Graphene as a subnanometre trans-electrode membrane
S. Garaj, W. Hubbard, A. Reina, J. Kong, D. Branton, and J.A. Golovchenko
"Graphene as a subnanometre trans-electrode membrane"
Nature 467, 190-194 (2010)
Entry by Meredith Duffy, AP 225, Fall 2011
Garaj et al. present atomically thin nanoporous graphene as a promising technology for the detection of single molecules, such as in DNA sequencing. First, by measuring trans-membrane conductance through pores of different diameters, they confirm the expected sub-nanometer thickness of the graphene sheet. Next, they provide computations to indicate the high spacial resolution of molecule detection potentially achievable due to the sheet's unusual thinness.
Methods, Results and Potential Applications
Graphene, a thin layer of graphite with high in-plane conductivity, was formed using chemical vapor deposition (CVD) on a nickel substrate. After the nickel was etched away with HCl, the graphene sheet was placed over a square hole in the SiN coating of a silicon frame. The graphene was then placed as a membrane between two compartments of a fluidic cell, each side of which was filled with an ionic solution and connected to a Ag/AgCl electrode (Figure 1).
The unmodified graphene sheet proved very insulating. Although applying a voltage between the electrodes did produce small currents, the conductance levels of the graphene were on the order of picosiemens for all ionic solutions tested. This conductance was likely due to ionic transport through defects in the graphene, not elecrochemical currents to/from the graphene itself. However, the addition of a single nanoscale-diameter pore increased the conductance of the membrane by 3-4 orders of magnitude, rendering the sheet itself essentially a total insulator by comparison. Noting that theory states that the conductance of a pore in an infinitely thin insulating sheet scales linearly with the pore diameter, the authors measured the conductance of pores of varying size from 5 to 23 nm and fitted the data to computer calculations that provided corrections to the relation between pore conductance and diameter for a thin (but not infinitely so) sheet (Figure 3). They thus determined the thickness L_IT of the graphene membrane to be 0.6 nm, which equates to only one or two atomic layers.
Although some molecule detection through nanopores is already possible with materials like silicon nitride (~30 nm thick), the unusual thinness of the graphene membrane is advantageous because it is on the same order of magnitude as nucleobases. Thus, as long as the DNA was single-stranded and remained linear, potentially only one base would be passing through the pore at a time. The pore would therefore be able to identify the base using its characteristic blockage vs. duration data as discussed above. Because the velocities of polymers through the pores (their "translocation speed") and electronic bandwidth requirements are too high to test this experimentally, the authors run a computation to determine the resolution of the pore.
Modeling a polymer as a cylinder with initial diameter of 22 Angstroms and a sudden decrease in diameter to 20 Angstroms, the authors passed this polymer through a 24-Angstrom pore. Defining the pore's resolution as "the distance over which conductance changes from 75% of its greatest value to 25% of that value," for a membrane thickness of 1.5 nm the resolution was found to be 7.5 Angstroms, whereas for the graphene's thickness of 0.6 nm the resolution was 3.5 Angstroms (Figure 5). Thus even for a conservative estimate of 1.5 nm membrane thickness, resolution is sub-nanometer, meaning graphene nanopores are theoretically capable of detecting molecules on this order of magnitude.
Although there are still many problems to be solved before high-throughput DNA sequencing can be achieved with nanoporous graphene membranes, including slowing polymer flow through pores to speed matching our detection rate capabilities, minimizing noise without thickening the device, and finally proving that pore blockage differences between nucleobases are strong enough to be distinguishable, the unique properties of the atomically thin membranes exhibit great potential and are sure to generate much interest for future research and development.