Gas separation membrane

Abstract

1. A composite membrane suitable for separating gases from a gas mixture, comprising a selective layer coated on a support, the selective layer comprising: a) a polymer matrix comprising an amine polymer; b) graphene oxide nanofillers; c) a transfer carrier selected from an ionic liquid or an amino acid salt.

Description

Technical Field

[0001] Summary of the Invention The present invention relates to a composite membrane for separating a gas from a gas mixture, preferably for separating carbon dioxide from a gas mixture containing the same, a process for producing this composite membrane, a process for separating a gas from a gas mixture, and the use of this composite membrane for separating a gas. In particular, the present invention provides a composite membrane comprising a polymer matrix, a graphene oxide nanofiller, and a selective layer having a mobile carrier.

Background Art

[0002] The global climate crisis after industrialization is mainly due to post-industrialization and anthropogenic factors. Among all the causes, the emissions of greenhouse gases over the past few decades, and most importantly, the significant increase in CO2 emissions, may be directly linked to the current scenario of global warming. The introduction of carbon capture, utilization, and storage (CCUS) corresponds to the most effective solution for transitioning to a more sustainable energy system in the next decade. However, the limitations regarding cost and technical feasibility in CO2 capture technologies (i.e., absorption, membrane separation, and adsorption) still remain the main issues to be addressed in the introduction of CCUS.

[0003] Polymer gas separation membranes have been widely explored for CO2 separation applications due to their low cost, high modularity, and easy scalability. Membrane materials with excellent permeation properties (permeability and selectivity), as well as good chemical and mechanical properties, can significantly increase the efficiency of the separation process. Conventional polymer membranes based on the solution-diffusion mechanism suffer from an inherent trade-off between permeability and selectivity, as shown by Robeson's upper bound.

[0004] One approach to overcome this trade-off is to use nanofillers in the polymer selectivity layer to form a hybrid membrane containing the nanofillers. Nevertheless, limited research has been reported on fabricating thin composite membranes of hollow fibers with hybrid selectivity layers for CO2 separation applications. Many of these solutions use thick membranes with nanofillers.

[0005] Conventional polymer membranes have various nanofillers to enhance CO2 permeation properties, but 2D nanofillers such as graphene oxide (GO) have been scarcely studied for membrane fabrication. Their high surface-to-volume ratio resulting from the 2D structure causes nanoscale property changes when dispersed in the polymer matrix. The effective dispersion of these platelets affects the reorientation of polymer chain packing, including changes in crystallinity, free volume fraction, and CO2 solubility, even in small amounts. Wang et al, Journal of Membrane Science 589(2019)117246, describe polyaniline-coated carbon nanotubes between graphene oxide (GO) layers in a PVAm membrane.

[0006] Another approach to overcome the permeability-selectivity "trade-off" is the use of CO2-reactive carriers in the polymer selectivity layer. These facilitated transport membranes move CO2 through additional reactive pathways rather than conventional polymers that follow the solution-diffusion mechanism.

[0007] The reactive carriers are typically amine groups fixed to the main chain of the polymer chain. Recently, small CO2-affinity molecules that react reversibly with CO2 have also been added to the polymer host matrix as "mobile carriers" to increase CO2 transport. These reactions require water to assist CO2 transport across the membrane. Dai et al. in J.Memb.Sci.2019, Vol 578, 61-68, describe the addition of amino acid salts to a PVA membrane.

[0008] The present invention relates to a separation membrane comprising a selective layer coated on a support. In particular, the present invention involves the incorporation of two-dimensional (2D) nanofillers, especially graphene oxide nanofillers, in the selective layer. Due to their large aspect ratio, when added as nanofillers in a composite membrane, 2D materials such as graphene oxide (GO) significantly affect the mechanical and transport properties. The presence of hydroxyl groups on the GO surface also confers hydrophilicity and increases surface interaction with CO2. Advantageously, hydrophilic GO nanosheets, when added to a facilitated transport matrix, induce disruption of polymer chains leading to distributed water channels with increased CO2 solubility. These effects on the gas separation properties of such hybrid membranes are highly dependent on the surface chemistry of the added nanosheets and their lateral dimensions.

[0009] The inventors of the present application have found that GO-based 2D nanoplatelets in a composite membrane for CO2 separation, in the form of thin film composites (TFCs) having excellent permeation properties even when having an ultrathin selective layer, provide a valuable class of membranes, hybrid facilitated transport membranes (HFTM). Previously, 2D nanoplatelets were thought to have an adverse effect on gas separation when dispersed in a polymer matrix due to their barrier properties, and the fabrication of TFC membranes with ultrathin selective layers (<500 nm) containing 2D nanoplatelets in facilitated transport membranes was considered difficult.

[0010] The inventors have surprisingly found that the use of mobile carriers, i.e., low molecular weight CO2 affinity components, in the selective layer further increases the performance, especially assisting in CO2 / CH4 separation and resisting the carrier saturation phenomenon. In particular, the mobile carriers act synergistically with modified graphene oxide in combination with an amine polymer matrix to improve the permeability and selectivity. The presence of the mobile carriers also reduces the saturation degree of the carriers.

[0011] Size-controlled 2D graphene oxide also has a positive effect on CO2 transport even at very low loading levels.

[0012] A surprising aspect of the present invention is that a composite membrane comprising a mobile carrier and a GO-based filler has an increased CO2 permeability and an increased CO2 flux under industrially relevant conditions compared to the same module without the mobile carrier. SUMMARY OF THE INVENTION

[0013] Viewed from one aspect, the present invention is a composite membrane suitable for separating a gas from a gas mixture, comprising a selective layer coated on a support, wherein the selective layer is a) a polymer matrix comprising an amine polymer, b) graphene oxide nanofillers, and c) a mobile carrier selected from an ionic liquid or an amino acid salt, and provides a composite membrane.

[0014] In a further aspect, the present invention is a composite membrane suitable for separating a gas from a gas mixture, comprising a selective layer coated on a hollow fiber support, wherein the selective layer is a) a polymer matrix comprising an amine polymer, b) porous graphene oxide or PEG-modified graphene oxide nanofillers, and optionally c) a mobile carrier selected from an ionic liquid or an amino acid salt, and relates to a composite membrane.

[0015] In a further aspect, the present invention is a composite membrane suitable for separating a gas from a gas mixture, comprising a selective layer coated on a flat sheet support, wherein the selective layer is a) a polymer matrix comprising an amine polymer, b) porous graphene oxide or PEG-modified graphene oxide nanofillers, and optionally c) a mobile carrier selected from an ionic liquid or an amino acid salt, and relates to a composite membrane.

[0016] In a further aspect, the present invention is a composite membrane suitable for separating a gas from a gas mixture, comprising a support, for example, a support of a hollow fiber or a flat sheet, on which a selective layer is coated, wherein the selective layer is a) a polymer matrix comprising an amine polymer; and b) a porous graphene nanofiller or a chemically modified graphene nanofiller, optionally, the chemically modified graphene nanofiller is graphene oxide having an organic unit grafted thereon, preferably, the organic unit is selected from a nitrogen and / or oxygen-containing organic unit, a polymer, or a nitrogen and / or oxygen-containing polymer, preferably, the chemically modified graphene nanofiller is a PEG-modified graphene oxide nanofiller, a chemically modified graphene nanofiller; and optionally c) a mobile carrier selected from an ionic liquid or an amino acid salt, and relates to a composite membrane.

[0017] Viewed from another aspect, the present invention is a process for forming a composite membrane, comprising (I) a) forming an aqueous solution comprising a polymer matrix comprising an amine polymer, b) a graphene nanofiller; and optionally c) a mobile carrier selected from an ionic liquid or an amino acid salt; and (II) casting the aqueous solution onto a support, for example, a flat sheet or a hollow fiber support. The present invention provides a process.

[0018] In particular, the process comprises (I) a) forming an aqueous solution comprising a polymer matrix comprising an amine polymer, b) a graphene nanofiller, and c) a mobile carrier selected from an ionic liquid or an amino acid salt; and (II) casting the aqueous solution onto a flat sheet support using a bar roller to apply a selective layer.

[0019] Viewed from another aspect, the present invention is a process for forming a composite membrane, comprising: (I) a) a polymer matrix comprising an amine polymer; b) a graphene oxide nanofiller; c) a mobile carrier selected from an ionic liquid or an amino acid salt, to form an aqueous solution; and (II) casting the aqueous solution onto a support using a dip coating method to produce a selective layer, preferably wherein the support is a hollow fiber support.

[0020] The present invention also provides a process for separating a gas from a gas mixture, for example, separating carbon dioxide from a gas mixture containing carbon dioxide, the process comprising passing the gas mixture through a composite membrane as previously defined herein. In one embodiment, the composite membrane comprises a hollow fiber support and a plurality of such composite membranes are present within a module.

