--- title: Dynamics-Internal Grain Selection url: https://www.emergentmind.com/topics/dynamics-internal-grain-selection type: topic --- # Dynamics-Internal Grain Selection “Dynamics-internal grain selection” (*Editor’s term*) denotes the class of grain-evolution processes in which the fate of grains, grain-boundary segments, or internal grain regions is determined by dynamical mechanisms such as grain boundary migration, lattice misorientation evolution, triple junction drag, defect transport, odd-elastic stresses, elastic anisotropy, or diffusion-limited atomic mobility, rather than by a purely static minimization picture. In the literature, this selection appears as preferential growth of elastically compliant orientations, persistence of low-energy or special boundaries, shape- and orientation-dependent strengthening, stochastic activation of soft grain-boundary regions, self-rotation and self-fission of odd grains, and reversible switching of mobile nanograins [2505.03957], [2411.15747], [1706.07887]. ## 1. Conceptual scope and governing variables Across continuum, phase-field, dislocation-dynamics, and data-driven descriptions, the common structure is an evolving competition among grains mediated by energy, mobility, and defect kinetics. In models with dynamic lattice misorientations and triple junction drag, the total grain boundary energy is written as \[ E(t)=\sum_j \sigma(\Delta^{(j)}\alpha)\,|\Gamma_t^{(j)}|, \] so grain selection is controlled by misorientation-dependent boundary energy, curvature, and triple-junction mobility [2105.07255]. In dislocation-dynamics simulations of copper, selection is expressed through the strengthening of grains with particular size, shape, and orientation combinations [1903.01708]. In elastically driven abnormal grain growth, reduction of elastic energy provides a thermodynamically plausible driving force for orientation-selective grain growth [2607.05298]. In soft metals during electrochemical deposition, grain competition is governed essentially by the difference in surface energy and strain energy between grains [2411.10839]. Large-ensemble threshold-dynamics simulations further show that anisotropic grain boundary energies and mobilities introduce a statistical preference for certain grain orientations and alter the overall grain size distribution [2309.09243]. | Mechanism class | Dominant variable | Selection outcome | |---|---|---| | Dynamic misorientation and triple junction drag | \(\sigma(\Delta \alpha)\), \(\eta\), \(\gamma\) | Low-energy boundaries persist; local texture selection | | Dislocation dynamics | \(d\), shape factor \(\psi\), orientation | Shape- and orientation-dependent Hall-Petch strengthening | | Elastic anisotropy | \(E_{hkl}\), \(\Delta Y^{el}\) | Compliant orientations grow preferentially | | Odd elasticity | \(\alpha_1,\alpha_2\), surface odd stress | Self-rotation, self-fission, reverse ripening, fragmentation | | Surface-energy and diffusion-barrier anisotropy | \(E_s\), \(E_a\), \(D_i\) | Texture selection during electrochemical growth | | Statistical anisotropy in GB character | \(\gamma_{ij}\), \(\mu_{ij}\), HAGB fraction | AGG, texture development, twin-boundary growth | This synthesis suggests that “selection” is not restricted to monotonic grain coarsening. Depending on the driving physics, it can mean enhanced growth, suppressed growth, fragmentation, locomotion, rotational alignment, or preferential survival of particular boundary characters. ## 2. Odd elasticity and anomalous grain selection A particularly explicit dynamics-internal selection mechanism is provided by the phase field crystal with transverse interactions (T-PFC), which extends standard PFC by adding transverse, nonconservative, nonreciprocal interactions to the density field \(\psi(\mathbf r,t)\): \[ \frac{\partial \psi}{\partial t} = \left[ (\nabla \psi)\times \nabla \left(\alpha_1 \nabla^2 \psi+\alpha_2 \nabla^4 \psi\right) \right]_z + \nabla^2 \frac{\delta F_{\text{PFC}}}{\delta \psi}. \] This lowest-order nonconservative flux breaks 2D parity symmetry and generates odd elasticity. In the amplitude expansion, the generalized elasticity tensor takes the form \(C_{ijkl}=C^{(\text e)}_{ijkl}+C^{(\text o)}_{ijkl}\), with \(C^{(\text e)}_{ijkl}=C^{(\text e)}_{klij}\) and \(C^{(\text o)}_{ijkl}=-C^{(\text o)}_{klij}\); the odd bulk modulus \(A\) and odd shear modulus \(K^{\text o}\) are proportional to the strength of transverse interactions and the amplitude of density modulations [2505.03957]. Within this framework, odd grains experience a net surface torque due to non-cancelling surface odd stress, producing spontaneous rigid-body self-rotation with a rotation frequency obeying \(\omega/\alpha_1 \sim N^{-s}\), where \(s\approx 1\). The same model identifies a distinct type of surface cusp instability induced by