Twin Planes Research Articles

Tuning Crystal Plane Orientation in Multijunction and Hexagonal Single Crystalline CsPbBr3 Perovskite Disc Nanocrystals

Two-dimensional-shaped CsPbBr3 platelet nanocrystals are widely studied for their bright high energy emission and self-assembly. These nanostructures are in orthorhombic phase, have a square shape, and have the vertical axis [001] perpendicular to the basal plane. Moreover, these are mostly single-crystalline structures with a continuous lattice and appear like slices of cube nanocrystals. In contrast, herein, multijunction and hexagonal single crystalline 2D discs of CsPbBr3 are reported to have all their vertical axes [100]. These are obtained by using the perovskite derivative of tetragonal Cs3MnBr5 as the parent material and subsequent B-site Pb(II) introduction in the presence of phenacyl bromide at different reaction temperatures. At low temperature, multijunction discs having random orientations of two horizontal axes [010] and [001] from one to another segment are observed. Orientations of planes remained random as both coherent and incoherent twin planes were observed at their boundaries. However, the number of junctions/segments was reduced at higher temperature, and finally hexagonal single crystalline discs remained as the ultimate product. Analysis suggested that the crystal nature of parent Cs3MnBr5 and temperature-dependent variation in the rate of Pb(II) insertions determined the nature of discs having randomly oriented or static planes in the entire nanostructure. Not only in 2D discs but also, 3D nanocrystals having similar segments with different orientations are formed upon Pb(II) exchange with Mn(II) alloyed cubic CsBr. Hexagonal single crystalline and segmented multijunction CsPbBr3 discs remain unique among 2D perovskites nanostructures, and their formation mechanism indeed introduced new fundamentals of the crystallization process of these emerging energy materials.

Evolution of deformation twinning mechanisms in magnesium from low to high strain rates

We present a systematic investigation of {101¯2} extension twinning mechanism in single crystal magnesium micropillars deformed over seven orders of magnitude of strain rate, from 10–4 to 500 s−1, revealing how the accommodation of newly formed twins evolves with and depends on the kinetic compatibility of interfacial processes when high deformation rates are imposed. By combination of post-mortem 3D Electron Backscattered Diffraction, Transmission Kikuchi Diffraction and Transmission Electron Microscopy techniques, this work unveils the progressive evolution of the accommodating twin mechanisms from low to high strain rate, correlating differences in mechanical behavior with differences in twin crystallography. Away from quasi–static conditions, simple considerations of twinning shear do not suffice to describe unconventional twin morphologies, requiring the competition between newly activated dislocations and lattice distortions for allowing the evolution of the twin boundary along non–invariant twin planes. Under shock compressions, the basal/prismatic transformation establishing a lattice misorientation of 90° entirely governs the parent → twin conversion. The results illustrated here confirm that some of the recent interpretations deduced by particular twin morphologies are not universally valid and that deformation twinning is not only stress- but also strongly time–controlled.

Elastic interaction energy analysis during twin plane formation in the martensitic transformation process of α"-martensite in Ti-Nb-Al

To understand the twin selection mechanism that occurs in the initial stage of the martensitic transformation of shape memory alloys, the stress state during twin plane formation was numerically analyzed by micromechanics. The self-accommodation microstructure of α''-martensite of the Ti-23Nb-3Al (at%) shape memory alloy, in which the lattice parameter was adjusted to eliminate the formation of internal twinning due to lattice-invariant deformation, was adopted for analysis. The two lattice correspondence variants (CVi and CVj) formed a twin plane. CVi was assumed to be spheroid, and the elastic strain energy during twin plane formation was evaluated by analyzing the elastic interaction energy when CVj formed around CVi. There are three types of twins ({011}α" compound twin, {111}α" type I twin, and <211>α" type II twin) in the self-accommodation microstructure of α"-martensite. Regardless of the CVi geometry (i.e., aspect ratio (1 to 0.01) and minor axis orientation), the {111}α" type I twin exhibited minimum elastic interaction energy, indicating that minimum elastic strain energy was generated during this twin plane formation. This result is consistent with the experimental results, which showed the selective formation of {111}α" type I twins in the initial stage of martensitic transformation.

