A Scalable Route to Nanoporous Large-Area Atomically Thin

2 days ago - An ideal roll-to-roll graphene CVD reactor would allow for (i) heating of the catalyst (e.g., Cu foil) in a reducing environment to clean...
0 downloads 11 Views 8MB Size
Download PDF
Research Article Cite This: ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

www.acsami.org

A Scalable Route to Nanoporous Large-Area Atomically Thin Graphene Membranes by Roll-to-Roll Chemical Vapor Deposition and Polymer Support Casting Piran R. Kidambi,*,†,‡ Dhanushkodi D. Mariappan,† Nicholas T. Dee,† Andrey Vyatskikh,†,§ Sui Zhang,†,∥ Rohit Karnik,† and A. John Hart*,† †

Department of Mechanical Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, United States Department of Chemical and Biomolecular Engineering, Vanderbilt University, Nashville, Tennessee 37235-1826, United States § Division of Engineering and Applied Sciences, California Institute of Technology, Pasadena, California 91125, United States ∥ Department of Chemical and Biomolecular Engineering, National University of Singapore, Singapore 117586 ‡

S Supporting Information *

ABSTRACT: Scalable, cost-effective synthesis and integration of graphene is imperative to realize large-area applications such as nanoporous atomically thin membranes (NATMs). Here, we report a scalable route to the production of NATMs via high-speed, continuous synthesis of large-area graphene by roll-to-roll chemical vapor deposition (CVD), combined with casting of a hierarchically porous polymer support. To begin, we designed and built a two zone roll-toroll graphene CVD reactor, which sequentially exposes the moving foil substrate to annealing and growth atmospheres, with a sharp, isothermal transition between the zones. The configurational flexibility of the reactor design allows for a detailed evaluation of key parameters affecting graphene quality and trade-offs to be considered for high-rate roll-to-roll graphene manufacturing. With this system, we achieve synthesis of uniform high-quality monolayer graphene (ID/IG < 0.065) at speeds ≥5 cm/min. NATMs fabricated from the optimized graphene, via polymer casting and postprocessing, show size-selective molecular transport with performance comparable to that of membranes made from conventionally synthesized graphene. Therefore, this work establishes the feasibility of a scalable manufacturing process of NATMs, for applications including protein desalting and small-molecule separations. KEYWORDS: graphene membranes, roll-to-roll graphene synthesis, chemical vapor deposition, nanoporous atomically thin membranes, polymer support casting

■

process for membrane and barrier applications,10 these studies have used small, fixed substrate reactors whereas commercial filtration membranes typically require surface areas of several square meters or more. The need for such large areas necessitates the development of continuous (e.g., roll-to-roll) graphene CVD methods. However, prior studies of roll-to-roll CVD graphene synthesis using various reactor designs,15−21 and roll-to-roll transfer methods,13,22 have focused primarily on characterization of electrical properties. Importantly, the quality requirements of graphene for membrane applications tend to be significantly different from electronics; for example, subnanometer- to nanometer-sized pinhole defects would remain relatively un-noticed in micrometer-sized electronic devices but could adversely impact membrane applications by allowing for leakage pathways with much higher flows than those from

INTRODUCTION Graphene, a single-atom-thick membrane of hexagonally bonded carbon atoms, is impervious (when defect-free) to even the smallest molecule, helium.1 Its unique combination of exceptional mechanical strength2,3 along with the opportunity for tunable nanometer-scale pore creation in the atomically thin hexagonal lattice4 makes graphene a promising material for improving the permeability, selectivity, and energy efficiency of membranes. Applications of these membranes include desalination,5 dialysis,6 fuel cells,7 and isotope separation,8 among others.9 Scalable, cost-effective graphene synthesis and integration routes are central to enabling commercialization of nanoporous atomically thin membranes (NATMs). Among graphene production methods, chemical vapor deposition (CVD) of graphene on polycrystalline Cu foil has become the preferred approach for high-quality monolayer graphene synthesis,10−12 albeit primarily for applications in electronics.13,14 While some recent studies have focused on optimizing the graphene CVD © XXXX American Chemical Society

Received: January 16, 2018 Accepted: March 6, 2018

A

DOI: 10.1021/acsami.8b00846 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 1. Roll-to-roll CVD reactor design and construction. (a) Overview of scalable processing route for NATM synthesis from CVD graphene. (b) Schematic diagram of the roll-to-roll graphene CVD reactor with distinct annealing zone and growth zone. Cu foil loaded on the roller in the input chamber is processed through the reactor and wound at the roller in the output chamber. (c) Optical images of the roll-to-roll graphene CVD reactor. The elevated small furnace seen in C is not part of the system. Schematic of (d) side view, (e) cross-sectional view, and (f) optical image of the transition from annealing to growth zone for the CVD reactor.

