Field of the invention

The present invention pertains generally to the field of gas separation filters, especially useful to the C0 ₂ capture from the flue gas. More particularly, the present invention is related to filter membranes for C0 ₂ capture and even more particularly, the present invention relates to composite filter membranes for C0 ₂ capture.

Background of the invention

The global CO2 emissions have risen rapidly in the last decade. CO2 concentration in the atmosphere exceeded 410 ppm in early 2018. The short- to medium-term solutions to limit the increase of CO2 concentration in the atmosphere include CO2 capture, use and sequestration (CCUS). To reduce the energy-penalty of capture, which is currently prohibitively-high, the development of an energy-efficient carbon capture route is vital to control CO2 emissions. Membrane-based CO2 separation is one of most promising and energy- efficient solution because membranes do not rely on the expensive thermal energy. Membranes have been shown to be more energy-efficient than the commercial amine-based CO2 scrubbing technology for the postcombustion carbon capture (CO2 capture from the flue gas), especially from the power plants, petrochemical refineries, and steel and cement industries, where the CO2 concentration in the flue gas is higher than 15%. The proposed performance target by the department of energy in the US is as follows: a CO2 permeance (defined as pressure-normalized flux) higher than 1000 gas permeation units (GPU, where 1 GPU = 3.35 x 10 ¹⁰ mole nr ² s ¹ Pa ^-1 ) and a CO2/N2 separation factor higher than 20.

Single-layer graphene is the thinnest molecular barrier, and is highly promising for the gas separation if molecular-sized pores can be incorporated with a high pore-density in the graphene lattice. Experiments and molecular simulations have indicated that nanoporous graphene film can yield orders of magnitude higher gas permeance than that from the conventional membranes, attributing to the ultrashort diffusion path of molecules through the atom-thick membrane. However, the incorporation of molecular-selective nanopores in otherwise impermeable graphene lattice, at a reasonably high pore-density, has proven to be a major challenge.

Recently, the inventors demonstrated an efficient molecular-sieving of H ₂ from CH ₄ using the single-layer graphene membrane by developing an ozone-based etching chemistry which allowed a control over pore-diameter in the range of 0.30-0.38 nm. Sieving C0 ₂ from N ₂ is more challenging (the difference in kinetic diameters of C0 ₂ and N ₂ is 0.03 nm, which is much smaller than that between the kinetic diameters of H ₂ and CH ₄ , 0.09 nm). Until now, there are no report on single-layer-graphene based separation of gas mixture comprised of C0 ₂ and N ₂ . Multi-layered film of graphene-oxide (GO) hosting C0 ₂ -philic molecules into the interlayered nanochannels have demonstrated the separation of C0 ₂ from N ₂ . However, the gas transport pathway in GO membranes is long and tortuous, resulting in a moderate C0 ₂ permeance of 250 GPU, lower than the capture target. Moreover, the long-term stability of GO membrane is questionable.

In this regard, a primary object of the invention is to solve the above-mentioned problems and more particularly to develop a high-performance graphene-based membrane yielding performance that meets or exceeds the C0 ₂ capture target.

Summary of the invention

The present invention relates to a single-layer graphene film chemically modified with a polymer film having a thickness of 5-100 nm, preferably 5-50 nm, more preferably 5-20 nm and even more preferably 10 nm, which yields a record performance in C0 ₂ /N ₂ separation, and which meets the C0 ₂ capture target.

In order to obtain this film, the lattice of single-layer graphene was etched to incorporate porosity with pore-size (defined by electron-density-gap in the pore) in the range of 0.2-10.0 nm, preferably, 0.3-5.0 nm and even more preferably 0.3-2.0 nm. Then, a layer of C0 ₂ -philic polymer (polymer with high permeability for C0 ₂ with respect to other gases) having a thickness of 5-100 nm, preferably 5-50 nm, more preferably 5-20 nm and even more preferably 10 nm, acting as a C0 ₂ -selective film, was chemically grafted on top of the graphene surface. The C0 ₂ -philic polymer was subsequently swollen with another low molecular-weight C0 ₂ -philic polymer. The resulting membranes yielded a high C0 ₂ permeance of 6100 GPU and an attractive C0 ₂ /N ₂ separation factor of 22.5.

