Tunable Ion Sieving of Graphene Membranes ... - ACS Publications

Subscriber access provided by UNIV OF NEWCASTLE

Communication

Tunable Ion Sieving of Graphene Membranes through the Control of Nitrogen-bonding Configuration Jun-ho Song, Hye-Weon Yu, Moon-Ho Ham, and In S. Kim Nano Lett., Just Accepted Manuscript • DOI: 10.1021/acs.nanolett.8b01904 • Publication Date (Web): 06 Aug 2018 Downloaded from http://pubs.acs.org on August 9, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Tunable Ion Sieving of Graphene Membranes through the Control of Nitrogen-bonding Configuration Jun-ho Song, † Hye-Weon Yu, † Moon-Ho Ham, †† In S. Kim*, † †

Global Desalination Research Center (GDRC), School of Earth Sciences and Environmental

Engineering, Gwangju Institute of Science and Technology (GIST), 123 Cheomdangwagi-ro, Buk-gu, Gwangju, 61005, Korea ††

School of Materials Science and Engineering, Gwangju Institute of Science and Technology

(GIST), 123 Cheomdangwagi-ro, Buk-gu, Gwangju, 61005, Korea

Corresponding Author *E-mail: [email protected]

1 ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 27

ABSTRACT

Graphene oxide (GO) membranes with notable ionic sieving properties have attracted significant attention for many applications. However, the swelling and unstable nanostructure of GO laminates in water results in enlarged interlayer spacing and low permeation cut-off, limiting their applicability for water purification and desalination. Herein, we propose novel nitrogendoped graphene (NG) membranes for use in tunable ion sieving that are made via facile fabrication by a time-dependent N-doping technique. Doping reaction time-associated variation in atomic content and bonding configurations strongly contributed to the nanostructure of NG laminates by yielding narrower interlayer spacing and a more polarized surface than GO. These nanostructural features subsequently allowed ion transport through the combined mechanisms of size exclusion and electrostatic interaction. The stacked NG membranes provided size-dependent permeability for hydrated ions and improved ion selectivity by 1-3 orders of magnitude in comparison to GO membrane. For ions small enough to move through the interlayer spacing, the ion permeation is determined by electrostatic properties of NG membranes with the type of Nconfiguration, especially polarized pyridinic-N. Due to these properties, the NG membrane functioned as an unconventionally selective graphene-based membrane with better ion sieving for water purification.

KEYWORDS: Nitrogen-doped graphene membrane; ion sieving; doping reaction time; electrostatic interaction; pyridinic-N.


Page 3 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Membrane-based filtration is regarded as an important technology, due to its comparatively environmentally friendly nature and relatively small energy requirement for water purification.1-4 The trade-off between water permeability and ion selectivity of conventional membranes, however, can restricted its utilization.5, 6 Recently, a number of research groups have studied graphene-based membranes for not only gas separation but also water treatment and desalination.1, 3, 7-14 The current focus on graphene has been initiated from its physicochemical properties, which include a high intrinsic strength, specific surface area, and electron and thermal mobility despite its atomic thickness.9, 15-19 Even with its advantages as a membrane, the use of pristine graphene involves risks and challenges including arduous output, difficult massproduction, and inordinately high prices.20-22 Even worse, its impermeable structure becomes an impediment for use as a satisfactory membrane due to the high density of the electron cloud formed by sp2-bonded carbon structures in its hexagonal honeycomb lattices.19 Researchers have now directed their attention to graphene oxide (GO), which is prepared from the chemical exfoliation of graphite and which forms a special angstrom-range waterway in time for stacking.14 This structure has the potential to facilitate a more rapid mass-production at a relatively lower price than pristine graphene.20,

21

GO, having various oxygen-containing

functional groups (e.g., hydroxyl, epoxy, carbonyl, carboxyl groups, etc.), can readily permeate water due to its surface hydrophilic property.23 However, its durability and stability cannot be guaranteed in water systems. Joshi et al.24 reported molecular sieving through stacked GO membranes having diverse cations and anions. They proved that the sieving effect in stacked GO membranes is affected by the ion hydrated radius, and that it has an especially sharp cutoff at ~4.5 Å. When GO membranes get drenched, GO in presence of functional groups releases, subsequently increasing between two adjacent GO sheets (called interlayer spacing and 3 ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 27