selective pores.10 Further, more detailed understanding of the parameter space of graphene CVD and key trade-offs specific to roll-to-roll implementation15−21 is necessary for scale-up and beneficial to all applications. An ideal roll-to-roll graphene CVD reactor would allow for (i) heating of the catalyst (e.g., Cu foil) in a reducing environment to clean the surface of oxides and impurities while allowing for catalyst grain growth, (ii) exposure to hydrocarbons in a reducing environment to induce graphene nucleation and subsequent domain growth, and finally (iii) appropriate handling of the processed foil to prevent damage to the graphene.15−21 The design requirements on the CVD reactor in turn follow directly from the process stages, that is, uniformity of temperature and gas-phase concentrations in the annealing and growth stages, adequate reactor length (residence time) to facilitate uniform graphene growth with full coverage, spatially confined heating (while annealing) and cooling (after CVD) of the Cu foil, controlled gas composition and pressure with isolation from ambient, and minimum moving parts within the reactor to simplify operations and maintenance. After graphene synthesis, the fabrication of NATMs requires (i) clean transfer of the synthesized graphene to an appropriate porous support with minimal contamination,6,9,10,23−26 (ii)

leakage sealing to minimize nonselective transport through tears introduced from graphene transfer and processing,6,9,26,27 and finally (iii) the creation of a narrow size distribution of nanopores in the leakage-sealed graphene2,5,6,9,27 to achieve molecular-size-based selectivity. Traditional membrane fabrication approaches, for example, polymer casting, phase inversion, and interfacial polymerization, used extensively for scalable manufacturing of roll-to-roll flat sheet reverse osmosis membranes, can indeed be effectively leveraged to address some of these challenges and aid the development of NATMs.26,28 Here, we demonstrate the feasibility of a complete end-toend scalable manufacturing approach for producing NATMs by combining optimized roll-to-roll CVD graphene synthesis with casting of a porous polymer support (summarized in Figure 1a). We use computational fluid dynamic (CFD) simulations to guide the design and construction of a roll-to-roll CVD reactor with split zones and demonstrate synthesis of uniform highquality monolayer graphene (ID/IG < 0.065) on Cu foil at speeds ≥5 cm/min. With use of the optimized graphene, we fabricate large-area (>3 cm2) NATMs by (i) solution casting ∼50 μm thick hierarchically porous polyether sulfone (PES) supports, (ii) leakage sealing using interfacial polymerization, and (iii) nanopore creation by pulsed oxygen plasma etching. B

DOI: 10.1021/acsami.8b00846 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 2. Computational fluid dynamics (CFD) design of the roll-to-roll graphene CVD reactor. (a) Schematic diagram of the CVD reactor. The plot of temperature (T) as a function of position (z) shows the heated zone of the reactor enclosed by the furnace (see Figure 1b). (b) Color-coded concentration profile maps for CH4 in the transition region from the annealing zone to the growth zone. The uniform colors indicate rapid mixing. Changes in (c) concentration of CH4 (orange) and H2 (blue), (d) gas velocity, (e) temperature, and (f) pressure as a function of the length of the heated zone of the reactor as seen by the moving Cu foil. Note a pressure of 4 Torr at the outlet and a slit size of 0.7−1.3 mm was used for CFD simulations. The slit size was increased to 5 mm for actual experiments.

a rectangular quartz sleeve centered within a cylindrical quartz tube, which is enclosed in a hot-wall tube furnace; this tube-intube geometry allows the foil to be heated in a first atmosphere (e.g., for annealing) and then make an isothermal transition to the growth atmosphere as it exits the downstream end of the sleeve. The Cu foil loaded on the roller in the input chamber passes into the rectangular sleeve which is held along the centerline of the surrounding cylindrical tube. Within the rectangular quartz sleeve, the foil slides over a flat quartz support which extends from the exit of the rectangular tube (growth zone, 30 cm long) to the end of the cylindrical tube. At the exit of the cylindrical tube, the foil enters the output chamber and winds on to the roller in the output chamber. A stepper motor in the output chamber drives the winding roller, pulling the foil through the system; there are no mechanisms within the tube assembly, ensuring minimal moving parts in the

The NATMs made using roll-to-roll CVD graphene show sizeselective molecular transport for diffusion-driven flow, with performance comparable to recently reported NATMs using CVD graphene synthesized via batch processing.6

■

RESULTS AND DISCUSSION

Roll-to-Roll CVD Reactor Design and Construction. To enable continuous synthesis of graphene, we designed and built a roll-to-roll CVD system with a tube-in-tube design (Figure 1, S1 of the Supporting Information) to allow the Cu foil to first be annealed in a reducing atmosphere and then rapidly transition into the hydrocarbon atmosphere for graphene growth. Such a design allows us to effectively leverage progress made in graphene CVD using batch-style furnaces with tube geometries,10−21 and explore the capabilities of a continuous feed process. Therefore, our roll-to-roll CVD reactor consists of C

DOI: 10.1021/acsami.8b00846 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 3. Optimizing CVD conditions for full graphene coverage. SEM images of the synthesized CVD graphene on Cu at 5 cm/min at 1000 °C as a function of CH4/H2 ratio: (a) 0.17, (b) 0.67, and (c) 1. Note 150 sccm H2 was introduced into the annealing zone while different flow rates of CH4 were introduced into the growth zone. (d) Low-magnification image (a) showing preferential graphene growth within certain crystallographic orientations of Cu. (e) Representative Raman spectrum for graphene synthesized for conditions shown in (b). (f) Coverage computed from SEM images using image thresholding in ImageJ software as a function of CH4/H2 ratio (g, h) Cracks and discontinuities in CVD graphene on Cu are formed when the graphene is on the outer edge of winding (also see Figure S3). (g,i) No cracks are seen when the direction of winding is changed and graphene is on the inner edge of the winding.

