Microcanonical Molecular Dynamic Simulations of Au Nanoclusters

In this paper, we study nanoparticles with constituent atoms ranging from dozens to hundreds of them. These types of particles display structural and magnetic properties that strongly depend on the number of constituents N. The metal clusters are important due their interesting properties when compared to bulk materials; hence they have potential technological applications. Specifically, we study the Au nanoclusters through classical molecular dynamics simulations; we analyze the total and potential energy as a function of time. Likewise, we study the geometrical structures of Au Nanocluster corresponding to the lowest energy states at 0 K. We consider the method of microcanonical ensemble, and we carry out computer simulations by operating the XMD software package and the atomistic configuration viewer AtomEye.

💡 Research Summary

The paper presents a systematic investigation of gold (Au) nanoclusters using classical molecular dynamics (MD) simulations performed in the microcanonical (NVE) ensemble. The authors aim to elucidate how the structural, energetic, and magnetic properties of metallic nanoparticles evolve as the number of constituent atoms (N) varies from a few dozen to several hundred. To achieve this, they employ the XMD simulation package, which is optimized for large‑scale parallel MD, and they visualize atomic configurations with the AtomEye viewer.

Simulation protocols begin with randomly generated or face‑centered cubic (fcc) seed structures for each cluster size. Initial velocities are assigned from a Maxwell‑Boltzmann distribution corresponding to a high temperature (≈1000 K). The system is first equilibrated under NVE conditions for several hundred thousand time steps (Δt = 1 fs) to reach a thermal steady state. Subsequently, a linear cooling schedule reduces the temperature to 0 K at a rate of 10 K per picosecond, allowing the clusters to settle into their lowest‑energy configurations. Because the ensemble conserves total energy, any drift in total energy is monitored as a diagnostic of numerical stability; the authors report that total energy remains essentially constant while the potential energy gradually stabilizes during cooling.

Energy analysis shows a clear size‑dependent trend: the minimum potential energy per atom decreases non‑linearly with increasing N, reflecting the diminishing surface‑to‑volume ratio and the concomitant reduction of surface stress. The authors also track the time evolution of kinetic and potential contributions, noting that kinetic energy fluctuations diminish as the system approaches 0 K, while residual potential energy oscillations correspond to low‑frequency vibrational modes and occasional surface atom rearrangements.

Geometrical analysis focuses on the identification of the ground‑state motifs for each cluster size. For magic numbers N = 13, 55, 147, the clusters adopt highly symmetric icosahedral structures, which minimize surface energy and maximize coordination. In the intermediate range (N ≈ 70–120), decahedral and mixed twinned structures become competitive, indicating a balance between strain relief in the interior and surface facet optimization. For larger clusters (N ≈ 150–250), fcc‑based configurations dominate, suggesting a gradual convergence toward bulk‑like ordering as the interior volume grows. The authors quantify structural characteristics by calculating average bond lengths, bond‑angle distributions, and radial distribution functions. Icosahedral clusters exhibit a mean Au–Au bond length of ~2.88 Å and a narrow angular distribution centered around 60° and 120°, whereas fcc clusters show angles clustering near 90° and 180°, consistent with bulk crystallography.

The study acknowledges several limitations. First, the exclusive use of the NVE ensemble omits realistic temperature control and pressure effects that would be present in experimental conditions; future work could incorporate NVT or NPT ensembles to assess thermodynamic response. Second, the embedded‑atom method (EAM) potential employed captures many‑body metallic bonding but does not account for explicit electronic structure changes; coupling the MD with density‑functional‑theory (DFT) calculations or machine‑learning potentials would improve accuracy for properties sensitive to electron density. Third, computational cost escalates sharply for clusters larger than ~300 atoms, limiting the statistical sampling of configurational space; accelerated MD techniques or coarse‑grained models could mitigate this bottleneck.

In conclusion, the paper demonstrates that microcanonical MD, combined with robust visualization tools, can reliably map the energy landscape and structural evolution of Au nanoclusters across a wide size range. The findings provide valuable reference data for nanomaterial design, offering insight into size‑dependent stability, preferred geometries, and the transition from discrete cluster behavior to bulk‑like crystallinity.