The main advantage of the above mentioned range of pore size from 0.3 to 2 nm is achieving high selectivity. This invention also provides a novel graphene transfer method allowing the synthesis of crack-free graphene film on to a porous substrate.

In order to enhance clarity of the present specification, the following key abbreviations for membranes developed here are as follows:

Abbreviation Description

NG Nanoporous single-layer graphene created using plasma

ONG NG after ozone functionalization

PONG Polymer layer (PEI or PEGBG) coated on ONG

PONG1 PONG prepared using PEI

PONG2 PONG prepared using PEGBA

PONG@Si PONG film transferred to Si substrate

PONGl@Si PONG@Si prepared using PEI

PONG2@Si PONG@Si prepared using PEGBA

SPONG/PTMSP PONG coated with PTMSP and swollen with PEGDE

SPONG1/PTMSP SPONG/PTMSP film where polymer layer is PEI

SPONG2/PTMSP SPONG/PTMSP film where polymer layer is PEG BA

SPONGl_X SPONG1/PTMSP membrane where plasma time for the NG was X s

SPONG2 X SPONG2/PTMSP membrane where plasma time for the NG was X s

Brief description of the drawings

Further particular advantages and features of the invention will become more apparent from the following non-limitative description of at least one embodiment of the invention which will refer to the accompanying drawings, wherein

Figure 1 is a schematic diagram representing the manufacturing process of the membrane of the present invention; Figures 2a and 2b represent Raman spectra of as-synthesized graphene and graphene exposed to various plasma times, and a plot of ID/IG and / ₂ D//G ratios as a function of plasma exposure time, respectively, c-h) Aberration-corrected transmission electron microscopy (TEM) images and the corresponding pore-size-distribution of NG prepared by plasma treatment for 4 s (c-e) and 6 s (f-h). The mean pore size corresponds to the arithmetic average of the pore diameters. The scale bars in the TEM images correspond to 2 nm.

Figure 3 represent an XPS spectra of oxidized graphene by ozone treatment of 20 min at room temperature. The binding energies of C-C, C-O, C=0, 0-C=0 are 284.4, 285.7, 286.8, 288.5 eV, respectively;

Figure 4 shows SEM images of the surface morphologies of a) ONG@Si, b) PONGl@Si, c) PONG2@Si, and d) SPONG1/PTMSP on Si wafer;

Figure 5 shows SEM images of the cross-sectional morphologies of a) bare PTMSP, b) SPONG1/PTMSP, and c) SPONG2/PTMSP on Si wafer;

Figure 6 shows a Thickness analysis of the PONG1 samples, a) Topography, and b) the corresponding height distribution of the as-prepared sample on the Si wafer, c) Topography of the trench, and d) the corresponding line profiles across the trench;

Figure 7 shows the gas separation performance of various membranes as a function of plasma time. The SPONG1 membranes under a) the single-gas permeation conditions, and b) the mixture-gas permeation conditions (CO2/N2 = 20/80 vol%). c) The SPONG2 membranes under the single-gas permeation conditions. Membranes were tested at 30 °C, and the pressure in the feed side was 2 bar;

Figure 8 shows the gas separation performance of various membranes as a function of the pore-size in graphene. CO2 permeance and CO2/N2 selectivities as a function of the mean pore size in graphene for (a) SPONG1 membranes (comparing single gas and mixture data), and (b) comparing single gas data for SPONG1 and SPONG2 membranes;

Figure 9 shows the CO2/N2 mixture separation performance of SPONG membranes compared with the state-of-the-art thin-film composite membranes. Facilitated transport membranes are not included because of their widely-known stability issues.

Detailed description of the invention

The present detailed description is intended to illustrate the invention in a non- limitative manner since any feature of an embodiment may be combined with any other feature of a different embodiment in an advantageous manner.