nanochannel) to permit small ions to penetrate. This permeability indicates that GO membranes having a large interlayer spacing are insufficient for attaining acceptable selectivity. To overcome these problems, GO must decrease the interlayer spacing by transforming its structure, by removing oxygen-containing functional groups via reduction. Ironically, the existence of negatively charged oxygen functional groups on GO sheets could induce electrostatic interactions with charged molecules or ions.25, 26 The resultant potential at the solute and membrane interface has a tendency to interrupt penetration. In these studies, controlling the interlayer spacing and electrostatic interaction potential are issues in the ion transport of graphene-based membranes. One method for satisfying the above conditions is nitrogen-doping (N-doping) on the GO via a hydrothermal treatment. N-doping is the method that nitrogen was incorporated into carbon network with three types of binding configurations including pyridinic-N, pyrrolic-N, and graphitic-N, thereby influencing electronic, physicochemical, and structure properties. The presence of an extra valence electron, and larger electronegativity of N (3.04 on the Pauling scale) than those of C (2.55 on the Pauling scale) could enable to generate polarization, charge distribution, and electron density by producing activated sites on the carbon network.27-31 The pyridinic-N-like bonding configuration, connecting adjacent two C atoms and donating one p electron to the π system, accompanied by the strongest electron deficiency for taking electrons despite being smaller,32 while maintaining electrical properties.33 Luo et al.31 reported that the electronegative N atoms degrade the electron density of C atoms in C-N bond, and transmute into polarized C (δ+) from the C atoms. The polarized pyridinic-N (δ-), due to high electrons density, may have a higher energy barrier that could be induced high repulsion interaction with other molecules moving towards the adjoining C (δ+) atoms. Xie et al.34 reported that the small 4 ACS Paragon Plus Environment

Page 5 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

atomic percent of pyridinic-N could improve the sheet resistance, as the pyridnic-N electrode exhibited a high initial charge capacity and conductivity. In addition, NG is able to provide a narrower interlayer spacing, with atomic structure deformation on the plane due to the configuration of nitrogen-bonding and a reduction that synchronizes with N-doping.27, 35-37 Chen et al.36 conducted molecular dynamic (MD) simulations that posited N-doped graphene sheets mainly having pyridinic-based structure preserved outstanding properties such as intrinsic strength, cyclic stability, and thermal stability. Although simulations, their results demonstrate that N-doped graphene membrane exhibit a higher water flux than polymeric RO membranes, and that pyridinic nanoporous graphene with partial N-doping, in particular, displaying excellent salt rejection. We propose that NG membrane might also promote the selective ion permeation capability with size exclusion by narrow interlayer spacing and attractive electrostatic interaction10, 12, 15, 38 (Figure 1a). To date, however, no study has been reported to demonstrate selective ion permeation and rejection through NG membranes. We investigate that tunable ionic sieving through stacked NG membranes through the control of bonding configuration and interlayer spacing.


Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 27

Figure 1. (a) Proposed concept of ion sieving through NG membranes, according to the combined effect of electrostatic interactions (attraction and repulsion) between the negatively polarized NG surface and hydrated ions (top), and size exclusion of hydrated ions through the interlayer spacing (bottom). (b) Schematic of the process of nitrogen-doping from GO. (c) Fabrication process of an NG membrane, digital photos of an NG nanosheets coated on an PTFE membrane, and SEM images (planar, and cross-section view) of an NG membrane. 6 ACS Paragon Plus Environment

Page 7 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

NG was synthesized using the hydrothermal treatment method by changing the doping reaction time, by varying the nitrogen-bonding configuration on NG (Figure 1b, and see Method in Supporting information). Figure 1c will be discussed after characterization. To verify N-doping on the graphene lattice for NG powders (NG-p-x, where x represents doping reaction time), high-performance X-ray photoelectron spectroscopy (HP-XPS; K-Alpha+, ThermoFisher Scientific Inc., USA) was used, and Al Kα (hν = 1486.6 eV) excitation was measured.


Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 27

Figure 2. Characterizations of GO and NG powder samples. (a) XPS survey of GO (GO-p = NG-p-0) and NG powders (NG-p-x). (b) Atomic percent of GO-p, and NG-p-x. (c) High resolution of N 1s spectra of NG-p-24 sample, as a representative. (d) Detailed deconvolution analysis of the nitrogen-bonding configuration in NG-p-x. As depicted in Figure 2a, the XPS survey patterns of the NG powders clearly indicated the existence of N 1s (around 400 eV) with C 1s (around 285 eV) and O 1s (around 582 eV), whereas there was no detectable N 1s peaks in the GO powder. In addition, the atomic percent of NG powders revealed successful N-doping, which was within 7.0–8.0% in the experimental range. N-doping by changing the doping reaction time at 100 °C in hydrothermal treatment was seen to be limited in terms of quantity, accompanied by exponential-like reduction, corresponding with C/O and N/C ratio (Figure 2b and S2). It is consistent with the tendency of the N-doping process to react to the defective GO with carbonyl, carboxylic and hydroxyl groups.27, 35-37, 39 A detailed deconvolution of N 1s showed that NG powders contained three major peaks: ~ 398.5 eV (pyridinic-N), ~ 399.9 eV (pyrrolic-N), and ~ 401.2 eV (graphitic-N) (Figure 2c and S3). Pyridinic-N is formed when nitrogen atoms bond with two C atoms at the edge of a carbon plane or defect of graphene, leaving a localized electron pair perpendicular to the graphitic π-network.40 Pyrrolic-N has achieved a dominant position in NG powders. Among nitrogen-bonding configurations, graphitic-N, being located at a high binding energy indicates that the nitrogen atoms substituted for carbon atoms on a basal plane. In particular, the pyridiniclike atomic structure contributed not only to an electron-donating ability leading to improved conductivity and capacitance but also hexagonal holes or defects that conserve cyclic stability, intrinsic strength, and thermal stability.27,

35, 41, 42

In the Figure 2d, the pyridinic-N content

increase up to 2.5 at%, with increases in the doping reaction time, whereas the pyrrolic-N, and 8 ACS Paragon Plus Environment

Page 9 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

graphitic-N show either a minimal decrease or irregularity. This implies that the high pyridnic-N with lone pairs produces a large amount of defects and active sites that induces the negative charge surface.43 These results suggest that longer doping reaction time affords more pyridinic-N incorporated into the carbon network, and our process is beneficial to the formation of pyridinicN. From the XPS results (Figure 2, and S2-4), we confirmed that not only the degree of Ndoping and reduction but also nitrogen-bonding configuration have been changed and relation with doping reaction times. We posit here that the resultant bonding configuration plays a role in inducing NG sheets to have surface charge from localized valence electrons and polarization that affect electrostatic interactions on the plane.


Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 27

Figure 3. Characterization of GO and NG-x membranes. (a) Normalized XRD spectra of buckypaper of NG-x in dry and wet states, compared with that of GO. (b) Interlayer spacing of NG-x membranes calculated based on XRD pattern. The vertical error bars correspond to a halfwidth from the diffraction peaks. The blue dash denote the carbon atom van der Waals diameter (Dvdw = 3.4 Å). (c) Sheet resistance of NG synthesized by doping reaction time. The value of GO is 9.05 x 106 ± 3.6 x 104 Ω/sq.

Following the vacuum filtration method, flexible NG membranes were prepared on PTFE (Figure 1c, and see Method in supporting information). The finished membranes were first characterized, and then tested for ion sieving. Previous studies settled that the interlayer spacing is the dominant separation factor for size exclusion in the stacked GO membrane.24, 44 They asserted that the ions and molecules move through the nanochannels and are sieved depending on channel size. As seen in Figure 3a, using an X-ray diffraction (XRD) analysis, we examined the interlayer spacing of each buckypapers in dry and wet states to verify the transition of nanochannel size in the stacked membranes. The XRD spectra of the samples was obtained using an 3D high resolution x-ray diffractometer (3D-HR-XRD; Empyrean, PANalytical, Netherlands) with a Cu Kα (λ=1.54 Å) source in the range of 5° to 50°. For wet state, the buckypapers were pre-soaked in deionized water. In case of GO, while the interlayer spacing increased to 10.80 Å from 9.68 Å when wet, the spacing of NG remained constant. The stacked GO layers likely disrupted the alignment when wet, indicating a peak shift toward the left. This shift is because the existence of oxygen-functional groups on the lattice allows water molecules to easily seep into GO layers, similar to a capillary phenomenon.24 On the other hand, in the NG samples, the GO peak at 9.13° disappeared and shifted right toward 25°, suggesting that the graphitic 10 ACS Paragon Plus Environment