reactor.15 The foil is installed by opening the top covers of the end chambers, which are made using standard-size Klein Flange (KF) vacuum connections. The reactor is mounted on rails that allow for easy assembly and disassembly of the tubes by sliding the end chambers. The moving foil sequentially receives exposure to the annealing (H2) and growth (CH4, H2) gas mixtures, while continuously translating through the furnace. The annealing gases enter the reactor through the input chamber, proceed into the annealing zone (50 cm heated length) through a custom designed KF flange (Figure S1b,c), enter the growth zone (30 cm heated length) through the 5 mm slit in the downstream end of the rectangular quartz, and are exhausted to the vacuum pump which is connected to the end chamber. The gases for growth bypass the annealing zone by flowing within the gap between the rectangular sleeve and the cylindrical tube, and only contact the Cu foil after it exits the sleeve. The simple yet versatile reactor design allows for operational flexibility in decoupling the annealing and growth conditions while maintaining an isothermal transition between the two treatment zones, which will be discussed further below. Computational fluid dynamics (CFD) simulation of the designed reactor was performed to verify the separation of annealing and growth atmospheres (Figure 2). Model geometry and boundary conditions are presented in Figure 2a (see Experimental Methods for detailed description). Two processing chambers, the inner rectangular sleeve and the outer

cylindrical tube, were separated by 0.7−1.3 mm slit at the exit of the rectangular chamber. He/H2 and He/H2/CH4 mixtures were modeled to enter the annealing and the growth chambers at room temperature, and to undergo thermal expansion upon entering the heated region of the reactor. Concentration profiles of H2 and CH4 were then determined by combining nonisothermal flow simulations with a convective and diffusive transport model. Velocity (Figure 2d) and pressure profiles (Figure 2f) above the substrate are consistent with the annealing gas entering the heated region of the rectangular sleeve, expanding and accelerating, followed by a transition to the outer tube through the thin slit at a high velocity (up to 50 m/s). The temperature profile above the substrate, shown in Figure 2e, demonstrates rapid temperature stabilization (within 5% from the target value at 4 cm from the edge of the furnace) and seamless thermal transition between the process zones. 2D concentration maps along the center of the reactor for CH4 (Figure 2b) and the corresponding concentration profiles above the substrate (Figure 2c) reveal high-concentration uniformity and rapid stabilization of chemistry above the substrate upon transition between the zones (95% of target methane concentration at 1.5 cm downstream to the slit) even at low simulated pressures (4 Torr). We note the temperature profile seen by the Cu foil (Figure 2e) is however different than the furnace temperature profile (Figure 2a), specifically toward the end of the growth zone. The higher temperatures in Figure 2e D

DOI: 10.1021/acsami.8b00846 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 4. Evaluation of key parameters affecting graphene quality and trade-offs for roll-to-roll graphene manufacturing. SEM images of (a) Cu foil after annealing, (b) graphene grown on Cu at 1000 °C, 5 cm/min, and (c) 10 cm/min with 150 sccm H2 in the annealing zone and 150 sccm CH4 in the growth zone. (e,f) Graphs showing corresponding representative Raman spectra (black). SEM images of graphene grown on Cu at 1000 °C, 150 sccm H2 + 50 sccm CH4 in the annealing zone, and 150 sccm CH4 in the growth zone at (h) 10 cm/min, (i) 15 cm/min, and (j) 20 cm/min. Corresponding representative Raman spectra on (l−n) monolayer regions (red) and multilayer regions (blue). (d) ID/IG and (g) I2D/IG ratios computed from five different Raman spot measurements for conditions corresponding to (b, c) (black), (h−j) (red), and graphene grown on Cu at 1000 °C, 20 cm/min, 150 sccm H2 + 50 sccm CH4 in the annealing zone and 200 sccm CH4 in the growth zone (green). (k) ID/IG and (o) I2D/IG ratios computed from five different Raman spot measurements for graphene grown on Cu at 1000 °C, 10 cm/min, 150 sccm H2 in the annealing zone and 150 sccm CH4 in the growth zone (black), 150 sccm H2 + 50 sccm CH4 in the annealing zone and 150 sccm CH4 in the growth zone (red), 150 sccm H2 + 25 sccm CH4 in the annealing zone and 125 sccm CH4 in the growth zone (orange), and 150 sccm H2 + 10 sccm CH4 in the annealing zone and 140 sccm CH4 in the growth zone (orange). Error bars show one standard deviation.

beyond the heated zone are attributed to heat transfer from the hot gases leaving the growth zone to the Cu foil. Roll-to-Roll (R2R) Graphene CVD Parameter Space. Using the R2R system, we systematically vary the key CVD process parameters and identify trade-offs for continuous manufacturing of high-quality large-area monolayer graphene. Before being loade into the R2R system, the Cu foil substrate is precleaned by dipping it in a premixed acid etchant followed by washing with deionized water, using a separate simple roll-toroll setup (see Experimental Methods; Figure S2). This step

removes oxides, adventitious carbon, and other impurities on the polycrystalline Cu foil surface, thereby minimizing the potential for heterogeneous nucleation which would be detrimental to high-quality graphene synthesis.10,29 Graphene growth by CVD on Cu proceeds through nucleation and subsequent growth of individual graphene domains that merge to form a continuous layer.10−12 While many hydrocarbon sources can be used for graphene synthesis, here we chose a combination of CH4 and H2 under lowpressure conditions, which is known to allow for a robust E

