Selective protein transport through ultra-thin suspended reduced graphene oxide nanopores

Dae-Sik Lee a, Seokhan Park b, Yong Duk Han c, Jae Eun Lee d, Hu Young Jeong e, Hyun C. Yoon c, Moon Yeon Jung a, Sang Ouk Kim *b and Sung-Yool Choi *ad
aElectronics and Telecommunications Research Institute (ETRI), 218 Gajeongno, Yuseong-gu, Daejeon, 34129, Republic of Korea. E-mail: sungyool.choi@kaist.ac.kr
bNational Creative Research Initiative Center for Multi-Dimensional Directed Nanoscale Assembly, Department of Materials Science and Engineering, KAIST, Daejeon 34141, Republic of Korea. E-mail: sangouk.kim@kaist.ac.kr
cDepartment of Molecular Science and Technology, Ajou University, Suwon 16499, Republic of Korea
dSchool of Electrical Engineering, Graphene/2D Materials Research Center, KAIST, Daejeon 34141, Republic of Korea
eUNIST Central Research Facilities(UCRF), UNIST, 50 UNIST-gil, Ulsan 44919, Republic of Korea

Received 17th March 2017 , Accepted 8th June 2017

First published on 13th June 2017


The nanoporous free-standing graphene membrane is of great interest in high performance separation technology. In particular, the separation of biological molecules with similar sizes is one of the key challenges in the purification of biomaterials. Here, we report a reliable, cost-effective, and facile method for the fabrication of a graphene-based nanosieve and its application in the separation of similar-size proteins. A suspended reduced graphene oxide (rGO) nanosieve with ultra-thin, large-area, well-ordered, and dense 15 nm-sized pores was fabricated using block copolymer (BCP) lithography. The fabricated 5 nm-ultrathin nanosieve with an area of 200 μm × 200 μm (an ultra-high aspect ratio of ∼40[thin space (1/6-em)]000) endured pressure up to 1 atm, and effectively separated hemoglobin (Hb) from a mixture of hemoglobin and immunoglobulin G (IgG), the common proteins in human blood, in a highly selective and rapid manner. The use of the suspended rGO nanosieve is expected to provide a simple and manufacturable platform for practical biomolecule separation offering high selectivity and a large throughput.


Introduction

Single-atom thick graphene and graphene oxide (GO) have been considered ideal membrane materials for ultra-thin nanosieves with nanopores owing to their high mechanical flexibility, chemical stability, and mechanical strength.1–4 Moreover, graphene-based nanosieves can modulate ions and molecules in a highly selective manner. These nanosieves are highly versatile and find applications in a variety of fields such as selective biological molecular sieving,5–7 selective ion penetration,8,9 gas separation,10,11 and water purification.12,13 Thus far, substantial research efforts have been devoted to improve the performance of graphene-based nanosieves.5–15 Nonetheless, scalable production of large-area graphene-based nanosieves with vertically straight nanopores, especially for protein separation, is still challenging, particularly as it requires precise control over the pore size, long-term mechanical stability, film uniformity, and mass-productivity.

Ultra-thin nanosieves with nanopores are highly sought after membrane materials as they offer versatility and outstanding performance far surpassing that of commercialized nanoporous membranes. Various approaches for the production of nanosieves include track etching,16–19 electrochemical etching,20–22 electroless plating,23,24 focused ion beam (FIB) etching,25–28 and self-assembly.29–43 FIB and electron-beam writing have been widely used for drilling nanopores directly into graphene for biological separation applications.5,7,14,15,41,44–47 Unfortunately, a mass-producible and cost-effective method for the fabrication of uniform nanopores on graphene-based nanosieves is still challenging. Moreover, a few graphene-based nanosieve devices have been reported with their electrical properties thus far5,48–51 but there is no report on the mass-production of graphene-based nanosieves for practical biological applications.

In this work, 5 nm ultrathin reduced graphene oxide (rGO) nanosieves with a dense and well-ordered array of nanosized pores were successfully fabricated by large-area scalable block copolymer (BCP) lithography.52 Polystyrene-block-poly(methyl methacrylate) copolymer (PS-b-PMMA) films were spin cast and self-assembled at the surface of a multilayer stack rGO film. The following pattern transfer of the self-assembled nanoscale morphology onto the underlying rGO layer by reactive ion etching (RIE) enabled the large-scale production of the rGO nanosieve. The resulting rGO nanosieve with well-ordered and dense nanosized pores was suspended in a SiN window (200 μm × 200 μm) and utilized for the highly selective separation of similar-sized proteins.

Experimental section

Preparation of the graphene oxide film

Graphene oxide (GO) was prepared from natural graphite (SP1 Bay Carbon) by a modified Hummer's method. A multilayer GO thin film was spin cast onto a silicon oxide substrate by controlling the GO composition in a methanol/water suspension. For uniform multilayer film deposition, a GO solution with a weight ratio of GO[thin space (1/6-em)]:[thin space (1/6-em)]methanol[thin space (1/6-em)]:[thin space (1/6-em)]water = 1[thin space (1/6-em)]:[thin space (1/6-em)]1750[thin space (1/6-em)]:[thin space (1/6-em)]100 was used. During spin casting, nitrogen gas was blown at the central region of the GO film. A large-area multilayer GO thin film was obtained on a 4′ silicon wafer surface by several repeated spin castings.