In the present application the following terms and expressions have the following meanings:

Pore size: electron-density-gap in graphene pore

C0 ₂ -philic polymer: these are polymer chains with high permeability for C0 ₂ with respect to other gases (for example, in this case N ₂ ).

Permeability = Solubility x Diffusivity, where Solubility is defined as amount of C0 ₂ solubilized by the polymer chains (wt%) as a function of C0 ₂ concentration in the gas phase and Diffusivity is a measure of transport rate of C0 ₂ when it hops from one site in polymer to another site. A preferred range of C0 ₂ permeability for this application is 10 -10000 barrer with C02/N2 selectivity of 10-1000.

Oxidative Groups: Here, oxidative groups specifically refer to epoxides, carbonyl, carboxylic, hydroxyl and carbonic groups.

Allowing crack-free transfer: This expression means transfer of graphene from Cu substrate to a porous substrate in a way such that there are no microscopic cracks (defined as cracks with width greater than 0.1 micron) with a tolerance limit for smaller cracks to be 10-100 ppm.

Low molecular weight: lower than 10000 g/moles.

PTMSP: an abbreviation for a polymer name (polytrimethylsilylpropyne).

We will first detail an example of a membrane manufacturing process of the present invention.

First, a single-layer graphene was synthesized on a copper foil (25 pm, 99.8% purity, Sigma) using the low-pressure chemical vapor deposition (CVD). Briefly, the copper foil was annealed at 1000 °C in a C0 ₂ atmosphere for 30 min to get rid of organic contaminations. Subsequently, C0 ₂ flow was switched off and H ₂ (8 mL/min) was introduced to anneal the copper. Then, CH ₄ (24 mL/min) flow was open for 30 min as a precursor for graphene crystallization.

Then this graphene was used to prepare an oxygen-functionalized nanoporous single- layer graphene (ONG). The nanopores on the single-layer graphene were created by an 0 ₂ plasma. A piece of the graphene/copper foil was placed in the plasma chamber in an 0 ₂ atmosphere. After the pressure reaches 50 mTorr, the 0 ₂ plasma was opened for a certain time (4-8 s) to etch the graphene lattice. The nanoporous graphene was oxygen-functionalized by placing graphene in an oxidizing atmosphere (21% 0 ₃ in 0 ₂ ) generated by an ozone generator (Absolute Ozone ^® Atlas 30), for 20 min at room temperature.

In the further step, the ONG was modified by C0 ₂ -philic polymer to create a PONG. More particularly, PONG was prepared by spin coating a dilute solution of C0 ₂ -philic polymer on top of ONG resting on the copper foil. Aqueous solution of polyethylenimine (PEI, 10 mg/ml) or polyethylene glycol) bis(amine) (PEGBA, 20 mg/ml) was used as the coating solution. The coating solution was added dropwise within 10 s while the graphene/copper substrate was spinning at 1000 rpm. Subsequently, the spin coating was carried out at 3000 rpm for 60 s.

Subsequently, we carried out a coating of PTMSP for graphene transfer (ONG/PTMSP or PONG/PTMSP). Here, a thin PTMSP layer was coated onto the top of ONG or PONG by spin coating. A thin layer of 1.25 wt% of PTMSP toluene solution was spread on the substrate, followed by spinning the substrate at 1000 rpm for 30 s, and then 2000 rpm for 20 s. The resulting film was dried in a closed dish for 12 h, and then dried in a vacuum oven for 12 h at room temperature.

The final step comprises a swelling of the composite film and transfer to a porous tungsten support (SPONG/PTMSP). Here., the copper foil acting as substrate to the ONG/PTMSP or the PONG/PTMSP film was removed by chemical etching by placing the films on a FeCU (1 M in water) bath for 30 min. Then, the underside of the floating film was rinsed with 0.1 M HCI solution for 1 h, and then on Dl water for 1 h to remove the residues. In the case of PONG/PTMSP film, the film was floated on a polyethylene glycol) dimethyl ether (PEGDE, average M _n = 500 ) aqueous solution (2 mg/mL) for 24 h to swell the PEI and PEGBA films with PEGDE. Finally, the film (referred to as SPONG/PTMSP) was scooped up using a porous tungsten support. The remaining water on the surface of the film was carefully removed by a bloating paper, followed by drying in a vacuum oven for 12 h. In the case of ONG/PTMSP film, the film was directly scooped up using a porous tungsten support. Once created, the SPONG membranes were characterized by several methods including Raman spectroscopy, X-ray photoelectron spectroscopy (XPS), Scanning electron microscope (SEM), Atomic force microscope (AFM) and Gas permeation measurement.