DOI: 10.1021/acsami.8b00846 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces window of process parameters.10−14 We initially focused on optimizing the ratio (defined as the ratio of flow rates measured in sccm when the gases are introduced into the system) of CH4/H2 to achieve complete coverage of graphene on Cu foil, which is essential for membrane applications. A translation speed of 5 cm/min was chosen to give at least 6 min of annealing time and 6 min of growth time, and the growth temperature was set to 1000 °C to achieve conditions that work well in a standard stationary tube furnace CVD system.10−14,30 Higher temperatures were avoided to prevent excessive foil deformation under tension and adhesion to the quartz contact surfaces. The H2 flow rate in the annealing zone was kept constant at 150 sccm (all gases enter the growth zone after the annealing zone), and the CH4 flow rate into the growth zone was varied to achieve CH4/H2 ratios from 0.17 to 1. While the CFD model validated our reactor design with a small slit size, a larger slit size (5 mm) was used in our experiments to ease assembly and alignment of the foil path. Scanning electron microscopy (SEM) images (Figure 3a−c) of roll-to-roll CVD graphene show that the coverage (darker regions in the SEM) increases with CH4/H2 ratio. Complete coverage is obtained for CH4/H2 = 1. A representative Raman spectrum (Figure 3e) for incomplete coverage conditions corresponding to Figure 3b indeed shows characteristic peaks of monolayer graphene: 2D (∼2700 cm−1) and G (∼1600 cm−1), with I2D/IG > 1. However, the Raman spectrum also shows a relatively high D peak (∼1350 cm−1), with ID/IG ∼ 0.5 indicating defects and dangling bonds from graphene domain edges in the incompletely merged film.31 Quantitative coverage information (Figure 3f) obtained by analyzing the SEM images (thresholding the darker regions in SEM images, see Experimental Methods) confirms the qualitative trend seen in the images; that is, increasing CH4 flow rate while keeping H2 flow rate constant increases graphene coverage by increasing nucleation density or by increasing growth rate, or a combination of both. These observations emphasize the delicate balance between graphene formation from catalytic dissociation of CH4 and etching by H2 or trace contaminants in H2.10−14,30 In addition to CH4/H2 ratio, the polycrystalline crystallographic orientations of the Cu in the foil also affect the graphene formation and etching process within each Cu grain; that is, nucleation and/or growth is favored on certain facets, while etching may be favored on others. An example of this is seen in Figure 3d where there is preferential/selective graphene growth only on certain grains in the polycrystalline Cu foil.11,29 While CH4/H2 = 1 at 1000 °C gave complete coverage, cracks and discontinuities in the graphene were found, primarily perpendicular to the direction of movement of the foil (Figure 3g,h; Figure S3). We hypothesized the cracks were caused by slight elongation of the foil at high temperature due to the tension applied to facilitate roll-to-roll processing. Such an elongation could occur from (i) winding the foil with graphene facing outward, thereby giving a slight tensile stress to the graphene (see Figure 3g,h),32 or (ii) tension in the foil due to bearing friction on the foil roller in the input chamber. To address these aspects, we added a lower-friction bearing to the roller in the input chamber and changed the direction of winding on the output chamber to place the graphene on the inner surface. After these changes, we produced crack-free, full coverage graphene on Cu, as shown in Figure 3g,i. Having now achieved uniform, crack-free monolayer graphene by the roll-to-roll CVD process, we explored trade-

offs within the parameter space, specifically in terms of process speed (foil translation rate) versus the quality of graphene produced. Here, quality was measured by acquiring Raman spectra from at least five different spots for each synthesis condition. A faster speed gives shorter annealing times, which may not give sufficient surface cleaning and also results in a smaller grain size of the Cu foil, because grain growth occurs in the annealing zone. Hence, we preannealed the entire Cu foil on the rollers using the roll-to-roll system at ∼1000 °C in 150 sccm H2 at ∼5 cm/min (in the absence of CH4 in the growth zone), giving an annealing time of 12 min. This established large grains (typically >500 um2) of Cu with a clean surface (Figure 4a) in all the following experiments. Roll-to-roll graphene growth at 5 cm/min with 150 sccm H2 in the annealing zone and 150 sccm CH4 in the growth zone (see Figure 4b) shows complete coverage of the Cu surface with graphene (identified by the wrinkles in the SEM image). A representative Raman spectrum (Figure 4e) shows a negligible D peak, I2D/IG > 1 (see Figure 4g), and ID/IG ∼ 0.065 (see Figure 4d), confirming the high quality of the synthesized monolayer graphene.10−12,31 Increasing the foil speed to 10 cm/min while keeping all other conditions constant also allows for complete coverage of the Cu surface (Figure 4c), but results in a clear increase in the D peak (Figure 4f) with ID/IG ∼ 0.43 (Figure 4d). This indicates insufficient growth time (since nucleation density is expected to be similar to that of the 5 cm/ min case, see Figure 4b,e) for healing of defects in graphene while the individual domains of graphene merge to form a continuous monolayer.33,34 Hence, increasing the roll-to-roll processing speed under identical conditions gives lower graphene quality, and we determined that 5 cm/min at 1000 °C provided the best conditions. Knowing the residence time required for full coverage limited the speed to 5 cm/min in the two-zone configuration; we also explored higher speeds (≥10 cm/min) by adding CH4 to the annealing zone along with H2, in addition to the normal CH4 flow in the growth zone. Adding CH4 into the annealing zone doubles the residence time of the foil in the presence of carbon (giving more time for defect healing in graphene34), but comes at the expense of having hydrocarbons present while the Cu foil is being heated. A higher nucleation density along with increased defects is typically expected for graphene grown at lower temperatures,35 which in turn can lead to smaller domains of graphene and/or the nucleation of multilayers. In Figure 4h−j we show SEM images of graphene on Cu synthesized at 10, 15, and 20 cm/min by exposure to CH4 in both zones of the roll-to-roll CVD reactor. The SEM images show complete coverage of the Cu surface with monolayer graphene (see Figure 4l−n, red) and some regions of multilayer graphene. Raman spectra (see Figure 4l−n, blue) acquired on these regions confirm the multilayer nature of the graphene with I2D/IG < 1.31 The Raman spectra corresponding to the monolayer regions show an increase in D peak with increasing speed (see Figure 4l−n, red), consistent with prior observations (see Figure 4d−f). However, we note that ID/IG ∼ 0.2 obtained for graphene grown at 10 cm/min with CH4 in the annealing zone is distinctly lower than the ID/IG ∼ 0.43 obtained for graphene grown under otherwise identical conditions but with CH4 only in the growth zone (see Figure 4d). These observations imply that the longer residence time indeed helps heal defects as the graphene domains merge.34 However, the presence of CH4 in the annealing zone during heating of the foil could give a F

