Main content

Metadynamics [1] is a computational method aimed at enhancing the sampling of the configurational space. To achieve this it adds a time dependent repulsive bias potential function of coarse-grained variables, called collective variables. This bias descourages the system from revisiting already sampled configurations.
This method makes possible the estimate of the free energy surface of systems that are not tractable in reasonable computer time with regular unbiased molecular dynamics. 

Over the years many improvements to Metadynamics have been proposed, and it has been applied to a variety of different problems. [2]

sketch of metadynamics
A sketch of the process of metadynamics. First the system evolves according to a normal dynamics, then a Gaussian potential is deposited (solid gray line). This lifts the system and modifies the free-energy landscape (dashed gray line) in which the dynamics evolves. After a while the sum of Gaussian potentials fills up the first metastable state and the system moves into the second metastable basin. After this the second metastable basin is filled, at this point, the system evolves in a flat landscape. The summation of the deposited bias (solid gray profile) provides a first rough negative estimate of the free-energy profile. Credits for the figure go to Ref. [3]
Page URL:
© 2017 Eidgenössische Technische Hochschule Zürich