Symposium NT3 : Carbon Nanofluidics

Carbon nanotubes have three key attributes that make them of great interest for novel membrane applications 1) atomically flat graphite surface allows for ideal fluid slip boundary conditions and extremely fast flow rates 2) the cutting process to open CNTs inherently places functional chemistry at CNT core entrance for chemical selectivity and 3) CNT are electrically conductive allowing for electrochemical reactions and application of electric fields gradients at CNT tips. In general, the transport mechanisms through CNT membrane are a) ionic diffusion is near bulk expectation with no enhancement from CNT b) gas flow is enhanced by ~1-2 order of magnitude due to specular reflection off of flat graphitic surface c) and pressure driven flux of a variety of solvents (H2O, hexane, decane ethanol, methanol) are 4-5 orders of magnitude higher than conventional Newtonian flow [Nature 2005, 438, 44] due to atomically flat graphite planes inducing nearly ideal slip conditions. Nearly all applications require chemical selectivity in what is allowed to pass across the membrane. However the act of placing selective functional chemistry at pore entrance or along the core of CNTs, dramatically/completely eliminates the enhanced flow effects by eliminating the near perfect slip boundary condition [ACS Nano 2011 5 3867]. Needed is a mechanism to pump chemicals through the pore where the selective chemistry is. This is routinely achieved in protein channels where permeates are accelerated through regions of precise functionality. Shown are electrostatically acivated biomimetic pumping systems. Also introduced here are biomimetic platforms using sequential enzymatic flow reactors, nanobubble valves for energy storage and monolayer catalytic reactors.

We perform an experimental study of ionic transport and current fluctuations inside individual Carbon Nanotubes (CNT). The conductance exhibits a power law behavior at low salinity, with an exponent close to 1/3 versus the salt concentration. This scaling behavior is rationalized in
terms of a model accounting for hydroxide adsorption at the (hydrophobic) carbon surface, leading to a density dependent surface charge. This is in contrast to boron nitride nanotubes which exhibit a constant surface conductance. Further we measure the low frequency noise of the ionic current in CNT and show that the amplitude of the noise scales with the surface charge, with data collapsing on a master curve for the various studied CNT at a given pH.

Using numerical simulation of carbon nanotubes so narrow to allow only a single file of water molecules, we find that flow is due to fast-moving density variations that deliver the flow in discrete molecular quantities. Single-file water does not freeze, and flowrates increase as temperature decreases. Our results build a new understanding of nanochannel flows and suggest new principles for the design of nanoscale devices.

Over 5000 high-resolution molecular dynamics simulations of a long armchair CNT with TIP3P water molecules were run over 1 to 300 K. A key parameter is the number of water molecules in excess over the number of carbon rings along the CNT length. When the number of water molecules in the CNT matches the number of hexagonal rings along the CNT axis, water molecules are in one-to-one registry with rings, yielding constant density along the CNT. With fewer or more molecules, water molecules distribute unevenly: most are in one-to-one registry with the CNT rings, but localized defects appear, within which water molecules are more widely or closely spaced. From one side of a defect to the other, a water molecule is displaced by 1/2 of a hexagonal ring. For (6,6) CNTs, the width of the defect is 15 nm.

With a force applied to all water molecules, molecules outside of defects are stationary, but as defects pass by, they advance 1/2 the width of the CNT ring. Thus, the upstream propagation of (expansion) defects leads to fluid flow. We found that defect velocity correctly predicts flowrates. An analogous phenomenon occurs in traffic jams: cars are mostly stationary but intermittently move forward into the gap opened by the advance of the preceding car. As cars shift forward, the gap propagates backward.

The Frenkel-Kontorova (FK) equations describe the dynamics of particles with inter-particle forces (that favor equal spacing of particles) and with particle-substrate forces (that favor a different equal spacing). The competition between these two length scales produce solitons, localized variations in particle density. The observed properties of the defects in CNTs match the properties of FK solitons, including: particle density follows a sech distribution; solitons convect mass, so as the number of solitons decreases, flowrate decreases; solitons propagate at all temperatures, so, single-file flow occurs at temperatures well below the freezing point of bulk water.

Our results introduce a new mechanism of rapid single-file flow in nanochannels. Transport through long CNTs occurs in discrete molecular units due to defect propagation. Defects are shown to be similar to FK solitons, which argues for competition between water-water and carbon-carbon spacing as the mechanism underlying rapid transport. As this competition is present in all CNTs and liquid molecules, defects may play a role in transport in larger CNTs and for diverse fluids.

Financial support from the Lillian Sidney Foundation and Northwestern University's QUEST HPC.

Fluid phase transitions inside single, isolated carbon nanotubes (CNT) are predicted to deviate substantially from classical thermodynamics and also allow the study of ice nanotube (ice-NT) properties. Herein, we measure, using two different techniques, the diameter dependent phase boundaries of ice-NTs within isolated CNTs 1.05, 1.06, 1.15, 1.24, and 1.52nm in diameter using Raman spectroscopy. The results reveal both an exquisite sensitivity to diameter and substantially larger temperature elevations of the melting transition than theoretically predicted by as much as 100°C. Dynamic water filling and reversible freezing transitions were marked by 2 to 5cm^-1 shifts in the radial breathing mode (RBM) frequency, revealing reversible melting at 138°C and 102°C for 1.05 and 1.06nm single and double-walled CNTs, respectively. A near-ambient phase change at 15°C was observed for 1.52nm CNT, whereas freezing inside 1.24nm tube was suppressed at -30°C. We find that the interior aqueous phase also decreases the axial thermal conductivity of the CNT reversibly by as much as 500%, allowing digital control of the heat flux. These extreme phase transitions enable the study of ice-NT at high temperatures and their potential utilization as novel phase change materials.