DOI: 10.1021/acsami.8b00846 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 5. Nanoporous atomically thin membranes (NATMs) from roll-to-roll CVD graphene. (a) Schematic overview of the membrane fabrication process using roll-to-roll CVD graphene on Cu. (b) Optical and (c, d) top-view SEM images of graphene on polyether sulfone (PES) support. (e) SEM image of a cross section of porous PES support scaffold with graphene on the top surface. (f) Diffusive permeance and (g) selectivity (ratio of permeance values) of KCl, L-Tr, vitamin B12, and lysozyme (∼4 nm, small protein) for graphene (G) + PES membrane, G + PES after interfacial polymerization (IP), and G + PES + IP after 60 s of oxygen plasma.

greater nucleation density,35 and the formation of defective graphene at lower temperatures prevents complete mitigation of defects. Increasing the flow of CH4 in the growth zone further beyond CH4/H2 = 1 (150 sccm H2 + 50 sccm CH4 in the annealing zone, 200 sccm CH4 in the growth zone) decreases the quality of graphene (green points, Figure 4d,g). However, graphene quality is correlated with the ratio of CH4/ H2 in the annealing zone, as seen by the decrease in the ID/IG ratio (Figure 4k,o). The coverage and quality of graphene from our roll-to-roll process are intrinsically linked to the gas composition and residence time in the growth zone, and healing of defects34 in the as-formed graphene emerges as the rate-limiting step. Further, heating in H2 and subsequent isothermal transition to CH4 is necessary to prevent low-temperature nucleation. We note that our findings are specific to this roll-to-roll system design and represent a set of conditions in a multidimensional parameter space; nevertheless, our study successfully captures the key trade-offs in continuous graphene synthesis while producing high-quality monolayer graphene (ID/IG ∼ 0.06) at speeds up to 5 cm/min. Nanoporous Atomically Thin Membranes (NATMs) from Roll-to-Roll CVD Graphene. Now, we use the highquality R2R graphene (ID/IG ∼ 0.065) grown at 5 cm/min to fabricate NATMs using scalable downstream processes. Here we successfully leverage traditional membrane fabrication approaches, for example, polymer casting, phase inversion, and interfacial polymerization, that are extensively used for rollto-roll manufacturing of polymer flat sheet reverse osmosis membranes,28 to form supports for NATMs. For use as a membrane, the CVD graphene must be removed from the growth substrate, that is, Cu foil, and placed on a

porous support using a clean process that does not leave residue on the graphene. Here, we use solution casting of polyether sulfone (PES) followed by a phase inversion reaction in water to form a hierarchically porous support directly on graphene on Cu, as shown in Figure 5a.26,36 After phase inversion, the Cu foil is etched (see Experimental Methods) to result in a large-area (>3 cm2) PES-supported graphene membrane (Figure 5b). SEM images (Figure 5c,d) show graphene (identified using wrinkles) suspended on ∼300−500 nm features that are pores in the PES support. These pores in the PES support are interconnected and branch out into larger, micrometer-scale pores further below, visible in the SEM cross section, Figure 5e. Also seen in the SEM images are regions that appear bright because graphene is absent (see Figure 5c,d). In these areas, the graphene has been damaged due to (a) transfer/handling, (b) abrasion/contact during winding (during synthesis) and unwinding (postsynthesis) on the rollers, and (c) van der Waals contact with the graphene in the preceding or succeeding windings on the roller. These damaged regions are problematic for membrane applications because they offer a direct route for transport of molecules, thereby short-circuiting the selective transport through the membrane. We seal such damaged regions using an interfacial polymerization process after transfer to the PES supports. In interfacial polymerization two monomers separated by graphene react only in regions where the graphene is damaged, forming polymer plugs and effectively sealing the membrane (see Figure 5a and Experimental Methods).6,9,26,27 Finally, we use a pulsed oxygen plasma etch6 to form nanopores in the graphene lattice and realize NATMs. The performance of the roll-to-roll graphene/PES-based NATM was evaluated for diffusion-driven transport using G

