Density functional theory based study of molecular interactions, recognition, engineering, and quantum transport in π molecular systems.

CONSPECTUS: In chemical and biological systems, various interactions that govern the chemical and physical properties of molecules, assembling phenomena, and electronic transport properties compete and control the microscopic structure of materials. The well-controlled manipulation of each component can allow researchers to design receptors or sensors, new molecular architectures, structures with novel morphology, and functional molecules or devices. In this Account, we describe the structures and electronic and spintronic properties of π-molecular systems that are important for controlling the architecture of a variety of carbon-based systems. Although DFT is an important tool for describing molecular interactions, the inability of DFT to accurately represent dispersion interactions has made it difficult to properly describe π-interactions. However, the recently developed dispersion corrections for DFT have allowed us to include these dispersion interactions cost-effectively. We have investigated noncovalent interactions of various π-systems including aromatic-π, aliphatic-π, and non-π systems based on dispersion-corrected DFT (DFT-D). In addition, we have addressed the validity of DFT-D compared with the complete basis set (CBS) limit values of coupled cluster theory with single, double, and perturbative triple excitations [CCSD(T)] and Møller-Plesset second order perturbation theory (MP2). The DFT-D methods are still unable to predict the correct ordering in binding energies within the benzene dimer and the cyclohexane dimer. Nevertheless, the overall DFT-D predicted binding energies are in reasonable agreement with the CCSD(T) results. In most cases, results using the B97-D3 method closely reproduce the CCSD(T) results with the optimized energy-fitting parameters. On the other hand, vdW-DF2 and PBE0-TS methods estimate the dispersion energies from the calculated electron density. In these approximations, the interaction energies around the equilibrium point are reasonably close to the CCSD(T) results but sometimes slightly deviate from them because interaction energies were not particularly optimized with parameters. Nevertheless, because the electron cloud deforms when neighboring atoms/ions induce an electric field, both vdW-DF2 and PBE0-TS seem to properly reproduce the resulting change of dispersion interaction. Thus, improvements are needed in both vdW-DF2 and PBE0-TS to better describe the interaction energies, while the B97-D3 method could benefit from the incorporation of polarization-driven energy changes that show highly anisotropic behavior. Although the current DFT-D methods need further improvement, DFT-D is very useful for computer-aided molecular design. We have used these newly developed DFT-D methods to calculate the interactions between graphene and DNA nucleobases. Using DFT-D, we describe the design of molecular receptors of π-systems, graphene based electronic devices, metalloporphyrin half-metal based spintronic devices as graphene nanoribbon (GNR) analogs, and graphene based molecular electronic devices for DNA sequencing. DFT-D has also helped us understand quantum phenomena in materials and devices of π-systems including graphene.