Ionic and molecular transport inside nanometre-scale geometries is distinct from micro- and macroscale transport due to the dominance of surface forces which leads to novel physical phenomena. Single-Walled Carbon nanotubes (SWCNTs) with their unique structural and physical properties appear as particularly interesting channels for understanding fluidic and ionic transport at the nanoscale and for the development of nanofluidic applications. The possibility of transporting ions and molecules through SWCNTs has already been reported by several groups although rationalizing the variety of experimental results still represents an important challenge and a crucial step for the rise of SWCNT-based nanofluidics. To reach this goal, a prerequisite is the fabrication of more robust SWCNT-based fluidic devices that allow the collection of a variety of experimental data (e.g. fluidic, electrical, and optical). With the recent advances in nanofabrication technology, the development of such devices becomes feasible but yet challenging.

Here, we will report on ion current studies performed in microfluidic devices incorporating one or several individual SWCNTs which have diameters in the 1-2 nm range and lengths of tens of microns. In such devices, the individual SWCNTs can be optically addressed allowing Raman studies of the charge transfer between the matrix and the SWCNT. The ionic currents experimentally measured are found to be in good agreement with standard models of ionic transport through charged and slippery nanochannels but much lower than the values previously reported by some groups. Ion current studies commonly reveal a non-ohmic voltage-activated behaviour supporting the existence of energy barriers for the entering or the transport of ions through small-diameter CNTs. A study of the selectivity of ion transport through SWNTs will be also shown for different cations.

Fast water transport inside carbon nanotubes (CNT) is a highly topical issue in fundamental research, while related applications are already under consideration in environmental and energy fields [1, 2]. A substantial set of simulation and theoretical works have pointed out two important parameters for filling hydrophobic nanochannels, namely curvature effects [3] or the strong modification of the hydrogen-bond (HB) network [4]. In contrast, experimental publications are rather scarce.
Here, we report the first experimental analysis of water structure and H-bond network during CNT filling [5-6]. Experiments were performed on single-walled carbon nanotubes (SWCNT) with diameters between 1 and 2 nm, as they are ultimate one-dimensional carbon nanochannels. We present in-situ monitoring of water filling of SWCNT at room temperature, using X-ray scattering (XRS) and infra-red (IR) spectroscopy. A systematic method is developed to determine the water radial density profile from XRS measurements, which reveals a progressive structuration with increasing the filling rate. Furthermore, IR spectroscopy gives evidence for a large proportion of loosely bonded water molecules inside SWCNTs, even for fully hydrated nanotubes. These results are discussed with respect to theoretical and simulation studies. The present experimental data provide a solid reference for the elaboration of an energetic model accounting for the properties of water in hydrophobic nanoconfinement. They could also be relevant with respect to the unimpeded permeation of water through graphene-based membranes [7], or in other fields such a biology, where the extreme permeability to water of aquaporins, those membrane proteins that form nanopores, is crucial for many physiological processes [8].
References:
[1] Lee, B.; Baek, Y.; Lee, M.; Jeong, D. H.; Lee, H. H.; Yoon, J.; Kim, Y. H. Nature Comm. 2015, 6, 7109
[2] Park, H.G.; Jung, Y. Chem. Soc. Rev. 2014, 43, 565
[3] Falk, K.; Sedlmeier, F.; Joly, L.; Netz, R. R.; Bocquet, L. Nano Lett. 2010, 10, 4067
[4] Joseph, S.; Aluru, N.R. Nano Lett. 2008, 8, 452
[5] Paineau, E.; Albouy, P. A.; Rouzière, S.; Orecchini, A.; Rols S.; Launois, P. Nano Lett. 2013, 4, 1751
[6] Dalla-Bernardina, S.; Paineau, E.; Brubach, J.B.; Judeinstein, P.; Rouzière, S.; Launois, P.; Roy, P. submitted
[7] Nair, R. R.; Wu, H. A.; Jayaram, P. N.; Grigorieva, I. V.; Geim, A. K. Science 2012, 335, 442
[8] Sui, H.; Han, B. G.; Lee, J. K.; Walian, P.; Jap, B. K. Nature 2001 , 414, 872

The maintenance and regulation of ion gradients, specifically the electrochemical proton gradient, ΔμH⁺, across biological membranes has been established as a necessary intermediate in biological energy transduction that is crucial for proper cellular function. In recent years, artificial membrane channels have attracted much attention due to the key role of proton channels in performing a number of specific functions in different cells. Carbon nanotube (CNT) structures are similar to biological channels with their smooth, narrow, hydrophobic inner pores. The use of single-walled carbon nanotubes as biomimetic components requires the ability to segment these microscopic structures down to biologically relevant proportions. To meet this need, we have developed a sonochemical method to produce shortened CNT porins (CNTPs). These CNTPs self-insert into lipid bilayers, creating channels that retain the ion transport properties characteristic to CNTs while assimilating to a biological environment. We demonstrate that 0.8 nm diameter CNTPs, which promote formation of true one-dimensional water wires, support proton transport with rates exceeding those in bulk water by an order of magnitude. Proton transport rates in these narrow CNT pores also exceed those of biological channels and Nafion. Surprisingly, the larger 1.5 nm diameter nanotube porins still showed proton transport rates comparable to bulk water. This work establishes small diameter CNT porins as a promising membrane proton conductor material and points to strong spatial confinement as a key factor in enabling efficient proton transport.

The equilibrium properties of nanoconfined fluids, such as the density and the self-diffusivity, can be substantially different from the bulk properties of that fluid. Understanding the physical basis for and magnitude of these anomalous fluid properties is key to many nanoscale engineering pursuits. These applications range from nanoscale drug delivery to enhanced oil recovery to higher-efficiency molecular simulation methods, and even potentially to constructing a sub-continuum model of nanoscale fluid flow.

We present here a theoretical description of these phenomena in the context of a dense Lennard-Jones fluid confined within a carbon nanotube (CNT). In particular, we show that the anomalous nanoconfined fluid density can be substantially lower than the bulk density, and that this reduced density is primarily due to repulsive interactions between the fluid and the CNT. Using a mean-field approach to describe the energetic landscape near the CNT wall, we obtain closed-form analytical results for the lengthscales associated with fluid layering near the fluid-CNT interface. Combined with empirical knowledge about the layered-fluid density, these results allow us to predict the equilibrium fluid density as a function of the CNT radius. This prediction is in excellent agreement with molecular dynamics simulations as well as several results from the literature. We also show how accurate knowledge of the anomalous density can inform models of simple nanoscale transport phenomena.