DOI: 10.1021/acsami.8b00846 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces solutes of different sizes: KCl (salt, K+,Cl− ∼0.66 nm), Ltryptophan (L-Tr, ∼0.7−0.9 nm), vitamin B12 (B12, ∼1−1.5 nm), and lysozyme (Lz, ∼3.8−4 nm). Ideally, the NATM would allow the smaller molecules to pass while blocking the larger molecules. Permeance and selectivity data for R2R graphene on PES, with and without interfacial polymerization, and with interfacial polymerization after 60 s of oxygen plasma, are shown in parts (f) and (g), respectively, of Figure5. The graphene on PES support (G + PES) shows size-selective transport because of intrinsic defects in graphene synthesized at 1000 °C in addition to nonselective leakage across tears in the graphene.10 The permeance values for all species, in particular Lz, reduce significantly after interfacial polymerization, consistent with the sealing of tears in graphene with polymer plugs. However, upon exposure of the leakage-sealed membranes to oxygen plasma, the permeance for species 98%, Alfa Aesar), and lysozyme (ultrapure grade, VWR International) were performed using 1 mM of the respective solutes in 0.5 M KCl on the feed and 0.5 M KCl on the permeate side.6,10,23,27 The increase in concentration of the solutes on the permeate side was measured using a UV−vis spectrometer (Agilent − Cary 60).6,10,23,27 ⎛ V × dC ⎞ P = ⎜ ΔC ×dAt ⎟ was used to compute permeance (P), where V is ⎝ ⎠ volume of the diffusion cell (7 mL), A is the membrane area (∼5 mm orifice of diffusion cell), ΔC is the concentration difference between the solute in feed and permeate cell, and dC is the rate of change of

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected] (A.J.H.). *E-mail: [email protected] (P.R.K.). ORCID

Piran R. Kidambi: 0000-0003-1546-5014 Nicholas T. Dee: 0000-0002-8633-3564 Rohit Karnik: 0000-0003-0588-9286 A. John Hart: 0000-0002-7372-3512 Notes

The authors declare the following competing financial interest(s): R.K. is a co-founder and has equity in a startup company aimed at commercializing graphene membranes.

■

ACKNOWLEDGMENTS Support to P.R.K., N.T.D., and A.J.H., and for graphene synthesis and characterization, was provided by the U.S. Department of Energy, Office of Science, Basic Energy Sciences, under Award No. DE-SC0010795. Support to D.M. and partial support to N.T.D. was provided by a grant from the Skoltech-MIT Next Generation Programme. Membrane fabrication and testing was supported by the U.S. Department of Energy, Basic Energy Sciences, Award No. DE-SC0008059 (to R.K.). SEM images in this work were acquired using facilities at the Center for Nanoscale Systems (CNS) at Harvard University, a member of the National Nanotechnology Infrastructure Network, supported by the National Science Foundation under NSF Award No. ECS-0335765, and the MRSEC Shared Experimental Facilities at MIT, supported by the National Science Foundation under Award No. DMR1419807.

■

REFERENCES

(1) Bunch, J. S.; Verbridge, S. S.; Alden, J. S.; van der Zande, A. M.; Parpia, J. M.; Craighead, H. G.; McEuen, P. L. Impermeable Atomic Membranes from Graphene Sheets. Nano Lett. 2008, 8 (8), 2458− 2462. (2) Zandiatashbar, A.; Lee, G.-H.; An, S. J.; Lee, S.; Mathew, N.; Terrones, M.; Hayashi, T.; Picu, C. R.; Hone, J.; Koratkar, N. Effect of Defects on the Intrinsic Strength and Stiffness of Graphene. Nat. Commun. 2014, 5, 3186. (3) Lee, C.; Wei, X.; Kysar, J. W.; Hone, J. Measurement of the Elastic Properties and Intrinsic Strength of Monolayer Graphene. Science 2008, 321 (5887), 385−388. (4) Koenig, S. P.; Wang, L.; Pellegrino, J.; Bunch, J. S. Selective Molecular Sieving through Porous Graphene. Nat. Nanotechnol. 2012, 7 (11), 728−732. (5) Surwade, S. P.; Smirnov, S. N.; Vlassiouk, I. V.; Unocic, R. R.; Veith, G. M.; Dai, S.; Mahurin, S. M. Water Desalination Using Nanoporous Single-Layer Graphene. Nat. Nanotechnol. 2015, 10 (5), 459−464. (6) Kidambi, P. R.; Jang, D.; Idrobo, J.-C.; Boutilier, M. S. H.; Wang, L.; Kong, J.; Karnik, R. Nanoporous Atomically Thin Graphene Membranes for Desalting and Dialysis Applications. Adv. Mater. 2017, 29 (33), 1700277. (7) Hu, S.; Lozada-Hidalgo, M.; Wang, F. C.; Mishchenko, A.; Schedin, F.; Nair, R. R.; Hill, E. W.; Boukhvalov, D. W.; Katsnelson, M. I.; Dryfe, R. A. W.; Grigorieva, I. V.; Wu, H. A.; Geim, A. K. Proton Transport through One-Atom-Thick Crystals. Nature 2014, 516 (7530), 227−230. (8) Lozada-Hidalgo, M.; Zhang, S.; Hu, S.; Esfandiar, A.; Grigorieva, I. V.; Geim, A. K. Scalable and Efficient Separation of Hydrogen Isotopes Using Graphene-Based Electrochemical Pumping. Nat. Commun. 2017, 8 (15215), 15215.