Nanotwin-induced ductile mechanism in thermoelectric semiconductor PbTe

•Pb-CTBs weaken ionic Pb-Te bonds to generate an easy dislocation source at CTBs •Twinning partial dislocations glide on (111) twin planes, leading to CTB migration •“Catching bond” induces high ductility and maintains integrity for nanotwinned PbTe Coherent twin boundaries (CTBs) with the lowest interfacial energy provide a strong phonon-CTB scattering source to suppress the lattice thermal conductivity needed for thermoelectric properties, but the impact on mechanical properties of PbTe remains unexplored. We construct nanotwinned structures with Pb- or Te-terminated CTB (Pb- or Te-CTB) along (111) plane and employ molecular dynamics simulations to examine structural evolution. We find that Pb-CTBs weaken ionic Pb-Te bonds to generate an easy dislocation source at CTBs. Due to nucleation and motion of partial dislocations on each Pb-CTB plane driven by shear load, Pb-CTBs gradually migrate to Te-CTBs, which is accompanied by breaking and re-forming of Pb-Te bonds. This “catching bond” maintains structural integrity while dramatically enhancing deformability of nanotwinned PbTe. Dislocations move from Te-CTBs toward twin lamellae, resulting in the structural slippage and fracture. These findings provide a theoretical strategy to improve the ductility of PbTe-based semiconductors through TB engineering. Coherent twin boundaries (CTBs) with the lowest interfacial energy provide a strong phonon-CTB scattering source to suppress the lattice thermal conductivity needed for thermoelectric properties, but the impact on mechanical properties of PbTe remains unexplored. We construct nanotwinned structures with Pb- or Te-terminated CTB (Pb- or Te-CTB) along (111) plane and employ molecular dynamics simulations to examine structural evolution. We find that Pb-CTBs weaken ionic Pb-Te bonds to generate an easy dislocation source at CTBs. Due to nucleation and motion of partial dislocations on each Pb-CTB plane driven by shear load, Pb-CTBs gradually migrate to Te-CTBs, which is accompanied by breaking and re-forming of Pb-Te bonds. This “catching bond” maintains structural integrity while dramatically enhancing deformability of nanotwinned PbTe. Dislocations move from Te-CTBs toward twin lamellae, resulting in the structural slippage and fracture. These findings provide a theoretical strategy to improve the ductility of PbTe-based semiconductors through TB engineering.

Study of icosahedral quasi-crystalline phase T2-Al6CuLi3 and transformation in 2A97 Al-Li alloy fabricated by laser additive manufacturing

Precipitated phases of 2A97 Al-Li alloy fabricated by laser additive manufacturing with thermal input are mainly composed of quasicrystal T2 phase gathered around grain boundary and TB phase in grain interior. Diffraction patterns of quasi-crystalline T2 phase show two-fold, three-fold and five-fold symmetry, respectively. Under condition of thermal cycling, T2 phase transforms into a face-centered cubic Y phase with a lattice constant of 2 nm. Y phase can grow into twins, and the twin plane is (111). After solution treatment at 515 °C for 1.5 h, δ' phase, T1 phase and TB phase are completely dissolved in matrix, while quasi-crystalline T2 phase transforms into orthorhombic O phase. The size of O phase is similar to that of T2, with length of about 400 nm and rougher surface.

1D Germanium Sulfide van der Waals Bicrystals by Vapor-Liquid-Solid Growth.

Defects in two-dimensional and layered materials have attracted interest for realizing properties different from those of perfect crystals. Even stronger links between defect formation, fast growth, and emerging functionality can be found in nanostructures of van der Waals crystals, but only a few prevalent morphologies and defect-controlled synthesis processes have been identified. Here, we show that in vapor-liquid-solid growth of 1D van der Waals nanostructures, the catalyst controls the selection of the predominant (fast-growing) morphologies. Growth of layered GeS over Bi catalysts leads to two coexisting nanostructure types: chiral nanowires carrying axial screw dislocations and bicrystal nanoribbons where a central twin plane facilitates rapid growth. While Au catalysts produce exclusively dislocated nanowires, their modification with an additive triggers a switch to twinned bicrystal ribbons. Nanoscale spectroscopy shows that, while supporting fast growth, the twin defects in the distinctive layered bicrystals are electronically benign and free of nonradiative recombination centers.

X-ray Imaging of Alloy Solidification: Crystal Formation, Growth, Instability and Defects.