Page 11 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

structure was restored after N-doping due to reduction. By increasing the doping reaction time, the interlayer spacing of NG in a wet state declined from 3.62 ± 0.392 Å to 3.55 ± 0.426 Å, indicating that the interlayer spacing depends on the reduction rate, even under swelling conditions. It implies that N-doping induce to demolish hydrogen bonds, and limit water intake in NG.44 Figure 3b shows that the interlayer spacing of 3.55 ± 0.426 Å for the wet NG-24 is similar to that of 3.4 Å for the graphite structure, which is the carbon atom van der Waals diameter (Dvdw).15 According to theoretical research, the π-orbitals of graphene form a dense and delocalized cloud that impedes any small ions that attempt to pass through a defect-free graphene lattice.19 Following the molecular dynamics, Cohen-Tanugi et al.45 suggested that critical pore size in nanoporous membrane is around 5.5 Å. In our stacked form, our finding that all the peaks in the NG samples were not pointed implies that the resultant products of N-doping had a poorer crystallite and amorphous structure than GO, which has a sharp peak with a high intensity. The broad and low-intensity curves of the NG membranes can be attributed to shortrange ordering, dislocations, crystal lattice defects, and impurities as a result of N-doping, consistent with previous results.46-49 To demonstrate the N-doping effect on surface charge, the sheet resistance of the NG membrane were simply measured using a four-point probe (CMT-SR2000, Changmin Tech Co., Ltd, Korea). Lihui et al.50 revealed that the sheet resistance is monotonically decreasing by increasing fixed charge density. Werner et al.51 also confirmed that the sheet resistance determined as a function of surface charge density. In this way, we could indirectly predict the electrostatic interaction between ions and the membrane surface charge, as an ionic selectivity. Figure 3c shows that the sheet resistance decreases of NG membranes with the doping reaction time and much smaller than that of the GO membrane, leading to a two orders of magnitude 11 ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 27

variation due to the N-doping effect. In spite of a similar reduction degree obtained from the XPS analysis (Figure 2b) and interlayer spacing obtained from the XRD analysis (Figure 3b), the sheet resistance of NG-24 displays better electrical properties than that of NG-18. The only difference is pyridinic-N content, similar to previous study.33 We confirmed that the NG with increasing doping reaction time also have a positive influence on the surface charge, which leads to enhance electrostatic interaction for ion selectivity. To investigate the selectivity properties of the NG membrane, ion permeation through the membranes by different doping reaction time is conducted using some representative ions (see Method in supporting information and Figure S1).


Page 13 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Figure 4. Ionic sieving. (a) Ion permeation rates through NG membrane under different doping reaction times. Ion permeation rates through the species used to penetrate were 0.1 M aqueous solutions of K3[Fe(CN)6], MgCl2, NaCl, and KCl. (Na2SO4 was used only in NG-18 and NG-24.) The hydrated diameters (DH) were taken from refs. 52-54. The dashed box shows the belowdetection limit for our measurements. (b) The permeation rates of each ions depend linearly on the interlayer spacing in a wet state without X-error. The Y-error bars indicate the standard deviation.


Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 27

We investigated the possibility that specific ions having different hydrated diameters (DH) could sieve through NG membranes based on the doping reaction time (Figure 4a). We supposed that ions in water capture a rigid space, given by their DH when considering size exclusion effects. The ions, which are small-size DH, permeate with nearly the same rate, whereas relatively large-size ions exhibit a low permeation rate, suggesting that GO membranes have ion sieving potential, as described in previous studies.24, 44, 55, 56 Figure 4a shows that different species had rates determined by DH, such as 6.62 Å for K+, 7.16 Å for Na+, 7.58 Å for SO42-, 8.56 Å for Mg2+,52 and 9.50 Å for [Fe(CN)6]3-

53

permeating through NG membranes

synthesized under different reaction times (e. g. NG-x). For GO membrane, the difference in permeation rate of monovalent cation and divalent cation is very slight. Contrariwise, trivalent [Fe(CN)6]3- ion having smaller hydrated diameter gap with Mg2+ than gap between Na+ and Mg2+ shows relatively low permeation rate. This confirms that our GO membrane was not significantly different from other GO membranes.24, 44, 56 In other word, it was analogous with the pattern of an exponential decay function, and the permeation rate for the smallest ion (K+) was between 101

~ 100 mol m-2 h-1. According to Joshi et al.,24 the rate of penetration does not show significant

dependence on ion charge since triply charged ions such as AsO43- (DH for AsO43-; 7.70 Å) permeate at nearly the same rate as singly charged Na+ or Cl- (DH for Cl-; 6.64 Å). Our result indicates that GO membrane with large interlayer spacing is not capable of sieving below 8.56 Å like size exclusion because of marginal differential value of permeation rate, regardless of ionic charge. The entire permeation rates of NG-x lead to a 1–3 orders of magnitude decrease from GO membranes, depending on the size of the DH. In general, as the doping reaction time increases, the permeation rate of each ion tends to decrease, similarly to exponential decay functions. The difference between each plot of GO and NG implies that there was a synergistic 14 ACS Paragon Plus Environment