dt

concentration with time.6,10,23,27 Selectivity was computed by taking the ratio of permeance between two species.6,10,23,27

■

Research Article

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsami.8b00846. Detailed schematic of the roll-to-roll reactor components, images of the experimental setup to clean oxides from the Cu foil surface, and SEM images of crack in graphene on Cu (PDF) I

DOI: 10.1021/acsami.8b00846 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces (9) Wang, L.; Boutilier, M. S. H.; Kidambi, P. R.; Jang, D.; Hadjiconstantinou, N. G.; Karnik, R. Fundamental Transport Mechanisms, Fabrication and Potential Applications of Nanoporous Atomically Thin Membranes. Nat. Nanotechnol. 2017, 12 (6), 509− 522. (10) Kidambi, P. R.; Terry, R. A.; Wang, L.; Boutilier, M. S. H.; Jang, D.; Kong, J.; Karnik, R. Assessment and Control of the Impermeability of Graphene for Atomically Thin Membranes and Barriers. Nanoscale 2017, 9 (24), 8496−8507. (11) Kidambi, P. R.; Ducati, C.; Dlubak, B.; Gardiner, D.; Weatherup, R. S.; Martin, M. B.; Seneor, P.; Coles, H.; Hofmann, S. The Parameter Space of Graphene Chemical Vapor Deposition on Polycrystalline Cu. J. Phys. Chem. C 2012, 116 (42), 22492−22501. (12) Kidambi, P. R.; Bayer, B. C.; Blume, R.; Wang, Z.-J.; Baehtz, C.; Weatherup, R. S.; Willinger, M.-G.; Schloegl, R.; Hofmann, S. Observing Graphene Grow: Catalyst−Graphene Interactions during Scalable Graphene Growth on Polycrystalline Copper. Nano Lett. 2013, 13 (10), 4769−4778. (13) Bae, S.; Kim, H.; Lee, Y.; Xu, X.; Park, J.-S.; Zheng, Y.; Balakrishnan, J.; Lei, T.; Ri Kim, H.; Song, Y. Il; Kim, Y.-J.; Kim, K. S.; Ö zyilmaz, B.; Ahn, J.-H.; Hong, B. H.; Iijima, S. Roll-to-Roll Production of 30-Inch Graphene Films for Transparent Electrodes. Nat. Nanotechnol. 2010, 5 (8), 574−578. (14) Li, X.; Cai, W.; An, J.; Kim, S.; Nah, J.; Yang, D.; Piner, R.; Velamakanni, A.; Jung, I.; Tutuc, E.; Banerjee, S. K.; Colombo, L.; Ruoff, R. S. Large-Area Synthesis of High-Quality and Uniform Graphene Films on Copper Foils. Science 2009, 324 (5932), 1312− 1314. (15) Polsen, E. S.; McNerny, D. Q.; Viswanath, B.; Pattinson, S. W.; Hart, A. J. High-Speed Roll-to-Roll Manufacturing of Graphene Using a Concentric Tube CVD Reactor. Sci. Rep. 2015, 5, 10257. (16) Kobayashi, T.; Bando, M.; Kimura, N.; Shimizu, K.; Kadono, K.; Umezu, N.; Miyahara, K.; Hayazaki, S.; Nagai, S.; Mizuguchi, Y.; Murakami, Y.; Hobara, D. Production of a 100-M-Long High-Quality Graphene Transparent Conductive Film by Roll-to-Roll Chemical Vapor Deposition and Transfer Process. Appl. Phys. Lett. 2013, 102 (2), 023112. (17) Zhong, G.; Wu, X.; D’Arsie, L.; Teo, K. B. K.; Rupesinghe, N. L.; Jouvray, A.; Robertson, J. Growth of Continuous Graphene by Open Roll-to-Roll Chemical Vapor Deposition. Appl. Phys. Lett. 2016, 109 (19), 193103. (18) Deng, B.; Hsu, P. C.; Chen, G.; Chandrashekar, B. N.; Liao, L.; Ayitimuda, Z.; Wu, J.; Guo, Y.; Lin, L.; Zhou, Y.; Aisijiang, M.; Xie, Q.; Cui, Y.; Liu, Z.; Peng, H. Roll-to-Roll Encapsulation of Metal Nanowires between Graphene and Plastic Substrate for HighPerformance Flexible Transparent Electrodes. Nano Lett. 2015, 15 (6), 4206−4213. (19) Yamada, T.; Ishihara, M.; Kim, J.; Hasegawa, M.; Iijima, S. A Roll-to-Roll Microwave Plasma Chemical Vapor Deposition Process for the Production of 294 Mm Width Graphene Films at Low Temperature. Carbon 2012, 50 (7), 2615−2619. (20) Chandrashekar, B. N.; Deng, B.; Smitha, A. S.; Chen, Y.; Tan, C.; Zhang, H.; Peng, H.; Liu, Z. Roll-to-Roll Green Transfer of CVD Graphene onto Plastic for a Transparent and Flexible Triboelectric Nanogenerator. Adv. Mater. 2015, 27 (35), 5210−5216. (21) Hesjedal, T. Continuous Roll-to-Roll Growth of Graphene Films by Chemical Vapor Deposition. Appl. Phys. Lett. 2011, 98 (13), 133106. (22) Jang, B.; Kim, C.-H.; Choi, S. T.; Kim, K.-S.; Kim, K.-S.; Lee, H.J.; Cho, S.; Ahn, J.-H.; Kim, J.-H. Damage Mitigation in Roll-to-Roll Transfer of CVD-Graphene to Flexible Substrates. 2D Mater. 2017, 4 (2), 024002. (23) Kidambi, P. R.; Boutilier, M. S. H.; Wang, L.; Jang, D.; Kim, J.; Karnik, R. Selective Nanoscale Mass Transport across Atomically Thin Single Crystalline Graphene Membranes. Adv. Mater. 2017, 29 (19), 1605896. (24) Boutilier, M. S. H.; Jang, D.; Idrobo, J. C.; Kidambi, P. R.; Hadjiconstantinou, N. G.; Karnik, R. Molecular Sieving Across

