Two-dimensional biomimetic potassium channels at an oil/water interface supported by graphene laminar membranes

Published: 13 September 2022| Version 2 | DOI: 10.17632/7rv8827cbd.2
Yong Chen,
Xiaoyuan Wang


In this work, a new strategy was proposed herein to mimic K+ channels embedded in biomembrane by using the graphene laminar membrane (GLM) composed of two-dimensional (2D) angstrom(Å)-scale channels to support a simple model of semi-biomembrane, namely oil/water (O/W) interface. Additionally, the ion-transfer voltammetry (ITV) was firstly employed to investigate the transmembrane behaviors of some small cations through GLM with comparison to the measurements of ionic conductivity (G). Amazingly, we found that K+ is completely preferred over Na+ and Li+ to transfer across the GLM-supported O/W interface although the monovalent ion selectivity of GLM in the W/GLM/W system is still low (K+/Na+~1.11 and K+/Li+~1.35). Moreover, the voltammetric responses for the ion transfer of NH4+ at the GLM-supported O/W interface are also observed as found that NH4+ can pass through the biological K+ channels. The underlying mechanism of as-observed K+-selective voltammetric responses is discussed and found to be consistent with the energy balance of cationic partial-dehydration (energetic cost) and hydrated cation-π interactions (energetic gain) as involved in biological K+ channels. This work provides a novel platform to not only understand the ion discrimination of graphene-based Å-scale channels from a new perspective by using ion-transfer voltammetry but also exploit the potential applications of 2D laminar membranes in the fields of biomimetic ion channel and electrochemical ion sensing or extraction. Fig. 3(d) and Fig.S3 a-c are the data of PXRD patterns of dry GLM/PET, PET, F-GLM (free-standing GLM) and the GLM/PET after immersion inside deionized water for 48 h. Fig.4. a-d, Fig. 5a,b, and Fig. S4 and Fig. S5a,b are the I-V characteristics and the corresponding Ionic conductance (G=ΔI/ΔV) values obtained by using GLM/PET or F-GLM under different conditions as described in every Figure. Fig.6a-d, Fig. 7b, c, and Fig. S6 and Fig. S7 are the CVs and the corresponding DPVs obtained by using GLM/PET or F-GLM under different conditions as described in every Figure.


Steps to reproduce

Graphene laminar membrane (GLM) was fabricated by using vapor filtration method with an organic support, namely porous polyethylene terephthalate (PET) membrane, which has been used as the hard template to synthesize nanoporous membranes and employed in the membrane-supported O/W interface due to its good resistance to organic solvent. The detailed setup employed in the fabrication of GLM by the vacuum filtration method is described as Fig. S1 and based on the following steps. Firstly, 2 mL of commercial 2 g/L n-methyl-2-pyrrolidone (NMP) monolayer graphene nanosheets dispersion (flake size ranging from 0.5 to 3.0 m, Nanjing XFNANO materials Tech Co., Ltd, China) was centrifuged at 5500 rpm for 30 minutes. The supernatant in the centrifuge tube was carefully removed and the sediment was re-dispersed in 40 mL isopropanol by sonication for 3 minutes to obtain the graphene mono-layered sheet dispersion (~0.1 g/L). Subsequently, as-prepared graphene mono-layered sheet dispersion (2 mL) was filtered through a PET membrane by vacumm filtration under the pressure of 10000 pa. After that, the PET membrane covered with graphene layers was then dried at 60°C in an oven overnight to obtain a graphene laminar membrane (GLM) on the PET support, in short GLM/PET. In order to better clarify the influence of 2D graphene Å-scale channels of as-prepared GLM/PET on the ion permeation and the ion transfer in the followed experiments, a commercial free-standing GLM (Nanjing XFNANO materials Tech Co., Ltd, China) with thickness of 25 m (abbreviated as F-GLM) was also investigated in this work as a comparison. The morphologies and microstructures of GLM/PET and F-GLM were examined by SEM (S-3400N, HITACHI, Japan) and XRD (XRD-6100, SHIMAOZU, Japan). Ion permeation measurements were conducted by using a custom-made cell, where a GLM was placed between two liquid reservoirs to form the W/GLM/W system. The reservoirs were filled with chloride solutions in chosen concentrations, including LiCl, NaCl, KCl, NH4Cl and MgCl2. During measurements of the reported concentration dependence, the sequence of used solution was from low to high concentration and the cell was thoroughly washed with DI water and the test solution. The I-V characteristics were measured by using a four-electrode system, namely the opposite electrode (CE) and reference electrode (RE) are platinum wire electrode and Ag/AgCl electrode respectively. CHI660D electrochemical workstation and linear sweep voltammetry (LSV) were used for ionic conductance test. The maximum applied voltage was limited to be less than 0.2 V in order to avoid the possible hysteresis. LSV was set: initial potential: -0.2 V, end potential: 0.2 V, scanning speed: 0.001 V /s, sampling interval: 0.001 V, resting time: 2 s, sensitivity:  The 10-7-10-5. The GLM-supported W/DCE interface is polarized using four-electrode potentiostats (CHI660D, CHI, USA). High-resistivity distilled water was used to prepare all aqueous solutions.


Shanghai Institute of Technology


Ion Channels, Graphene, Interfacial Electrochemistry