped samples to different hydrogen concentrations within the α-phase, the pop-in load decreased with the increase in hydrogen concentration and multiple pop-ins were observed on the load–displacement curves. The interaction between the dissolved hydrogen atoms and the newly formed dislocation loops, resulting in the reduction of their line energy, is evaluated and explained based on the novel thermodynamic defactant concept.

Fatigue behavior of thin Cu films with thicknesses between 50 nm and 3.0 μm on polyimide substrates has been studied using cyclic tensile testing. The results show a clear length scale effect on both fatigue damage morphology and fatigue life. Extrusions decrease in size and number while the number of cracks increases with decreasing film thickness. The change in fatigue damage with film thickness suggests a fatigue mechanism transition from dislocation mediated extrusion formation to crack formation controlled behavior. The very clear increase in fatigue life with decreasing film thickness at constant applied strain is attributed to an increase in yield stress.

A comprehensive study on the relationship between yield strength, relative density and ligament sizes is presented for nanoporous Au foams. Depth-sensing nanoindentation tests were performed on nanoporous foams ranging from 20% to 42% relative density with ligament sizes ranging from 10 to 900 nm. The Gibson and Ashby yield strength equation for open-cell macrocellular foams is modified in order to incorporate ligament size effects. This study demonstrates that, at the nanoscale, foam strength is governed by ligament size, in addition to relative density. Furthermore, we present the ligament length scale as a new parameter to tailor foam properties and achieve high strength at low densities.

The effect of loading frequency and microstructure on thermal fatigue damage in 200 and 300 nm thick polycrystalline sputtered Cu lines on Si substrates has been investigated. Alternating currents were used to generate temperature cycles (with ranges from 100 to 300 °C) and thermal strains (with ranges from 0.14% to 0.42%) in the Cu lines at frequencies of 0.2 and 20 kHz. Fatigue loading caused the development of severe surface roughness that was localized within individual grains. Raising the loading frequency accelerated damage formation and failure. The frequency effect is believed to result from differences in the concentration of defects created by the deformation-induced motion of dislocations to the grain boundaries.

Systematic investigations of fatigue damage and dislocation structures in thin Cu films with different thicknesses (0.2-3.0 mu m) and grain sizes (0.3-2.1 mu m mean diameter) were carried out using focused ion beam microscopy and transmission electron microscopy. The morphologies of fatigue-induced extrusions, cracks, and dislocation structures were studied and found to be controlled by film thickness and grain size. When either of these length scales is decreased below roughly 1 mu m, the typical dislocation wall and cell structures found in fatigued coarse-grained bulk materials no longer develop and are replaced by individual dislocations. Similarly, the typical surface damage of fatigued bulk metals, such as extrusions and cracks near extrusions, is gradually suppressed and replaced by damage that is localized at interfaces, such as cracks, grooves, and voids along grain and twin boundaries. This gradual transition from damage characteristic of bulk metals to damage localized at interfaces is attributed to constraints on dislocation activity at submicrometer length scales. Based on the experimental results and a theoretical analysis of extrusion formation, a mechanistic map of fatigue damage behavior is proposed that summarizes this length scale dependence. (c) 2006 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

Microelectronic devices experience thermal cycles with amplitudes as large as 100 K during normal use. Differences in the thermal expansion coefficients of the different materials comprising the device lead to strain changes during thermal cycling. We demonstrate here that cyclic thermal strains lead to surface damage formation and failure in copper lines during the application of an alternating electrical current. The presence of soft coatings like photoresist on the Cu lines does nothing to inhibit damage formation in the copper lines. Thus, thermal fatigue of Cu interconnects may be a serious reliability threat to devices containing soft interlevel dielectric materials.

The kinetics of θ phase (Al2Cu) precipitation in passivated, bamboo Al-0.2 at.% Cu segments was studied in-situ in an SEM during electromigration. The electrical current was used to reversibly transport Cu back and forth along the segment and allowed for a systematic study of Cu diffusion and precipitate nucleation, growth, and dissolution. The observations allowed an independent determination of diffusion with and without an electric current and showed that both interface and lattice diffusion are important in Cu precipitation in bamboo segments. It was also found that a substantial supersaturation of Cu is required to nucleate the θ phase. A simple model for electromigration-induced motion of Cu can be used to explain the data and predicts that an interconnect with Cu both in solution and in periodically spaced precipitates will have optimal reliability.

Tensile creep experiments on melt-spun ribbons of Pd40Ni40P19Si1 were performed near the glass transition temperature. Changes in the shear viscosity during relaxation were measured and equilibration of the glass to the metastable super-cooled liquid state was observed. The relaxation behavior varied considerably, depending on the thermal history of the sample: monotonie increases and decreases as well as crossover changes in the viscosity were measured. The results are interpreted with phenomenological and atomistic models for flow, and a qualitative modification of the atomistic defect model is suggested to account for the crossover behavior.

Tensile creep measurements on melt-spun ribbons of Pd 40 Ni 40 P 19 Si 1 , one of the most crystallization-resistant metallic glasses, were performed over a wide temperature range. The equilibrium viscosity was reached by both increasing and decreasing viscosity relaxation and was well described by a Fulcher-Vogel relation. The viscosity relaxation kinetics could not be entirely described by flow defect models. The newtonian viscous behavior of the glass near T g was verified and a zero creep effect due to surface tension was observed.