The Raman characterization was carried out on graphene/copper using a Renishaw micro-Raman spectroscope (457 nm, 2.33 eV, 50x objective). More than 10 spectra were obtained using the mapping method for every sample. The ID/IG and / ₂ D//G ratios were calculated by analyzing the Raman data using MATLAB. The background was subtracted from the Raman data for the calculation.

XPS was used to confirm the functionalization of graphene (ONG). The XPS analysis was performed using a Mg Ka X-ray source (1253.6 eV) and a Phoibos 100 (SPECS) hemispherical electron analyzer with a multichanneltron detector.

SEM was used to observe the surface or cross-sectional morphologies of ONG, ONG/PTMSP, and SPONG/PTMSP films, using FEI Teneo SEM with an operating voltage of 1-5 kV and a working distance of 2.5-6 mm. For the cross-section imaging, the sample was coated with a 10-nm-thick iridium layer. No conductive coating was used for observing surface morphologies.

AFM was used to detect the surface topography and thickness of PEGBA or PEI layers. For this, we transferred the graphene to a silica wafer by the conventional poly(methyl methacrylate) or PMMA based wet-transfer technique. Post-transfer, PMMA was removed by acetone wash and a heat treatment at 400 for 2 h in an Ar/H ₂ atmosphere. Subsequently, the porosity and oxygen-functionalization in graphene were generated by the similar method as described above. This was followed by spin-coating of PEI or PEGBA (same conditions as used before). These films are referred to as PONG@Si.

The single-gas and mixture gas permeation measurements were carried out in a homemade permeation cell similar to the ones described in Zhao, J. et al. Etching gas-sieving nanopores in single-layer graphene with an angstrom precision for high-performance gas mixture separation. Sci. Adv. 5, eaavl851 (2019). All flow rates were controlled by the mass flow controllers (MFC). The pressure of the feed side was controlled by adjusting the backpressure regulator. The transmembrane pressure difference was maintained at 1 bar. Ar was used as the sweep gas, carrying the permeate gas to a calibrated mass spectrometer (MS, Hiden Analytical, HPR-20) for the real-time analysis of the permeate concentration. The feed and sweep gas lines, and the membrane cell were heated in a convection oven to precisely control the temperature. The permeate composition data was recorded and averaged after the steady-state condition was attained (typically 30 minutes after changing the conditions). The permeances, J _it of the gas i were calculated by Equation 1

Ji = Xd(A DR _Ϊ ) Equation 1 where X _t is the molar flow rate of component /, A is the membrane area, and DRi is the transmembrane pressure difference for component /. The separation selectivity a _tj of the two components (/ and j, where / is the faster-permeating component) was calculated by Eq. 2 where C _£ is the concentration of component / in a given stream.

In order to summarize the example above, we can say that the composite graphene membrane is fabricated as follows.

First, the basic step (step 0) consists in synthesizing monolayer graphene on a copper foil, preferably by CVD, then the following process is carried out as illustrated in Figure 1:

1) incorporating nanopores in the graphene lattice, preferably by the oxygen plasma treatment,

2) functionalizing the graphene surface, preferably by 0 ₃ treatment,

3) coating graphene with a C0 ₂ -philic polymer layer,

4) coating the resulting film in with PTMSP to improve the mechanical strength of the film allowing crack-free transfer,

5) removing Cu, 6) swelling the C0 ₂ -philic polymer layer with PEGDE, and

7) transferring the resulting film onto a porous support.