[0021] The present invention also provides the use of a composite membrane as defined above in the separation of a gas from a gas mixture, for example, the separation of carbon dioxide from a gas mixture containing carbon dioxide.

[0022] The features of the aspects and / or embodiments shown herein can be used individually and in combination in all aspects and embodiments of the present invention that are technically feasible, unless otherwise indicated.

[0023] Definitions The following definitions apply: The term composite membrane means the presence of a support and a selective layer thereon. The selective layer itself may also be referred to as a membrane. The composite membrane facilitates gas separation, such as the separation of carbon dioxide from a gas mixture.

Best Mode for Carrying Out the Invention

[0024] The present invention relates to a composite membrane for gas separation. The composite membrane is suitable for separating gases from a gas mixture and, in one embodiment, comprises a selective layer coated on a support, the selective layer comprising a) a polymer matrix comprising an amine polymer, and b) a graphene oxide nanofiller, and c) a mobile carrier selected from an ionic liquid or an amino acid salt.

[0025] Support Gas separation membranes can typically take two forms, supported or unsupported. This membrane is supported on a support. As described below, the support can be in the form of a flat sheet or a hollow fiber support. Both of these support types are included in the present invention.

[0026] Suitable supports are known in the art and most are porous to the gas being transported. Typically, therefore, the support is porous. Suitable supports include polyethersulfone (PES), polytetrafluoroethylene (PTFE), polypropylene, sulfonated polysulfone, polyvinylidene fluoride (PVDF), polyacrylonitrile (PAN) and related block copolymers, cellulose-based such as cellulose acetate (CA), polyimide, polyetherimide (PEI), aliphatic polyamide, polyetheretherketone (PEEK), polyphenylene oxide (PPO), and polysulfone (PSf). Such supports are commercially available from suppliers such as Osmonics. In a preferred embodiment, the support is PVDF, and in particular, the support is in the form of a flat sheet. When the support is a hollow fiber, both PSf and, in particular, PPO are preferred, especially PPO.

[0027] Most of these supports are typically ultrafiltration supports, and the pore sizes in the supports are on the order of 20 to 1000 angstroms, although it is more common to express the pore size in terms of the molecular weight cut-off value.

[0028] In some embodiments of the present invention, the use of a microporous support structure is also within the scope of the present invention. Such supports have much larger pore sizes, e.g., 0.10 to 10 μm, and allow very rapid gas transport therethrough. Although it is not common to express the pore sizes of these supports in terms of MWCO terminology, in the present invention, microporous supports are considered to have MWCO values in excess of 100,000.

[0029] The microporous support can be formed from any suitable material, including those described above in connection with ultrafiltration supports and inorganic materials such as ceramics (alumina, zirconium oxide), glass membranes such as silica, etc. These can be made by sintering, sol-gel or leaching techniques known in the art.

[0030] Conventionally, in gas separation membranes using a selective layer polymer, the use of these microporous supports has been assumed to be impossible because the pores of the support are very large and the polymer simply collapses into the pores. This can be overcome by using high molecular weight polymers in the selective layer, which have been found to have not only excellent permeability and selectivity but also excellent mechanical strength. Alternatively, pore fillers such as 3M (trademark) Fluorinert (trademark) Electronic Liquid FC-72 can be used to fill the pores to prevent penetration of the casting solution. This is a low viscosity, low VOC fluorinated compound. The use of fluorinated hydrocarbons is preferred as pore fillers. The mechanical strength of the high Mw polymer allows the use of supports with micron-sized pores without filling problems.

[0031] The molecular weight cut-off (MWCO) of the support is preferably kept as high as possible. The MWCO is essentially a measure of the pore size in the support, with a larger MWCO value representing a higher pore size. In the present invention, the MWCO is preferably greater than 20,000, for example, at least 35,000, preferably greater than 50,000, more preferably at least 60,000, particularly at least 75,000. In a highly preferred embodiment, the MWCO is at least 100,000. In fact, the present invention enables the use of supports having an MWCO of up to 300,000, for example, from 30,000 to 300,000. In one embodiment, the MWCO may be less than the molecular weight Mw of the polymer from the selectivity layer cast on top.

[0032] When the composite membrane of the present invention is prepared using a high molecular weight selectivity layer, it has been found that the problem of "fouling" is minimized even when using a high molecular weight support. This then enables the use of a support with a high MWCO, and thus can lead to an improvement in permeability and selectivity.

[0033] Without wishing to be bound by theory, when using a porous support with larger pores, regardless of whether it is an ultrafiltration or microfiltration support, the increased pore size not only reduces the mass transfer resistance to the gas to be separated, but also changes the support separation mechanism itself. An ultrafiltration support with a low pore size (low MWCO) can, for example, exhibit selectivity for nitrogen via Knudsen diffusion and not exhibit selectivity for carbon dioxide.

[0034] As described in more detail below, by using a high Mw selectivity layer polymer, it is possible to use a porous support that has larger pores and as a result has a low mass transfer resistance to the gas molecules separated by the selectivity layer without affecting the mechanical stability.

[0035] In a preferred embodiment, the support may have a porous lower layer with a thin and high-density upper layer. By high density is meant that there are no pores in the high-density upper layer.

[0036] The high-density upper layer preferably has a thickness of 60 nm or less, for example, a thickness of approximately 40 nm or less. However, it is within the scope of the present invention for the high-density layer to have a greater thickness, for example, 100 to 1000 nm, for example, 200 to 700 nm, for example, 600 nm.

[0037] The support having a high-density upper layer is preferably a hollow fiber support and ideally can be formed from PPO. The high-density upper layer is formed during the spinning process. In this case, no pore filler is required.

[0038] Polymer matrix The selective layer contains an amine polymer. The polymer matrix may contain one or more polymers. "Polymer matrix" means the polymer constituent a) in the selective layer of the composite membrane. The polymer matrix contains an amine-based polymer, for example, a polymer having a hydrocarbon backbone with pendant amine groups or a polyamine (i.e., having an amine group in the main chain). Preferably, the polymer matrix contains a polymer having a hydrocarbon backbone with pendant amine groups. More preferably, this is a polyallylamine-based or polyvinylamine-based polymer. Polyallylamine-based or polyvinylamine-based polymers include modified ones, for example, polyallylamine ("PAA") or polyvinylamine ("PVAm") structurally modified with amino groups. Polyallylamine-based polymers are particularly preferred.

[0039] In a specific embodiment, the polymer matrix contains a steric hindrance polymer of the following formula (I),

Chemical formula

[0040] The integer m is typically 0-6, preferably 0-2, more preferably 0-1, and most preferably 1.

[0041] The integer n is used to indicate the polymeric nature of the structure, and the value of n is typically such that the polymer matrix has a Mw as defined below or in the claims. This definition of n is valid for the following structure definitions.

[0042] Steric hindrance of amine-based polymers in the solid phase has been demonstrated to increase the gas permeation performance of facilitated transport membranes.

[0043] In certain embodiments, the polymer matrix comprises a sterically hindered polyallylamine (''SHPAA'') of formula (II),

Chemical formula

Chemical formula

[0044] Typically, SHPAA(A) is used with a flat sheet support, and SHPAA(B) is used with a hollow fiber support. In certain embodiments, SHPAA(B) is preferred.

[0045] Similarly, the selective layer polymer can be a sterically hindered polyvinylamine polymer (SHPVAm) of formula (III),

Chemical formula

[0046] The polymer matrix can include a single polymer or a combination of two or more polymers. In certain embodiments, the polymer matrix consists of at least one polymer. Typically, the polymer matrix comprises at least 60% by weight, preferably at least 70% by weight, more preferably at least 80% by weight, even more preferably at least 85% by weight of an amine-based polymer as defined herein or in the claims. Typically, the polymer matrix comprises 60 - 99% by weight, preferably 70 - 98% by weight, 80 - 95% by weight of an amine-based polymer. In some embodiments, the amine polymer is the only polymer in the polymer matrix.

[0047] Typically, the polymer matrix occupies at least 50% by weight of the selective layer, for example, at least 60% by weight, at least 70% by weight, at least 80% by weight, at least 90% by weight, or at least 95% by weight of the selective layer. Typically, the polymer matrix occupies 50 - 99% by weight of the selective layer, preferably 60 - 95% by weight of the selective layer, more preferably 70 - 95% by weight of the selective layer.

[0048] In certain embodiments, the amine polymer (e.g., SHPAA) can be combined with another polymer, particularly an oxygen-containing polymer such as polyvinyl alcohol (PVA). The oxygen-containing polymer can contain oxygen-based functional groups in the main chain of the polymer or as pendant functional groups. The oxygen groups are preferably hydroxyls. The oxygen groups are preferably pendant. The use of PVA is preferred.

[0049] The combination of the amine polymer and the second polymer, particularly PVA, results in a reduction in the brittleness of the selective layer and excellent transport properties due to the water-swellable nature of PVA.

