The performance of density functional and wavefunction-based methods for 2D and 3D structures of Au10

The transition from 2D to 3D structures in small gold clusters occurs around 10 atoms. Density functional theory predicts a planar D2h structure for Au10 in contrast to recent second-order Møller-Plesset perturbation theory calculations, which predict a 3D C2v arrangement. The validity of the use of single-reference second-order Møller-Plesset theory for near metallic systems remains, however, questionable. On the other hand, it is less than clear how well density functional approximations perform for such clusters. We, therefore, decided to carry out quantum chemical calculations for Au10 using a variety of different density functionals as well as wavefunction-based methods including coupled cluster theory to compare the different energetically low lying 2D and 3D cluster isomers. The results are perhaps not encouraging showing that most computational methods do not predict correctly the energetic sequence of isomers compared to coupled cluster theory. As perturbative triple corrections in the coupled cluster treatment change the order in cluster stability, the onset of 2D to 3D transition in these gold clusters remains elusive. As expected, second-order Møller-Plesset theory is not suitable for correctly describing such systems.