The membranes prepared using PEI are referred to as SPONGl_X, where X represents plasma time in second. Accordingly, the membranes prepared using PEGBA are referred to as SPONG2-X.

Single-layer graphene can be synthesized on the copper foil preferably by the CVD method. Figure 2a and 2b show the Raman spectroscopy data which displays that the as- synthesized graphene has an ID/IG ratio of 0.08, and an / ₂ D/7G ratio of 3.26, confirming that the as-synthesized graphene comprised of a low-density of the intrinsic defects, and is primarily single-layer.

The 0 ₂ plasma can be utilized as an effective method to incorporate nanopores in the graphene lattice. The evolution of pore-density with the plasma time is characterized by the Raman spectroscopy, as shown in Figure 2a and 2b. With 4s of plasma, the ID/IG ratio, indicating the extent of disorder in graphene lattice, increases to 1.7, while the / ₂ D//G ratio decreases to 0.48. With further increasing time of plasma treatment, both ratios decrease. This behaviour indicates that the graphene treated with 4 s of plasma transitions from the "nanocrystalline graphite stage" into a "sp ³ -amorphous carbon stage", where ID is proportional to the probability of finding the sp ² carbon ring. According to the analysis of the ID/IG ratio at 4 s, the distance between defects appears to be smaller than 5 nm, and the density of defects appears to be larger than 1.27xl0 ¹² cm ² . With increasing the plasma-etching time, the size and density of the defects increases. For graphene with 8 s of etching, the distance between defects is smaller than 3.5 nm, and the density of defects is larger than 2.6xl0 ¹² cm ² .

Aberration-corrected high-resolution transmission electron microscopy (HRTEM) revealed that NG indeed comprised of a high-density of nanopores, consistent with the Raman analysis (Figure 2c-h). Typically, the presence of polymer contaminations/residues on the graphene lattice can lead to pore-nucleation and/or expansion under the electron beam. Therefore, for an accurate estimation of the pore-size-distribution, a contamination-free transfer technique was used. As a result, extremely stable imaging conditions were obtained where neither pore-nucleation nor expansion was observed during imaging. Based on HRTEM observations, the nanopores had a lognormal distribution, with mean pore size (arithmetic average of the pore diameters) of 1.8±1.2 and 2.4±1.5 nm after 4 and 6 s of plasma treatment, respectively (Figure 2c-h). Several large nanopores (up to 5 and 8 nm in the 4 and 6 s samples, respectively) were also present. Both zigzag and armchair pore-edge configurations could be observed. Overall, at 4 and 6 s, the pore-densities were 2.1 x 10 ¹² and 2.3 x 10 ¹² cm ² , respectively, corresponding to porosities of 6.8 and 13%, respectively. The order of magnitude for the pore-density obtained by electron microscopy is consistent with the defect-density from the Raman analysis. While new pores were nucleated as the etching time was increased, the mean pore size increased at a linear rate (4 L/s) with respect to the etching time. Based on this, the mean pore size in NG etched for 8 s is estimated to be around 3.2 nm.

The nanoporous graphene is preferably further treated by ozone to functionalize it with abundant oxygen-containing groups, as indicated by the XPS data (Figure 3). According to peak area ratios, the concentration of C-C, C-O, C=0, 0-C=0 were 57.6, 28.6, 9.0, 4.8 wt%, respectively. Therefore, the percentage of carbon sites functionalized with an oxygen group was as quite high (42.4%). The oxygen-containing groups were mainly composed of epoxy, hydroxy, carbonyl, and carboxy groups.

The functional groups, grafted on graphene as a result of the 0 ₃ treatment, improve the interaction with the subsequent C0 ₂ -philic polymer layer via hydrogen-bonding and electrostatic interactions. The surface and the cross-sectional morphologies of ONG and ONG coated with C0 ₂ -philic polymer can be observed by SEM (Figure 4). Overall, the surfaces appear to be quite uniform and smooth indicating a uniform coating of the polymer layers on the graphene. The cross-sectional morphologies in Figure 5 show that the thicknesses of ONG/PTMSP, PONG1/PTMSP, and PONG2/PTMSP are 400, 220, and 120 nm, respectively. Surprisingly, the thickness of PONG1/PTMSP and PONG2/PTMSP are much smaller than that of ONG/PTMSP film. This can be attributed to the fact that coating of PEI or PEGBA films on graphene modifies the surface-wetting properties, leading to only a thinner PTMSP film when PTMSP is coated on top of PONG1 or PONG2, compared to when PTMSP is directly coated on

ONG.

