In collaboration with the CASSIOPEE beamline and theoreticians from the Universidad Autónoma de Madrid (Spain), a research group from Nancy (Institut Jean Lamour) specialized in microscopy and spectroscopy of surfaces and interfaces was interested in the physical properties of alkali-metal ultra-thin films deposited on strongly boron-enriched Si(111) substrate. This material presents a triangular geometry and has been proposed few years ago as the prototype of a 2D Mott1 insulator leading potentially to exotic magnetic and superconducting properties.
The Nancy research group has first evidenced the quadrupling of the surface unit cell resulting in a 2√3x2√3R30° surface reconstruction first interpreted as the realization of a bi-polaronic insulator2 : indeed, in the strong electron-phonon coupling limit, the pairing of polarons i.e. electrons «dressed» by their lattice deformation, should lead to a strongly distorted insulating phase. This exotic state of matter should be detectable in solids by a strong charge modulation leading to the alternation of empty and doubly-occupied sites with singular spectral signature such as temperature broadening of photoemission peaks.
Figure 1 : (a) optimized structure from DFT Wien2k calculations ; (b) high resolution core levels photoemission measurements on B-1s; (c) comparison of calculated density of states with ARPES and STS measurements; (d) STM images integrated on unoccupied (V=+1,2 V) and occupied (V=-1,2 V) states for a 0,3 nA tunnel current
By combining photoemission spectroscopy, scanning tunneling microscopy/spectroscopy and ab initio calculations we have determined the structural model of the 2√3x2√3R30° surface reconstruction. In fine, the giant lattice distortion is shown to be responsible for the band gap opening and hence for the insulating nature of this semiconducting interface.
Calculations give evidence for a unit cell containing two trimers of potassium adatoms leading to a giant vertical distortion for only a part of Si adatoms (fig 1-a). As a consequence, an inhomogeneous doping of Si-B molecular dangling bonds takes place in quantitative agreement with the chemical shifts observed in core levels photoemission spectroscopy (fig 1-b). Then, three filled bands separated from a single empty one by a large gap of 1 eV have been deduced from the structure in agreement with photoemission and scanning tunneling spectroscopy measurements (fig 1-c). To the end, comparison of measured and calculated STM images evidences the dephasing of the local density of states across the gap revealing the distinct spatial localization of the occupied and unoccupied electronic states (fig 1-d). Calculated charge transfers from the DFT approach are nevertheless too small showing that even if the electron-phonon coupling should play a role, this interface has not to be considered as a bipolaronic insulator.
Therefore this alkali/Si(111) material is definitely a strongly distorted band insulator closing the controversy on its fundamental ground state
1 – A Mott insulator is a material which should be a metal from the band theory but behaves as an insulator with a charge localization induced by electron-electron Coulomb repulsion. Charge carrier doping of these materials (mainly transition metal oxides) leads in some cases to exotic superconducting properties not explained up to know.
2 – An electron propagating into an ionic crystal shares a lattice distortion. The quasi-particle consisting of an electron dressed by a lattice distortion is called a «polaron». Two neighboring polarons can condensate into a «bipolaron» due to their mutual attraction leading to an insulating ground state. Bipolaronic mechanism has been proposed as one way to get room temperature superconductivity but this is controversial.
Publication :
Giant Alkali-Metal-Induced Lattice Relaxation as the Driving Force of the Insulating Phase of Alkali-Metal/Si(111):B. L. Chaput, C. Tournier-Colletta, L. Cardenas, A. Tejeda, B. Kierren, D. Malterre, Y. Fagot-Revurat, P. Le Fèvre, F. Bertran, A. Taleb-Ibrahimi, D. Trabada, J. Ortega et F. Flores. Phys. Rev. Lett. 107, 187603 (2011)