Centimeter-Scale Single-Layer Nanoporous Graphene Membranes. ACS Nano 2017, 11 (6), 5726−5736. (25) Madauß, L.; Schumacher, J.; Ghosh, M.; Ochedowski, O.; Meyer, J.; Lebius, H.; Ban-d’Etat, B.; Toimil-Molares, M. E.; Trautmann, C.; Lammertink, R. G. H.; Ulbricht, M.; Schleberger, M. Fabrication of Nanoporous Graphene/polymer Composite Membranes. Nanoscale 2017, 9 (29), 10487−10493. (26) Qin, Y.; Hu, Y.; Koehler, S.; Cai, L.; Wen, J.; Tan, X.; Xu, W. L.; Sheng, Q.; Hou, X.; Xue, J.; Yu, M.; Weitz, D. Ultrafast Nanofiltration through Large-Area Single-Layered Graphene Membranes. ACS Appl. Mater. Interfaces 2017, 9 (11), 9239−9244. (27) O’Hern, S. C.; Jang, D.; Bose, S.; Idrobo, J. C.; Song, Y.; Laoui, T.; Kong, J.; Karnik, R. Nanofiltration across Defect-Sealed Nanoporous Monolayer Graphene. Nano Lett. 2015, 15 (5), 3254−3260. (28) Lee, K. P.; Arnot, T. C.; Mattia, D. A Review of Reverse Osmosis Membrane Materials for desalinationDevelopment to Date and Future Potential. J. Membr. Sci. 2011, 370 (1−2), 1−22. (29) Braeuninger-Weimer, P.; Brennan, B.; Pollard, A. J.; Hofmann, S. Understanding and Controlling Cu-Catalyzed Graphene Nucleation: The Role of Impurities, Roughness, and Oxygen Scavenging. Chem. Mater. 2016, 28 (24), 8905−8915. (30) Blume, R.; Kidambi, P. R.; Bayer, B. C.; Weatherup, R. S.; Wang, Z.-J.; Weinberg, G.; Willinger, M.-G.; Greiner, M.; Hofmann, S.; Knop-Gericke, A.; Schlögl, R. The Influence of Intercalated Oxygen on the Properties of Graphene on Polycrystalline Cu under Various Environmental Conditions. Phys. Chem. Chem. Phys. 2014, 16 (47), 25989−26003. (31) Ferrari, A. C.; Basko, D. M. Raman Spectroscopy as a Versatile Tool for Studying the Properties of Graphene. Nat. Nanotechnol. 2013, 8 (4), 235−246. (32) Na, S. R.; Wang, X.; Piner, R. D.; Huang, R.; Willson, C. G.; Liechti, K. M. Cracking of Polycrystalline Graphene on Copper under Tension. ACS Nano 2016, 10 (10), 9616−9625. (33) Chen, J.; Shi, T.; Cai, T.; Xu, T.; Sun, L.; Wu, X.; Yu, D. Self Healing of Defected Graphene. Appl. Phys. Lett. 2013, 102 (10), 103107. (34) Lee, J.; Baek, J.; Ryu, G. H.; Lee, M. J.; Oh, S.; Hong, S. K.; Kim, B. H.; Lee, S. H.; Cho, B. J.; Lee, Z.; Jeon, S. High-Angle Tilt Boundary Graphene Domain Recrystallized from Mobile Hot-WireAssisted Chemical Vapor Deposition System. Nano Lett. 2014, 14 (8), 4352−4359. (35) Li, X.; Magnuson, C. W.; Venugopal, A.; An, J.; Suk, J. W.; Han, B.; Borysiak, M.; Cai, W.; Velamakanni, A.; Zhu, Y.; Fu, L.; Vogel, E. M.; Voelkl, E.; Colombo, L.; Ruoff, R. S. Graphene Films with Large Domain Size by a Two-Step Chemical Vapor Deposition Process. Nano Lett. 2010, 10 (11), 4328−4334. (36) Kidambi, P. R.; Zhang, S.; Chen, Q.; Warner, J.; Kong, J.; Karnik, R. Facile Fabrication of Functional Atomically Thin Membranes by Bottom-Up Synthesis of Nanoporous Single-Layer Graphene. Manuscipt in preparation. (37) http://www.thermo.com.cn/Resources/200802/productPDF_ 31131.pdf. (38) Dalwani, M.; Zheng, J.; Hempenius, M.; Raaijmakers, M. J. T.; Doherty, C. M.; Hill, A. J.; Wessling, M.; Benes, N. E. Ultra-Thin Hybrid Polyhedral Silsesquioxane−polyamide Films with Potentially Unlimited 2D Dimensions. J. Mater. Chem. 2012, 22 (30), 14835.

J

DOI: 10.1021/acsami.8b00846 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX