Aqueous proton transfer across single-layer graphene
Summary¶
Achtyl et al. combine interface-specific second-harmonic generation (SHG) on graphene-on-fused-silica with plane-wave DFT (PBE/PAW, NEB) and ReaxFF molecular dynamics to argue that aqueous protons can traverse monolayer graphene reversibly during bulk pH cycling, communicating with silanol acid–base chemistry beneath the sheet. After arguing against macroscopic pinhole transport via SEM and diffusion-order-of-magnitude estimates, they attribute permeation to rare atomic defects. The modeling reports comparatively low barriers (~0.61–0.75 eV) for water-mediated Grotthuss-like transfer across hydroxyl-terminated defect motifs, while pyrylium-like ether terminations are predicted to block exchange; helium/hydrogen transfer is argued to be disfavored, supporting proton selectivity.
Methods¶
Experiments (interface electrochemistry + SHG). Well-characterized single-layer graphene on fused silica is exposed to aqueous electrolytes at room temperature while the bulk pH is cycled between about 3 and 10 at fixed 1 mM ionic strength. A dual-pump flow cell (Fig. 1a) runs near 0.9 ml s⁻¹ as documented on the opening pages of papers/Achtyl_NatureComm_2015.pdf. Interfacial second-harmonic generation (SHG) with ~120 fs pulses (energies below the graphene damage threshold cited in the article) tracks silanol acid–base speciation at the buried silica/water interface. SEM and proton-diffusion scaling estimates (developed in the article) are used to argue against macroscopic pinholes as the dominant transport route.
Static QM / DFT (barriers and defect models). Plane-wave DFT within the GGA–PBE approximation uses projector-augmented wave (PAW) potentials. The Methods section of papers/Achtyl_NatureComm_2015.pdf specifies a 400 eV cutoff for calculations on systems without explicit water solvation, whereas solvated supercells use 283 eV cutoffs for C and O (with other elements treated as stated in the article). The surface Brillouin zone is sampled with a Monkhorst–Pack mesh reported as 3×3×1, and electronic energies are converged to within 1×10⁻⁵ eV in the same section. Nudged elastic band (NEB) calculations map proton-transfer pathways for oxygen- versus hydroxyl-terminated defect motifs referenced in Fig. 2 and Table 1.
DFT dispersion correction: N/A — the Methods excerpt used for this note does not name an explicit DFT-D/vdW correction beyond the PBE functional; see the article/SI for any updates.
MD application (ReaxFF). Reactive MD uses the ReaxFF implementation described in the Methods section of papers/Achtyl_NatureComm_2015.pdf on a (6×6) periodic graphene sheet solvated on both sides (~15.01 × 17.83 Å in-plane, 30 Å normal). Simulations are run in the NVT ensemble with a 0.25 fs timestep and Berendsen thermostat (100 fs coupling constant) to control the total system temperature, with longer aggregate sampling windows reported for rare-event statistics (consult the article/SI for exact trajectory lengths).
Force-field training: N/A — the publication applies an established ReaxFF parameterization for C/H/O/water/graphene rather than reporting a new fit in this manuscript.
Electric field / bias during MD: N/A — the MD discussion focuses on neutral defect motifs; biased electrochemistry is represented experimentally rather than as constant-potential AIMD here.
Findings¶
SHG traces with graphene overlaying silica match control traces on bare silica within experimental uncertainty, yet still track silanol acid–base chemistry expected only if protons communicate through the sheet—supporting permeation without requiring macroscopic tears. After ruling out macroscopic pinholes, the authors attribute transport to rare native defects. DFT/NEB and ReaxFF sampling converge on 0.61–0.75 eV barriers for water-mediated Grotthuss-like proton hops through hydroxyl-terminated defects, whereas pyrylium-like ether bridges suppress exchange. Helium and H\(_2\) permeation barriers remain comparatively unfavorable in the models highlighted, supporting selective aqueous proton transport. The combined spectroscopy + modeling narrative contrasts defect terminations that enable versus block proton exchange; laser-averaged SHG cannot assign single-defect structures without correlative microscopy or fabrication studies.
Limitations¶
- Rare-defect mechanisms require knowledge of actual defect populations in CVD graphene, which can vary by synthesis and transfer.
- Extract is early pages; full simulation setup and statistics live later in the PDF/SI.
Relevance to group¶
Adri C. T. van Duin is among the co-authors; the work pairs precision interface spectroscopy with atomistic modeling of defect-mediated proton transport (consult the full article for the simulation model details used beyond the p1–2 extract).