The coating of PTMSP on top of PONG facilitates the transfer of graphene from Cu to a porous substrate, without inducing any cracks or tear. A 100% success rate (defined by the number of successfully transferred samples normalized by the number of attempts) can be achieved. The PTMSP film was selected due to the following reasons:

(i) PTMSP is one of the most permeable polymers. The standalone PTMSP film on the porous support yields a C0 ₂ permeance of 9097 GPU with a CO2/N2 selectivity of 10.2. This high gas permeance ensures that the CO2 permeance of composite film would not be limited by the PTMSP layer;

(ii) PTMSP was observed to be stable in the highly reactive etching solution, and did not develop cracks or pinholes; and

(iii) PTMSP provides sufficient mechanical support (rigidity) to the floating graphene film to eliminate cracks and tear in the film.

To confirm the thickness of the PEI and PEGBA films, the topography of the PONG samples is measured with or without scratches using the atomic force microscopy (AFM). The PONG samples for AFM were prepared by transferring ONG to a Si wafer, followed by coating PEI or PEGBA film on top of graphene. Figure 6a shows that the surface of PONG1 is relatively smooth. Several pm-scale uncovered areas were present attributing to the fact that the ONG does not cover the Si wafer completely due to cracks and tear, and as a result, the PEI coating solution does not wet the domains devoid of ONG. The height distribution data revealed a bimodal height distribution, with a modal spacing of about 8 nm (Figure 6b). Assuming that the bottom of the uncovered area to be Si wafer, the modal spacing indicates the thickness of the PEI layer to be around 8 nm. To further confirm the thickness, we created trenches in the PONG1 by gently scratching the sample's surface (Figure 6c, d), and measured the height profile. The step height across the trench was ca. 8-10 nm (Figure 6c, d), in good agreement with the result from the previous method.

The gas separation performance is evaluated using a homemade gas permeance setup, and the data is shown in Tables 2, 3 and Figure 7. The gas separation performance from all SPONG1-X membranes shows a consistent trend; with increasing the time of O ₂ plasma treatment from 4 to 8 s, the CO ₂ permeance increases (from 1030 to 11640 GPU) while the CO ₂ /N ₂ and CO ₂ /CH ₄ selectivity decreases (41.9 to 14.7 and 18.6 to 7.5, respectively, Figure 7a). This indicates a strong role of graphene nanopores in the separation of CO ₂ from N ₂ and CH ₄ . In general, the pore-size-distribution (PSD) in plasma-treated graphene can be characterized by two kinds of pores; the pores across which the molecular transport is primarily in the temperature-activated mode and the pores across which the transport is primarily in the effusive mode. The former represents pores that are commensurate with the size of molecule that transverse the pores. For C0 ₂ transport, this corresponds to pores that have an electron-density-gap similar to or slightly smaller than the kinetic diameter of C0 ₂ (0.33 nm). The effusive transport takes place from the pores that are large enough such that molecule does not experience any activation barrier while translocating the pore. The permeation coefficient (permeance per pore) is higher in the effusive transport compared to that in the activated transport attributing to the activation barrier and the loss in entropy at the transition state in the case of activated transport. Therefore, a high degree of C0 ₂ /N ₂ separation can be achieved if pores with an electron-density-gap in the range of 0.33 to 0.36 nm can be incorporated in graphene (kinetic diameter of N ₂ is 0.36 nm). Also, increasing the mean-pore-size is expected to increase permeance and reduce separation selectivity. Based on this, the observed decrease in C0 ₂ /N ₂ and C0 ₂ /CH ₄ selectivity with increasing plasma time strongly indicates that permeance increase and selectivity loss was primarily because of the increase in the mean-pore-size in graphene.

