Possible n/p-type conductivity of two-dimensional graphene oxide by boron and nitrogen doping: Evaluated via constrained excitation

As the first-principles calculations using the supercell approximation give widely scattered results in a two-dimensional charged system, making the evaluation of defect ionization energy difficult, here an alternative constrained excitation is applied to overcome this problem for defect analysis. As an example in graphene oxide with 50% oxygen coverage (according to the popular epoxy-chain-plus-hydroxyl-chain model), the structures, stabilities, and electronic properties of nitrogen and boron dopants are investigated. Generally, boron prefers to replace carbon in the sp³ region as an acceptor while nitrogen has a tendency to substitute the sp² carbon close to the boundary between the sp² region and the sp³ region as a donor. Their ionization energies are 0.24-0.42 eV for boron and 0.32-0.67 eV for nitrogen. However, a special case of nitrogen doped in the boundary-sp³ carbon can change to be an acceptor with the assistance of its neighboring (epoxy) oxygen "Lift-off," leading to the shallowest ionization energy of 0.12 eV and the best candidate for p-type conductivity. The present study offers the detailed pictures of boron and nitrogen defects in graphene oxide for the potential n- and p-type conductivity.