Crystallization of Size-Selective Nanopores in Graphene for Gas Separation

Nanoporous single-layer graphene (N-SLG) membranes, owing to their single-atom thinness, have the potential to exceed the permeance and selectivity limits of gas separation membranes. However, two key issues in the top-down N-SLG synthesis need to be addressed to achieve scalable, high-performance membranes: a) reproducible synthesis of high-quality SLG film by chemical vapor deposition (CVD) on a low-cost Cu substrate, and b) introducing high-density of pores with a narrow pore-size-distribution (PSD). This dissertation addresses these issues by developing a method to prepare smooth and oriented Cu foil by a facile approach to obtain high-quality SLG membranes. On the fundamental science front, it explores two novel methods of tuning the PSD in graphene for gas separation. Low-cost Cu foils are rough, and result in membranes with large nonselective intrinsic vacancy defects which hampers their application in gas separation. Herein, we conduct a systematic high-temperature annealing study on two separate, commercial, low-cost Cu foils leading to their transformation to smooth Cu(111), decreasing their surface roughness by ~ 2-fold. The smooth, oriented Cu foils yielded SLG with a significantly lower defect density with ID/IG ratio decreasing from 0.18 ± 0.02 to 0.04 ± 0.01. The intrinsic defects in these SLG films were H2 selective with H2 permeance reaching 1000 gas permeation units (GPU; flux normalized with transmembrane pressure difference) and attractive H2/CH4 and H2/C3H8 selectivities of 13 and 26, respectively. To decouple the pore nucleation and expansion in SLG, we utilize CO2 as a mild etchant to expand the existing pores at temperatures ranging 750 – 1000 °C. CO2 could uniformly expand the intrinsic pores in SLG in a controlled manner, down to a few Ã/min, without nucleating new defects in the SLG basal plane. Furthermore, we revealed two distinct kinetic zones for the reaction of CO2 with graphene edges, with the transition happening around the pore diameter of ~ 2 nm. Etching rate of the larger expanded-pores was constant and independent of the pore size. The expansion was thermally activated with an activation energy of 2.71 eV, consistent with the literature based on ab-initio calculation for CO2 dissociative chemisorption on zigzag edges. In comparison, the etching rate was an order of magnitude slower in smaller pores indicating that geometrical confinement in smaller pores play an important role. An exponential relation between the density of expanded pores and etching temperature, with an activation energy of 3.58 eV, was observed. Next, we develop a novel method to fabricate N-SLG with a high pore-density while maintaining a narrow PSD. We exposed the highly porous SLG (treated by O2 plasma) with a broad PSD to graphene CVD condition in presence of both CH4 and CO2. The pore expansion (as a result of etching) and shrinkage (as a result of growth) reached a comparable rate. Moreover, CO2 suppressed the graphene grain nucleation, leading to high-quality graphene synthesis. Membranes with H2 permeance reaching 10000 GPU and H2/C3H8 selectivity of 26 were fabricated by optimizing the CH4/CO2 ratio at 800°C.In summary, we address the current obstacles in SLG membrane development by utilizing CO2 in graphene CVD environment to tune the PSD of SLG membranes. Moreover, a simple annealing method to optimize the morphological and crystallographic properties of Cu foils for high-quality SLG synthesis is proposed.