Abstract:
Niobium is widely used in quantum computing devices and charged particle accelerators, while it is in superconducting state. In these applications the niobium surface is subjected to very different electric field strengths and the performance of both is limited by losses that occur in the native oxide which inevitable forms on its surface. Current research indicates that these losses occur by a two-level system (TLS) mechanism either in the surface oxide layer, or at its interfaces. However, the exact physical location and microstructure of the sites that host the TLS are not known. To gain this knowledge, tools and techniques require development. Here, electron energy loss spectroscopy (EELS) in a high-resolution transmission electron microscope is used and extended through the use of model fitting. The developed technique allows localized thickness - composition parameters to be extracted from relatively low-density EELS data in a manner that reduces background and other systematic effects from consideration. Using this technique in future studies of the ~5 nm thick surface oxide layer will enable more reliable and rapid comparisons of oxide layers formed under differing conditions, which may thus have different loss characteristics. This in turn will help isolate the hosts of the loss mechanisms in these materials, and so drive performance improvements in niobium devices where energy loss on its surface is critical.