Page 15 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

effect, involving not only size exclusion due to diminished interlayer spacing, but also improved electrostatic interaction with the polarized NG surface. In particular, a noticeable change took place in the ion permeation rate for NG-18 and NG-24. The larger species such as [Fe(CN)6]3and Mg2+ presented at levels below the limits of detection, suggesting that large ions had considerable difficulty in penetrating through those membranes. For example, NG-24 allowed the 7.58 Å SO42- ion to pass through it, whereas it did not allow Mg2+, with a diameter of 8.56 Å to pass. Based on Figure 4a, we note that NG-24 was efficacious for selective ion permeation according to the size of the hydrated ion, and we carefully deduced that the cut-off size for ionic sieving by hydrated ion size was around 8.0 Å. We also plotted the permeation rate by interlayer spacing to verify the function of interlayer spacing on size exclusion. The most interlayer spacing had a small gap among NG-x membranes, and this derivation could be overlapped, especially around 3.6 Å, corresponding with Figure 3b. As depicted in Figure 4b, the observed ion permeation rates for all ions showed a linear dependence, decreasing by 1–3 orders of magnitude as the interlayer spacing decreased. The interlayer spacing for the wet NG membrane is much smaller than the theoretical DH of the small species. Nevertheless, the results indicate that a narrow interlayer spacing has a low permeation rate for different species, instead of perfect barrier, with only large ions not being moved through an interlayer spacing of 3.56 ± 0.326 Å for NG-18 and 3.55 ± 0.426 Å for NG-24. Chong et al.57 insisted that imperfections of the laminar structure and microstructural defects generated by the reduction of GO resulted in an increased permeation rate due to the diminished length of the transport passage, corresponding to the broad and low intensity of the XRD peak. Aba et al.58 also reported that defects caused by shrinkage of the GO membrane in the drying process could serve as a short-cut to induce a high permeation rate. From the result, we insist that it is not perfect to pursue size exclusion theory to describe ion 15 ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 27

transport through graphene-based membranes.15 This has led to an investigation of the N-doping effect that includes amorphous structure and deformation of structures produced by nitrogenbonding configuration in planes, corresponding to XRD and XPS analyses.35, 59 Even if the result show linear proportional relation between interlayer spacing and ion permeation rate, these considerations further confirm that NG membranes have other causes that influence the permeation rate, than just the interlayer spacing.

Figure 5. Correlation analysis (a) between the ion permeation rate and the anion/cation valence ratio (Z-/Z+), (b) between content of pyridinic-N like bonding configuration and permeation rate: 16 ACS Paragon Plus Environment

Page 17 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

black for Fe(CN)63-, red for Mg2+, magenta for SO42-, green for Na+, blue for K+ and insert for sheet resistance. To gain insight into the impact of ion permeation through NG membranes, we carried out a comparison analysis using the anion/cation valence ratio (Z-/Z+) and nitrogen-bonding configurations. As shown in Figure 5a, the permeation rates of NG membranes with different Z/Z+ were analyzed. In general, high values of Z-/Z+ correspond to a strong net repulsion force, indicating that electrostatic interactions dominated the ionic sieving, in accordance with previous studies.9, 38 However, it was demonstrated that the permeation of the Mg2+ cation, with a small Z/Z+, through the NG membranes could be predominantly attributed to its large hydrated ion size rather than electrostatic interactions. The differential between the permeation rate of NG-18 and NG-24 with valence ratio of 2.0 and 1.0 indicates that electrostatic forces via the negatively polarized NG sheet were more significant in NG-24 than in NG-18, as the interlayer spacing was the similar. We confirmed again that relatively large ions were predominantly influenced by size exclusion rather than by the valence ratio, because a valence ratio of 0.5 slowed a slower rate than other low values. We believed that pyridinic-N inducing negative polarization could also assist the selective performance, based on electrostatic interaction. Typically, the negative charges on the functional groups of GO electrostatically repel anions and attract cations.38 In a simulation study,60 the adsorption energies of adsorbates on N-doped graphene pores were higher than on F-doped graphene pores because the N-doped graphene pores were more negatively charged than F atoms. High adsorption energies make it hard for Na+ ions to pass through the Ndoped graphene pores due to strong interactions between the pores and the ions. In the Figure 5b, the observed ion permeation rate revealed the significantly dependence on pyridinic-N. Moreover, the sheet resistance linearly decreased proportional to pyridinic-N. The relevance of 17 ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 27

other nitrogen-bonding configurations are inferior to that of pyridinic-N. The others are supposed to less correlation with permeation rate, and sheet resistance because there is little or no change with increasing content (Figure S5). We confirmed that the pyridinic-N content could differentiate the ion permeation rate between NG membranes having similar interlayer spacing. In other words, the electrostatic interactions in NG membranes, which resulted from the polarization and high adsorption energy of pyridinic-N is more influential as a mechanisms than size exclusion by interlayer spacing, except for a concurrent effect in which size exclusion was predominant for large hydrated ions. To demonstrate the potential for ionic sieving in NG membranes, NG and NG membranes were successfully prepared using facile methods (i.e., hydrothermal treatment for NG, and vacuum filtration for NG membranes). The NG obtained, having different N contents and nitrogen-bonding configurations, depends on the doping reaction time and could be used to determine characteristics that could influence ion permeation. When an NG membrane is soaked in deionized water, its interlayer spacing remained stationary around 3.56 Å, whereas the GO membrane increased to 10.80 Å from an initial value of 9.68 Å. This result is based on our doping method, which along with reduction contributes to N-doping. In addition, we conducted practical ion permeations through time-dependent NG membranes using different hydrated diameters of ions, and characterized our results. By the size of the ion-hydrated diameter, the results appears to follow size exclusion, and small ion indicate small width of decrease with increasing doping reaction time. Interlayer spacing in NG membranes is smaller than the target ions in wet state. Still, relatively small ions penetrated the NG membranes. Although this result may be considered incoherent, it is occurred because that NG membrane was composed of amorphous and imperfect structure with microstructural defects due to N-doping through a 18 ACS Paragon Plus Environment

Page 19 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

hydrothermal process. The ion permeation does not perfectly correlate with size exclusion using interlayer spacing. In this regard, to explain selectivity mechanism in the NG membranes, we found the correlation between nitrogen-bonding configuration, permeation rate, and sheet resistance. Of particular interest, one of the nitrogen-bonding configurations, pyridinic-N, could provide an answer for key of selectivity in NG membranes. Tuning the content of pyridinic-N in NG membrane allows to improve the ionic sieving based on its size exclusion and electrostatic interaction. NG membranes as an ion selective membrane opens a new chapter in graphenebased membrane technologies for water purification such as ion exchange, electrodialysis, and desalination.


Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 27

ASSOCIATED CONTENT Supporting Information The Supporting Information is available free of charge on the ACS Publications website at DOI: Material and Method, additional XPS spectra, relation with other N-configuration, and supplementary discussion (PDF)

AUTHOR INFORMATION Corresponding Author *E-mail: [email protected] ORCID In S. Kim: 0000-0002-6016-5267

ACKNOWLEDGMENT This research was supported by a grant (18IFIP-C071145-06) from the Plant Research Program funded by the Ministry of Land, Infrastructure and Transport of the Korean government. The authors would also like to give special thanks to Mr. Gyung-Gap Jeong in JCRF for assisting with XRD measurements, Ms. Mihee Jang in GDRC for ion chromatography measurements, and Mr. Chang-Min Kim in GIST for the helpful suggestions on this study.


Page 21 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

REFERENCES 1.

Vlassiouk, I. V. Nature nanotechnology 2017, 12, 1022-1023.

2.

Devanathan, R. Nature nanotechnology 2017, 12, 500-501.

3.

Jiang, Y.; Biswas, P.; Fortner, J. D. Environmental Science-Water Research &

Technology 2016, 2, 915-922. 4.

Subramani, A.; Jacangelo, J. G. Water Res 2015, 75, 164-187.

5.

Werber, J. R.; Deshmukh, A.; Elimelech, M. Environmental Science & Technology

Letters 2016, 3, 112-120. 6.

Park, H. B.; Kamcev, J.; Robeson, L. M.; Elimelech, M.; Freeman, B. D. Science 2017,

356. 7.

Yang, E.; Kim, C. M.; Song, J. H.; Ki, H.; Ham, M. H.; Kim, I. S. Carbon 2017, 117,

293-300. 8.

Yang, E.; Ham, M. H.; Park, H. B.; Kim, C. M.; Song, J. H.; Kim, I. S. Journal of

Membrane Science 2018, 547, 73-79. 9.

Sun, P.; Wang, K.; Zhu, H. Adv Mater 2016, 28, 2287-2310.

10.

You, Y.; Sahajwalla, V.; Yoshimura, M.; Joshi, R. K. Nanoscale 2016, 8, 117-119.

11.

Deng, J.; You, Y.; Bustamante, H.; Sahajwalla, V.; Joshi, R. K. Chem Sci 2017, 8, 1701-

1704.


Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

12.