[0050] Viewed from another aspect, the present invention is a composite membrane suitable for separating a gas from a gas mixture, comprising a selective layer coated on a support, wherein the selective layer a) a polymer matrix comprising an amine polymer and PVA, b) graphene oxide nanofiller, and c) a mobile carrier selected from an ionic liquid or an amino acid salt, and provides a composite membrane.

[0051] This advantage is most pronounced when the second polymer, such as PVA, is present in an amount of 2 to 20 wt%, preferably 5 to 15 wt%, based on the total weight of the amine polymer and the second polymer. This ensures low brittleness without impairing the separation characteristics of the selective layer.

[0052] The weight average molecular weight (Mw) of the polymer matrix used in the present invention can range from 10,000 to 3,000,000, preferably from 20,000 to 750,000, more preferably from 30,000 to 500,000, and even more preferably from 80,000 to 300,000 (unless otherwise indicated, molecular weights are expressed in g / mol herein).

[0053] The weight average molecular weight (Mw) of the selective layer polymer is typically at least 50,000. Preferably, the Mw of the selective layer polymer is at least 100,000.

[0054] The weight average molecular weight (Mw) of the amine polymer is typically at least 50,000. Preferably, the Mw of the selective layer polymer is at least 100,000, for example, 80,000 to 300,000.

[0055] The weight average molecular weight (Mw) of the oxygen polymer such as PVA is typically at least 50,000. Preferably, the Mw of the selective layer polymer is at least 100,000, for example, 80,000 to 300,000.

[0056] It has been found that using a higher Mw results in obtaining the strength of the selective layer. This enables the use of a support having a much higher MWCO. Thus, in certain embodiments, the present invention provides a selective layer polymer having an Mw of at least 100,000 and having a support with an MWCO of at least 60,000.

[0057] It has also been observed that even when using a higher molecular weight polymer, this does not result in a reduction in permeability or selectivity. The use of a higher Mw polymer means that the actual selective layer used tends to be denser than a selective layer formed from a lower Mw polymer. Surprisingly, the inventors have found that even at a higher density, the value of the permeability of the composite membrane remains very high and the gas selectivity is good.

[0058] A further benefit of using a higher Mw polymer matrix relates to water uptake. Higher Mw polymers have more densely packed molecular chains, which means more densely packed amino groups. This results in better water uptake, especially compared to lower Mw polymers that promote the reactivity of amino groups towards carbon dioxide.

[0059] One skilled in the art can also expect that the increased water uptake causes swelling of the selective layer, and thus, in a thicker selective layer, it is obvious that gas crossover is more difficult, resulting in lower permeability values. However, any swelling that occurs is limited and offset by the increased carbon dioxide transport facilitated by the higher water uptake.

[0060] Therefore, the combination of a high molecular weight polymer matrix and a high MWCO support provides a composite membrane with excellent properties.

[0061] A further benefit of using a higher Mw polymer matrix is their ability to withstand higher pressures. Membranes of the prior art have conventionally been used at low gas pressures. However, flue gases from industrial plants can be at relatively high pressures, for example up to 15 bars, and ideally, any composite membrane should be able to perform gas separation for such higher pressure gases. In particular, it is preferred that the permeability and selectivity obtained at higher gas pressures are not reduced (or not significantly reduced) compared to operation at lower pressures. A further feature of the present invention is that the claimed composite membrane can handle gases under pressure, for example at pressures up to 20 bars, for example up to 15 bars, for example 2 - 15 bars, or 2 - 10 bars.

[0062] Crosslinking a high Mw polymer causes densification of the selective layer, making it more difficult for carbon dioxide to contact the amine groups in the polymer, and thus, in practice, results in a reduction in permeability and selectivity. However, for low molecular weight polymers that do not "fill" the pores in the support, crosslinking is typically essential to provide a selective layer with sufficient strength.

[0063] When using a higher Mw polymer matrix, the higher Mw provides sufficient strength to the selective layer to overcome the filling problem, so the requirement for crosslinking using a crosslinking agent no longer exists. Also, despite the use of a higher Mw polymer matrix that causes overall densification of the selective layer compared to a lower polymer matrix, no reduction in permeability or selectivity caused by the use of a higher Mw polymer matrix is observed. In fact, the opposite is observed and the membrane actually functions better than the crosslinked control. Therefore, it is particularly preferred that the selective layer is not crosslinked using an external crosslinking agent.

[0064] Heat treatment of the composite membrane can provide advantageous properties, especially when the composite membrane operates at elevated gas pressures, such as above 10 bars.

[0065] Heat treatment means exposing the composite membrane (i.e., the membrane on the support) to heat to generate strength therein. Suitable heat treatment conditions include heating up to 50 - 150 °C, such as 80 - 120 °C, particularly 90 - 110 °C. This heat treatment step is not considered a crosslinking step because no external crosslinking agent is used, but it probably imparts additional strength to the composite membrane by promoting intermolecular interactions between polymer chains and between the polymer matrix and the porous support.

[0066] It will be apparent that the heat treatment of the selective layer is carried out while it is supported. Without wishing to be limited by theory, the heat treatment step is also thought to modify the support and thus enable improved permeability values. It may improve the interaction between the support and the high - density layer of the amine polymer.

[0067] Graphene oxide nanofiller The selective layer of the composite membrane of the present invention also contains graphene oxide nanofillers, particularly 2D graphene oxide fillers. The term 2D means that one of the dimensions of the filler is very small, for example, 10 nm or less. Thus, the graphene oxide filler can be considered to be in the form of flakes or a plane having a width and height but a very low thickness. Thus, graphene oxide (GO) is a two-dimensional material, and thus the terms nanofiller, nanoplatelet, or nanosheet can be used interchangeably herein. Also, the term graphene oxide includes both physically or chemically modified graphene oxide.

[0068] The membrane of the present invention is typically a "hybrid" membrane. Thus, the graphene oxide nanofillers are typically dispersed in a polymer matrix containing an amine polymer.

[0069] The terms "nanofiller", "nanoplatelet", or "nanosheet" indicate that the average dimension of the graphene oxide is 1000 nm or less, for example, in the range of 10 to 1000 nm, preferably 100 to 1000 nm, more preferably 300 to 900 nm, and even more preferably 400 to 800 nm. These dimensions refer to the average lateral dimension, i.e., within the 2D plane of the nanosheet. These dimensions can be measured using atomic force microscopy (AFM).

[0070] Surprisingly, it has been found that the size of the GO flakes affects the gas permeation performance, and the best results are observed at an average dimension in the range of 400 to 800 nm. An average particle size greater than 1000 nm (i.e., a "micro" filler) results in a decrease in performance. The average thickness of the graphene oxide nanosheet is typically 10 nm or less, for example, 2.0 nm or less, particularly 1.0 nm or less. In certain embodiments, the nanosheet is the thickness of at least one layer of graphene oxide, for example, the thickness of at least two layers. In certain embodiments, the graphene oxide nanofiller is in the form of a single layer of graphene oxide.

[0071] In certain embodiments, the nanofillers are size-controlled, i.e., the particles are uniform or substantially uniform in their size distribution. The size distribution profile is typically unimodal. Typically, at least 75%, preferably at least 90%, more preferably at least 95%, even more preferably at least 99% of the nanofiller particles are within ±50%, preferably within ±25%, more preferably within ±10% of the average dimension. Alternatively, at least 75%, preferably at least 90%, more preferably at least 95%, even more preferably at least 99% of the nanofiller particles can have a (mean) lateral dimension of less than 1000 nm. Alternatively, at least 75%, preferably at least 90%, more preferably at least 95%, even more preferably at least 99% of the nanofiller particles can have a (mean) lateral dimension in the range of 10 to 1000 nm. Alternatively, at least 75%, preferably at least 90%, more preferably at least 95%, even more preferably at least 99% of the nanofiller particles can have a (mean) lateral dimension in the range of 100 to 1000 nm. Alternatively, at least 75%, preferably at least 90%, more preferably at least 95%, even more preferably at least 99% of the nanofiller particles can have a (mean) lateral dimension in the range of 300 to 900 nm. Alternatively, at least 75%, preferably at least 90%, more preferably at least 95%, even more preferably at least 99% of the nanofiller particles can have a (mean) lateral dimension in the range of 400 to 800 nm.

[0072] Different suppliers of GO provide dispersions with different flake sizes. The flake size is typically optimized to obtain the best performance. The inventors used GO from a commercial supplier, Graphene-XT, for a particular hollow fiber composite membrane and made a unique nanofiller for a flat sheet composite membrane. Nanofillers from any source can be exfoliated by sonication to obtain single layers of GO. Different sonication times can be used to control the flake size and affect the gas permeation performance. Using sonication to control the flake size is potentially valuable. Longer sonication times result in smaller flakes. Sonication for 4 - 8 hours seems to result in an ideal GO flake size.