We also show that this energetic landscape and associated density profile can be used to explain the anomalous diffusive transport observed in such systems. In particular, the presence of a ring structure near the fluid-CNT interface implies that fluid molecules near the CNT wall are not free to move as in the bulk of the fluid and thus exhibit non-Fickian diffusive behavior. By constructing different models for diffusion in the near-wall and the bulk regions of the CNT, we show that we can approximately predict the overall anomalous diffusive behavior of the nanoconfined fluid. We demonstrate that these results are in agreement with molecular dynamics simulations and exhibit qualitative agreement with several experimental results. Finally, we show how some aspects of both the density and diffusion theories can be extended to describe nanoconfined water.

Related publication:

Why are Fluid Densities So Low in Carbon Nanotubes? G.J. Wang and N.G. Hadjiconstantinou. Physics of Fluids, Vol. 27, No. 5, 052006 (2015).

Due to their high symmetry and large aspect ratio, carbon nanotubes (CNTs) are expected to display new and exciting transport regimes that are distinct from other nanopore systems. Some unique transport properties have been theoretically predicted and experimentally reported. However, a main challenge to appropriately exploit these unique properties in different applications is a better understanding of their origin. In order to study the transport of liquids and solutes through CNT, it is mandatory first to fabricate nanotube-integrating fluidic devices well-controlled, robust and allowing a variety of experimental studies. With the recent advancements witnessed in the nanofabrication technology, the development of such fluidic devices becomes feasible although still challenging. The most basic fluidic setup is made of two two fluidic reservoirs only connected by one or more open CNTs. Electrophoretic transport phenomena can then be studied by recording the ionic current under an applied voltage across two electrodes placed in each reservoir filled with an electrolyte solution. Consequently, the fabricated devices must be designed not only to measure the transport through the nanotube, but also to ensure that this transport is occurring solely through the tube and not through the barrier material or through defects in the barrier. The fabrication of such systems has already been demonstrated by several groups (Lee et al., 2010, Science 329:1320; Liu et al., 2010, Science 327:64) which have established different approaches relying on different fluidic platforms. The developed platforms differed greatly in terms of barrier materials, of CNT number, length and orientation, and of synthesis and opening methods.
In this context, we will present a detailed fabrication protocol of CNT-based microfluidic devices that are suitable for studying ionic and molecular transport under an applied electrical potential, pressure, and concentration gradient. The developed protocol allows the design of microdevices that are easy to fabricate, to reproduce and to use. The fabricated devices also grant the ability to perform in situ optical characterizations (Raman spectroscopy, fluorescence or optical imaging) in the visible range.

Membrane has been widely used human’s life as well as industry in water ultrafiltration system and gas separation. Membrane technology is required to control surface energy when it filters out impurities in organic solvents or water. These membranes need high permeance to enable processing within a reasonable time. Desalination is one of the main technology for producing fresh water from seawater and other saline water sources. The membrane properties greatly affect the water productivity and energy costs in the reverse osmosis (RO) desalination processes. Recent years, significant research efforts devoted to developing high-performances with improved flux and salt rejection. In particular, many researchers are interested in it with possibility of dramatically raised fluid flows of liquids and gases in nano-channel with several diameters. Based on a reverse osmosis (RO) and nano-filtration (NF) membrane, water flux (or permeability) and salt (or NaCl) rejection are inversely correlated. This study focuses on high permeability by using AAO membrane with hydrophilic and hydrophobic surface. In order to make hydrophobic surface carbon layer was deposited on inner wall of the AAO membrane by non-catalytic using CVD process. The AAO membranes were slowly heated up to 800^oC at 10/min under a mixture of acetylene and argon for 50 min. A systematic analysis about water flux using AAO membranes with different pore diameter from 40 nm to 90 nm has been conducted. The results have shown that permeability, though higher than for hydrophilic materials such as carbon coated AAO membranes, can be observed for hydrophobic material as well. The results also confirm that the water flux increases with increasing diameter along the theoretical value, and organic solvent permeability is increased with increasing diameter. These results offer a better understanding of fluid transport properties in nano-channels providing important information for various areas of research, including drug delivery, material separations and lab-on-a-chip devices. This work also presents the possibility to develop high permeation hydrophilic membranes.

Membrane prepared by creating nanoscale pores in a single layer of graphene, could be used as an effective separation membrane due to its chemical and mechanical stability, its flexibility and, most importantly, its one-atom thickness. Theoretical studies have indicated that the performance of such membranes should be superior to state-of-the-art polymer-based filtration membranes, and experimental studies have recently begun to explore their potential. We provide experimental evidence that single-layer porous graphene can be used as a desalination membrane. Nanometre-sized pores were created in a graphene monolayer using an oxygen plasma etching process, which allows the size of the pores to be tuned. The resulting membranes exhibit a salt rejection rate of nearly 100% and rapid water transport.

Reduced graphene oxide (rGO) is an attractive electrode material for supercapacitors because of its unique intrinsic properties including large specific surface area, high electrical conductivity, and superior electrochemical and mechanical stabilities. However, practical applications of rGO as active material in supercapacitors have been hindered to some extent by the strong interaction between rGO sheets. In most cases, rGO sheets were easily restacked into irregular graphite-like aggregates during the processes of assembling them into bulk electrodes, greatly reducing their ion-accessible surface area, and thereby deteriorating their performances in supercapacitors. To address this problem, we will introduce the methods for self-assembling rGO sheets into different three-dimensional architectures including hydrogels, “porous pottery”, thin films with fully accessible surface areas, etc. The supercapacitors showed high specific capacitances, good rate capability and excellent stability.

Biological systems control transport of ions or small molecules across membranes using ion channels that form highly efficient and selective pores in lipid bilayers. Inner pores of carbon nanotubes (CNTs) have unusual transport properties that have the potential to replicate many of the features of the biological transport. Although several experimental approaches demonstrated the transport through macroscopically long CNTs, there is still a need for a facile approach for building CNT nanopores that replicate most of the transport properties and geometry of membrane channels. We will show that pores formed by ultra-short carbon nanotubes (CNTs) assembled in the lipid membranes can come remarkably close to that goal. These CNT porins (CNTPs) can transport water, protons, ions, and DNA, and can even display stochastic gating behavior common for biological ion channels. CNTPs also provide a flexible and controllable system for studying the effects of channel confinement on the efficiency of molecular transport using both bulk scale and single-pore scale measurements. Overall, CNT porins represent a simplified biomimetic system that is ideal for studying fundamentals of nanofluidic transport and for building engineered mesoscale material systems.