self-generated surface odd stress. This instability nucleates and emits dislocations at cusp tips, drives their self-propulsion into the grain, and can produce full fragmentation of a single grain into smaller self-rotating grains. The critical size for this instability obeys \(R_c \sim |\alpha_1|^{-\beta}\) with \(\beta \approx 2.5\) [2505.03957]. The resulting selection laws are anomalous. At low \(\alpha_1\), normal Ostwald ripening prevails, with large grains growing and small grains shrinking. At high \(\alpha_1\), large grains above the instability threshold undergo self-fission, lose mass to smaller grains, and may shrink or match in size, yielding a transition from normal to reverse Ostwald ripening. In polycrystals, increasing \(\alpha_1\) drives a transition from conventional grain coarsening to persistent grain self-fragmentation, with the characteristic grain size scaling as a negative power of the transverse interaction strength. The paper distinguishes this mechanism from reverse ripening in active fluids driven by negative pseudotension and from misfit-stress-driven reverse ripening; here the mechanism is odd-elasticity-driven and tied to surface instability [2505.03957]. The same transverse-interaction field becomes a control parameter for grain locomotion. Spatial modulation of \(\alpha_1(\mathbf r)\) selects among self-rotation, self-translation, and self-rolling, so a single crystallite can be steered through pre-designed spatial profiles of transverse interaction. In this setting, grain selection is inseparable from grain kinematics: the same odd stresses that select grain size and fragmentation pathway also select the mode of motion. ## 3. Grain-boundary network evolution, misorientation dynamics, and triple-junction drag A second major lineage treats selection as a consequence of coupled grain-boundary motion, evolving lattice misorientations, and triple junction kinetics. In the curvature-based model with dynamic misorientations and triple junction drag, the governing equations are \[ \left\{ \begin{aligned} v_n^{(j)} &= \mu\,\sigma(\Delta^{(j)}\alpha)\,\kappa^{(j)},\\ \frac{d\alpha^{(j)}}{dt} &= -\gamma\Bigl[ \sigma_\theta(\Delta^{(j+1)}\alpha)|\Gamma_t^{(j+1)}| - \sigma_\theta(\Delta^{(j)}\alpha)|\Gamma_t^{(j)}| \Bigr],\\ \frac{d\mathbf a}{dt} &= \eta\sum_{k=1}^3 \sigma(\Delta^{(k)}\alpha) \frac{\mathbf b^{(k)}(0,t)}{|\mathbf b^{(k)}(0,t)|}. \end{aligned} \right. \] These equations encode weighted mean-curvature motion, grain rotation through misorientation relaxation, and triple-junction drag [2105.07255]. The physical consequence is that grain selection is controlled simultaneously by geometry and crystallography. Grain boundaries with higher energy and unfavorable curvature shrink, while boundaries of lower energy and special misorientation are more likely to persist or expand. Increasing the misorientation-relaxation rate \(\gamma\) sharpens selection by promoting low-misorientation states. Triple-junction mobility \(\eta\) controls how effectively this energetic preference can be realized: small \(\eta\) yields collective, slowed evolution with strong junction constraint, whereas large \(\eta\) approaches a curvature-dominated regime in which energetically unfavorable grains are more selectively removed. The same framework reports that, in steady-state experimental and simulated grain boundary character distributions, the distribution is inversely related to the grain boundary energy density, and the simulated steady state is closely matched by a Boltzmann distribution \(\rho_D(\Delta \alpha)=Z_D^{-1}\exp[-\sigma(\Delta \alpha)/D]\) [2105.07255]. When curvature effects are relaxed to isolate misorientation dynamics and triple-junction drag, the reduced ODE system \[ \frac{d\alpha^{(j)}}{dt} = - \left[ \sigma'(\Delta\alpha^{(j+1)})|\mathbf b^{(j+1)}| - \sigma'(\Delta\alpha^{(j)})|\mathbf b^{(j)}| \right], \qquad \frac{d\mathbf a}{dt} = \sum_{j=1}^3 \sigma(\Delta\alpha^{(j)})\frac{\mathbf b^{(j)}}{|\mathbf b^{(j)}|} \] makes the local selection rule especially transparent. The equilibrium analysis shows that misorientations go to zero at steady state, so the mechanism produces local texture selection, while the full reduced system admits local existence, uniqueness, continuous dependence on initial data, and monotone energy dissipation [1903.11512]. A more general variational description is provided by the unified framework based on the dynamic Frank-Bilby equations. There the state variables are grain-boundary velocity \(\mathbf v\), defect density \(\mathbf B\), and misorientation \(\theta\), constrained by the dynamic Frank-Bilby relation \[ \dot{\theta}\cos\frac{\theta}{2}\mathbf n - 2\sin\frac{\theta}{2}\frac{d}{ds}\left(\mathbf v\times \hat{\mathbf z}\right) + \mathbf B_t^0 =0, \] with evolution determined from the Onsager principle \(\min(Q+\dot E)\) subject to that constraint. The theory accommodates low-angle dislocation-mediated motion, high-angle disconnection-mediated motion, coupled motion, sliding, and grain rotation within a single variational structure, and recovers previously available models as limiting cases [2411.15747]. The tricrystal theory of coupled grain-boundary migration, grain rotation, sliding, and junction motion shows how these ingredients alter actual selection pathways. Finite junction mobility slows shrinkage and rotation, modifies grain shape, and can determine whether the embedded grain disappears by shrinking to zero area or by rotating to match a neighboring grain. External shear stress adds translation to rotation and migration, while grain-boundary diffusion is required for shape accommodation and coherency [1410.3002]. A recurrent implication across these models is that selection in a grain network cannot be reduced to curvature alone; it is a rate-dependent outcome of coupled interface, orientation, and junction dynamics. ## 4. Size, shape, orientation, and elastic anisotropy as selection variables Dislocation-dynamics simulations of copper provide a dynamics-driven internal grain selection rule in which grain size, shape, and crystallographic orientation directly determine flow resistance. Across grain sizes of \(1.25\)–\(10\,\mu\text m\), three orientations \([100]\), \([111]\), and \([135]\), and three shapes (cube, plate, needle), the flow stress follows the Hall-Petch law \[ \sigma_y=\sigma_0+K d^{-1/2}, \] with a slope \(K\) that varies strongly with orientation and shape. \(K\) is largest for \([135]\) and is largest for needle-shaped grains, smaller for plates, and smallest for cubes. The simulations further show a dislocation density storage rate scaling as \[ \frac{d\rho}{d\gamma}\approx \frac{4}{bd}, \] and a generalized effective size \(d_{\text{eff}}=\psi^2 d\) that collapses data for different shapes onto a universal line. In this formulation, grains that offer fewer or more constrained paths for slip and dislocation escape accumulate more back-stress and are selectively stronger [1903.01708]. Elastic loading can also select grains by orientation. In the phase-field and micromechanical study of abnormal grain growth in ultrafine-grained Ni thin films, the driving force is elastic energy reduction. Under constant uniaxial stress \(\sigma_0\), the potential energy is \[ \Pi=-\frac{\sigma_0^2}{2E}V, \] so growth of elastically softer grains reduces the system energy. For FCC Ni, \(E_{100}\) texture at \(250^\circ\text C\), mixed \(<111>\) and \(<113>\) growth-direction texture with increasing off-axis tilt at \(450\)–\(600^\circ\text C\), and dominant \(<220>\) texture with finer nanocrystalline morphology at \(750^\circ\text C\). The mechanism is attributed to a combination of competitive growth and roughness/shadowing rather than to a strictly geometric or stress-based model for the \(<200>\) tilt [2312.13160]. These studies place internal grain selection on a common footing: microstructure is dynamically sorted by the interaction of geometry, stored stress, defect accumulation, and anisotropic elastic response. ## 5. Surface energy, diffusion barriers, and electrochemical grain selection In electrochemical growth of soft metals such as Li and Na, grain selection is formulated as a competition between surface energy anisotropy and atomic mobility-related intrinsic strain energy. The thermodynamic theory writes the energy difference between two grains as \[ \Delta U_{12} = \Delta I_{\text{surface}} + \Delta F_{\text{strain}}, \qquad \Delta I_{\text{surface}} = \frac{E_{s,1}-E_{s,2}}{h}. \] The intrinsic strain is linked to atomic mobility by \[ E_{\text{intrinsic}} = \epsilon_c + (E_T-\epsilon_c)\exp\!\left(-\beta D_i/(LR)\right), \] with diffusion coefficient \[ D_i=D_0\exp\!\left(-\frac{E_{a,i}}{k_B T}\right). \] Grains with lower surface energy are favored thermodynamically, while grains with lower diffusion barrier and higher self-diffusion coefficient more effectively relieve strain and are favored kinetically [2411.10839]. The associated phase-field model evolves orientation fields through \[ \frac{\partial \phi_q}{\partial t} = - L_q(T)\frac{\delta F}{\delta \phi_q}, \] with mobility \(L_q\) and gradient coefficient \(K_q\) parameterized by diffusion barriers and surface energies obtained from DFT. This produces a dynamically explicit selection rule. In solid-state batteries under high stack pressure, load stress-induced surface energy anisotropy can dominate, and Li(001) grains are selected despite their higher diffusion barrier. In liquid-electrolyte cells, or when pressure-induced anisotropy is relieved, diffusion-barrier anisotropy dominates and Li(101) grains prevail. For Na, lower surface-energy anisotropy favors diffusion-oriented grain