Bismuth-stabilizedc(2×6)reconstruction on a InSb(100) substrate: Violation of the electron counting model

Abstract

By means of scanning tunneling microscopy/spectroscopy ͑STM/STS͒, photoelectron spectroscopy, and first-principles calculations, we have studied the bismuth ͑Bi͒ adsorbate-stabilized InSb͑100͒ substrate surface which shows a c͑2 ϫ 6͒ low-energy electron diffraction pattern ͓thus labeled Bi/ InSb͑100͒c͑2 ϫ 6͒ surface͔ and which includes areas with metallic STS curves as well as areas with semiconducting STS curves. The firstprinciples phase diagram of the Bi/InSb͑100͒ surface demonstrates the presence of the Bi-stabilized metallic c͑2 ϫ 6͒ reconstruction and semiconducting ͑4 ϫ 3͒ reconstruction depending on the chemical potentials, in good agreement with STS results. The existence of the metallic c͑2 ϫ 6͒ phase, which does not obey the electron counting model, is attributed to the partial prohibition of the relaxation in the direction perpendicular to dimer rows in the competing reconstructions and the peculiar stability of the Bi-stabilized dimer rows. Based on ͑i͒ first-principles phase diagram, ͑ii͒ STS results, and ͑iii͒ comparison of the measured and calculated STM and photoemission data, we show that the measured Bi/ InSb͑100͒c͑2 ϫ 6͒ surface includes metallic areas with the stable c͑2 ϫ 6͒ atomic structure and semiconducting areas with the stable ͑4 ϫ 3͒ atomic structure.

Unusual Bi-Containing Surface Layers of III–V Compound Semiconductors

“…On the other hand, the absence of the smaller CLSs in the measurements of course does not exclude the presence of the phases producing only the smaller shifts. For the presented InSb(100)c(2 × 6)-Bi case, the CLS analysis is complicated by the presence of other reconstructed phases, different from the Model-2, as shown in [39]. However, the CLSs of the Model-2 do not disagree with the CLS measurements, in consistency with the above findings.…”

“…The Bi-induced III-V surfaces are usually produced at low temperatures, as compared to many other III-V surfaces, because Bi is easily re-evaporated from the surfaces. Figure 10.8 shows an example comparison of the measured and calculated STM images for an InSb(100)c(2 × 6)-Bi surface [39]. It can be seen that the Model-2 reproduces best the measured filled-state image in Fig.…”

“…In contrast, the InSb(100)c(2 × 6)-Bi was found (Figs. 10.8 and 10.9) [39]. The Bi experiments on GaSb(100) and GaP (100) substrate have not yet been reported.…”

“…This is in agreement with the experimental results as the (2 × 1) reconstruction is not seen on the InSb(100) surface. Instead, the (4 × 3) reconstruction is stabilized [39].…”

Architecture of PTCDA molecular structures on a reconstructed InSb(001) surface

“…A structural model for surfaces of these compounds (the ζ model) was proposed by Lee et al 42 and Kumpf et al, 43,44 and was subsequently supported by a few independent theoretical and experimental studies [45][46][47][48] and recently was proved to be the ground state under certain experimental conditions. 49,50 According to this model, the surface top bilayer exhibits mostly a 4 × 1 symmetry and the c(8 × 2) superstructure is introduced by A III atom dimerization in the second bilayer.…”