Proton exchange membranes (PEMs) with both high selectivity and high permeance are of great demand in hydrogen-based applications, especially in fuel cells. Although graphene membranes have shown high selectivity of protons over other ions and molecules, the relatively low permeance of protons through perfect pristine graphene restricts its practical applications. Inspired by the nitrogen-assisted proton transport in biological systems, we introduced N-doping to increase the proton permeance and proposed a type of N-doped graphene membranes (NGMs) for proton exchange, which have both high proton permeance and high selectivity. Compared to the state-of-the-art commercial PEMs, the NGMs show significant increases in both areal proton conductivity (2−3 orders of magnitude) and selectivity of proton to methanol (1−2 orders of magnitude). The work realized the controllable tuning of proton permeance of the graphene membrane with N-doping and developed a new type of graphene-based PEMs with high performance for practical applications.

Membranes are key components in chemical purification, biological separation, and water desalination. Traditional polymeric membranes are subjected to a ubiquitous trade-off between permeance and selectivity, which significantly hinders the separation performance. Nanoporous atomically thin membranes (NATMs), such as graphene NATMs, have the potential to break this trade-off. Owing to their uniqueness of two-dimensional structure and potential nanopore structure controllability, NATMs are expected to have outstanding selectivity through molecular sieving while achieving ultimate permeance at the same time. However, a drastic selectivity discrepancy exists between the proof-of-concept demonstrations and scalable separation applications in graphene membranes. In this paper, we offer a possible solution to narrow this discrepancy by tuning the pore density and pore size separately with two successive plasma treatments. We demonstrate that by narrowing the pore size distribution, the selectivity of graphene membranes can be greatly increased. Low-energy argon plasma is first applied to nucleate high density of defects in graphene. Controlled oxygen plasma is then utilized to selectively enlarge the defects into nanopores with desired sizes. This method is scalable, and the fabricated 1 cm2 graphene NATMs with sub-nanometer pores can separate KCl and Allura Red with a selectivity of 104 and a permeance of 1.1 × 10–6 m s–1. The pores in NATMs can be further tuned from gas-selective sub-nanometer pores to a few nanometer size. The fabricated NATMs show a selectivity of 35 between CO2 and N2. With longer enlargement time, a selectivity of 21.2 between a lysozyme and bovine serum albumin can also be achieved with roughly four times higher permeance than that of a commercial dialysis membrane. This research offers a solution to realize NATMs of tunable pore size with a narrow pore size distribution for different separation processes from sub-nanometer in gas separation or desalination to a few nanometers in dialysis.