Barbara Jones, IBM Almaden Research Center
A Scanning Tunneling Microscope (STM) can these days almost routinely place atoms on a conducting surface and image them, as well as see some larger electronic density patterns. Less routine are the calculations to model these atomic-scale systems, which are multi-day, even multi-week computations performed on supercomputers at IBM and elsewhere. I describe our first principles calculations of the unusual charge and spin properties of singles and pairs of magnetic atoms on a complex surface composed of a layer of insulator on a metal. Different magnetic elements, although close on the periodic table, can have unexpectedly different behavior. When two magnetic atoms are close to one another they interact, with complex and interesting results. Our calculated pairing interaction energy shows an excellent match to experimental values. We moreover show that the coupling can be decomposed into three different pairing interactions and how to extract the values of each separately. The ultimate goals of the experimentalists for these systems are future nanomemories or atomic quibits. I will conclude with some comments about the role of first-principles calculations for nanostructures.