Thermal or chemical reduction of graphene

We used chemical and thermal treatments for the reduction of the GO film. Thermal treatment was performed at various temperatures from 200 to 1000 °C at 100 °C intervals. Chemical reduction was carried out using hydrazine monohydrate vapor exposure at various temperatures from 20 to 200 °C and the temperature was changed in steps of 20 °C.

Self-assembled PS-b-PMMA nanotemplate

The PS-b-PMMA thin film was spin-cast onto the rGO film. The BCP with PS and PMMA blocks having a molecular weight 46–21 kg mol−1 was spin-cast onto the surface of the rGO film from a toluene solution. Thermal annealing was conducted at 250 °C for 3–5 h to obtain a self-assembled morphology. The PMMA component was selectively degraded by UV radiation to reduce the resistivity by using the reactive ion etching (RIE) process. The PMMA domains in the BCP film could be selectively removed by dry O2 plasma RIE (50 W, 40 sccm, 100 s). The resulting nanoporous PS template was applied as a lithographic mask for the subsequent pattern transfer to the rGO film on a SiN supporting membrane with an ultra-thin gold film, which was deposited by the e-beam evaporation process (15 nm).

Chemical patterning of graphene with selective RIE

The chemical patterning of graphene was performed via selective oxidation or selective etching using dry O2 RIE (100 W, 40 sccm) with a lithographic and nanoporous PS template mask. The residual BCP nanotemplate was removed by strong sonication in acetone.

Fabrication of the SiN membrane

In order to form an etch mask for bulk-micromachining, a 300 nm thick SiN layer was formed on both sides of a 600 μm-thick double polished 6′ silicon (100) wafer by low-pressure chemical vapor deposition (LPCVD). After the deposition, the SiN layer on the top side of the substrate was patterned by the RIE process for forming 0.5–2 μm-sized pores and the SiN layer on the other side of the substrate was also patterned by the RIE process for opening a backside etching window. As an etch mask for this process, a 1.8 μm thick positive photoresist was patterned on both sides of the silicon substrate by the conventional UV-lithography. Next, the backside of the wafer patterned with SiN was exposed to an anisotropic silicon etchant, KOH at 85 °C until the SiN layer underneath the deposited SiN layer was revealed. Finally, a 15 nm thick gold layer was deposited on the top side of the substrate by using a conventional e-beam evaporator to improve the adhesion with the graphene.

Fluorescence labeling of proteins

An Alexa Fluor®488 Protein Labeling Kit and Alexa Fluor®350 Protein Labeling Kit were acquired from ThermoFisher Inc. Human hemoglobin (Hb, molecular weight (MW) = 64 kDa) and immunoglobulin G (IgG, MW = 160 kDa) from human serum were obtained from Sigma Aldrich Inc.

The Hb and IgG were fluorescence labeled with the Alexa-Fluor®488 dye and Alexa-Fluor®350 dye, according to the instructions provided by the manufacturer. In short, a protein solution at a concentration of 2 mg mL−1 in standard phosphate-buffer saline (PBS) was mixed with 50 μL of 1 M sodium bicarbonate solution and the provided vial of Alexa-Fluor® dye, after which it was reacted for 1 h at room temperature. After the reaction, the protein–dye conjugate was passed through gel separation columns (Bio-Rad BioGel P-30 fine exclusion purification resin, that is, porous polyacrylamide beads), which are designed to separate the free dye from those proteins for which MW > 40 kDa. Thus, the separation of the labeled proteins, Hb with a MW of 64 kDa, and IgG with a MW of 160 kDa, from the unincorporated dye, as well as the successful collection of the labeled proteins using the separation column were possible. The resulting fluorescence-labeled protein (Hb and IgG) solutions were then diluted to a concentration of 8.0 μM and 7.2 μM in PBS, respectively.

Characterization

The nanoscale morphology was characterized using a Hitachi S-4800 scanning electron microscope (SEM) with a field emission source of 1 kV. The advancing water contact angle was measured with a Surface Electro Optics PNX 150. X-ray photoelectron spectroscopy (XPS) data were obtained using a Thermo VG Scientific ESCA 2000. Bright-field TEM (BFTEM) and high-resolution transmission electron microscopy (HRTEM) images were obtained by using a probe-side aberration-corrected transmission electron microscope (JEOL JEM-2100F) operating at an accelerating voltage of 200 kV. The cross-sectional TEM sample was prepared by using a focused ion beam (FIB) (Helios Nano Lab 450). Before the FIB processing, both sides of the weak graphene membrane were supported by carbon ink.

Results and discussion

Fabrication of reduced graphene oxide nanopores

The overall process for the fabrication of a nanoporous graphene membrane exploiting BCP lithography is schematically described in Fig. 1(a). First, a highly pure aqueous dispersion of GO was produced from natural graphite by a modified Hummer's method, which was then purified by dialysis. A multilayer GO film was then directly spin-cast from the aqueous dispersion onto a silicon oxide substrate. Unlike chemical vapor deposition (CVD) or other graphene film preparation methods, this straightforward solution processing of the graphene-based film enables the formation of a wrinkle-free and highly uniform film over a large area. It is noteworthy that this highly flat film morphology is crucial for achieving large-area uniformity during the subsequent BCP lithography process and pattern transfer by RIE. After deposition, the GO thin film was chemically and thermally reduced, resulting in a 5 nm thick rGO film. A PS-b-PMMA thin film was spin-cast onto the surface of the rGO film and thermally annealed at 250 °C for 3–5 h. Since the rGO surface has an identical interfacial tension to both PS and PMMA components, the BCP thin film self-assembled into a vertical nanoscale cylindrical morphology at the surface of the rGO film. This offered a nanopattern template with an ultrathin, large-area and a sharp pattern edge.
image file: c7nr01889d-f1.tif
Fig. 1 Fabrication procedure for nanopores on rGO, and the structural perspective view of the proposed graphene ultra-thin nanosieve without and with the SiN support.

