Two-dimensional electron liquids (2DELs) have increasing technological relevance for ultrafast electronics and spintronics, yet significant gaps in their fundamental understanding are exemplified on the prototypical SrTiO3. We correlate the exact SrTiO3(001) surface structure with distinct 2DELs through combined microscopic angle-resolved photoemission spectroscopy and non-contact atomic force microscopy on truly bulk-terminated surfaces that alleviate structural uncertainties inherent to this long-studied system. The SrO termination is shown to develop a 2DEL following the creation of oxygen vacancies, unlike the intrinsically metallic TiO2 termination. Differences in degeneracy of the 2DELs, with nearly the same band filling and identical band bending, are assigned to polar distortions of the Ti atoms in combination with spin order, supported with the extraction of fundamental electron-phonon coupling strength. These results not only resolve the ambiguities regarding 2DELs on SrTiO3 thus far, but also pave the way to manipulating band filling and spin order in oxide 2DELs in general.

• Publication type
• Journal Article MeSH

Determination of the atomic structure of solid surfaces typically depends on comparison of measured properties with simulations based on hypothesized structural models. For simple structures, the models may be guessed, but for more complex structures there is a need for reliable theory-based search algorithms. So far, such methods have been limited by the combinatorial complexity and computational expense of sufficiently accurate energy estimation for surfaces. However, the introduction of machine learning methods has the potential to change this radically. Here, we demonstrate how an evolutionary algorithm, utilizing machine learning for accelerated energy estimation and diverse population generation, can be used to solve an unknown surface structure-the (4×4) surface oxide on Pt3 Sn(111)-based on limited experimental input. The algorithm is efficient and robust, and should be broadly applicable in surface studies, where it can replace manual, intuition based model generation.

• Keywords
• Density Functional Calculations, Machine Learning, Structure Elucidation, Surface Chemistry,
• Publication type
• Journal Article MeSH

Determination of the atomic structure of solid surfaces typically depends on comparison of measured properties with simulations based on hypothesized structural models. For simple structures, the models may be guessed, but for more complex structures there is a need for reliable theory-based search algorithms. So far, such methods have been limited by the combinatorial complexity and computational expense of sufficiently accurate energy estimation for surfaces. However, the introduction of machine learning methods has the potential to change this radically. Here, we demonstrate how an evolutionary algorithm, utilizing machine learning for accelerated energy estimation and diverse population generation, can be used to solve an unknown surface structure-the (4×4) surface oxide on Pt3Sn(111)-based on limited experimental input. The algorithm is efficient and robust, and should be broadly applicable in surface studies, where it can replace manual, intuition based model generation.

Machine learning techniques can be implemented to accelerate surface structure determination based on density functional theory. The application of such an algorithm is demonstrated here for a surface oxide on Pt3Sn(111) which had eluded determination by experimental methods.

• Keywords
• Density Functional Calculations, Machine Learning, Structure Elucidation, Surface Chemistry,
• Publication type
• Journal Article MeSH