The carbon atoms at the pore-edge are oxygen-functionalized and highly polar attributing to a) their generation by the plasma treatment and b) subsequent graphene functionalization by ozone. These oxygenated polar groups strongly interact with the C0 ₂ - philic polymer layers (PEI, PEGBA, and PEGDE) via electrostatic and van der Waal's interaction. Such strong interaction shrinks the effective electron-density-gap, especially for the large- pores that can accommodate the polymer chains. The polymer-modified pore yields higher C0 ₂ /N ₂ or C0 ₂ /CH ₄ selectivities compared to that from the unmodified pores. Therefore, when the pore size is increased significantly at the higher plasma time (8 s), the effectiveness of the polymer in shrinking the pore size is reduced. As a result, the SPONGl-8s membrane yields a C0 ₂ /N ₂ selectivity of 14.7, close to that from pure PTMSP membrane (10.2). This phenomenon indicates that the porous graphene film in the SPONGl-8s membrane contributed only slightly to the overall selectivity, attributing to that the pore size at 8 s plasma exposure is too big. In this case, it is very likely that larger pores are not fully covered with PEI film because of the relatively weaker interactions between the copper under the graphene pore and PEI. Another important feature of the SPONG membrane is the ultrathin functional polymer film on top of the NPG. The PEGDE swollen PEI film contains a high-density C0 ₂ -philic groups (amino and ethylene oxide), leading to the enhancement of C0 ₂ sorption and C0 ₂ separation capability. Especially, the linear PEG oligomer (M _n =500), brings a high-density of ethylene oxide groups, and has a very low crystallinity at room temperature, and thus affords a high intrinsic-free-volume and CO2 solubility, as demonstrated by Shao and coworkers. The branched PEI polymer possesses a strong intramolecular hydrogen bonds, resulting in a low free volume and CO2 permeability. The impregnation of PEGDE into the PEI network can break the intramolecular hydrogen bonds, increasing the free volume. As a result, the ultrathin PEGDE swollen PEI film has a high CO2 permeance and good CO2/N2 separation capacity. For example, the SPONGl-6s (membrane M3) shows a remarkably high CO2 permeance of 5470 GPU with a CO2/N2 ideal selectivity of 25.2.

The mixed gas permeation measurement can be performed using a feed mixture of 20/80 mol% for CO2/N2 at 30 °C and 2 bar (Table 1 and Figure 7b). Comparing the data from the single-gas and mixture gas, we see that the selectivity is nearly the same, while the gas permeance is decreased by around 1% for the mixture gas. For the SPONGl-4s membrane, the CO2 permeance decreased from 1030 to 970 GPU while the selectivity decreased from 41.9 to 41.1 when we switched to the mixed gas feed.

Figure 7c shows the single-gas separation performance for the SPONG2 membranes. In comparing to SPONG1, these membranes yield a lower CO2 gas permeance but a higher CO2/N2 ideal selectivity (Table 2). For example, the SPONG2-4s membrane yielded a CO2 permeance of 620 GPU with a CO2/N2 ideal selectivity of 57.2.

Table 2. C0 ₂ permeance, CO /N and CO /CH ideal selectivity of SPONG membranes under single-gas permeation condition at 30 °C, 2 bar.

CO ₂ /N ₂ ideal CO ₂ /CH ₄ ideal

Membrane type Nomenclature CO permeance (GPU)

selectivity selectivity

SPONG1-4S Ml 1030 41.9 18.6

SPONG1-4S M2 1060 36.6 15.8

SPONG1-6S M3 5470 25.2 10.9

SPONG1-6S M4 6210 20.4 9.3

SPONG1-6S M5 4360 24.8 10.0 SPONG1-6S M6 3680 28.9 10.9

SPONG1-8S M7 11650 14.7 7.5

ONG-8s M8 3910 11.6 5.9

SPONG2-4S M9 620 57.2 20.5

SPONG2-6S M10 1240 35.1 14.4

SPONG2-8S Mil 4370 20.8 9.3

SPONG2-8S M12 8280 17.1 8.6

C0 ₂ /N ₂

Membrane type Nomenclature C0 ₂ permeance (GPU)