Synchrotron and laboratory-based X-ray imaging techniques have been increasingly used for in situ investigations of alloy solidification and other metal processes. Several reviews have been published in recent years that have focused on the development of in situ X-ray imaging techniques for metal solidification studies. Instead, this work provides a comprehensive review of knowledge provided by in situ X-ray imaging for improved understanding of solidification theories and emerging metal processing technologies. We first review insights related to crystal nucleation and growth mechanisms gained by in situ X-ray imaging, including solute suppressed nucleation theory of -Al and intermetallic compound crystals, dendritic growth of -Al and the twin plane re-entrant growth mechanism of faceted Fe-rich intermetallics. Second, we discuss the contribution of in situ X-ray studies in understanding microstructural instability, including dendrite fragmentation induced by solute-driven, dendrite root re-melting, instability of a planar solid/liquid interface, the cellular-to-dendritic transition and the columnar-to-equiaxed transition. Third, we review investigations of defect formation mechanisms during near-equilibrium solidification, including porosity and hot tear formation, and the associated liquid metal flow. Then, we discuss how X-ray imaging is being applied to the understanding and development of emerging metal processes that operate further from equilibrium, such as additive manufacturing. Finally, the outlook for future research opportunities and challenges is presented.

Automated Crystal Orientation Mapping in py4DSTEM using Sparse Correlation Matching.

Crystalline materials used in technological applications are often complex assemblies composed of multiple phases and differently oriented grains. Robust identification of the phases and orientation relationships from these samples is crucial, but the information extracted from the diffraction condition probed by an electron beam is often incomplete. We have developed an automated crystal orientation mapping (ACOM) procedure which uses a converged electron probe to collect diffraction patterns from multiple locations across a complex sample. We provide an algorithm to determine the orientation of each diffraction pattern based on a fast sparse correlation method. We demonstrate the speed and accuracy of our method by indexing diffraction patterns generated using both kinematical and dynamical simulations. We have also measured orientation maps from an experimental dataset consisting of a complex polycrystalline twisted helical AuAgPd nanowire. From these maps we identify twin planes between adjacent grains, which may be responsible for the twisted helical structure. All of our methods are made freely available as open source code, including tutorials which can be easily adapted to perform ACOM measurements on diffraction pattern datasets.

Modification mechanism and tensile property of Al-9Si-0.4Mg-0.1Cu alloy

In this work, eutectic Si modification grade formula was established and effect of Sr element on modification level and mechanical properties of Al-9Si-0.4Mg-0.1Cu alloy was studied. Thermal analysis, SEM, EPMA and TEM were applied to analyze cooling curves, microstructure and Si twinning of the alloys. The results showed that precision and accuracy of eutectic Si modification grade formula were significantly improved and quantitative evaluation modification level of eutectic Si could be achieved. Modification level of eutectic Si increased first and then decreased with Sr content increasing. Though twin plane re-entrant edge (TPRE) mechanism and impurity induced twinning (IIT) mechanism were effective under certain conditions, they cannot fully explain the modification mechanism of eutectic Si in Al-9Si-0.4Mg-0.1Cu alloy. Sr element caused constitutional undercooling at the front of solid-liquid interface, promoted the formation of fine Al 2 Si 2 Sr phases and nano-precipitates, resulting in the changes of size and morphology of eutectic Si. With the improvement of eutectic Si modification level, the UTS and El of the alloys increased. While, adding excessive Sr led to the formation of coarse Al 2 Si 2 Sr, which was detrimental to the improvement of modification level and mechanical properties of alloy. • Modification level function suitable for quantitatively evaluate modification grade of eutectic Si was established. • Fine Al 2 Si 2 Sr phases and nano-precipitates change the size and morphology of eutectic Si. • Coarse Al 2 Si 2 Sr reduces the modification level of eutectic Si and the mechanical properties of alloy.

Impact of precipitated phases on the microstructure and mechanical properties of eutectic Sn58Bi alloy