Carbon-based nanostructures are attractive materials for nanofluidic manipulation and transport. When fluids are confined in nanometer scale pores, the size of the fluid molecule becomes comparable to the size of the pore. At this scale, the fluid physics can be quite different from its macroscopic counterpart. For example, the translational and rotational motions of fluid molecules can be vastly different under confinement. In this talk, we will address a number of fundamental issues when fluids are confined in thick and ultra-thin nanopores. First, we will discuss the multiphase structure of water confined in nanopores. Second, we will address the spatial variation and anisotropic diffusion of water in confined nanotubes. Third, we will investigate dynamics of a single water molecule in extreme confinement. Fourth, we will investigate water and ion transport through ultra-thin membranes and finally, we will present a number of applications of nanoporous membranes including water desalination, energy conversion, protein recognition, DNA sequencing, etc.

Graphene oxide (GO) membranes, formed by stacking micrometer-sized single GO sheets together, have recently attracted wide attention due to their great potential in water desalination and selective molecular sieving. This new type of membranes consists of numerous parallel and interconnected uniform-height 2-D graphene nanochannels with nearly frictionless surfaces. Further developments of these membranes, including enhancing their mass transport rate and/or molecular selectivity, rely on the understanding of fundamental transport mechanisms through individual graphene nanochannels, which has not been studied experimentally before due to fabrication and measurement difficulties.

Herein we report the fabrication of 2-D single graphene nanochannels devices and the study of ion transport through graphene nanochannels for the first time. Graphene nanochannels were prepared by standard MEMS fabrications process and wet transfer of graphene. A modified anodic bonding technique was developed to seal graphene nanochannels with well-defined geometry and uniform channel height (~20nm). Ion transport in such graphene nanochannels was studied using DC conductance measurement and compared with that in silica nanochannels. For the same geometry, our results show that graphene nanochannels and silica nanochannels exhibit the same conductance at high concentrations (10^-2M~1M) and the conductance is roughly proportional to the bulk concentration inside the reservoirs, indicating negligible changes of ionic concentrations and electrophoretic mobilities in graphene nanochannels compared with silica nanochannels. However, at low concentrations (10^-6M~10^-4M) where surface-charge-governed ion transport occurs, ionic conductance of the graphene nanochannels is always higher than that of the silica nanochannels, which suggests that there is either a higher ionic concentration, or additional conductance path in graphene nanochannels.

Our further control experiment with graphene strips replacing nanochannels confirms that the reduction/oxidation reactions occurring at graphene/solution interface will contribute significantly to the ionic conductance. According to a 2-D Poisson-Nernst-Planck Model we built, even after considering the possible conductance enhancement brought by water slip boundary at graphene/solution interface and the charge redistribution inside channel due to surface charge on graphene surface, the conductance induced by redox reactions is still comparable to the channel ionic conductance. As it is of great challenge to decouple the redox reactions from the electrokinetic flow, the previously reported enhanced ion transport studies in GO membranes and carbon nanotubes need to be reassessed. This work expands the graphene compatibility with micro/nanofabrication and also sheds light on the ion transport mechanism in carbon nanofluidics.

Graphene oxide (GO) is an attractive material for the development of nanofluidic devices and macroscopic membranes based on nanofluidic properties. One example is low cost GO membranes for ultrafiltration of brackish, processed, and coolant water streams that operate at low energy intensity and yield high salt rejection. The ability to cofacially stack this two dimensional material results in the formation of nanoscopic flow channels defined by the average separation distance between the GO sheets that can be tailored using chemical methods. These channels have been shown to transmit significant water flow due to capillary pressures that exceed 1000 Pa while rejecting dissolved salts. The sheet surface and edge oxygen moieties that line these channels play a significant role in the rate of water transmission within the membrane. The charge density induced by these moieties result in channels that have been shown to be rectifying for ionic transport in the presence of an electric field. These attributes create an opportunity to use ionic conductivity (DC) and electrochemical impedance spectroscopy (AC) methods to probe the structure of the channels and to correlate structure with efficiency for salt rejection. In this paper, we employ scanning ionic conductance microscopy as a means of characterizing the local transport properties of GO membranes. A scanning dual barrel capillary is applied to correlate morphological information with measurements of ion transmission through the membrane. We correlate these measurements with macroscopic transport measurements to determine optimum channel structure to yield rejection for a variety of mono- and multivalent cation salts.

Sandia National Laboratories is a multi-program laboratory managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Corporation, for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-AC0494AL85000.

Graphene, a single layer of graphite, represents the first two dimensional atomic crystal. It consists of carbon atoms covalently bonded in a hexagonal chicken wire lattice. This unique atomic structure gives it remarkable electrical, mechanical, and thermal properties. However, it is the mechanical properties of this material that fascinate our group the most. It is the thinnest and strongest material in the world as well as being impermeable to all standard gases. This high strength, extreme flexibility, and unprecedented barrier properties make graphene an intriguing starting material for membrane separations. Graphene acts as a barrier for gases and liquids and represent the thinnest membrane possible (one layer of atoms) with the smallest pore sizes attainable (single atomic vacancies), and unprecedented mechanical stability. In this talk, I will provide an overview of our recent experimental work on gas and liquid ion transport through angstrom sized pores in suspended graphene membranes. These measurements help elucidate the fundamental molecular and ionic transport mechanisms in this unique material.

The use of nanoporous graphene-based films for water filtration and reverse osmosis desalination has recently emerged as a promising approach to reduce the energy consumption and capital costs of a treatment plant. Nanoporous graphene (NPG) is a two-dimensional material with very high mechanical strength, favorable surface properties, and the potential to exhibit water permeability two orders of magnitude greater than that of the membranes used commercially today. In this talk, I will highlight our computational and experimental work on the design, understanding, performance, and impact of NPG-based materials for filtration applications.

Nanoporous graphene membranes have significant potential to advance membrane technologies for gas separation, water desalination and nanofiltration. Understanding the mechanical strength of porous graphene is critical because membrane separations often involve high pressures. We studied the burst strength of CVD graphene membranes placed on porous supports at applied pressures up to 100 bar by monitoring the gas flow rate across the membrane as a function of pressure. Increase of gas flow rate with pressure allowed for extraction of the burst fraction of graphene as it failed under increasing pressure. SEM and AFM images acquired before and after the burst test were in good agreement with the gas flow rate measurements, and revealed that wrinkles in graphene were prone to failure whereas non-wrinkled areas could sustain high pressure. We also studied the effect of sub-nanometer pores on the ability of graphene to withstand pressure. The results showed that porous graphene membranes can withstand pressures comparable to or even higher than the >50 bar pressures encountered in water desalination, with non-porous CVD graphene exhibiting even higher mechanical strength. Our study shows that porous polycrystalline CVD graphene has ultra-high burst strength under applied pressure, suggesting the possibility for its use in high-pressure membrane separations.

Recently graphene oxide has been attracted due to the selective permeation of water, gas and ions. This membrane has been regarded as a promising candidate for water purification, gas harvesting and ion filtering. However, the exact mechanism of the permeance has not been clearly proved. In this talk, we present the graphene oxide with deformed shape. According the change in the interlayer distance and pore size, we tested the selective and permeation of the membrane. The permeation and selectivity of the membrane is definitely dependent of the interlayer space and pore size. Those results are very helpful to figure out the exact pathway of the molecules. In addition, we will discuss also other effects to explain more plausible mechanism to the permeation.

Investigation of the transport properties of ions confined in nanoporous carbon is generally difficult due to the stochastic nature and distribution of multiscale complex and imperfect pore structures within the bulk material. Here we demonstrate a combined approach of experiment and simulation to describe the structure of complex layered graphene membranes, which allows their use as a unique porous platform to gain unprecedented insights into nano-confined transport phenomemna across the entire sub-10 nm scales. By correlation of experimental results with simulation of concentration-driven ion diffusion through the cascading layered graphene structure with sub-10 nm tuneable interlayer spacing, we are able to construct a robust, representative structural model that allows the establishment of a quantitative relationship between the nano-confined ion transport properties in relation to the complex nanoporous structure of the layered membrane. This correlation reveals the remarkable effect of the structural imperfections of the membranes on ion transport, and particularly the scaling behaviours of both diffusive and electro-kinetic ion transport in graphene-based cascading nanochannels as a function of channel size from 10 nm down to sub-nanometre. Our analysis shows that a range of ion transport effects that were previously observed in simple one-dimensional nanofluidic systems will translate themselves into bulk, complex nanoslit porous systems in a very different manner and the complex cascading porous circuities can enable new transport phenomena that are unattainable in simple fluidic systems.

Graphene membrane, having a staggered multilayer structure, is demonstrated to be a promising membrane for energy storage and liquid separation. The superior property of graphene membrane is owing to the exotic behaviour of fluid confined in the graphene nanochannel (< 10 nm). Unlike the 1-D nanochannel in lab-on-a-chip devices, the graphene membrane has a unique cascading nano-slit system. Understanding of ion transport in graphene membranes is very limited.

In this talk, a quantitative and statistical representative microstructure model is obtained for graphene membranes by correlating diffusion permeation experimental results with continuum simulation results. This task is achieved by taking advantages of the tuneable graphene membrane platform recently developed in our team. The crucial role of the pinhole defects in graphene sheets is revealed. Based on this structure model, comprehensive continuum simulations were performed to study the electrokinetic properties of ion transport inside graphene membranes. Comparison with direct experimental measurements leads to an interesting scaling law that correlates the relative conductance with channel size.

We find some novel electric double layer (EDL) structures, such as EDL caused by external electric field (coined as binary boundary layer (BBL)) and EDLs at the pore aperture regions. Influences of these EDL structures and the channel surface charges on the driving force distribution and ion concentration inside the cascading nanoslit system are carefully studied. With the obtained information, ion transport in graphene membranes are analysed.

To understand the unusual ion transport behaviour observed by the experiments at molecular scale, atomistic simulations for ion electro-kinetic flow through the membranes with and without surface charge were performed. The EDL structures, ion concentrations, and ion transport properties were carefully studied and compared with continuum simulations. For graphene membranes with zero surface charge, our MD simulations showed a strong BBL caused by external electric field. As a result, a novel polarized electro-osmosis flow (EOF) phenomenon is observed. Unlike the conventional EOF, the polarized EOF has two flows in opposite directions next to the two opposite walls of a slit, respectively. With surface charges, the ion transport is a combination of ion electrophoresis, conventional EOF, and the polarised EOF. In small slits, the conventional EOF is dominant for enhancing electrokinetic conductivity. Our MD simulations provide novel physical insights that are missed in continuum simulations, which should be considered in future continuum simulations and models.

Nanofluidics and nanofiltration have emerged quite recently as an intriguing interdisciplinary science, with applications to sensing, desalination and efficient energy storage and conversion technologies. Water confined in carbon nanotubes (CNTs) exhibits unexpected properties such as fast conduction rates and a variety of structural and phase transitions. Recently, pores in graphene also receive a significant attention as an ultimate membrane for various gas and fluid filtering applications. In this talk, I will present a thermodynamic, structural, and nonequilibrium analysis of water confined/flowing through the CNTs and graphene pores using molecular dynamics simulations.