[0073] Furthermore, graphene oxide can be designed to improve gas permeation in composite membranes. Both physically and chemically modified GO nanoplatelets with various properties can be synthesized and successfully dispersed in a polymer matrix. These physically and chemically modified GO nanoplatelets can also be subjected to sonication to control the flake size beforehand.

[0074] In certain embodiments, the graphene oxide nanofiller is physically modified, for example using hydrogen peroxide, to make it porous. Such modification can occur at elevated temperatures, for example 100 - 250 °C. Thus, in certain embodiments, graphene oxide is porous. Typically, the average pore size is 1 - 200 nm, for example, 2 - 20 nm.

[0075] Physical modification to make GO porous is expected to create defects in a plane perpendicular to the direction of gas transport while maintaining the 2D morphology of the original platelet. More generally, the increase in pore content and defects across the thickness generated during the processing causes nanoscale changes in the polymer packing, resulting in improved permeation.

[0076] The GO nanofiller of the present invention can be subjected to both ultrasonic treatment for controlling the particle size and physical modification for imparting pores to the nanofiller.

[0077] In another embodiment, graphene oxide is chemically modified to have organic units (preferably oxygen and / or nitrogen-containing organic units) grafted thereon, preferably a polymer, preferably a nitrogen and / or oxygen-containing organic polymer, more preferably an oxygen-containing organic polymer, for example, a polyether, for example, a polyalkylene polyol such as polyethylene glycol (PEG) grafted thereon. Thus, in certain embodiments, graphene oxide is modified with polyalkylene glycol groups such as PPG (polypropylene glycol) and / or PEG (polyethylene glycol) groups. The nitrogen-containing polymer can be, for example, a polyamine polymer, for example, polyethyleneimine (PEI). The polyether can include, for example, a polyether backbone having pendant PEG groups (for example, based on polyglycerol such as hexaglycerol). When the term polyalkylene polyol is used, a mixture of polyalkylene polyols (for example, PPG and PEG) can be used. An example of an oxygen and nitrogen-containing polymer is a PEG / PEI hybrid.

[0078] Oxygen and / or nitrogen introduce polarity to the grafted polymer, which can be beneficial for membrane performance. In certain embodiments, oxygen is in the repeating unit (for example, ether) and / or at the end (for example, -OH), and typically in the repeating unit. In certain embodiments, nitrogen is in the repeating unit (for example, amine) and / or at the end (for example, a terminal amine group such as -NH2), and typically at the end.

[0079] In certain embodiments, the graphene oxide nanofiller is a polymer-modified graphene oxide nanofiller. Preferably, the polymer grafted onto the graphene oxide nanofiller is selected from or comprises a polyamine and / or a polyether, preferably a polymer selected from or comprising polyethyleneimine and / or a polyalkylene glycol, preferably a polymer selected from or comprising PEI, PPG, and / or PEG, preferably a polymer selected from or comprising PPG and / or PEG, preferably a polymer selected from or comprising PEG, preferably a polymer comprising PEG.

[0080] The organic unit or polymer grafted onto the graphene oxide nanofiller can be branched or linear. In the case of a branched nitrogen-containing polymer, such as branched polyethyleneimine, the polymer can have primary, secondary, or tertiary amine groups, as is well known in the art.

[0081] Other groups (such as terminal groups, e.g., terminal -NH2 groups) can be present in the polymer units grafted onto the graphene oxide nanofiller. For example, if a PEG-containing polymer is grafted onto graphene oxide, this PEG unit can have other groups, such as a terminal -NH2 group, or the following 8-arm PEG:

Chemical Formula

[0082] An organic group (e.g., PEG) can be grafted onto graphene oxide via any typical coupling reaction, such as an EDC coupling reaction. Thus, a linker group can be present between the graphene oxide and the organic moiety. Typically, the molecular weight of the grafted group (e.g., a nitrogen- and / or oxygen-containing polymer) ranges from 1,000 to 500,000, preferably from 1,000 to 100,000, preferably from 2,000 to 50,000, preferably from 5,000 to 20,000 g / mol. The molecular weight of the grafted group is typically given as the number average molecular weight M n as. A preferred example is a commercially available 8-arm PEG having an M n of 10,000.

[0083] The GO nanofillers of the present invention can be subjected to both ultrasonic treatment for controlling the particle size and chemical modification.

[0084] In certain embodiments, the graphene oxide nanofillers are either porous (i.e., physically modified) or have PEG groups grafted thereon (i.e., chemically modified).

[0085] The amount of graphene oxide nanofillers in the selective layer is typically less than 5 wt%, preferably less than 1 wt%, more preferably less than 0.5 wt%. Suitable ranges include 0.05 wt% to 5.0 wt%, preferably 0.1 to 1.0 wt%, preferably 0.1 to 0.5 wt%, or preferably 0.1 to 0.3 wt%. These weight percentages are determined on a dry weight basis.

[0086] The inventors have surprisingly found that graphene oxide can efficiently increase the CO2 permeability by about 200% compared to that of an unmodified selective layer, without a significant change in selectivity, at very low loadings (e.g., down to 0.2 wt% nanofiller). At low loadings, graphene oxide (regardless of whether it is graphene oxide, porous graphene oxide, or chemically modified graphene oxide) effectively disrupts the packing of polymer chains while increasing CO2 adsorption in the matrix and reorienting the distribution of water.

[0087] In certain embodiments, the nanofillers are aligned or substantially aligned in the selective layer, i.e., the planes formed by the 2D shape of the nanofiller particles are parallel or substantially parallel, preferably parallel or substantially parallel to the plane formed by the selective layer. In other words, the nanofiller particles are typically aligned or parallel along their two larger dimensions. Alternatively stated, the nanofillers can be coplanar or substantially coplanar, e.g., coplanar or substantially coplanar with the plane formed by the selective layer. As used herein, "substantially" aligned, parallel, or coplanar typically means that 75% or more, preferably 90% or more, e.g., 95% or more of the nanofiller particles deviate by ±45° or less, preferably ±25° or less, more preferably ±10° or less from the plane formed by the selective layer. Alignment within the plane of GO is due to a decrease in the surface free energy of the GO-based filler and results in improved gas permeation. Further, in many cases, it may be necessary to have nanofillers that align with a selective layer in the range of hundreds of nanometers in thickness, since the lateral dimensions could otherwise potentially exceed the thickness of the selective layer.

[0088] In the case of porous graphene oxide, the nanofiller particles typically have pores that are evenly distributed. Typically, the pores in each nanofiller particle are evenly distributed throughout each layer of the graphene oxide. Thus, in certain embodiments, each layer of graphene oxide in the nanofiller particles is porous, and preferably, each layer of graphene oxide has a similar number of pores (e.g., within ±50%, such as within ±25%, or within ±10% of the average number of pores per layer of the particles). In certain embodiments, the number of pores per 100 nm 2 varies by up to ±50%, such as up to ±25%, or such as up to ±10%. The even distribution of pores can be achieved by generating the porous nanoparticles in a dispersed state (e.g., hydroxide treatment on already cleaved monolayers in a GO dispersion). Creating pores in a dispersed state ensures that pores are created in all monolayers of the graphene oxide nanofillers. When porous graphene oxide is produced in a solid state (e.g., thermal annealing of aggregated GO powder), then the porosity is typically not uniform. Thus, typically, the porosity in graphene oxide is generated in a dispersion. In certain embodiments, the pores in graphene oxide are not created by solid state treatment of graphene oxide (e.g., thermal annealing of solid GO).

[0089] Mobile carrier To increase the number of reaction sites for CO2 to interact in the selective layer, low molecular weight CO2 affinity components can be added. These diffuse through the polymer matrix and improve permeation and are thus referred to herein as mobile carriers. The use of mobile carriers in the selective layer increases performance and in particular aids in CO2 / CH4 separation and resists carrier saturation phenomena. Mobile carriers are typically dispersed in a polymer matrix containing an amine polymer.

[0090] A surprising aspect of the present invention is that a composite membrane containing a mobile carrier has an increased CO2 permeability and an increased CO2 flux under industrially relevant conditions compared to the same module without the mobile carrier.

[0091] In use, the amine groups of the polymer matrix reversibly react with CO2 in the presence of water that transports CO2 across the composite membrane. Polymers such as polyvinylamine and polyallylamine contain a high density of amine groups relative to the hydrocarbon content in the repeating units of the polymer, but their effects contributing to the increase in CO2 transport depend on the access of CO2 to the amine groups and the proximity to form continuous channels for the reaction and movement of CO2. However, the amine groups in the polymer matrix are locked in the polymer, restricting their mobility.

[0092] To increase diffusivity, a mobile carrier that is also a CO2 affinity is added to the matrix. The addition of these mobile carriers not only increases the density of the CO2 affinity moieties in the polymer matrix but also increases the mobility of the CO2 reactive species and thus increases the CO2 diffusivity in the host matrix. Important characteristics of such mobile carriers include (1) low molecular weight (higher mobility), (2) high CO2 uptake capacity, and (3) the ability to form a weak bond with CO2 that improves the transport of CO2 through the water-swellable composite membrane matrix (reversible CO2 association / dissociation) and promotes its release on the permeate side.

[0093] The mobile carrier is an ionic liquid or an amino acid salt.

[0094] An ionic liquid is a salt that is liquid at 25 °C and atmospheric pressure. Room temperature ionic liquids often contain bulky and asymmetric organic cations based on heterocycles such as 1-alkyl-3-methylimidazolium, 1-alkylpyridinium, fluorosulfonyl-trifluoromethanesulfonylimide (FTFSI) N-methyl-N-alkylpyrrolidinium, and ammonium ions. Phosphonium cations are also possible. A wider range of anions are used, ranging from simple halides to inorganic anions such as tetrafluoroborate and hexafluorophosphate, and small or large organic anions such as bistriflimide, acetate, cyanamide, triflate, or tosylate.

[0095] Suitable ionic liquids include [Emin][OAc], [Emim][Cl], [Emim][dicyanamide], and 1-butyl-3,5-dimethylpyridinium bromide. In certain embodiments, the ionic liquid comprises 1-ethyl-3-methylimidazolium ([Emim]) or 1-butyl-3-methylimidazolium as the cation. 1-Ethyl-3-methylimidazolium acetate ([Emim][OAc]) is particularly preferred. Typically, the ionic liquid has a melting point in the range of 25 to 100 °C.

[0096] An amino acid salt is a salt of a compound containing a COOH group and an amino group that can be primary, secondary, or tertiary. The salt preferably forms with the acid moiety, i.e., the salt is the cation and the amino acid forms the anion.

[0097] Suitable amino acid salts are salts of any naturally occurring amino acid, e.g., any essential amino acid, preferably salts of Gly, Arg, Cys, or Pro. Preferred amino acid salts are salts of proline (e.g., ι-proline) such as the potassium salt of proline (i.e., "ProK"). The cation in the salt is ideally an alkali metal.

[0098] The room temperature ionic liquid [Emim][OAc] reacts with CO2 via the carbene pathway to form a carbene-CO2 adduct. One of the main advantages of using [Emim][OAc] as a mobile carrier is that its carbene pathway interaction with CO2 does not affect the viscosity of the solution and can be beneficial in reducing the mass transfer resistance of the selective layer during CO2 adsorption.

[0099] The carbene-CO2 adduct (shown in Scheme A below) has also been reported to have faster diffusion.

Chemical formula

[0100] The secondary amino acid ProK (potassium L-prolinate) reacts with CO2 to form the carbamate and bicarbonate / carbonate species represented by Scheme B above).

[0101] PZEA-Sarc is an amino acid salt containing one primary amine, two secondary amines (one from sarcosine), and one tertiary amine. CO2 interacts with this mobile carrier to form primary and secondary monocarbamates (Scheme C above).

[0102] The mobile carrier is typically present in the selective layer in an amount of 1.0 to 40% by weight, preferably 2.0 to 30% by weight, more preferably 5.0 to 25% by weight (dry weight). Particularly preferred ranges for ionic liquids are 2.0 to 40% by weight, preferably 5.0 to 15% by weight, and particularly preferred ranges for amino acid salts are 5.0 to 40% by weight, preferably 15 to 25% by weight.

[0103] Viewed from another aspect, the present invention is a composite membrane suitable for separating a gas from a gas mixture, comprising a selective layer coated on a support, wherein the selective layer is a) a polymer matrix comprising at least 50% by weight of an amine polymer, b) 0.05 to 5.0% by weight of a graphene oxide nanofiller, c) 1.0 to 40% by weight of a mobile carrier selected from an ionic liquid or an amino acid salt, and provides a composite membrane.

[0104] In one embodiment, the present invention is a composite membrane suitable for separating a gas from a gas mixture, comprising a selective layer coated on a support of a hollow fiber or a flat sheet, wherein the selective layer is a) a polymer matrix comprising an amine polymer and preferably PVA, b) a porous graphene oxide or PEG-modified nanofiller, and optionally, c) a mobile carrier selected from an ionic liquid or an amino acid salt, and relates to a composite membrane.

[0105] All of the preferred embodiments discussed above are also applicable to this embodiment.

[0106] Other optional layer components The selective layer of the present invention preferably consists essentially of a polymer matrix, a graphene oxide nanofiller, and a mobile carrier. Thus, typically, these materials are the only materials used in the selective layer other than any necessary additives at trace levels, such as stabilizers, antioxidants, or residual solvents. The combination of the polymer matrix, the graphene oxide nanofiller, and the mobile carrier preferably forms at least 95% by weight of the selective layer, such as at least 98% by weight, particularly at least 99% by weight of the selective layer. The selective layer of the present invention preferably consists of a polymer matrix, a graphene oxide nanofiller, and a mobile carrier.

[0107] Support formation The support can be in the form of a flat sheet or a hollow fiber. Techniques for making these supports are known in the art.

[0108] The hollow fiber process typically involves dissolving a support material in a suitable solvent to form a solution, and then spinning the solution to form hollow fibers. In the spinning process, the support solution is fed to a spinneret by the force of a pump and then extruded. The bore liquid is passed through the center of the spinneret to ensure that the fibers formed are hollow. The fibers flow out from the base of the spinneret and finally enter a coagulation bath. However, there is an air gap between the base of the spinneret and the coagulation bath. The presence of the air gap allows for the evaporation of the solvent and also allows the fibers to stretch and straighten under their own weight. This hollow fiber spinning technique is well known to those skilled in the art.

[0109] When there are hollow fibers, it will be understood that the selective layer can be formed on the outside or inside of the fiber (however, it is preferably not both). The external coating can be simply carried out using spray or dip coating in a solution containing elements for the selective layer. The internal coating of the hollow fiber involves circulating the solution inside the hollow fiber lumen and then drying it as in the case of an external dip coating. The procedure is repeated until a thin, defect-free layer of the selective layer polymer is formed. It is preferred to coat on the inside.

[0110] In a preferred embodiment, the support can have a porous lower layer with a thin, high-density upper layer. The support with a high-density upper layer is preferably a hollow fiber support and can ideally be formed from PPO or PSf. The high-density upper layer is formed during the spinning process.

[0111] Composite membrane formation The first step in forming the composite membrane of the present invention involves casting a solution containing a polymer matrix, graphene oxide nanofiller, and a mobile carrier onto a support. The support can typically be in the form of a flat sheet or a bundle of hollow fibers. Casting of the solution is carried out using known techniques. There are various options available for coating the support with a thin film, including dip coating, vapor deposition, spin coating, and spray coating. These techniques are considered "casting" according to the present invention.

[0112] When the support is a hollow fiber, the term "casting" typically means immersion or spray coating of the hollow fiber support. When the selective layer is located inside the hollow fiber, the term "casting" includes the above process.

[0113] The casting solution is typically aqueous, but other solvents can be used.

[0114] The casting solution preferably has a solids content of 0.01 to 20% by weight, preferably 0.05 to 10% by weight, preferably 0.05 to 5.0% by weight, more preferably 0.1 to 3.0% by weight (as used herein, "solids" refers to the content of only the polymer matrix and graphene oxide nanofiller in the casting solution). The mobile carrier is likely to dissolve in the casting solution.

[0115] For casting on a flat sheet support, the casting solution preferably has a solids content of 0.1 to 10% by weight, preferably 0.5 to 5.0% by weight. For casting on a hollow fiber support, the casting solution preferably has a solids content of 0.01 to 1.0% by weight, for example, 0.05 to 0.5% by weight. Clearly, the relative amounts of the polymer matrix, graphene oxide nanofiller, and mobile carrier in the casting solution depend on the desired concentration in the selective layer polymer.

[0116] When the solution is cast on a flat support, a bar coating method is used to create the selective layer. The bar coating method flattens the selective layer, aligns the nanofillers, and optionally reduces the thickness of the selective layer. A Meyer bar is typically used, which is shown in Figure 3.

[0117] Composite membrane The thickness of the selective layer varies depending on the concentration of the solute in the casting solution, with higher concentration solutions giving thicker membranes. However, the thickness can also be adjusted using a casting knife or reduced using the bar coating method. Thus, in typical embodiments, the membrane is made by casting. In particular, the polymer matrix components are typically dissolved in the casting solution prior to evaporation. The polymer matrix containing the amine polymer is typically not formed, for example, by interfacial polymerization.

[0118] It will be apparent that the polymer matrix containing the amine polymer, the graphene oxide nanofiller, and the mobile carrier all typically form a monolayer.

[0119] The thickness of the selectivity layer of the present invention can be less than 100 μm, preferably less than 10 μm, more preferably less than 1 μm, and even more preferably less than 500 nm. Typically, the thickness of the selectivity layer is in the range of 20 nm to 100 μm, preferably 50 nm to 10 μm, preferably 100 nm to 5 μm, more preferably 100 nm to 1 μm, and even more preferably 100 nm to 500 nm. A thinner selectivity layer tends to have a higher transmittance value but lower strength. A selectivity layer with a thickness of less than 200 nm is particularly preferred.

[0120] It will also be understood that any selectivity layer should ideally be defect-free.

[0121] The thickness of the support on which the selectivity layer can be carried can vary, but this can be on the order of 50 to 500 μm, for example, approximately 100 μm. However, it will be understood that the present invention encompasses both the use of a flat support and the use of a hollow fiber support. When the support is a hollow fiber support, the thickness of the support is regarded as the wall thickness of the hollow fiber. The support should be porous.

[0122] After the formation of the selectivity layer on the support, the solvent is removed, for example, by evaporation. This can be achieved using gentle heat, for example, about 60 °C, if necessary.

[0123] To avoid any possibility of loss of the selectivity layer forming material to the support, there is usually a reasonable difference between the average molecular weight of the selectivity layer polymer and the molecular weight cut-off of the support structure. Such a difference can be greater than about 10,000, for example, greater than about 15,000, for example, greater than about 20,000, and particularly greater than 50,000. Alternatively, a pore filling material can be used before casting using a casting solution containing the selectivity layer components.

[0124] Next, if desired, the formed selectivity layer can be crosslinked. Crosslinking can be carried out chemically using a crosslinking agent such as glutaraldehyde or ammonium fluoride. However, as described above, it is preferred that the selectivity layer is not crosslinked.

[0125] This is also the stage of the manufacturing process where the composite membrane can be heat-treated.

[0126] The resulting membrane functions as a fixed-site carrier (FSC) for the transport of gases, such as carbon dioxide, due to the high concentration of amino groups.

[0127] The composite membrane of the present invention can be formed into a module for use in a gas separation system.

[0128] Use Tests have shown that the composite membrane of the present invention can be used for at least 800 hours without any significant loss of activity, which forms a further aspect of the present invention.

[0129] The composite membranes of the present invention operate most effectively when they are wet. Therefore, prior to use of the composite membranes, they can be swollen, for example, in the presence of water in the form of steam. Ideally, the composite membranes of the present invention should operate in a wet environment, such as at a relative humidity of at least 75%, for example, a humidity of 75 - 100%.

[0130] Therefore, the process for the production of the composite membranes of the present invention preferably further comprises the step of contacting the composite membranes with water, such as water vapor, and / or the step of operating the membranes in a wet environment.

[0131] The presence of water vapor in the composite membrane is assumed to promote the transport of carbon dioxide across the composite membrane.

[0132] Examples of gases that can be separated from a gas mixture using the composite membrane of the present invention include carbon dioxide having various components such as nitrogen, methane, carbon monoxide, oxygen, volatile organic compounds, or hydrogen. Separation of mixtures containing hydrogen is also envisioned. These gases can occur in any situation, such as industrial and domestic gas streams.

[0133] During use, the gas mixture to be separated typically flows across the composite membrane under pressure. The temperature used can vary, but is typically in the range of 10 - 90°C, preferably 20 - 65°C. Temperatures above 40°C, above 50°C, or above 55°C are preferred. However, it is possible to operate at even higher temperatures, and separation at temperatures above 100°C can provide improved results.

[0134] Preferably, the composite membrane is used to separate carbon dioxide from nitrogen or methane. In this latter regard, the composite membrane of the present invention can thus have applications in fields where these gases are present in mixtures such as flue gas, biogas (e.g., biogas upgrading), natural gas (e.g., natural gas upgrading), syngas, or optionally sweetening of natural gas.

[0135] The pressure at which the gas mixture is applied to the composite membrane is important because it affects the flow across the composite membrane and potentially its selectivity. Thus, the supply pressure can be in the range of 0.5 - 100 bars, for example, 1.0 - 20 bars, particularly 1.5 - 15 bars. The supply pressure can be in the range of 1 bar (typical flue gas) - 80 bar (typical natural gas). The composite membrane of the present invention is most useful for applications at pressures below 10 bar.

[0136] Particularly when the supply gas is at low pressure, e.g., 1.0 - 5 bars, it can be advantageous to operate the composite membrane using a vacuum on the permeate side of the membrane. This can improve the performance of the composite membrane.

[0137] The composite membrane of the present invention preferably exhibits a selectivity of at least 20, more preferably at least 50, particularly at least 100, and most particularly at least 150. The selectivity is measured as described in the examples.

[0138] The transmittance in GPU (1 GPU = 10 -6 cm 3 (STP) cm -2 s -1 cmHg -1 = 3.35×10 -10 mol m -2 s -1 Pa -1 ) is preferably at least 700, preferably at least 800, preferably at least 1000, and preferably at least 1050. Preferred ranges include 1050 to 6000, preferably 1080 to 4000.

[0139] The composite membrane of the present invention has been found to have an increased flux compared to a membrane that does not contain nanofillers and does not contain mobile carriers. In certain embodiments, the membrane of the present invention has a CO2 flux of more than 300 NLm -2 h -1 super, preferably more than 350 NLm -2 h -1 super, more preferably more than 370 NLm -2 h -1 super. Preferably, the membrane of the present invention has a CO2 flux in the range of 370 to 1000 NLm -2 h -1 .

[0140] Here, the present invention will be further described with reference to the following non-limiting examples and drawings.

Brief Description of the Drawings

[0141]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Example

[0142] Materials Poly(allylamine hydrochloride) (Mw = 120,000 - 200,000) was purchased from Thermo Fisher Scientific, Sweden, purified, and modified to sterically hindered polyallylamine.

[0143] For the study of hollow fibers, graphene oxide powder (2.5 wt% in water) was supplied by Graphene-XT, Italy and used as a diluted dispersion.

[0144] Polyvinyl alcohol (Mw = 89,000 - 98,000, 89% hydrolyzed), 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide, N-hydroxysulfosuccinimide, 8-arm-poly(ethylene glycol)-NH2 (hexaglycerol core, Mn = 10,000) were used as received from Sigma-Aldrich. A polyvinylidene fluoride (PVDF) ultrafiltration membrane (50k MW) with a polypropylene (PP) substrate was obtained from Synder Filtration, USA.

[0145] 3M (trademark) Fluorinert (trademark) Electronic Liquid FC-72 was used as received from Kemi-Intressen, Sweden.

[0146] L-proline Reagentplus (registered trademark) (≥99 wt%), 1-(2-aminoethyl)piperazine (99 wt%), 1-ethyl-3-methylimidazolium acetate (97 wt%), and sarcosine (N-methylglycine) (98 wt%) were purchased from Sigma-Aldrich.

[0147] Poly(p-phenylene oxide) (PPO) hollow fibers for use in a hollow fiber support having an inner diameter of 350 μm and an outer diameter of 540 μm were obtained from Parker A / S Norway.

[0148] The CO2 / N2 mixture (10 vol% CO2 in N2), CO2 / CH4 mixture (40 vol% CO2 in CH4), N2, and CH4 (99.95%) used for the permeation tests were supplied by AGA, Norway. Hydrogen peroxide (H2O2, 30% in water) used for the modification of GO was supplied by Sigma Aldrich, Norway.

[0149] Characterization methods Chemical changes to the nanofiller were monitored by Fourier transform infrared (FTIR) spectroscopy using a Thermo Nicolet Nexus spectrometer equipped with a smart endurance reflection cell in attenuated total reflectance mode with a diamond crystal. An average of 16 scans with a resolution of 4 cm -1 was used to construct spectra in the range of 4000 cm -1 and 800 cm -1 .

[0150] The surface chemical composition of the synthesized GO was analyzed using X-ray photoelectron spectroscopy (XPS, XPS-Theta Probe, Thermo Fisher Scientific Co., USA) with a monochromatic Al Kα source having a C correction of 284.5 eV.

[0151] The film morphology was analyzed by Field Emission SEM APREO (FEI, Thermo Fisher Scientific, USA) equipped with an in-lens detector in immersion mode. Before analysis, the samples were sputter-coated with 8 nm Pd / Pt alloy.

[0152] Polymer Polymer Example 1 - Synthesis of Sterically Hindered Polyallylamine (SHPAA) - for flat sheet membranes Sterically hindered polyallylamine was obtained by modifying polyallylamine purified with 2-bromopropane. As shown in Reaction Scheme 1, polyallylamine reacts with 2-bromopropane in methanol under reflux conditions at 50 °C in the presence of stoichiometric amounts of KOH to give poly-N-isopropylallylamine.

Chemical formula

[0153] The poly-N-isopropylallylamine prepared herein has an estimated Mw of 120 - 250 K.

[0154] Polymer Example 2 - Synthesis of Sterically Hindered Polyallylamine (SHPAA) - for hollow fibers Polyallylamine hydrochloride was purified by reacting it with an equivalent amount of KOH in MeOH-precipitated KCl. Subsequently, the purified PAA was modified to poly-N-isobutylallylamine by reaction with an equivalent amount of 2-bromobutane and KOH in MeOH at 50 °C (Scheme 2). The resulting polymer was purified by separating the precipitated KCl crystals and subsequently drying at 60 °C in an N2 atmosphere.