The PMMA domains in the BCP film could be selectively etched by dry O2 plasma RIE (50 W, 40 sccm, 100 s). Nanoscale perforation of the underlying graphene film was performed by pattern transfer of the BCP morphology employing O2 RIE (100 W, 40 sccm). The resulting nanoporous rGO film overlaid with the PS template mask was drifted and transferred to a Si device (1.5 cm × 1.5 cm) with and without a stress-reduced 100 nm thick SiN membrane with 0.5–2 μm wide micropores, which act as a mechanical supporter, as shown in Fig. 1(b) and (c). The residual BCP nanotemplate was removed by ultrasonication in acetone. The morphology of the nanoscale rGO nanosieve was characterized using scanning electron microscopy (SEM) and high resolution transmission electron microscopy and HRTEM.

Schematics for fabrication of the graphene nanosieve on the Si device are shown in Fig. 2(a) and (b). A SiN membrane with 0.5–2 μm wide windows was used as a mechanical support. The detailed process is described in the Experimental section. Briefly, the overall process consists of three stages: the formation of the BCP nanotemplate, nanoscale perforation of the rGO film by RIE, and membrane transfer onto a Si device. This straightforward process allows for low cost production and needs minimal processing time based on the large-area scalable self-assembly principle. The projection diagram of the graphene nanosieve is shown in Fig. 2(c). The graphene nanosieve opens on the SiN window with an area of 200 μm × 200 μm. Our nanotemplate formed by BCP lithography has hexagonally arranged, dense, and perpendicular PMMA cylinders in the PS matrix. The replicated rGO nanosieve replicates the close-packed and dense nanoporous morphology, as shown in Fig. 2(d) (average diameter of graphene nanopores is 15 nm and the center-to-center distance between the neighboring nanopores is 43 nm). The areal density of the graphene nanopores was ∼8 × 1010 cm−2, which is much larger than that of common anodized aluminum oxide (AAO) membranes.53


image file: c7nr01889d-f2.tif
Fig. 2 Fabrication procedure for rGO-based ultra-thin nanosieves with the SiN support (a) and without the SiN support (b), projection diagram of the nanosieves (c), top SEM view (d), and cross-sectional BFTEM and HRTEM images of nanosieves showing the rGO film with a thickness of ∼5 nm (e).

As mentioned above, the separation of similar-sized biomolecules is a challenging task. Commercial filter membranes suffer from low permeability and significant sample loss arising from the non-uniform pore size and quite long channel length with several tens of micrometers. In this regard, an ultra-thin graphene-based membrane with uniformly-sized nanopores is highly advantageous. Fig. 2(e) shows the cross-sectional view of the nanosieve by TEM analysis. A 5 nm-thick nanoporous rGO film suspended over a 2 μm-wide SiN window is shown. The TEM image clearly shows the contrasts between the rGO layer, gold layer, and SiN support. The HRTEM image of the multilayer rGO film confirms a typical graphene layer with a spacing of ∼3.4 Å.

Nanopore size control

A number of processing parameters such as the hydrogen gas flow rate during the thermal reduction or hydrazine monohydrate vapor pressure in the chemical reduction may influence the reduction of the GO film. Above the critical temperatures, the underlying graphene oxide film was sufficiently reduced to have a neutral interfacial tension to PS-b-PMMA BCPs. Several seconds or minutes of reduction were sufficient to induce the nanodomains perpendicular to the surface. Nanodomains perpendicular to the rGO surface were achieved over the entire BCP film at temperatures above 600 °C and 120 °C for thermal and chemical treatments, respectively. The average diameter of the nanopores could be controlled from 10 to 30 nm with a standard deviation ranging from 1.5 to 2.5 nm (Fig. S1).

Few nanometers’ scale of nanopore size can be controlled by etching process conditions such as power, flow rate and etching time after the formation of a self-assembled nanostructure. Here, we have mainly controlled the etching time. The molecular weight of BCPs can change not only the nanopore size dramatically but also the center-to-center distance.

Here, the center-to-center distance between the neighboring nanopores was about 43 nm. The open porous area of the nanosieve could be controlled to be as high as 40%, compared to that of the AAO with a porosity of 10% commonly,53 which is greatly desired for a high throughput separation. The photographs of the fabricated nanosieve device, combined with the CMOS-compatible semiconductor processes, are shown in Fig. S2.