A new type of low-melting-point heterogeneous alloy (LMH), based on Sn58Bi alloy (SBE), was prepared using Cu, Zn, Ag, and Sb microalloying and rapid solidification. The addition of Cu and Zn elements in equal mass ratios resulted in the Cu5Zn8 phase being dispersed in the Sn55Bi3Cu3Zn (LMH3) alloy. Subsequently, the additional Cu reacted with Sn to form a nano-sized η′-Cu6Sn5 phase at the interface of the Sn-Bi eutectic phase. Sn53Bi3Cu3Zn2Ag2Sb (LMH32) was prepared by adding appropriate amounts of Ag and Sb based on LMH3, leading to the precipitation of Ag3Sn and SbSn phases in the SBE matrix. The different precipitated phases in turn produced different types of fiber textures that existed in the matrix phase of SBE, LMH3, and LMH32. In addition, compared to LMH3, the presence of Ag3Sn and SbSn phases in LMH32 caused the formation of Bi twins identified via electron backscattered diffraction. The twin planes (101̅1) and (11̅02) had corresponding rotation axes and rotation angles of 93.8°[7̅431] and 90.2°[5̅415̅], respectively. Nanoindentation results indicated that LMH32 exhibited higher creep resistance than Sn55Bi3Cu3Zn (LMH3) and SBE. The precipitated phase significantly refines the grain size and eutectic lath structure of the matrix phase (Sn and Bi). During the deformation process, the fine lath structure and the interfacial η′-Cu6Sn5 nanophase can hinder the slippage of the eutectic interface. Moreover, the presence of the precipitated phase forms a large strain gradient in the alloy matrix, which produces a significant back-stress hardening effect and induces substantial dislocation nucleation. The formation of more dislocations made the creep mechanism of LMHs depend on dislocation climbing.

Nanotwin orientation on history-dependent stress decay in Cu nanopillar under constant strain

Cu with nanotwin (NT) possesses great electrical, mechanical, and thermal properties and has potential for electronic applications. Various studies have reported the effect of NT orientation on Cu mechanical properties. However, its effect on Cu stress-relaxation behavior has not been clarified, particularly in nano-scale. In this study, Cu nanopillars with various orientations were examined by a picoindenter under constant strain and observed by in situ TEM. The angles between the twin plane and the loading direction in the examined nanopillars were 0°, 60°, to 90°, and a benchmark pillar of single-crystal Cu without NT was examined. The stress drops were respectively 10%, 80%, 4%, and 50%. Owing to the interaction by NT, the dislocation behavior in nanopillars was different from that in bulk or in thin film samples. Especially, the rapid slip path of dislocations to go to the free surface of the nanopillar induced a dislocation-free zone in the 0° nanopillar, which led to work-softening. On the contrary, a high dislocation density was observed in the 90° nanopillar, which was generated by dislocation interaction and obstruction of dislocation slip by twin planes, and it led to work-hardening. The findings reveal the NT orientation in Cu nanopillars affected stress relaxation significantly.

Fibonacci Stoichiometry and Superb Performance of Nb&amp;lt;sub&amp;gt;16&amp;lt;/sub&amp;gt;W&amp;lt;sub&amp;gt;5&amp;lt;/sub&amp;gt;O&amp;lt;sub&amp;gt;55&amp;lt;/sub&amp;gt; and Related Super-Battery Materials

In this contribution, two important crystallographic concepts for the formation of a series of block structures associated with channeling have been compared: chemical twinning and crystallographic shear. Twin planes respectively shear planes besides formed channels serve as a sink for charge carriers or, when the oxidation state of metal ions can be reduced, as a reservoir for intercalated lithium. In this way, Wadsley-Roth shear phases such as niobium tungsten oxide exhibit channels for ultra-fast lithium-ion diffusion. They are in focus as anode material for super-batteries, superb in terms of energy respectively power density, charging time, cycle life and safety. It should be noted that the transition metal to oxygen ratio TM/O = 21/55 of the title compound is a Fibonacci number quotient. Also, the crystal lattice can be traced back to Fibonacci geometry. When replacing only 0.0213 tungsten atoms in the formula with less expensive titanium, a TM/O ratio of 0.381966 =ϕ2 can be adapted besides an average valence electron concentration of 2⋅ϕ-2, where represents the most irrational number of the golden mean. The additional disorder caused by even such small titanium replacement and accompanied oxygen vacancies could fasten up the already high lithium diffusion further. Ultrasonic treatment may be applied besides thermal cycling to prepare phase-pure of the highest electrochemical performance. A replacement of oxygen by some fluorine atoms is an obvious synthesis possibility, but the higher binding energy expected between lithium and fluorine in contrast to oxygen may rather hinder than promote lithium diffusion.

