Onset Of Grain Growth Research Articles

Texture Evolution in Al(Cu) Interconnect Materials

ABSTRACTAn investigation of the microstructural evolution of Al(Cu) thin films deposited on a variety of interlevel dielectric (ILD) layers was performed. A combination of X-ray texture measurements and scanning electron microscopy (SEM) was employed to link the texture behavior of the as-deposited Al(Cu) films at different thicknesses to the observed morphological development within the films. Three regimes of texture were revealed, corresponding to (1): Al(Cu) island growth and individual island coalescence, (2): fully coalesced film and the onset of grain growth and (3): extensive grain growth. The first and last of these regimes exhibited offset (111) texture, in which the maximum diffracted intensity from Al (111) is offset from the substrate normal. However, the position of maximum offset texture differed between the first and third stages of growth, indicating that two different mechanisms were responsible. The offset (111) texture observed in the third regime of Al(Cu) film microstructure was due to the faceting of grain surfaces. The time required for the films to reach these three stages depended on the effective diffusivity of the Al atoms on the ILD surfaces, which differed in chemistry and topography.

An examination of the interparticle contact area during sintering of W-0.3 Wt Pct Co

As a powder compact sinters, its microstructure evolves. One way to quantify the scale of the microstructure is to consider the interparticle contact area. This study examines two known models for calculating the interparticle contact area: the classic two-sphere model and the Voronoi cell model. Both models have particular assumptions about the microstructure that make them not applicable for treating densification to near full density with concurrent grain growth. The classic two-sphere model assumes a regular packing of particles and a perfectly spherical particle geometry and neglects an increasing particle coordination number with sintering. The Voronoi cell model assumes that the scale of the microstructure remains constant; i.e., as long as the compact is densifying, grain growth does not occur. We propose a modified Voronoi cell that accounts for an increasing grain size, making it applicable to a general case where grain growth occurs during sintering. The three models are compared to the interparticle contact area data, obtained by stereology techniques, for W-0.3 wt pct Co sintered from green state to near full density. The original Voronoi cell model fits the data only at low temperatures, before the onset of grain growth. Below approximately 90 pct relative density, the two-sphere model with an assumed coordination number of six (coordination number in a green compact) and the modified Voronoi cell model provide a good fit to the data. At higher densities, both models overestimate the interparticle contact area.

Nanostructured Cu-Bi alloys prepared by co-evaporation in a continuous gas flow

Abstract A new preparation process for nanostructured alloys utilizing co-evaporation in a continuous gas flow has been developed. Inert gas is injected into a preparation chamber through a nozzle in order to establish a convectional flow that carries crystallites of one element above an evaporation source of a second element so that the atoms of the second source condense at the crystallite's surfaces in the form of a thin coating. In this manner Cu crystallites were coated with Bi, resulting in Bi-doped grain boundaries after consolidation of the Bi-coated Cu crystallites. The maximum overall concentration of Bi without formation of crystalline Bi was 10 at.%, corresponding to one monolayer of Bi atoms distributed along the grain boundaries. The average grain size in the consolidated samples was found to depend on the Bi concentration. Annealing experiments showed that the onset of grain growth was shifted to higher temperature for Bi-doped samples.

X-ray diffraction characterization of defect behavior in nanocrystalline nickel during annealing

Nanocrystalline nickel samples were annealed while collecting high angle X-ray diffraction data in order to correlate changes in strain arid defect content with changes in grain size.The microstrain broadening magnitude decreased rapidly prior to any grain growth, while the Debye-Waller parameter was not reduced significantly prior to the onset of grain growth.

Preparation of low resistivity Cu–1 at. %Cr thin films by magnetron sputtering

Copper is an attractive metal for very large scale integrated interconnections because of its low bulk resistivity of 1.7 μΩ cm. Alloying of the copper may be required to improve the mechanical stability, adhesion, electromigration lifetime, and resistance to corrosion. However, the addition of chromium to copper increases its bulk resistivity at an initial rate of 4 μΩ cm per atomic percent at 0 °C. We have investigated the effects of sputtering parameters and post-deposition annealing on the resistivity of Cu–1 at. %Cr thin films deposited by magnetron sputtering on thermally oxidized silicon wafers. Resistivities as low as 2.3 μΩ cm have been obtained. The resistivities of as-deposited films decreased with increasing thickness, from 10 to 2.3 μΩ cm, corresponding to film thicknesses of 100 to 1200 nm, respectively. The thicker films were heated in situ by increasing the number of passes the substrate makes in front of the target in the side-scan sputtering system. Ex situ resistivity versus temperature analysis, measured in purified He up to 600 °C on the deposited films, reveals that there are two drops in resistivity; at about 200 °C and 400 °C. Transmission electron microscopy shows the onset of grain growth at 125 °C–200 °C and the development of a bimodal grain size distribution at 400 °C–500 °C. The final resistivities of the films ranged from 2.3 to 4.5 μΩ cm. In general, the final resistivity of the in situ heated Cu–1 at. %Cr films is lower than that of the ex situ heated films. This suggests that Cr surface diffusion during heated deposition is more effective in lowering the resistivity than Cr bulk diffusion during post-deposition annealing. This model is supported by Auger electron spectroscopy, which reveals surface segregation of Cr during annealing. Heating during deposition allows low resistivity (<2.5 μΩ cm) films to be achieved at a low substrate temperature (200 °C–250 °C), whereas post-deposition annealing requires about 600 °C to achieve the same values.

