Adaptive quantum design of atomic clusters

Adaptive quantum design identifies the best broken-symmetry configurations of atoms and molecules that enable a desired target function response. In this work, numerical optimization is used to design atomic clusters with specified quasiparticle densities of states. The dominant self-assembled building blocks of these engineered quantum systems are found to depend on the symmetry of the target function. For example, particle-hole symmetric spectra can be constructed from a dilute configuration of atomic dimers, whereas more complex structures such as trimers and quadrumers are required for asymmetric target functions. The convergence of the optimization algorithms depends on the shape of the target function, the density of atoms, and constraints due to substrates and boundary conditions. Hybrids of steepest descent methods, simulated annealing, and genetic algorithms are found to be most efficient.