Effects of deformation-induced martensitic transformation on cryogenic fracture toughness for metastable Si8V2Fe45Cr10Mn5Co30 high-entropy alloy

In this study, fracture toughness was evaluated at room and cryogenic temperatures for a metastable Si8V2Fe45Cr10Mn5Co30 (at.%) alloy, and roles of deformation-induced martensitic transformation (DIMT) on fracture behavior were investigated with respect to types of transformation mechanism and martensite morphology. The as-annealed alloy consisted mainly of face-centered-cubic (FCC) phase with athermal hexagonal-close-packed (HCP) martensite. The DIMT from FCC to body-centered-cubic (BCC) via intermediate HCP resulted in an excellent combination of strength and ductility by high strain-hardening effect. Considerably high fracture toughness and ductile-dimpled fracture mode appeared at room temperature, but the fracture toughness significantly deteriorated at cryogenic temperature. The BCC martensite exhibited a thick-plate morphology as a resultant product of stress-induced transformation from low-stability FCC, and the plates had an alternating variant characterized by a ladder martensite, based on a 60°-BCC-twin relationship with {112} twin planes. This plate-type BCC martensite had the fewer phase boundaries and geometrically necessary dislocations than the lath-type BCC martensite. In addition, cleavage {001}BCC-planes of two martensite blocks having the alternating variants were slightly deviated, thereby leading to the much lower toughness along with the brittle fracture mode of lamellar cleavage facets at cryogenic temperature. Thus, the present work suggests that the proper control of FCC stability is required to enhance the resistance to low-temperature embrittlement for securing cryogenic applications.

Shape Control Beyond the Seeds in Gold Nanoparticles

In typical seed-mediated syntheses of metal nanocrystals, the shape of the nanocrystal is determined largely by the seed nucleation environment and subsequent growth environment (where “environment” refers to the chemical environment, including the surfactant and additives). In this approach, crystallinity is typically determined by the seeds, and surfaces are controlled by the environment(s). However, surface energies, and crystallinity, are both influenced by the choice of environment(s). This limits the permutations of crystallinity and surface facets that can be mixed and matched to generate new nanocrystal morphologies. Here, we control post-seed growth to deliberately incorporate twin planes during the growth stage to deliver new final morphologies, including twinned cubes and bipyramids from single-crystal seeds. The nature and number of twin planes, together with surfactant control of facet growth, define the final nanoparticle morphology. Moreover, by breaking symmetry, the twin planes introduce new facet orientations. This additional mechanism opens new routes for the synthesis of different morphologies and facet orientations.

Twin boundary reversibility characteristics in α-Fe

Understanding the grain boundary deformation dynamics is very crucial to designing materials with stable microstructures. With this quest, the deformation behavior of coherent twin boundary under cyclic shear loading has been studied in α-Fe using molecular dynamics simulations to understand the influence of strain amplitude and temperature. Twin boundary exhibited shear coupled migration along with cyclic irreversibility character at lower temperatures and gained almost perfect reversibility at higher temperatures of around 1500 K. TB exhibited more sliding than migration with an increase in temperature. The stress associated with the migration of twins was observed to fluctuate around its average value. This was caused by layer by layer propagation of twin boundary through the activity of 1/6 111 = 1/12 111 + 1/12 111 type edge partial dislocations on {112} twin plane. Complete twin reversibility and complex TB migration were noticed in the presence of multiple parallel twin nanowires. Twin migration was observed at the larger size nanowires with a proper combination of boundary conditions and strain rates; otherwise, it was absent. The influence of shear strain amplitude in offsetting the twin boundary was found to be minimal.

Evolving dislocation cores at Twin Boundaries: Theory of CRSS Elevation

Superior mechanical response of twinnable materials fundamentally arises from an elevation of Critical Resolved Shear Stresses (CRSS) due to Dislocation-Twin Boundary (D-TB) reactions. These reactions exhibit rich variety with several possible outcomes and exhibit complex dependence on microstructural properties, causing state-of-the-art models to adopt a case-by-case simulation of each reaction relying on empirical potentials or twin-interaction parameters. We develop an analytical “Evolving Dislocation Core” (EDC) model devoid of empiricism, capable of predicting the CRSS-elevation for any reaction, given the microstructural properties (elastic constants, twin crystallography, etc.). The approach is fundamentally rooted in energy-minimization within a fully-anisotropic framework revealing the evolution of dislocation cores with progression of the reaction. The core-structure of complex dislocations (e.g. stair-rod) in the reaction is proposed, for the first time in literature, as a non-planar composite of disregistries distributed on slip and twin planes. The model is applied to multiple slip-incorporation reactions in several Face-Centered-Cubic (FCC) materials (Pb, Ag, Cu, Ni-Co alloys and Ni-Ti alloys and high-entropy alloy FeNiCoCrMn). The predicted CRSS-elevations show agreement with atomistic simulations (Ni) and experiment (FeNiCoCrMn). The model further establishes a strong correlation of the elevation with unstable stacking/twinning fault energy and the magnitude of the sessile dislocation's Burgers vector, while revealing poor correlation with the stable intrinsic stacking fault energy which is a common benchmark. Thus the analytical EDC model developed in this study advances understanding of slip-twin interactions on multiple fronts while serving as an effective predictive model for CRSS-elevation instrumental in materials design.

