Asynchronous multicanonical basin hopping method and its application to cobalt nanoclusters

Abstract

The multicanonical basin hopping (MUBH) method, which uses a multicanonical weight in the basin hopping (BH) Monte Carlo method, was found to be very efficient for global optimization of large-scale systems such as Lennard-Jones clusters containing more than 150 atoms. We have implemented an asynchronous parallel version of the MUBH method using the message passing interface (MPI) to take advantage of the full usage of multiprocessors in either a homogeneous or heterogeneous computational environment. Based on the intrinsic properties of the Monte Carlo method, this MPI implementation used the task parallelism to minimize interthread data communication. For a Co nanocluster consisting of N atoms, we have applied the asynchronous multicanonical basin hopping (AMUBH) method (for 181 < N < or = 200), together with BH (for 2 < or = N < 150) and MUBH (for 150 < or = N < or = 180), to search for the molecular configuration of the global energy minimum. AMUBH becomes the only practical computational scheme for locating the energy minimum within realistic computational time for a relatively large cluster.

Global minima of transition metal clusters described by Finnis–Sinclair potentials: A comparison with semi-empirical molecular orbital theory

“…The initial geometries were uniformly scaled before each relaxation according to the mean nearest-neighbour distance. For Fe 55 , the structure of the corresponding Al 55 cluster is a low-lying minimum for the 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 , the Morse ρ = 3 global minimum and the FS global minimum each relaxed rapidly to a low-lying local minimum with the alternative potential. The aluminium and FS global minima correspond to the same structure.…”

“…Here, the databases obtained for different parameterizations of the Sutton-Chen (SC) potential [46] and for the Morse potential as a function of the range parameter are particularly useful [24,47]. The other entries for metal clusters in the CCD include data for sodium [48] (Gupta [49,50] and Murrell-Mottram (MM) [51][52][53] potentials), aluminimum [54] (glue potential [55]), lead [45,56] (glue [57] and Gupta [49,50] potentials), and noble metals [58], along with more systematic results for the Gupta potential [59,60]. Recent successes for structure prediction include a combination of theory and experiment to characterize gold clusters [61].…”

Curved-line search algorithm for ab initio atomic structure relaxation

“…The metal cluster is an important subfield related to catalysts [4,10,18]. A lot of works have been done in searching of the optimal cluster structures, and the density of local minima in energy [4,5,10,18,19,[26][27][28]. For metallic systems, we found that the N -body Gupta force field [29] is a very good general potential.…”

“…In higher dimension, the situation can be more complicated, making it impossible to find the natural degree of freedom (e.g., the rotation angle). One such example is a metal cluster [4,18,19] (which will be studied in this paper), where hundreds of steps might be needed to relax a structure while there is no obvious natural (or say internal) degree of freedom to speed up the convergence. To overcome these problems, one needs to do the minimization steps along guided curved lines following the energy valleys.…”

“…Note, the above procedure maintains the feature of CG, when close to the minimum. We will use metal cluster [4,18,19] to demonstrate our CLS algorithm. The metal cluster potential is intrinsically high dimensional due to their long range atom-atom interaction.…”