Mechanical properties and microstructural characterization of Al-0.5%Cu thin films

Using a wafer curvature technique we have studied the variation of stress with tem-perature in Al-0.5%Cu thin films deposited on oxidized silicon wafers. Concurrently, the microstructural changes in the films induced by the thermal cycling inherent to this technique were studied with in-situ transmission electron microscopy heating experi-ments. On heating an as-sputtered film a stress drop occurs, corresponding to the onset of grain growth. The in-situ TEM experiments indicate that the extent of grain growth is significantly altered by the presence of compressive stresses in the film. During cool-ing, dislocation loops nucleate on {111} planes inclined to the film surface, although the grain size plays an important role in determining the extent to which this mechanism can account for the deformation. A native oxide can influence the stress levels in the film by pinning one end of the dislocation loops. Upon cooling below 200° C a rapid increase in stress occurs. Although this increase has been attributed to hardening due to the precipitation of excess copper, no evidence of precipitate-dislocation interactions were observed.

The effects of strain on grain misorientation texture during the grain growth incubation period

Abstract The effects of post-recrystallization strains of 2% and 7% on the microtexture and grain misorientation texture of an austenitic stainless steel are investigated. The time to onset of grain growth is decreased by two to three orders of magnitude after the additional strain. An electron back-scattering technique is used to collect grain orientations as a function of strain and post-strain annealing time and the results show that the proportions of coincident site lattice (CSL) boundaries increase during the post-strain anneals. Most of the CSLs are potentially ‘special’ types and tend to occur in clusters, with an increased tendency for clustering as the end of the grain growth incubation period is approached. It is therefore proposed that special boundary clustering is instrumental in the initiation of anomalous grain growth.

Textural transformations in thermomechanical treatment

1. To the dislocation structure of a certain kind, produced by HTMT, there corresponds a certain type and sharpness of the texture. The texture of hot-deformed austenite is inherited in martensitic transformation. 2. The texture of cold forming is inherited in primary recrystallization at the stage of patenting; it is reproduced in primary recrystallization under conditions of rapid heating of austenitic steels and in multiple α+K→γ→α+K transformations, e.g., under conditions of patenting. This is the result of oriented onset of grain growth at the early stages of development of recrystallization (in situ) in the field of retained stresses and in the endeavor to compensate misorientations. 3. The correlation between the structural state and the crystallographic texture can be used for checking thermomechanical treatment.

Local texture changes associated with grain growth

Recently it has been shown that grain growth is accompanied by macroscopic texture changes. Such changes will necessarily produce modifications in the proportions of ‘special’ grain boundaries and consequently may affect significantly the kinetics of grain growth and particularly anomalous growth. After new contributions from computer modelling processes are taken into account, it becomes clear that the onset of grain growth for a fixed driving force is a complex function of particle pinning, effects from solutes, surfaces and residual strains, grain-size distribution and texture-boundary mobility. The aim of this paper is to provide insight into the relation between grain texture-grain misorientation texture and anomalous grain growth in the alloy Nimonic PE16 where, for the particular heat-treatment conditions employed, particle pinning may be ignored. Macroscopic orientation measurements reveal only the overall texture; the present work considers microstructural regions of anomalous growth and texture measurements arecollated on a grain-specific basis (microtexture). From this grain-specific data the grain-misorientation textures (GMT) (i. e. the grain-boundary counterparts of an inverse pole figure) are also computed. The experimental results demonstrate that two types of microtexture may exist for different regions of anomalous growth within the same specimen. These differences are rationalized by noting that further grain growth-inducing heat treatments cause a change from the first texture to the second. Clear evidence is thus provided that there are micro textural changes associated with anomalous grain growth. The grain-boundary textural data show some deviation from a random distribution, particularly for boundaries between small grains (SS boundaries) as compared to boundaries between large and small grains (LS boundaries). These differences between the SS and ls groups extend to the proportions of geometrically special boundaries in each group, with the SS group containing a higher proportion of special (therefore higher than average mobility and lower energy) boundaries than the ls group. This trend was particularly apparent for boundaries that fell into the second texture group. The implication from these results is that in this case anomalous growth will tend to stagnate because the boundaries of small grains surrounding the large grain (SS boundaries) are on average of lower energy than the boundaries that border the large grain (LS boundaries). Finally, it is suggested, by analogy with primary recrystallization, that the definition of grain growth in terms of an energy-reduction criterion should be extended to encompass the sum of all available sources of energy minimization. Hence both grain-boundary textural rearrangements and grain-boundary migration are included.