The evolution of local stress during deformation twinning in a Mg-Gd-Y-Zn alloy

In this study, we perform in-situ tensile testing and high-resolution electron backscatter diffraction (HR-EBSD) to monitor the twin local stresses at different stages: (i) before twinning, (ii) twins that have propagated into the parent crystal, and (iii) thickening of the twin lamella. The resolved shear stress (RSS) at the twin interface is calculated during deformation twinning, which indicates that the RSS evolution is a dynamic process. Negative shear stress is observed during the propagation stage, which provides evidence for the twinning shear transformation-induced stress reversal. Furthermore, a local Schmid factor (LSF) is used to evaluate the extent to which the local stress is dominated by the RSS on the identified twin plane. The experimental results indicate that the LSFs evolve with strain and are at their maximum level before twinning. The initiated twins are correlated with the highest LSFs compared with the unactivated variants. In addition, the product of the Luster Morris m′ value times the LSF of the twin variant may reveal a threshold for the slip-induced twinning process.

Twin boundary formation at a grain-boundary groove during the directional solidification of InSb

We investigated twin boundary formation during directional solidification of the compound semiconductor InSb. We directly observed a groove formed at the junction of a solid–liquid interface and a Σ9 grain-boundary by optical microscopy. When the depth of the grain-boundary groove reached about 100 μm, rapid growth occurred to fill the groove. We found that two facet planes of the grain-boundary groove became twin boundary planes. Calculations of the change in free energy associated with nucleation at the grain-boundary groove indicates that the preferential twin nucleation occurs at the bottom of the groove, consistent with the direct observations in this study.

The interplay between twinning and cation inversion in MgAl2O4-spinel: Implications for a nebular thermochronometer

Abstract We report a first-principles-based thermodynamic investigation of the interplay between cation inversion and twinning in MgAl2O4 spinel (MAS). We examine the atomic-scale structure of (111) twins and characterize the local octahedral and tetrahedral distortions. We observe that the asymmetric nature of polyhedral distortions about the (111) twin plane causes anisotropy in cation inversion energies near the planar fault. The predicted enthalpies and entropies of inversion reveal that in comparison to the Kagome layer, the anti-site occupancies of Al and Mg, i.e., cation inversion, on the mixed-cation-layer near the twin boundary are more favorable and stable in the entire range of temperature of twin stability. Structurally, such a stable inversion is necessitated by the minimization in the polyhedral distortions, especially by the octahedral distortion, which exhibits a reduction of four orders of magnitude relative to the polyhedra with no inversion. The fundamental understanding obtained on the thermodynamics of the twin-cation inversion interplay in conjunction with the kinetics of inversion was used as a basis for developing a thermochronometer for deducing the temperature of twinning in MAS. This work serves as an important steppingstone for experimental characterization of MAS structures within a host of Earth and planetary materials. In the case of the latter, our results enable the use of planar faults, such as twins, as important markers for deducing the physical and chemical landscape that MAS experienced in its evolution and transport within the solar protoplanetary disk.

Atomic-Scale Investigation of Electromigration with Different Directions of Electron Flow into High-Density Nanotwinned Copper through In Situ HRTEM

Electromigration (EM) has aroused substantial attention with the shrinkage of modern devices. EM is an interaction between electron carriers and atoms, meaning that it arises from atomic behaviour. Both the number of twin boundaries (TBs) and the direction of electron flow are crucial factors affecting EM behaviour. In this study, EM phenomena with perpendicular and parallel electron flow were investigated by high resolution transmission electron microscopy (HRTEM). Twin planes could limit the growth direction of voids in the parallel case. The evolution of smooth surfaces into step-like surfaces resulting from the triple point retarding EM has been demonstrated. With the measurement of the resistance, it could be concluded that the specimen with electron flow perpendicular to the TBs had better EM restriction. Columnar grains can confine the position of the voids. This study demonstrates the dynamic evolution of the surface structure, providing insight into the fabrication of interconnections in the integrated circuits industry.