Because of their exceptional electrical, thermal, mechanical, and fluidic properties, carbon nanotubes provide ideal 1-D nanofillers for composite fabrication. CNT type, orientation, and concentration dictate the nanocomposite properties and, for several applications (e.g. thermal interfaces, dry adhesives, and films with anisotropic electrical conductivity), vertical alignment of nanotubes is critical to meet performance requirements. In particular, to exploit CNT as fast mass-transport conduits in highly permeable and selective membranes, small-diameter SWNTs need to be organized in high density, vertically-aligned arrays that span the entire composite thickness. Unfortunately, achieving high loading of long, well-oriented CNTs in a large scale membrane remains a challenge.
Here, we demonstrate cm², free-standing, flexible SWNT/parylene membranes with well-aligned nanotubes at loadings 10x greater than previously reported films. To fabricate these membranes, 1-3 nm SWNT forests with number densities surpassing 10¹²/cm² are grown on a 4-inc wafer with a LPCVD process. A following highly conformal coating with parylene-N produces pinhole-free composites that, after opening the CNT pores, are tested for their transport performances. Pressure-driven water and gas permeability measurements confirm that narrow SWNTs sustain flow rates exceeding predictions of continuum hydrodynamics models for liquids and Knudsen theory for gases by several orders of magnitude. In addition, these membranes fully reject test analytes with dimensions larger than the SWNT diameters without loss of their outstanding water permeances. Finally, recorded water-vapor diffusion rates surpass state-of-art breathable fabrics at all relative humidities and match those of ePTFE. These results provide a step forward toward realizing the promises of CNTs for next generation high-performance membranes.

This work performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.

Carbon nanotube (CNT) membranes offer an ideal platform to test physical mechanisms of unique nanofluidic transport and to devise breakthrough design of membrane technology for setting new trade-off between permeance and selectivity. In this work, we explore the osmotically driven transport of water and salt ions through a CNT membrane using a titania matrix, designed for seawater desalination applications. Our forward-osmosis (FO) characterization shows that intrinsic properties of CNT pores, namely highly water-permeable microporous channels, and the densely charged titania surface altogether provide salt rejection as high as polymeric FO membranes at one-order-of-magnitude larger water permeance. This freestanding CNT membrane turns out to be a great solution for eliminating an internal concentration polarization (ICP) problem, thereby maintaining the FO membrane performance even at high salinities. It is electrostatic repulsion that is responsible for the rejection of salt ions by microporous CNT channels embedded in the highly charged matrix, supported by clear dependency on solution pH. Our findings provide the ground for an optimal design of FO membranes toward improved performance.

Recently, we developed ultrathin nanofiltration membranes of diamond-like carbon (DLC). DLC membrane deposited on a porous inorganic substrate has many pores of about 1 nm and they are very stable for various organic solvents, acids and bases, and redox reagents. Small organic solvents, such as ethanol, toluene, chloroform, hexane, etc. could permeate through the 1 nm pores surprisingly fast and the membranes showed the rejection more than 90% for azobenzene. The permeation rate increased inversely proportional to the viscosity of the solvents, indicating that Darcy's Law is applicable to 1 nm pores. Separation performance of these DLC membranes was three orders of magnitude higher than that of commercially available organic-solvent-resistant nanofiltration membranes. Pore size of DLC membranes could be decreased to 0.5 to 0.6 nm by choosing the deposition parameters, then the membranes showed rejection properties for MgCl2 or NaCl. Some membranes showed high rejection rates for branched alcohols and some showed very high flux for toluene. One unclear point is thickness dependence of water permeation. Many carbon-based membranes do not show a simple increase of the flux with decreasing thickness. We have carefully studied the structure, size distribution, and chemical and physical properties of sub-nanometer pores in DLC membranes. Their potential in separation science will be discussed.

Carbon Nanomembranes (CNMs) are extremely thin (~1nm), synthetic two-dimensional (2D) layers or sheets with tailored physical, chemical or biological function. With their two opposing surfaces they interface and link different environments by their distinct physical and chemical properties, which depend on their thickness, molecular composition, structure and the environment on either side. Due to their minute nanometer thickness and 2D architecture, they can be regarded as "surfaces without bulk" separating regions with different gaseous, liquid or solid components and controlling any materials exchange between them [1].
A universal scheme for the fabrication of functional CNMs is presented [2,3]. Its first step is the formation of a monolayer of aromatic molecules on a solid surface. This precursor layer is exposed to electrons or UV, which leads to a dehydrogenation, followed by a 2D cross-linking between neighboring molecules. The cross-linked monolayer is then released from the surface, forming a self-supporting CNM with properties that are determined by the precursor molecules. CNMs can then be further processed, for example perforated or surfaces functionalized [4]. It will be shown that CNMs can be engineered with a controlled thickness, elasticity, conductivity and permeability [5]. CNMs are also tested as ballistic membranes for the separation of gases and liquids [6]. Helium ion microscopy, spectroscopic methods and functional tests are applied to investigate the structure and composition as well as permeation properties of the CNMs.
[1] D. Anselmetti, A. Goelzhaeuser, Ang. Chem. Int. Ed., DOI: 10.1002/anie.201406789 (2014).
[2] A. Turchanin, A. Goelzhaeuser, Prog. Surf. Sci., 87, 108 (2012).
[3] P. Angelova, H. Vieker, N. Weber, D. Matei, O. Reimer, I. Meier, S. Kurasch, J. Biskupek, D. Lorbach, K. Wunderlich, L. Chen, A. Terfort, M. Klapper, K. Muellen, U. Kaiser, A. Goelzhaeuser, A. Turchanin, ACS Nano 7, 6489 (2013)
[4] Z. Zheng, C. T. Nottbohm, A. Turchanin, H. Muzik, A. Beyer, M. Heilemann, M. Sauer, A. Goelzhaeuser, Angew. Chem. Int. Ed. 49, 8493 (2010).
[5] X. Zhang, C. Neumann, P. Angelova, A. Beyer, and A. Goelzhaeuser, Langmuir 30, 8221 (2014).
[6] M. Ai, S. Shishatskiy, J. Wind, X. Zhang, C.T. Nottbohm, N. Mellech, A. Winter, H. Vieker, J. Qui, K.J. Dietz, A. Goelzhaeuser, A. Beyer, Adv. Mater. 26, 3421 (2014).