Protein separation

To evaluate the possibility of practical use of the nanosieve, we carried out a simple separation experiment as developed by Striemer et al.33 In a typical experiment, 5 nm-thick nanosieves with an area of 200 μm × 200 μm were tested with and without a supporting SiN membrane. Since the pores at the top of the nanosieve were less than 15 nm in diameter, the nanosieve was expected to selectively allow the permeation of molecules a few nanometers in size. Fluorescent dyes (free Alexa Fluor® 488, molecular weight (MW) = 643, free Alex Fluor® 350, MW = 618, hydrodynamic diameter (D) ∼ 1 nm) and two proteins were used in this experiment: Alexa Fluor® 350-labeled human Immunoglobulin-G [IgG, MW = 160[thin space (1/6-em)]000, D ∼ 15 nm] and Alexa Fluor® 488-labeled hemoglobin [Hb, MW = 64[thin space (1/6-em)]000, D ∼ 8 nm].

Since it is generally known that reduced graphene oxide is highly hydrophobic in nature, there is a need to produce a nanopore membrane with hydrophilic qualities for water molecule wetting.52 We achieved this by applying three oxygen plasma treatment steps, as follows. We employed oxygen plasma reactive ionic etching (50 W, 40 sccm, 100 s) to etch the PMMA, followed by oxygen plasma reactive ionic etching (100 W, 40 sccm) to perforate the graphene oxide film, and finally oxygen plasma treatment (50 W, 50 sccm, 30 s) before performing the protein separation test. We believe that the oxygen plasma treatment produced an rGO surface with at least a degree of hydrophilicity. This is because the oxygen plasma changes some of the reduced graphene oxide into graphite oxide, a layered material consisting of hydrophilic oxygenated graphene sheets (graphene oxide sheets) with oxygen functional groups on their basal planes and edges.54,55 The water dispersity and hydrophilicity of GO have mainly been attributed to the ionizable edge –COOH groups. Moreover, the basal plane contains a network of hydrophobic polyaromatic islands of unoxidized benzene rings, or sp2-subdomains.56,57 Therefore, the GO should be regarded as being an amphiphile with a largely hydrophobic basal plane and hydrophilic edges.

Fig. 3 is a schematic diagram of the permeation test. Hb was labelled with two moles of dye per mole of protein and IgG was labelled with six moles of dye per mole of protein. The nanosieve was placed on a 100 μm thick spacer mounted on a Petri dish. The space between the nanosieve and the bottom of the Petri dish was filled with 0.7% (w/v) agarose hydrogel in 50× Tris-acetate-EDTA (TAE) buffer solution. Then, 5 μL of the mixture containing the fluorescent dye, Alexa Fluor® 350-labeled IgG and Alexa Fluor® 488-labeled Hb was dropped onto the nanosieve. Finally, the permeation of the fluorescent dye and fluorophore-labeled proteins through the nanosieve was monitored by using a fluorescence microscope. Since Alexa Fluor® 488, and Alexa Fluor® 350 have different fluorescence emission properties, we could easily make out which among these permeated and which did not. The fluorescence signal emitted in the Hb permeation test by free Alexa Fluor® 488 was measured using a fluorescence microscope to compare the performances of the graphene nanosieve (pore size ∼15 nm, and thickness ∼5 nm) with and without the SiN support (open rate = 10%) and the commercialized nanosieve (Millipore™, pore size ∼100 nm and thickness ∼30 μm) using 100 μg ml−1 Alexa Fluor® 488-labeled hemoglobin with a diameter of ∼8 nm in 0.1 M PBS containing 0.05% Tween.


image file: c7nr01889d-f3.tif
Fig. 3 Schematic diagram of the rGO nanosieve device for the protein separation test.

Fig. 4(a) shows the quantitative results of the permeation test using our graphene nanosieve and commercially available nanosieve. Fig. 4 indicates the integrated fluorescence intensities obtained from the measured areas in Fig. 3. For the fluorescence intensity variation of Hb with time, the increasing rate on the commercial nanosieve was quite slow despite the large pore size of 100 nm and the thickness of 30 μm. The increasing rate of fluorescence intensity in the graphene nanopores was considerably faster. It is obvious that our graphene nanosieve shows much faster permeation than commercialized nanosieves. Besides, the increase in the fluorescence intensity was about 5–6 times higher when the SiN supporter was supplementarily removed. The open porosity of the graphene nanosieve without the SiN support (about 40%) was much larger than that of the nanosieve with SiN (about 10%). The ultra-thin graphene nanosieve with refined pore size control and high open porosity could facilitate the high throughput of protein penetration. The clogging (or binding) of proteins in the membrane surface or nanosieves is an important issue. A major parameter for preventing the fouling or binding of proteins on substrates is the control of the surface properties of the membrane, given that the strong anti-clogging or antifouling effect of the superhydrophilic or hydrophilic surface of proteins is well known. Therefore, the oxygen plasma pre-treatment of a membrane could reduce the anti-clogging or anti-fouling effect in microfluidic devices. It was previously reported that clogging or blocking by particles occurs naturally in cases where the particle size is larger than 10% of the smallest dimension in the system.58 However, an increase of the porosity from 10% to 40% and an extremely thin membrane thickness of 5 nm can produce a very high flux, such that we could expect to attain a reduction in the blocking or clogging in the nanopores. Therefore, the oxygen plasma treatment to produce a hydrophilic surface and the use of an ultra-thin nanomembrane with precise porosity control allow selective and efficient protein separation without the sticking or clogging of proteins.


image file: c7nr01889d-f4.tif
Fig. 4 Hb permeation test, (a) comparison between the graphene nanosieve device (pore size ∼15 nm, and thickness ∼5 nm) with the SiN support (open rate = 10%, filled squares) and that without the support (open rate = 40%, filled triangles), and the commercialized nanosieve (Millipore™, pore size ∼100 nm, and thickness ∼30 μm) (filled circles) using 100 μg ml−1 Fluor® 488-labeled Hb with a diameter of ∼8 nm in 50× TAE buffer solution.

Fig. 5 and ESI video show the optical images of the permeation test in a mixture of free Alexa Fluor 488 dye, Hb and IgG using the graphene nanosieve without the SiN support taken by using a fluorescence microscope. Also, optical images of the permeation test and fluorescence intensity are shown in Fig. S3 and S4, as a function of penetration time with the SiN membrane support. There is a little fluorescence increase of the labeled IgG at the center of the supporting membrane, in the case of existence of the optically transparent SiN membranes. It seems that the Alexa Fluor® 350 exciting light usually gives rise to signal interference, such as scattering or reflections, as observed by the CCD detectors. Since graphene is ultra-thin and transparent, the transmitting fluorescence is observed over the entire window area. The fluorescence intensity is measured at the edge of the open window. While the fluorescence emitted by the free Alexa Fluor 488 dye gradually became stronger, the fluorescence from IgG protein decreased (ESI Videos 1 and 2). The blue intensity decreases with time, for which there would appear to be two causes. First, the stronger photo-bleaching due to the extended exposure of the Alexa Fluor® 350 dyes to the excitation light, which is typically UV rays with a wavelength of up to 346 nm and thus a higher excitation energy could decrease the blue intensity over time. For Alexa Fluor® 488 green emission, excitation light with 490 nm visible rays with a lower energy is utilized. Thus, the photo-bleaching of Alexa Fluor® 350 blue seems to be much greater than that of Alexa Fluor® 488 green. In contrast, the Hb proteins with Alexa Fluor® 488 dyes completely penetrate through and diffuse through the graphene nanopores, such that there is an increase in the green intensity, even though there is some extinction caused by photo-bleaching. The photobleaching could be reduced by limiting the exposure of the fluorochrome to intense illumination (using neutral-density filters) coupled with the appropriate use of commercially available antifade reagents that can be added to the mounting solution.59 It was also observed that an air bubble was accidentally trapped between the graphene film and the agarose hydrogel and gradually moved down (ESI Video 3). This indicates that our nanosieve is mechanically strong and resilient enough to endure the bubble stimulations.


image file: c7nr01889d-f5.tif
Fig. 5 Permeation test with the mixture of Hb and IgG, (a) fluorescence photographs of Alexa Fluor® 488-labeled Hb (100 μg ml−1) and Alexa Fluor® 350-labeled IgG (100 μg ml−1) permeated through the graphene nanosieve as a function of permeation time and (b) kinetics of the permeation process for Alexa 488 Fluor dye-labeled Hb and Alexa Fluor® 350-labeled IgG through the nanosieve in accordance as a function of permeation time along with the line from A to A′ (the y-axis indicates the value obtained by the integration of the fluorescence intensities measured from A to A′). (c) Fluorescence intensities of Alexa Fluor® 488-labeled Hb and Alexa Fluor® 350-labeled IgG as a function of permeation time.

Fig. 5(b) describes the distribution of the permeated dye-labeled Hb and IgG as a function of permeation time along the line from A to A′. The dye-labeled Hb and IgG permeated through the entire area of the nanosieve uniformly and continuously without the occurrence of any mechanical crack or breakdown. Fig. 5(c) shows the integrated fluorescence intensities of Hb and IgG as a function of permeation time. While the fluorescence intensity of Hb increased rapidly, that of IgG decreased for 20 min. Obviously, Hb permeated through the graphene nanosieve, whereas the IgG protein was perfectly blocked because of its precipitation, as indicated by the darkened areas in the optical images. Thus, based on the experimental observations we can conclude that the fabricated nanosieve has a highly selective permeation for the two similar-sized proteins. Selectivity and permeation speed are two critical properties for membrane performance. The thickness, porosity, and pore size are key parameters affecting the properties. The thickness and porosity of the membrane affect the flux passing through the membranes. The flux passing through the membranes can be enhanced by increasing the pore diameter and reducing the thickness. The thickness and pore size ratio of the commercial membrane to the nanopore membrane are 6000 and 6, respectively. Thus, we believe that the extremely low thickness affects the permeation speed required to enable a high throughput. However, the control of the pore diameter is also important for selective transport. Thus, the extreme reduction of the thickness with precise pore size control allows selective and efficient protein separation without restricting the flow. In this regard, our suspended rGO nanosieve is expected to offer high throughput separations for many different nanoscale materials and molecules. The mechanical stability of the graphene nanosieve is a very important factor for practical applications. To test the mechanical stability of the graphene nanomembrane suspended over an open hole, atomic force microscopy (AFM) nanoindentation has been employed.60,61 The excellent mechanical properties of the monolayer graphene oxide (an ultrahigh Young's modulus of 207.6 GPa and a prestress of 76.8 MPa) were reported.60 The testing of the mechanical properties of the multiple-layer graphene oxide nanopore membrane would be possible by employing the AFM nanoindentation method. Additionally, the extremely thin graphene layer is nearly invisible and fragile to physical shock imposed by a pointed instrument such as a metal tweezer, and therefore requires handling with extreme care.

Diffusion of molecules from regions of higher concentration to regions of lower concentration is a dominant mechanism that the molecules in this work permeate through the nanosieve, since no pressure was applied across the nanosieve. In order to investigate the diffusion profile of Alexa 488 dye in water, a numerical simulation was performed using CFD-ACE+, which is a general and commercialized computational fluidic dynamics and multi-physics solver. The diffusion coefficient of the Alexa Fluor® 488 was set up as 430 μm2 s−1, and an isotropic diffusion model was employed. Fig. S5 shows the simulation results. The distribution of the permeated Alexa Fluor® 488 with regard to the permeation time shows similar shapes to that in Fig. 5(b).

Conclusions

We have presented a straightforward, facile and robust method based on BCP lithography for the fabrication of an ultra-thin rGO nanosieve with 15 nm wide vertical straight pores and an open areal ratio approaching 40%. PS-b-PMMA was spin cast onto the rGO film and self-assembled into a nanopattern template. Pattern transfer by RIE effectively replicated the BCP nanotemplate morphology onto the underlying rGO layer. Unlike previous approaches employing serial pore formation processes such as e-beam lithography, our large-area scalable fabrication exploiting the parallel molecular self-assembly principle offered a viable nanomembrane technology for selective and high throughput separation. The pore size was controllable over a wide range of 10–30 nm. The open areal ratio was as high as 40%, and the areal density of nanopores was about ∼8 × 1010 cm−2. The nanosieve could be fabricated in a large area compared with its 5 nm-ultrathin thickness with an area of 200 μm × 200 μm (an ultra-high aspect ratio of ∼40[thin space (1/6-em)]000) and could endure a mechanical pressure up to 1 atm and effectively separated Hb from a mixture of Hb and IgG in a very rapid and selective manner. Taking advantage of large-area scalable BCP based nanofabrication, our nanoporous membrane should be broadly useful as selective protein transport for various kinds of proteins by controlling the nanopore size. Furthermore, it can be useful for various medical devices, fuel cells, and water purification systems.

Acknowledgements

S.-Y. C. acknowledges the support from the Creative Research Program of ETRI (17ZB1310), the Global Frontier Research Center for Advanced Soft Electronics (CASE 2011-0031640), and the Creative Materials Discovery Program (NRF-2016M3D1A1900035). D.-S. L. acknowledges the support from the basic research program of ETRI (10RC1110) and the National Research Foundation of Korea under research projects (NRF-2017M3A9F1033056). S. O. K. acknowledges support from the Multi-Dimensional Directed Nanoscale Assembly Creative Research Initiative Center (2015R1A3A2033061).

