Lattice mold technique for the calculation of crystal nucleation rates

Abstract

We present a new simulation method for the calculation of crystal nucleation rates by computer simulation. The method is based on the use of molds to induce crystallization in state points where nucleation is a rare event. The mold is a cluster of potential energy wells placed in the lattice positions of the solid. The method has two distinct steps. In the first one the probability per unit volume of forming a sub-critical crystal cluster in the fluid is computed by means of thermodynamic integration. The thermodynamic route consists in gradually switching on an attractive interaction between the wells and the fluid particles. In the second step, the frequency with which such cluster becomes post-critical is computed in Molecular Dynamics simulations with the mold switched on. We validate our method with a continuous version of the hard sphere potential and with the sodium chloride Tosi-Fumi model. In all studied state points we obtain a good agreement with literature data obtained from other rare event simulation techniques. Our method is quite suitable for the study of both crystal nucleation of arbitrarily complex structures and the competition between different polymorphs in the nucleation stage.

Tuning the Glass Transition: Enhanced Crystallization of the Laves Phases in Nearly Hard Spheres

“…We thus find that the seeding approach is versatile and robust 22,23 -it not only enables one to determine the nucleation barrier and nucleation rate, but also locate the regime in the phase diagram where spontaneous nucleation may occur and provides information on how a crystal nucleus grows and melts. Our observation of spontaneous nucleation of the LP in a system of soft spheres is important and intriguing for two reasons.…”

“…To investigate the effect of particle softness on the nucleation of the LPs, we determine the nucleation barrier height, critical nucleus size, and nucleation rate using the seeding approach 22,23 . This technique involves inserting a crystalline seed of a pre-determined shape and size into a metastable fluid.…”

Overcoming the Time Limitation in Molecular Dynamics Simulation of Crystal Nucleation: A Persistent-Embryo Approach

“…If the nucleus melts the harmonic potential is gradually enforced. The strategy to adjust the spring constant to zero before reaching the critical nucleus size ensures the dynamics of the system is unbiased at the critical point, which is an advantage of this approach compared to others such as the lattice mold method [13]. A schematic of the simulation configuration is shown in Fig.…”

Homogeneous ice nucleation rates for mW and TIP4P/ICE models through Lattice Mold calculations

“…Admittedly LM is more expensive than Seeding, but still computationally efficient. Previously, LM has been successfully validated for the calculation of nucleation rates of Hard Spheres and the Tosi-Fumi NaCl model [77], and here is adopted to measure ice nucleation rates within a temperature range that includes both micrometer and nanometer-size water droplet experiments [17, 20-22, 24-26, 28, 29, 31]. We first apply the LM technique to the case of ice nucleation in the monoatomic water (mW) model [78], a coarse-grained description of water that mimics the hydrogen-bonding structure of water by means of a non-bonded angular dependent term which encourages tetrahedral configurations.…”