Theoretical Study of Guest Molecule Doping Effect on the Electronic Structures of Metal-Organic Framework

Abstract

Metal-Organic Framework (MOF) are three-dimensional nanoporous materials that are self-organized by coordination bonds between metal ions and organic ligands, and can provide functions such as magnetism and catalytic properties[1]. Talin et al. reported that the conductivity of a wide-gap MOF can be improved by adding an electron acceptor TCNQ to HKUST-1, which is one of most familier MOFs[2]. However, the details of the electronic structures of the MOF have not been clarified, and the guidelines for controlling the electrical conductivity in the MOF have not yet been proposed.In this study, we made use of first-principles molecular dynamics calculations to refer to the electronic structures of MOF. At the same time, the effect of guest molecule doping on the electronic structures of MOF was investigated by taking the type and density of guest molecules introduced into MOF as parameters.The first-principles molecular dynamics program PHASE/0[3] based on density functional theory was used to perform calculations by supercomputer systems. The MOF used in the calculation was MOF-5 consisting of Zn (II) and terephthalate dianion. Electron acceptors of TCNQ or F4-TCNQ, which has higher redox potential than that of TCNQ, were introduced into the pores of MOF-5 as guest molecules (Fig.1(A,C)). The changes in the electronic structues of MOF-5 caused by the introduction of the guest molecules were investigated as functions of the arrangement and number of guest molecules.As a result of the calculation, MOF-5 alone was found to be a wide gap material with an energy gap of 3.58 eV, and this gap value was confirmed to be consistent with experiments by Alvaro et al. [4] and calculations by Yang et al. [5]. There are two types of pores in MOF-5 (Fig.1(B)): pores surrounded by the ligand face facing inward (Pore1) and pores surrounded by the (hydrogen attached) ligand edges (Pore2). In this study, only the former (Pore1), which gave a larger energy gain when guest molecules were introduced, was examined. When one TCNQ was introduced as a guest molecule, a level (LUMO level of TCNQ) appeared at 1.09 eV above the valence band maximum (VBM) of MOF-5. The LUMO level became lower and close to the VBM with the increasing number of TCNQ, and reached at 0.75 eV above the VBM when three TCNQs were doped (Fig.2(A)). This is presumed to be caused by the formation of bonding and antibonding states as a result of condensation of TCNQ in the pores. Accordingly, introducing impurity levels by guest molecules effectively lower the electroexcitation barrier of 3.5 eV MOF-5’s wide band gap. In addition, when the guest molecule was replaced by F4-TCNQ, a level was intlroduced near the VBM as in the case of TCNQ at 0.74 eV above the VBM for one F4-TCNQ and at 0.36 eV for three F4-TCNQ (Fig.2(B)). Therefore, the larger the number of molecules introduced into the pores, the closer the level derived from the guest molecular doping approaches the VBM of MOF-5. Doping gusest molucules with higher redox potential is more effective in lowering the electroexcitation barrier.In conclusion, it was shown that impurity levels by increasing the density of guest molecules in the pores of MOF-5 were expected to work as acceptor levels. We will also discuss the dynamics of the holes injected into MOF-5, in particular, the possibility of their contribution to an electric current.