Chemical formula

[0155] The prepared polymer has an estimated Mw of 120 - 250K.

[0156] Mobile carrier Synthesis of mobile carrier Equivalent amounts of ι - proline and KOH were dissolved in DI water to form a solution with 10 wt% total solids. The solution was then stirred at high speed overnight at room temperature to form potassium L - prolinate (ProK).

[0157] Similarly, equivalent amounts of 1 - (2 - aminoethyl)piperazine and sarcosine were stirred at room temperature in the calculated amount of DI water to obtain 37.7 wt% 2 - (1 - piperazinyl)ethylamine sarcosinate (PZEA - SARC).

[0158] 1 - Ethyl - 3 - methylimidazolium acetate ([Emim][OAc]) was dissolved in DI water to form a 10 wt% solution and stirred overnight at room temperature.

[0159] Nanofiller The nanofiller for the example of the flat sheet membrane was prepared as follows.

[0160] GO Example 1 Synthesis of graphene oxide Graphene oxide for use in flat sheet membranes was synthesized by the modified Hummer's method. 10 g of graphite powder was mixed with 450 ml of sulfuric acid while stirring at 5 °C for 1 hour. Then, 30 g of potassium permanganate was added and stirred for 30 minutes, resulting in a color change from black to dark green. The solution was further heated to a maximum of 40 °C for 1 hour. 450 ml of deionized water was added dropwise while taking care to avoid a sharp increase in temperature. This solution was characterized by a color change to dark brown at this point. Then, the temperature was maintained at 95 °C for 30 minutes, followed by the addition of 300 ml of 10% hydrogen peroxide solution and then stirring for 15 minutes. The color change to light brown indicated the successful synthesis of graphene oxide. Then, GO was purified multiple times through a Whatman glass microfiber filter using approximately 5 L of 10% hydrochloric acid and subsequently washed in 3 L of acetone. Then, the filtered GO cake was dried under vacuum at 40 °C for 2 days to obtain graphene oxide flakes named GO herein.

[0161] 2 mg / ml of GO -1 A solution was prepared and tip sonicated for a maximum of 3 hours, followed by bath sonication for 30 minutes. AFM analysis revealed the presence of flakes in this solution with lateral dimensions of 1 μm or more.

[0162] GO Example 2 - Physical Modification of Graphene Oxide To physically modify GO flakes for better diffusion of permeates, random pores were introduced by the hydrothermal treatment of GO Example 1 using hydrogen peroxide. Using 1 M NaOH, the pH of 75 ml of 1 mg / ml -1 GO solution (diluted from the previously prepared 2 mg / ml -1 solution) was adjusted, and the pH was adjusted to 10 using 1 M NaOH solution. The mixture was stirred at high speed for 5 minutes, followed by bath sonication for 10 minutes. Then, 10 ml of 3% diluted hydrogen peroxide solution was added to the mixture, and the solution was stirred at high speed for 10 minutes, followed by bath sonication for 10 minutes. Then, the resulting mixture was treated in a Teflon autoclave at 180 °C for 6 hours and then cooled to room temperature.

[0163] The resulting dispersion of pGO (porous graphene oxide) in water had a concentration of about 1 mg / ml. The flakes were expected to inherit the same lateral dimensions as the GO from Example 1. -1 (pGO).

[0164] GO Example 3 - Chemical Modification of Graphene Oxide PEG groups were grafted onto the GO surface using an EDC coupling reaction. The synthesized GO dispersion in water is acidic. However, to activate multiple sites for PEG grafting for amide bond formation, 20 ml of the 4 mg / ml GO dispersion from Example 1 was treated with an equal volume of 3 M NaOH and subsequently bath sonicated at 25 °C for 1 hour to introduce additional carboxylic acid groups on the GO surface. This reaction enabled the hydrolysis of the ester in the GO to carboxylic acid groups. Then, dilute HCl was added to neutralize the solution, followed by dilution to 1 mg / ml to obtain a dispersion of carboxylated GO in water. Next, 100 mg of NHS and 150 mg of EDC were added to the GO-COOH dispersion, followed by bath sonication in ice for 30 minutes to activate the catalyst. Then, 200 mg of 8-arm PEG was added to the mixture, and the solution was stirred at room temperature for 24 hours. The solution was then centrifuged at 7000 rpm to remove aggregates, and then the dispersion was dialyzed in water using a Dialysis membrane Spectra / Por® 3 to remove the catalyst, salts, and other unreacted components. The residual dispersion had a GO-PEG concentration of about 1 mg / ml. -1 -1 -1

[0165] The flakes were expected to inherit the same lateral dimensions as the GO from Example 1.

[0166] The nanofiller for the hollow fiber composite membrane example was prepared as follows.

[0167] GO Example 4 - For GO / pGO Nanosheet - Hollow Fiber Composite Membrane An important parameter of GO flakes that affects gas permeation performance is the flake size (lateral dimension). Different suppliers of GO provide dispersions with different flake sizes. For the hollow fiber composite membrane experiments, the inventors used GO from a commercially available supplier, Graphene-XT.

[0168] The as-received GO dispersion was first diluted to 1 mg g -1 in solution and subsequently the pH was adjusted to 10 using 1 M NaOH. The diluted solution was sonicated in a bath sonicator at 25 °C for 30 minutes. The dispersion was then subjected to an ultrasonic dis-integrator (Vibra-Cell™ Ultrasonic Liquid Processor) at 60% amplitude in an ice bath with 3-second pulses followed by 2-second breaks. This procedure was carried out to simultaneously exfoliate and control the size of the GO flakes by varying the operating time.

[0169] The sonication was carried out for 3, 6, or 9 hours, and the resulting GO flakes were designated as GO3, GO6, and GO9, respectively.

[0170] Single layers of GO in aqueous dispersion were obtained using the sonication-assisted exfoliation procedure. To ensure the reproducibility of the methodology, the concentration of the GO dispersion was kept constant at 2 mg mL -1 and the sample volume was maintained at 300 mL in all procedures. The sonication process imparted random fragmentation of the 2D nanosheets induced by mechanical disruption of defective sp 3 regions. These random shreds were followed by crack propagation, resulting in a reduction in flake size.

[0171] AFM analysis revealed the presence of large flakes with lateral dimensions greater than 1 μm for GO3. Subsequent sonication yielded smaller flakes for GO6 and GO9, which decreased to the range of 400 - 800 nm and less than 500 nm, respectively. All samples were then subjected to hydrothermal treatment for the introduction of random pores.

[0172] The size-controlled GO dispersion was also hydrothermally treated to introduce random non-selective pores. The GO dispersion was mixed with a 3 wt% H2O2 solution, and the mixture was vigorously stirred for 10 minutes, followed by bath sonication for 10 minutes. Immediately thereafter, the mixture was treated at 180 °C for 6 hours in a Teflon autoclave. The resulting pGO dispersions derived from the GO3, GO6, and GO9 samples were named pGO3, pGO6, and pGO9, respectively.

[0173] The successful introduction of non-selective pores in the GO nanosheets by hydrothermal treatment was confirmed by representative S(T)EM imaging of GO3 and pGO3.

[0174] Representative imaging of the pGO flakes shows a further reduction in flake size after hydrothermal treatment. This size reduction is confirmed by the relative increase in the presence of carbonyl groups (observed from FTIR) exposed along the edges of pGO compared to GO.

[0175] The chemical changes of the GO nanoplatelets during the hydrothermal treatment process were studied using FTIR spectroscopy. The sonication procedure had little effect on the chemical structure of the GO and pGO nanoplatelets.

[0176] However, recognizable peak changes appeared between GO and pGO.

[0177] Support Flat sheet The support of the flat sheet was PVDF as a flat sheet. It had an MWCO of 50,000. The PVDF support was first washed with tap water at 45 °C for 1 hour, followed by washing with DI water for 30 minutes to remove the pore protecting agent. The support was dried overnight at room temperature before coating with the casting solution using a bar coater (see Figure 3). The pores were filled with the pore filler FC-72 to avoid penetration of the casting solution into the pores.

[0178] Hollow fiber support PPO was used as the hollow fiber support. It had a MWCO of 30,000 - 50,000. To fabricate the hollow fiber support, a PPO support made by conventional hollow fiber spinning technology was vertically suspended in a sealed state using a paper clip that also created fiber tension and avoided loosening. The fibers were washed twice with DI water to remove any dust particles that might be attached to the surface and then dried at room temperature.

[0179] Composite membrane formation Flat sheet membrane A 4 wt% PVA solution in water was prepared by dissolving PVA pellets in deionized water at 80 °C for 4 hours under reflux conditions. The modified SHPAA solution in methanol (Polymer Example 1) was dried under vacuum at 60 °C overnight to remove the residual solvent. The resulting as - received polymer was then dissolved in water at room temperature for 24 hours to obtain a 6 wt% solution.