Notes and references

  1. S. Park and R. S. Ruoff, Nat. Nanotechnol., 2009, 4(4), 217–224 CrossRef CAS PubMed.
  2. C. Lee, X. Wei, J. W. Kysar and J. Hone, Science, 2008, 321(5887), 385–388 CrossRef CAS PubMed.
  3. I. W. Frank, D. M. Tanenbaum, A. M. Van der Zande and P. L. McEuen, J. Vac. Sci. Technol., B, 2007, 25(6), 2558–2561 CAS.
  4. C. Gomez-Navarro, M. Burghard and K. Kern, Nano Lett., 2008, 8(7), 2045–2049 CrossRef CAS PubMed.
  5. S. Garaj, W. Hubbard, A. Reina, J. Kong, D. Branton and J. A. Golovchenko, Nature, 2010, 467(7312), 190–U73 CrossRef CAS PubMed.
  6. Z. S. Siwy and M. Davenport, Nat. Nanotechnol., 2010, 5(10), 697–698 CrossRef CAS PubMed.
  7. G. F. Schneider, S. W. Kowalczyk, V. E. Calado, G. Pandraud, H. W. Zandbergen, L. M. K. Vandersypen and C. Dekker, Nano Lett., 2010, 10(8), 3163–3167 CrossRef CAS PubMed.
  8. S. P. Koenig, L. D. Wang, J. Pellegrino and J. S. Bunch, Nat. Nanotechnol., 2012, 7(11), 728–732 CrossRef CAS PubMed.
  9. P. Z. Sun, M. Zhu, K. L. Wang, M. L. Zhong, J. Q. Wei, D. H. Wu, Z. P. Xu and H. W. Zhu, ACS Nano, 2013, 7(1), 428–437 CrossRef CAS PubMed.
  10. H. W. Kim, H. W. Yoon, S. M. Yoon, B. M. Yoo, B. K. Ahn, Y. H. Cho, H. J. Shin, H. Yang, U. Paik, S. Kwon, J. Y. Choi and H. B. Park, Science, 2013, 342(6154), 91–95 CrossRef CAS PubMed.
  11. D. E. Jiang, V. R. Cooper and S. Dai, Nano Lett., 2009, 9(12), 4019–4024 CrossRef CAS PubMed.
  12. E. N. Wang and R. Karnik, Nat. Nanotechnol., 2012, 7(9), 552–554 CrossRef CAS PubMed.
  13. D. Cohen-Tanugi and J. C. Grossman, Nano Lett., 2012, 12(7), 3602–3608 CrossRef CAS PubMed.
  14. S. Liu, Q. Zhao, J. Xu, K. Yan, H. L. Peng, F. H. Yang, L. P. You and D. P. Yu, Nanotechnology, 2012, 23(8) CrossRef PubMed , 085301.
  15. C. A. Merchant, K. Healy, M. Wanunu, V. Ray, N. Peterman, J. Bartel, M. D. Fischbein, K. Venta, Z. T. Luo, A. T. C. Johnson and M. Drndic, Nano Lett., 2010, 10(8), 2915–2921 CrossRef CAS PubMed.
  16. I. Vlassiouk, P. Y. Apel, S. N. Dmitriev, K. Healy and Z. S. Siwy, Proc. Natl. Acad. Sci. U. S. A., 2009, 106(50), 21039–21044 CrossRef CAS PubMed.
  17. S. Metz, C. Trautmann, A. Bertsch and P. Renaud, J. Micromech. Microeng., 2004, 14(3), 324–331 CrossRef CAS.
  18. P. Y. Apel, P. Ramirez, I. V. Blonskaya, O. L. Orelovitch and B. A. Sartowska, Phys. Chem. Chem. Phys., 2014, 16(29), 15214–15223 RSC.
  19. A. Eguizabal, M. Sgroi, D. Pullini, E. Ferain and M. P. Pina, J. Membr. Sci., 2014, 454, 243–252 CrossRef CAS.
  20. A. Thormann, N. Teuscher, M. Pfannmoeller, U. Rothe and A. Heilmann, Small, 2007, 3(6), 1032–1040 CrossRef CAS PubMed.
  21. H. U. Osmanbeyoglu, T. B. Hur and H. K. Kim, J. Membr. Sci., 2009, 343(1–2), 1–6 CrossRef CAS.
  22. H. Masuda, H. Yamada, M. Satoh, H. Asoh, M. Nakao and T. Tamamura, Appl. Phys. Lett., 1997, 71(19), 2770–2772 CrossRef CAS.
  23. K. B. Jirage, J. C. Hulteen and C. R. Martin, Science, 1997, 278(5338), 655–658 CrossRef CAS.
  24. C. R. Martin, M. Nishizawa, K. Jirage, M. S. Kang and S. B. Lee, Adv. Mater., 2001, 13(18), 1351–1362 CrossRef CAS.
  25. M. J. Kim, M. Wanunu, D. C. Bell and A. Meller, Adv. Mater., 2006, 18(23), 3149–3153 CrossRef CAS.
  26. M.-Y. Wu, R. M. M. Smeets, M. Zandbergen, U. Ziese, D. Krapf, P. E. Batson, N. H. Dekker, C. Dekker and H. W. Zandbergen, Nano Lett., 2009, 9(1), 479–484 CrossRef CAS PubMed.
  27. A. J. Storm, J. H. Chen, X. S. Ling, H. W. Zandbergen and C. Dekker, Nat. Mater., 2003, 2(8), 537–540 CrossRef CAS PubMed.
  28. H. D. Tong, H. V. Jansen, V. J. Gadgil, C. G. Bostan, E. Berenschot, C. J. M. van Rijn and M. Elwenspoek, Nano Lett., 2004, 4(2), 283–287 CrossRef CAS.
  29. J. K. Holt, H. G. Park, Y. M. Wang, M. Stadermann, A. B. Artyukhin, C. P. Grigoropoulos, A. Noy and O. Bakajin, Science, 2006, 312(5776), 1034–1037 CrossRef CAS PubMed.
  30. S. Y. Yang, I. Ryu, H. Y. Kim, J. K. Kim, S. K. Jang and T. P. Russell, Adv. Mater., 2006, 18(6), 709–712 CrossRef CAS.
  31. S. Y. Yang, J.-A. Yang, E.-S. Kim, G. Jeon, E. J. Oh, K. Y. Choi, S. K. Hahn and J. K. Kim, ACS Nano, 2010, 4(7), 3817–3822 CrossRef CAS PubMed.