Many amphiphilic diblock and triblock copolymers readily self-assemble into lamellar membranes and single bilayers. Unlike their lower molecular weight, biological and synthetic lipid counterparts, polymer membranes are structurally tough and mechanically robust. Polymer membranes are typically thicker, have significantly lower water permeability characteristics, and offer bending resistance over a wide range making them attractive, tunable alternative for demanding applications requiring exposure to physical stresses. Some examples include fieldable biosensors, drug delivery, and filtration. Previous attempts to engineer the permeability of these membranes include (1) insertion of protein channels to improve permeability, (eg. Aquaporins), (2) changing the polymer constituent to an inherently leaky polymer (eg. polystyrene-b-polyisocyanoalanine(2-thiophene-3-yl-ethyl)amide (PS-PIAT)), and (3) removal of lipids from a mixed lipid-polymer vesicle following cross-linking and hydrolysis . Here, we report a new strategy for improving permeability of polymer-membranes by incorporation of carbon nanotube porins into the polymer matrix. The CNT pores exhibit the excellent transport properties inherent to carbon nanotubes, but are hydrophobic and can be tailored to match the hydrophobic thickness of polymer or lipid bilayers thus acting as protein channel mimics. We report the permeability of diblock co-polymer membranes with and without these CNT pores and characterize the enhancement of permeability and ion transport characteristics of these novel membrane constructs.

Fast water transport through carbon nanochannels, including carbon nanotubes, graphene, and graphene oxides, has received considerable attention recently. In contrast, we reported impeded water transport through interstitial space of vertically-aligned multi-walled carbon nanotubes (VAMWNTs) [1-4]. The intrusion pressure of water was ~23 times higher than that of oil. The mechanism was based on superhydrophobicity and superoleophilicity of VAMWNTs [1]. The superhydrophobicity of VAMWNTs was also employed to retard water vapor transport [2]. Although the characteristic channel dimension was ~300 times greater than the target molecule size, water vapor transport was effectively suppressed at room temperature while other gas molecules exhibited high permeability [2]. The mechanism was explained by the capillary theory [3]. As the nanochannel size decreased, water vapor transport was further inhibited by the increased backward Laplace pressure. The water rejection rate was as high as 90%, and the separation factor (He/H₂O) was 10 [3]. The water vapor transport could be further reduced by enhancing condensation at the entry region of VAMWNTs [4]. Water vapor condensation was dramatically enhanced at a decreased temperature. This resulted in mushroom-shaped liquid meniscus at the entry region of VAMWNTs and increased backward Laplace pressure. Finally, the condensed water droplets on the tip of superhydrophobic VAMWNTs could be easily removed [4]. The recent progress and the transition mechanism from impeded to fast water transport will also be discussed. [1] Carbon, 48, 2192 (2010) [2] ACS Nano, 6, 5980 (2012) [3] Nanotechnology, 26, 235701 (2015) [4] Nanoscale, 7, 14316 (2015)

The investigation on the selective mass transport properties of graphene-based materials remains to be a hot topic in recent years. Numerous novel mass transport properties in relation to the unique structure of graphene have been discovered, which are not possible in the state of art commercial membrane materials.
Our group has devoted to research on selective mass transportation through GO membranes since 2012. Despite the significant progresses on the uncovering of the novel mass transport properties of graphene-based membranes, their inherent properties still need to be enhanced when it comes to the application as practical filtration and separation membranes, for example, the mechanical strength for resisting high pressure, the long-time stability, etc. which needs more efforts to be devoted into this specific research area.
This presentation deals with the latest developments, including experimental discoveries and theoretical results, on understanding the novel mass transportation through graphene-based membranes, to cast a glance to the future applications in the fields of filtration, separation, water desalination, proton conductor and energy storage, etc. Based on the progresses made by our group and others, the latest developments on understanding the novel mass transportation through graphene-based membranes, including perfect graphene lattice, nanoporous graphene and GO membranes are overviewed in relation to their potential applications and a summary and outlook is provided to point out the opportunities and challenges in this arising field. The aspects discussed in this presentation may enable researchers to have an intimate knowledge of the present state of this topic, to better understand the mass transport mechanism and to optimize the structural design of graphene porous materials toward controllable membranes production.

Atomically thin graphene’s high tensile strength and ability to retain pore structure allow it to stand out as promising material for next-generation separation membranes. Inspired by molecular dynamics simulations predicting high permeability and selectivity of graphene for water over solutes, experimental approaches to investigate water and molecular transport demonstrated the possibility of harnessing monolayer graphene as water purification membranes. Nevertheless, experimentally measured water permeability across single-layer graphene had been lower than the upper limits provided by the theoretical predictions and requires optimization of the porosity. Herein, we present two-step process to introduce a high density of sub-nm pores in graphene and enhance the inherent water permeability of graphene membranes. We study the combination of gallium ion bombardment to nucleate initial defect sites and oxygen plasma to etch the defects into water-selective pores in single layer graphene on porous polycarbonate track etch supports. The measured water flux driven by forward osmosis indicated that the nanoporous graphene exhibits high water permeability comparable to or exceeding conventional nanofiltration membranes. Rejection of molecules over 1 nm and multivalent ions further demonstrate the feasibility of high-flux graphene nanofiltration membranes.

Graphene oxide is a chemically rich 2D material that can be assembled into membranes with high water vapor permeation and low permeation for other gases. This unusual behavior may be useful for recovery of fresh water from contaminated sources, drying of biofuels and flue gas streams, and other industrially important applications if the membranes can be made stable under use conditions and if scalable methods can be developed to produce graphene oxide membranes at low cost and high volumes. The hierarchical structure of graphene oxide membranes gives rise to the outstanding water separation behavior must also be preserved in any effective fabrication method. Initial literature reports of graphene oxide membrane production describe vacuum filtration of low-concentration suspensions to produce membranes limited to a few square inches. Here we report progress in the development and demonstration of methods compatible with continuous production of functional graphene oxide membrane toward industrial application of this exciting material. In our experience, graphene oxide synthesis methods that maintain large-diameter graphene oxide flakes (greater than 100 micrometers) and achieve high oxygen-to-carbon ratios (above 40%) assist in the production of stable aqueous suspensions and produce high quality membrane films. Our team is using advanced simulation methods to understand the effects of oxidation group content and placement in flake assembly on water transport through our graphene oxide membranes. To confirm commercially-relevant performance of our scalable membranes we are determining water vapor permeance behavior in gas mixtures as a function of membrane thickness, both on polymer-supported and on freestanding membranes.

