[Paper Review] Kinetic Monte-Carlo simulations of sintering

Frank Westerhoff, R. Zinetullin+1 more|arXiv (Cornell University)|Mar 14, 2005

Material Dynamics and Properties7 references3 citations

TL;DR

This study presents a 3D kinetic Monte Carlo (KMC) simulation framework that explicitly models atomic displacements to account for elasticity in sintering of nano-particle aggregates. It reveals that sintering of two particles with different radii follows an r⁴ scaling with reduced radius, deviating from the classical N⁴/³ law, and demonstrates that compressive stress accelerates sintering by lowering activation energy due to strain-dependent diffusion barriers.

ABSTRACT

We simulate the sintering of particle aggregates due to surface diffusion. As a method we use Kinetic Monte-Carlo simulations in which elasticity can explicitly be taken into account. Therefore it is possible to investigate the shape relaxation of aggregates also under the influence of an external pressure. Without elasticity we investigate the relaxation time and surface evolution of sintering aggregates and compare the simulations with the classical Koch-Friedlander theory. Deviations from the theoretical predictions will be discussed.

Motivation & Objective

• To investigate sintering dynamics of nano-particle aggregates beyond the classical Koch-Friedländer theory, particularly for non-uniform particle sizes.
• To examine the influence of elasticity on sintering by allowing atoms to displace from lattice sites to minimize elastic energy.
• To assess the validity of the Koch-Friedländer surface evolution model under varying conditions, especially when the characteristic diffusion length changes.
• To determine how external stress (compression/tension) affects sintering kinetics through strain-dependent activation energies.
• To validate the simulation framework against theoretical predictions and explore the limitations of constant vs. variable relaxation time assumptions.

Proposed method

• Employed a 3D kinetic Monte Carlo (KMC) method on an fcc lattice, allowing atoms to displace from lattice sites to minimize total elastic energy using conjugate gradient optimization.
• Calculated activation energies for atomic hops using a strain-dependent model: $ E_{\text{sp}} = E_{\text{sp,0}} + \alpha \frac{1}{2}(E'_{\text{b,i}} + E'_{\text{b,f}}) $, where $ E'_{\text{b}} $ is strain-modified binding energy.
• Used a Lennard-Jones potential to model interatomic interactions and computed binding energies up to fourth nearest neighbors.
• Applied periodic boundary conditions in the x-direction to simulate external stress, with system size in x controlling the strain.
• Tracked the average squared radius perpendicular to the symmetry axis as a measure of system relaxation toward equilibrium.
• Used a fixed attempt frequency $ \nu $ and exponential rate $ q = \nu \exp(-\beta E_a) $, with $ \beta = 1/k_B T $, to determine hopping probabilities.

Experimental results

Research questions

• RQ1 How does the equilibration time for sintering two particles of unequal size scale with their reduced radius?
• RQ2 To what extent does the classical Koch-Friedländer theory accurately describe surface evolution during sintering when the characteristic diffusion length changes?
• RQ3 How does external compressive or tensile stress influence the sintering rate through elastic effects on atomic diffusion barriers?
• RQ4 Does the assumption of a constant relaxation time $ \tau $ in the Koch-Friedländer model remain valid during the early stages of sintering when particle size and diffusion length evolve?
• RQ5 What is the impact of atomic displacement (elastic relaxation) on the sintering dynamics compared to fixed-lattice KMC simulations?

Key findings

• The equilibration time for two sintering particles scales as $ \tau \propto r^4 $ with the reduced radius $ r = \left( \frac{1}{R_1} + \frac{1}{R_2} \right)^{-1} $, contradicting the classical $ N^{4/3} $ scaling.
• The Koch-Friedländer model with a variable $ \tau $ provides a better fit to surface evolution during early sintering, when the average diffusion length increases.
• A constant $ \tau $ is more appropriate in later stages when diffusion length stabilizes, but the value of $ \tau $ is smaller than in the early regime.
• Compressive stress reduces the activation energy for surface diffusion due to strain-dependent binding energy, leading to faster sintering kinetics.
• Under compressive strain, the relaxation process is faster, as indicated by a steeper initial slope in the normalized $ R^2(t) $ curve, while both compressive and tensile cases asymptotically approach similar equilibrium states.
• The final cylindrical morphology under periodic boundary conditions is likely metastable and subject to Rayleigh instability, with thermal fluctuations causing $ R^2 $ to increase again at long times.