Obtaining a faster convergence to the thermodynamic limit for metals using structure factor twist averaging

As metals play a large role in material design applications such as catalysis and surface chemistry, there is a crucial need for efficient and accurate wavefunction-based calculations. However, there is still a large computational cost due to larger finite size errors associated with methods such as coupled cluster that currently prohibit wide application to metals. While methods such as twist averaging can be used to reduce the finite size errors through use of offsets to k-point grids called twist angles, these methods often add an additional computational time cost due to requiring multiple calculations. Here, I will show how we use our new structure factor twist averaging method to address this issue through finding a single twist angle that represents the twist averaged system. With this special twist angle, we can achieve the desired reduction in finite size errors while providing a cost reduction up to two orders of magnitudes. I will demonstrate the effectiveness of our method through showing several applications to various metallic systems as well as a range of insulators and semiconductors.