Permeation through nanometre-pore materials has been attracting unwavering interest due to fundamental differences in governing mechanisms at macroscopic and molecular scales, the importance of water permeation in living systems, and relevance for filtration and separation techniques. Graphene-based materials can have well-defined nanometer pores and can exhibit low frictional water flow inside them, making their properties of interest for filtration and separation. I will discuss our most recent researches on the permeation properties of graphene based membranes and its prospect for various applications.

We find that electricity can be generated in graphene as it is dipped - ‘waved’ - in and out of an ionic solution, such as seawater, producing the so called waving potential¹. We also find that drawing a droplet of ionic solutions over the surface of graphene can induce electric potential in the graphene, the so called drawing potential². The results extend centuries’ old theories of electrokinetic effects and help understand the behaviour of carbon nanomaterials in liquids, which has been subject to conflicting reports for over a decade³. The drawing potential increases linearly with the drawing velocity and can be scale up by drawing multiple droplets, and has been used to demonstrate energy harvest from dropping droplets, sense the handwriting on graphene and stimulate a sciatic nerve of a frog. The waving potential, which is proportional to both the speed and the width of the graphene sheet, can also be scaled up by connecting multiple graphene sheet in series or parallel for possible applications in self-powered functional sensors such as tsunami monitors to wave energy harvest, monitors and remote ocean devices.
We also found that gas flow can induce electric voltage proportional to the square of Mach number in graphene sheets⁴.
The speaker would also like to share their understanding and theoretical and experimental research experience in other graphene related two-dimensional materials^5-7 and our most recent progress along the line.

References
1. Jun Yin, Zhuhua Zhang, Xuemei Li, Jin Yu, Jianxin Zhou, Yaqing Chen & Wanlin Guo, Waving potential in graphene. Nature Communications 5, 3582 (2014 May 6).
2. Jun Yin, Xuemei Li, Jin Yu, Zhuhua Zhang, Jianxin Zhou & Wanlin Guo, Generating electricity by moving a droplet of ionic liquid along graphene. Nature Nanotechnology 9 (5), 378-383 (2014 May).
3. J. Yin, Z. H. Zhang, X. M. Li, J. X. Zhou, W.L. Guo, Harvesting Energy from Water Flow over Graphene? Nano Lett., 2012, 12 (3), 1736–1741. http://nanotechweb.org/cws/article/tech/57149.
4. Wanlin Guo et al. Applied Physics Letters 99, 073103 (2011)；100, 183108 (2012).
5. Xiaofei Liu and Wanlin Guo, et al., Nature Communications 4, 1776 (2013).
6. Wanlin Guo*, Xiaofei Liu, Nature Nanotech. 9, 413 (2014).
7. Xiaofei Liu, Douxing Pan, Yuanzhou Hong, Wanlin Guo*, Bending Poisson effect in two-dimensional crystals. Physical Review Letters 112 (20), 205502 (2014).

Graphene nanopores can act as molecular sieves for gas molecules, providing high separation factors. Owing to its atomic thickness, the permeance of nanoporous graphene has the potential to exceed that of existing gas separation membranes by orders of magnitude. These attributes make nanoporous graphene an ideal membrane material for gas separation, with potential applications including natural gas purification, hydrogen production, and carbon capture. However, constructing selectively permeable, macroscopic area, single layer graphene membranes has been limited by material quality. Membranes produced by transferring macroscopic areas of CVD graphene have micron and nanometer scale defects. These imperfections produce non-selective gas leakage that hinders overall membrane performance. We conduct gas permeance experiments on single and few layer graphene membranes to understand the leakage pathways in large area graphene membranes. We then investigate approaches to seal or mitigate the effects of micron and nanometer scale defects in graphene. Finally, we explore methods of creating a high density of nanopores that are selectively permeable to certain gas molecules. Membrane structure is characterized by microscopy and modeling, while membrane performance is assessed based on gas permeance and selectivity measurements. This work provides insight into gas flow through nanoporous graphene membranes and guides their future development.

Porous 2D membranes have recently drawn great attention because the atomically thin membranes promise ultimately high fluxes. This high flux can be translated to low mass-transport impedance caused primarily by a very short path of interaction (friction) between transporting particles and membrane material. The extraordinary strength and chemical stability of graphene have enabled demonstration of this efficient transport mechanism across the ultrathin membrane material. However, missing is a reliable manufacturing option for large-scale porous graphene manufacturing that can satisfy both parallel process of perforation and size-cut-off control of pores. Here, we present a reliable, parallel-perforation method for graphene membrane manufacturing via block copolymer nanolithography. We will show size-cut-off control of pores in an ultrafiltration (>10 nm) range on graphene of tens of square centimeters in area. Transport characterization confirms ultrapermeable property of porous graphene for gases and liquids. Our successful, area-scalable manufacturability may advance membrane-technology applications.

Atomically thin membranes have generated a lot of interest in filtration and separation applications since they offer the ideal opportunities to study and elucidate the complex dynamics and transport phenomena at the nanometer length scales. 2D materials like graphene offers the minimum theoretical membrane resistance along with the opportunity to tune pore sizes at the nanometer scale in complete contrast to solution-diffusion of molecules through conventional membranes.
Amongst the several reported methods of 2D material synthesis, chemical vapor deposition (CVD) has emerged as the preferred route towards scalable, cost effective synthesis. Here we show selective molecular transport through sub-nanometer diameter pores in graphene grown via CVD. We develop a simple, cost effective and fast characterization technique to assess the quality of membrane grade 2D materials. A combination of pressure driven and diffusive transport measurements shows evidence for size selective transport behavior which can be used for separation by dialysis for applications such as desalting of biomolecular or chemical solutions.
Kidambi et al. (manuscript in preparation)
O’Hern et al. Nano Letters (2015).
Kidambi et al. Chemistry of Materials (2014).
Boutilier et al. ACS Nano (2014).
Kidambi et al. Nano Letters (2013).
O’Hern et al. Nano Letters (2013).
O’Hern et al. ACS Nano (2012).