separation factor

SPONG1-4S Ml 970 41.1

SPONG1-4S M2 910 38.8

SPONG1-6S M3 4940 25.8

SPONG1-6S M4 6040 21.4

SPONG1-6S M5 4030 22.6

SPONG1-8S M7 10940 15.2

SPONG2-8S Mil 4200 20.8 The strong dependence of gas separation performance on the graphene pore size is reflected by Figure 8. An increase in the mean pore size in graphene led to increase in the C0 ₂ permeance albeit with a decrease in the CO2/N2 selectivity. The permeance trend can be attributed to the fact that graphene porosity (area covered by vacancy defects) increased with increasing the pore size (refer to the characterization of graphene in the later section). CO2 permeation was not detected when the graphene lattice was not etched (Figure 8b). On the other hand, graphene pores larger than 2 nm become increasingly difficult to be effectively masked by the ultrathin C0 ₂ -philic layer (supplementary Note II). As a result, a tradeoff between the CO2 permeance and the CO2/N2 selectivity was observed as a function of the mean pore size. Based on this, highest performance is obtained when the pore-size is below 2 nm and when the porosity is highest.

Figure 9 shows the comparison of the SPONG membranes with the state-of-the-art membranes for the carbon capture (separation of the CO2/N2 mixture). The target area indicates that the desired membrane should have a high CO2 permeance (>1000 GPU) and moderate separation factor (>20) to achieve an economical CO2 capture from flue gas. The outstanding separation performance of the SPONG membranes is clearly established by the comparison in Figure 9. The SPONGl-6s membrane shows an exceptionally high CO2 permeance of 4940 GPU with a moderate CO2/N2 separation factor of 25.8 at 30 °C. At 40 °C, the CO2 permeance increased to 6100 GPU while the CO2/N2 separation factor, 22.5, remained in the target area. This CO2 permeance is nearly 60 fold higher than that from the commercial cellulose acetate membrane, while their CO2/N2 separation factors are comparable. ⁶

According to the techno-economic analysis model proposed by Merkel, a CO2/N2 selectivity dwelling between 20 and 30 is adequate for CO2 capture. For selectivity beyond 30, the capture cost does not change significantly. However, at a fixed selectivity, enhancing the CO2 permeance can remarkably decrease the capture cost by cutting down the membrane area. Thus, a moderate selectivity (20-30) in combination with the highest possible CO2 permeance is highly attractive for the cost-effective CO2 capture.

In summary, we report a hybrid single-layer graphene/polymer membrane, where a C02-philic polymer layer having a thickness of 5-100 nm, preferably 5-50 nm, more preferably 5-20 nm and even more preferably 10 nm added on top of graphene nanopores provided an attractive separation selectivity. The membrane achieved an outstanding CO2 capture performance, that is, CO2 permeance is 6100 GPU with a CO2/N2 separation factor of 22.5, mainly attributing to two reasons: (i) the single layer graphene with a high density of nanopores combined with the ultrathin polymer can afford a high gas permeance due to the short diffusion pathway; and (ii) the CC philic polymer can increase the solubility and adsorption of CO2, thus affording a high selectivity of CO2/N2. The approach of hybrid membranes involving single-layer graphene with another CC philic layer could pave the way for the large-scale commercialization of the membranes for carbon capture.

In addition to the above, further experimental assays on separation performance of functionalized nanoporous graphene have been ran and have shown the following results:

In table 4 below, one has tested the separation performance of polymers with intrinsic microporosity-1 (PIM-1) functionalized nanoporous graphene, at 2 bar and room temperature, where “polymers with the intrinsic microporosity" means microporous polymers where microporosity refers to pore-size from 0.2 to 2 nm:

Table 4

In table 5 below, one has tested the separation performance of ionic liquid functionalized nanoporous graphene, at 2 bar and room temperature:

Table 5