Page 22 of 27

Xu, Q.; Xu, H.; Chen, J. R.; Lv, Y. Z.; Dong, C. B.; Sreeprasad, T. S. Inorganic

Chemistry Frontiers 2015, 2, 417-424. 13.

Qu, X.; Alvarez, P. J.; Li, Q. Water Res 2013, 47, 3931-3946.

14.

Liu, G.; Jin, W.; Xu, N. Chemical Society reviews 2015, 44, 5016-5030.

15.

Wang, L.; Boutilier, M. S. H.; Kidambi, P. R.; Jang, D.; Hadjiconstantinou, N. G.;

Karnik, R. Nature nanotechnology 2017, 12, 509-522. 16.

Perreault, F.; Fonseca de Faria, A.; Elimelech, M. Chemical Society reviews 2015, 44,

5861-5896. 17.

Weiss, N. O.; Zhou, H.; Liao, L.; Liu, Y.; Jiang, S.; Huang, Y.; Duan, X. Adv Mater

2012, 24, 5782-5825. 18.

Zhu, Y.; Murali, S.; Cai, W.; Li, X.; Suk, J. W.; Potts, J. R.; Ruoff, R. S. Adv Mater

2010, 22, 3906-3924. 19.

Berry, V. Carbon 2013, 62, 1-10.

20.

Stankovich, S.; Dikin, D. A.; Dommett, G. H.; Kohlhaas, K. M.; Zimney, E. J.; Stach, E.

A.; Piner, R. D.; Nguyen, S. T.; Ruoff, R. S. Nature 2006, 442, 282-286. 21.

Park, S.; Ruoff, R. S. Nature nanotechnology 2009, 4, 217-224.

22.

Bonaccorso, F.; Bartolotta, A.; Coleman, J. N.; Backes, C. Adv Mater 2016, 28, 6136-

6166.


Page 23 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

23.

He, H. Y.; Klinowski, J.; Forster, M.; Lerf, A. Chemical Physics Letters 1998, 287, 53-

56. 24.

Joshi, R. K.; Carbone, P.; Wang, F. C.; Kravets, V. G.; Su, Y.; Grigorieva, I. V.; Wu, H.

A.; Geim, A. K.; Nair, R. R. Science 2014, 343, 752-754. 25.

Hu, M.; Mi, B. Environ Sci Technol 2013, 47, 3715-3723.

26.

Sun, P.; Zhu, M.; Wang, K.; Zhong, M.; Wei, J.; Wu, D.; Xu, Z.; Zhu, H. ACS nano

2013, 7, 428-437. 27.

Wang, H. B.; Maiyalagan, T.; Wang, X. Acs Catalysis 2012, 2, 781-794.

28.

Panchakarla, L. S.; Subrahmanyam, K. S.; Saha, S. K.; Govindaraj, A.; Krishnamurthy,

H. R.; Waghmare, U. V.; Rao, C. N. R. Advanced Materials 2009, 21, 4726-4730. 29.

Majeed, S.; Zhao, J. M.; Zhang, L.; Anjum, S.; Liu, Z. Y.; Xu, G. B. Nanotechnology

Reviews 2013, 2, 615-635. 30.

Wu, P.; Du, P.; Zhang, H.; Cai, C. Phys Chem Chem Phys 2013, 15, 6920-6928.

31.

Luo, Z. Q.; Lim, S. H.; Tian, Z. Q.; Shang, J. Z.; Lai, L. F.; MacDonald, B.; Fu, C.;

Shen, Z. X.; Yu, T.; Lin, J. Y. Journal of Materials Chemistry 2011, 21, 8038-8044. 32.

Jalili, S.; Vaziri, R. Molecular Physics 2011, 109, 687-694.

33.

Song, J. H.; Kim, C. M.; Yang, E.; Ham, M. H.; Kim, I. S. Rsc Advances 2017, 7,

20738-20741.


Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

34.

Page 24 of 27

Xie, Y.; Chen, Y.; Liu, L.; Tao, P.; Fan, M.; Xu, N.; Shen, X.; Yan, C. Adv Mater 2017,

29, 1702268. 35.

Luo, G.; Liu, L.; Zhang, J.; Li, G.; Wang, B.; Zhao, J. ACS applied materials &

interfaces 2013, 5, 11184-11193. 36.

Chen, Q.; Yang, X. N. Journal of Membrane Science 2015, 496, 108-117.

37.

Li, X.; Wang, H.; Robinson, J. T.; Sanchez, H.; Diankov, G.; Dai, H. J Am Chem Soc

2009, 131, 15939-15944. 38.

Sun, P.; Ma, R.; Deng, H.; Song, Z.; Zhen, Z.; Wang, K.; Sasaki, T.; Xu, Z.; Zhu, H.

Chem Sci 2016, 7, 6988-6994. 39.

Wang, W. W.; Dang, J. S.; Zhao, X.; Nagase, S. J Phys Chem C 2016, 120, 5673-5681.

40.

Li, Q.; Zhu, W. L.; Fu, J. J.; Zhang, H. Y.; Wu, G.; Sun, S. H. Nano Energy 2016, 24, 1-

41.

Guo, H. L.; Peng, S.; Xu, J. H.; Zhao, Y. Q.; Kang, X. F. Sensors and Actuators B-

9.

Chemical 2014, 193, 623-629. 42.

Wen, Q.; Wang, S.; Yan, J.; Cong, L.; Chen, Y.; Xi, H. Bioelectrochemistry 2014, 95,

23-28. 43.

Lin, Y. C.; Teng, P. Y.; Yeh, C. H.; Koshino, M.; Chiu, P. W.; Suenaga, K. Nano Lett

2015, 15, 7408-7413.


Page 25 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

44.

Abraham, J.; Vasu, K. S.; Williams, C. D.; Gopinadhan, K.; Su, Y.; Cherian, C. T.; Dix,

J.; Prestat, E.; Haigh, S. J.; Grigorieva, I. V.; Carbone, P.; Geim, A. K.; Nair, R. R. Nature nanotechnology 2017, 12, 546-550. 45.

Cohen-Tanugi, D.; Grossman, J. C. Nano Lett 2012, 12, 3602-3608.

46.

Singh, S. K.; Dhavale, V. M.; Kurungot, S. ACS applied materials & interfaces 2015, 7,

442-451. 47.

El-Deen, A. G.; Boom, R. M.; Kim, H. Y.; Duan, H.; Chan-Park, M. B.; Choi, J. H. ACS

applied materials & interfaces 2016, 8, 25313-25325. 48.

Duan, X.; Ao, Z.; Sun, H.; Indrawirawan, S.; Wang, Y.; Kang, J.; Liang, F.; Zhu, Z. H.;

Wang, S. ACS applied materials & interfaces 2015, 7, 4169-4178. 49.

Dhavale, V. M.; Singh, S. K.; Nadeema, A.; Gaikwad, S. S.; Kurungot, S. Nanoscale

2015, 7, 20117-20125. 50.

Lihui, G.; Yibin, Z.; Tietun, S. Solar Energy Materials and Solar Cells 1996, 43, 325-

333. 51.

Werner, F.; Cosceev, A.; Schmidt, J. Energy Procedia 2012, 27, 319-324.

52.

Nightingale Jr, E. The Journal of Physical Chemistry 1959, 63, 1381-1387.

53.

Fornasiero, F.; Park, H. G.; Holt, J. K.; Stadermann, M.; Grigoropoulos, C. P.; Noy, A.;

Bakajin, O. Proceedings of the National Academy of Sciences of the United States of America 2008, 105, 17250-17255.


Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

54.

Page 26 of 27

Volkov, A. G.; Paula, S.; Deamer, D. W. Bioelectrochemistry and Bioenergetics 1997,

42, 153-160. 55.

Koenig, S. P.; Wang, L.; Pellegrino, J.; Bunch, J. S. Nature nanotechnology 2012, 7,

728-732. 56.

Hong, S.; Constans, C.; Surmani Martins, M. V.; Seow, Y. C.; Guevara Carrio, J. A.;

Garaj, S. Nano Lett 2017, 17, 728-732. 57.

Chong, J. Y.; Aba, N. F.; Wang, B.; Mattevi, C.; Li, K. Sci Rep 2015, 5, 15799.

58.

Aba, N. F. D.; Chong, J. Y.; Wang, B.; Mattevi, C.; Li, K. Journal of Membrane Science

2015, 484, 87-94. 59.

Geng, D. S.; Yang, S. L.; Zhang, Y.; Yang, J. L.; Liu, J.; Li, R. Y.; Sham, T. K.; Sun, X.

L.; Ye, S. Y.; Knights, S. Applied Surface Science 2011, 257, 9193-9198. 60.

Gai, J. G.; Gong, X. L.; Wang, W. W.; Zhang, X.; Kang, W. L. Journal of Materials

Chemistry A 2014, 2, 4023-4028.


Page 27 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Table of Contents (TOC) graphic


Tunable Ion Sieving of Graphene Membranes ... - ACS Publications

Recommend Documents