selection even in solid-state conditions [2411.10839]. The practical significance lies in the coupling between texture selection and electrochemical performance. Fast-diffusion orientations promote smooth, dense, columnar deposits, better surface healing during stripping, and higher critical current density, whereas slow-diffusion textures increase interfacial resistance and lower operational current thresholds before failure. The development of an amorphous \(\text{Li}_x\text{Si}_{1-x}\) seed layer is presented as an interfacial means to reduce lattice mismatch, relieve strain-induced surface energy anisotropy, and shift selection toward the fast-diffusion texture [2411.10839]. This class of models makes explicit that grain selection during deposition is not simply a post-growth coarsening problem. It is built into the growth law itself through orientation-dependent surface energetics and transport kinetics. ## 6. Statistical, microscopic, and experimental signatures Large-scale threshold-dynamics simulations show how anisotropic grain-boundary energies and mobilities reshape ensemble statistics during grain growth. With curvature motion \(v=-m\gamma\kappa\), anisotropy produces abnormal grain growth, heavy tails in normalized grain-area distributions, bimodal area-fraction distributions, and a statistical preference for specific orientation groups. In the texture-focused simulations, both texture development and twin grain-boundary growth become more pronounced when the initial microstructure has a dominant fraction of high-angle grain boundaries. The simulations therefore identify initial HAGB content as a condition that sharpens later selection of low-energy special boundaries [2309.09243]. At a finer structural level, machine-learned softness provides a scalar predictor for the local propensity of an atom to rearrange in a polycrystal. Grain interiors, stacking faults, and twin boundaries have low or moderately negative softness, whereas high-energy grain boundaries show a wide range of higher softness values, with center-most grain-boundary atoms typically the softest and most dynamically active. The rearrangement probability obeys \[ P_R(S)=e^{\Sigma(S)}e^{-\Delta E(S)/k_B T}, \] and the analysis finds that the increase in rearrangement probability with softness is driven primarily by the entropic prefactor \(e^{\Sigma(S)}\), not by a large reduction in the energy barrier. This gives an atom-by-atom selection criterion for dynamically active internal grain regions and suggests that entropy variations dominate the dynamics in grain boundaries [1803.01416]. Experimental electrical noise measurements in RuO\(_2\) nanowires show that entire nanocrystalline grains can act as granular two-level systems, repeatedly switching between metastable coordinate states. Random telegraph noise and resistivity histograms allow extraction of relaxation times, grain sizes, and barrier heights, with \[ \Delta\rho_{\text{grain}}\approx N_{\text{grain}}\times \delta\rho_i, \qquad \frac{1}{\tau_0}=f_0\exp\!\left(-\frac{V_B}{k_B T}\right). \] This identifies a hierarchy of metastabilities in which a predominant mobile grain can dominate the measured fluctuations. Selection here takes the form of dynamical observability: among many grains, the mobile nanograin with the largest electrical signature effectively governs the device-scale response [1706.07887]. A different form of pathway selection appears in two-dimensional binary hexagonal materials. Phase-field crystal simulations of h-BN reveal a dual behavior of positive and negative coupling modes during grain rotation and shrinkage: for intermediate misorientations, the same initial angle can lead to either rotational direction depending on microstructural details, and grains may even switch direction during evolution. This behavior is absent in single-component graphene and is attributed to sublattice ordering, inversion-symmetry breaking, the energetic distinction between heteroelemental and homoelemental bonding, and the availability of multiple defect core structures and transformations [2112.11553]. Taken together, these statistical and microscopic studies show that dynamics-internal grain selection has several experimentally resolvable signatures: non-Gaussian grain-size statistics, evolving grain-boundary character distributions, local softness hot spots, metastable nanograin switching, and defect-structure-dependent branch selection in rotational dynamics. A common misconception is that selection must always appear as smooth average coarsening. The literature instead shows that it may appear as abnormal tails, branch switching, metastable intermittency, or reversible grain-scale motion, depending on which internal dynamical variable is rate-limiting. Source: https://www.emergentmind.com/topics/dynamics-internal-grain-selection