  32. W. Li, N. A. W. Bell, S. Hernandez-Ainsa, V. V. Thacker, A. M. Thackray, R. Bujdoso and U. F. Keyser, ACS Nano, 2013, 7(5), 4129–4134 CrossRef CAS PubMed.
  33. C. C. Striemer, T. R. Gaborski, J. L. McGrath and P. M. Fauchet, Nature, 2007, 445(7129), 749–753 CrossRef CAS PubMed.
  34. T. R. Gaborski, J. L. Snyder, C. C. Striemer, D. Z. Fang, M. Hoffman, P. M. Fauchet and J. L. McGrath, ACS Nano, 2010, 4(11), 6973–6981 CrossRef CAS PubMed.
  35. D. Z. Fang, C. C. Striemer, T. R. Gaborski, J. L. McGrath and P. M. Fauchet, Nano Lett., 2010, 10(10), 3904–3908 CrossRef CAS PubMed.
  36. T. A. Desai, D. J. Hansford, L. Leoni, M. Essenpreis and M. Ferrari, Biosens. Bioelectron., 2000, 15(9–10), 453–462 CrossRef CAS PubMed.
  37. W. H. Fissell, A. Dubnisheva, A. N. Eldridge, A. J. Fleischman, A. L. Zydney and S. Roy, J. Membr. Sci., 2009, 326(1), 58–63 CrossRef CAS PubMed.
  38. X. Peng, J. Jin, Y. Nakamura, T. Ohno and I. Ichinose, Nat. Nanotechnol., 2009, 4(6), 353–357 CrossRef CAS PubMed.
  39. C. Acikgoz, X. Y. Ling, I. Y. Phang, M. A. Hempenius, D. N. Reinhoudt, J. Huskens and G. J. Vancso, Adv. Mater., 2009, 21(20), 2064–2067 CrossRef CAS.
  40. D.-H. Choi, Y. D. Han, B.-K. Lee, S.-J. Choi, H. C. Yoon, D.-S. Lee and J.-B. Yoon, Adv. Mater., 2012, 24(32), 4408–4413 CrossRef CAS PubMed.
  41. D.-S. Lee, H.-W. Song, C.-G. Choi and M. Y. Jung, J. Biomed. Opt., 2014, 19(5) Search PubMed , 051211.
  42. B. H. Kim, D. O. Shin, S.-J. Jeong, C. M. Koo, S. C. Jeon, W. J. Hwang, S. Lee, M. G. Lee and S. O. Kim, Adv. Mater., 2008, 20(12), 2303–2307 CrossRef CAS.
  43. P. Mansky, Y. Liu, E. Huang, T. P. Russell and C. J. Hawker, Science, 1997, 275(5305), 1458–1460 CrossRef CAS.
  44. B. M. Venkatesan, D. Estrada, S. Banerjee, X. Jin, V. E. Dorgan, M.-H. Bae, N. R. Aluru, E. Pop and R. Bashir, ACS Nano, 2012, 6(1), 441–450 CrossRef CAS PubMed.
  45. M. D. Fischbein and M. Drndic, Appl. Phys. Lett., 2008, 93(11) CrossRef CAS , 113107.
  46. B. M. Venkatesan and R. Bashir, Nat. Nanotechnol., 2011, 6(10), 615–624 CrossRef CAS PubMed.
  47. Y. Liu, X. Dong and P. Chen, Chem. Soc. Rev., 2012, 41(6), 2283–2307 RSC.
  48. J. Bai, R. Cheng, F. Xiu, L. Liao, M. Wang, A. Shailos, K. L. Wang, Y. Huang and X. Duan, Nat. Nanotechnol., 2010, 5(9), 655–659 CrossRef CAS PubMed.
  49. M. Kim, N. S. Safron, E. Han, M. S. Arnold and P. Gopalan, Nano Lett., 2010, 10(4), 1125–1131 CrossRef CAS PubMed.
  50. X. Liang, Y.-S. Jung, S. Wu, A. Ismach, D. L. Olynick, S. Cabrini and J. Bokor, Nano Lett., 2010, 10(7), 2454–2460 CrossRef CAS PubMed.
  51. D. H. Lee, J. E. Kim, T. H. Han, J. W. Hwang, S. Jeon, S.-Y. Choi, S. H. Hong, W. J. Lee, R. S. Ruoff and S. O. Kim, Adv. Mater., 2010, 22(11), 1247–1252 CrossRef CAS PubMed.
  52. B. H. Kim, J. Y. Kim, S.-J. Jeong, J. O. Hwang, D. H. Lee, D. O. Shin, S.-Y. Choi and S. O. Kim, ACS Nano, 2010, 4(9), 5464–5470 CrossRef CAS PubMed.
  53. K. Nielsch, J. Choi, K. Shwirn, R. B. Wehrspohn and U. Gösele, Nano Lett., 2002, 2(7), 677–680 CrossRef CAS.
  54. D. R. Dredyer, S. Park, C. W. Bielawski and R. S. Ruoff, Chem. Soc. Rev., 2010, 39, 228–240 RSC.
  55. Y. Wang, Z. Li, J. Wang, J. Li and Y. Lin, Trends Biotechnol., 2011, 29(5), 205–212 CrossRef CAS PubMed.
  56. X. Liu, L. Cao, W. Song, K. Ai and L. Lu, ACS Appl. Mater. Interfaces, 2011, 3, 2944–2952 CAS.
  57. H. Lee, N. Son, H. Y. Jeong, T. G. Kim, G. S. Bang, J. Y. Kim, G. W. Shim, K. C. Goddeti, J. H. Kim, N. Kim, H.-J. Shin, W. Kim, S. Kim, S.-Y. Choi and J. Y. Park, Nanoscale, 2016, 8, 4063–4069 RSC.
  58. C. Wiles and P. Watts, Chem. Commun., 2011, 47, 6512–6535 RSC.
  59. C. Boudreau, T.-L. Wee, Y.-R. Duh, M. P. Couto, K. H. Aedakani and C. M. Brown, Sci. Rep., 2016, 6, 30892 CrossRef CAS PubMed.
  60. J. W. Suk, R. D. Piner, J. An and R. S. Ruoff, ACS Nano, 2010, 4(11), 6557–6564 CrossRef CAS PubMed.
  61. C. Lee, X. Wei, J. W. Kysar and J. Hone, Science, 2008, 321(385), 385–388 CrossRef CAS PubMed.

Footnote

Electronic supplementary information (ESI) available. See DOI: 10.1039/c7nr01889d

This journal is © The Royal Society of Chemistry 2017