[0180] For the flat sheet support, a casting solution concentration of about 1 wt% “solid” was used. The SHPAA / PVA blend polymer solution consisted of 90 wt% SHPAA and 10 wt% PVA based on the total polymer “solid” present in the solution. The % amount of nanofiller (Examples GO1 - GO3) was measured relative to the total amount of polymer and nanofiller present in the casting solution. For example, 0.5 wt% GO in the SHPAA / PVA blend indicates that the amount of GO is 0.5% of the total “solid” content, i.e., of the polymer and nanofiller in the solution.

[0181] The casting solution contained a 1 wt% blend of nanofiller [0.5 wt% and 99.5 wt%], and the SHPAA and PVA polymer blend was applied to a PVDF support to create a selectivity layer with a thickness of less than 200 nm.

[0182] In this example, the selectivity layer was applied by the bar - coating method as schematically represented in Figure 3.

[0183] Hollow fiber composite membrane The modified, purified, and dried SHPAA (Polymer Example 2) was dissolved in DI water to obtain a 6 wt% solution, and the polymer solution was stirred at room temperature for at least 2 days to obtain a transparent polymer solution. In the case of PVA, a 4 wt% solution was prepared by dissolving PVA pellets in DI water at 80 °C for 4 hours under reflux conditions.

[0184] To prepare the casting solution, the calculated amount of the polymer solution was added to DI water and diluted to a total “solid” casting solution concentration of 0.15 wt%. As described in Equations 1 and 2 respectively, the amount of the mobile carrier was measured as the ratio of the polymer phase, while the amount of the nanofiller (Example GO4) was measured from the perspective of the total solid content.

Equation

[0185] Coating of the thin selective layer is achieved by dip-coating the fibers using the casting solution at a constant low speed (in the range of 6 - 8 cm s -1 in both directions with a 30-minute time interval during the continuous coating procedure. Coating in the opposite direction ensures a defect-free selective layer. Additionally, the low viscosity of the casting solution due to the low solid content leads to a uniformity of the thickness of the selective layer that is independent of the coating speed and the filler loading. The hollow fibers are then dried at room temperature and subsequently at 60 °C under vacuum for 2 hours to remove the residual solvent components. The resulting hollow fibers exhibit a shiny appearance due to the presence of the ultrathin selective layer coating. The thickness of the selective layer is approximately 200 nm.

[0186] To assemble the coated hollow fiber composite membrane into a module, a small number of fibers (in the range of 2 - 5) were carefully inserted into a pre-assembled stainless steel hollow fiber module designed using 1 / 4 inch 3 / 8 inch Swagelok™ fittings. The ends were then sealed using an epoxy adhesive. The bore side of the fibers was opened by knocking off the cured adhesive on the extension.

[0187] Composite membrane form Flat sheet composite membrane A stable dispersion of GO-based fillers having both PVA and SHPAA / PVA blend matrix was obtained over the entire range of 0.2 wt% - 1 wt% GO filler loading. The blend polymer solution of SHPAA / PVA consisted of 90 wt% SHPAA and 10 wt% PVA based on the total polymer “solids” present in the solution.

[0188] The casting solution concentration was maintained at 1 wt% solids. Representative cross-section SEM imaging of the 0.2 wt% pGO filled selective layer reveals the presence of an ultra-thin selective layer with a thickness of less than 200 nm on the PVDF porous support. The surface image of the selective layer shows a visible difference between the neat polymer and the one filled with nanofillers.

[0189] The neat polymer layer shows a relatively smooth surface, but dark patches are observed in the composite membrane samples, which may be associated with flakes of aligned GO-based fillers. No obvious protrusions or aggregations of nanofillers are observed from the smooth surface, which confirms the alignment of GO nanoflakes along their two larger dimensions parallel to the coating surface. The in-plane alignment of GO is due to the reduction of the surface free energy of the GO-based fillers and the mechanically forced alignment of the thin 2D flakes to the tangent of the cylindrical surface of the bar at the contact points. Figure 2 shows the cross-section SEM images of (A) the porous PVDF support (B) the composite membrane containing SHPAA / PVA with 0.2 wt% pGO.

[0190] Laboratory scale gas permeation performance Flat sheet membrane with a facilitated transport SHPAA polymer matrix The facilitated transport membrane transports CO2 via a reactive pathway in addition to the solution diffusion mechanism. The effect of facilitated transport is brought about by amine groups attached to the main chains of the SHPAA polymer matrix that reversibly react with CO2 in the presence of water.

[0191] Gas permeation performance The composite membrane was evaluated for gas permeation performance using a wet mixed gas permeation test rig. The feed consisted of a 90 / 10 v / v CO2 / N2 mixture, or a 40 / 60 v / v CO2 / CH4 mixture. The flow rate of the feed was 300 ml min in the CO2 / N2 test -1 and 400 - 600 ml min in the CO2 / CH4 test -1 . The difference in the feed flow rates mainly compensated for the difference in membrane area and targeted a very low stage cut of less than 0.5%. The sweep gas for the CO2 / N2 test was CH4 and N2 was used for the CO2 / CH4 test. In both cases, the feed and sweep gas streams were humidified with a bubble tank in front of the membrane module. The shell side of the membrane was used for the feed gas and the bore side of the fibers was used as the permeate / sweep side. The pressure on the feed side was maintained constant at 1.7 bar in the CO2 / N2 test and varied between 2 - 20 bar in the CO2 / CH4 test. The pressure on the sweep side was kept at 1.02 bar. The operating temperature was maintained at 35 °C for all tests. The outlet gas compositions on both the feed and sweep sides were continuously monitored using pre-calibrated gas chromatographs (490 Micro GC, Agilent for the CO2 / N2 test and MG5, SRI Instruments Inc. for the CO2 / CH4 test). The permeability of component "i" was obtained using the following equation, [Number] where the total permeate flow V p was measured at the outlet in ml s using a bubble flow meter in steady state -1 . JPEG0007697969000011.jpg7170 and yi indicate the mole fractions of water and permeating species in the permeate stream, respectively. The partial pressures p i,f , p i,r , and p i,p of species "i" in the feed, retentate, and permeate are in cmHg -1 . The permeability of a component is expressed in GPU, where 1 GPU = 10 -6 cm 3 (STP) cm -2 s -1 cmHg -1 = 3.35x10 -10 mol m -2 s -1 Pa -1 . The separation factor is calculated using the permeabilities of the individual components according to the equation.

Equation

[0192] As shown in Figure 5, the separation performance of the neat SHPAA / PVA blend membrane is a CO2 permeability of 383 GPU and a CO2 / N2 separation factor of 55.

[0193] The separation performance of composite membranes of flat sheets of SHPAA / PVA with GO-based fillers is summarized in Figure 5. At a low loading of 0.2 wt%, both GO and pGO effectively disrupt the packing of the polymer chains while increasing the CO2 adsorption in the matrix and reorienting the water distribution. Consequently, the CO2 permeability increases sharply to approximately 455 GPU for 0.2 wt% GO and 610 GPU for 0.2 wt% pGO. The selectivity of membranes containing both GO and pGO at a loading of 0.2 wt% rapidly decreases to approximately 34.

[0194] A reduced CO2 permeability of about 250 GPU with a CO2 / N2 separation factor of about 37 was observed in membranes with a low loading of GO-PEG. This performance reduction can be explained by the effect of the filler loading. Membranes with GO-PEG are characterized as a classical case of a rigid interface between the nanofiller and the polymer matrix. Due to the strong interaction between the -OH groups and the amine-containing facilitated transport matrix, there is a densely packed polymer interface between the GO surface and the adjacent polymer matrix. The closely packed chains that form the volume of the rigid interface in combination with the GO barrier properties result in a reduced CO2 permeability at higher loadings, while significantly increasing the CO2 / N2 selectivity. Thus, at a high loading of 1 wt% filler, the CO2 / N2 separation factor increased sharply to about 90, while reducing the CO2 permeability to as low as 205 GPU.

[0195] A similar effect is seen with higher loadings of GO nanoplatelets, where the tortuous path for N2 permeation due to the alignment of multiple layers of GO results in an increased CO2 / N2 separation factor of about 65 at a loading of 1 wt%. Thus, an optimal loading of the GO-based filler for improved permeation is observed at 0.2 wt%, beyond which the influence of the GO barrier properties, and thus the tortuosity for gas permeation, appears simultaneously. At this loading, the improved permeation of CO2 due to the disruption of polymer chains and increased adsorption brought about by the high aspect ratio nanoplatelets counteracts the resistance caused by the additional tortuosity of the impermeable platelets.

[0196] In the case of pGO, the presence of non-selective pores reduces the tortuosity, but the presence of a bulk distribution of small-sized impermeable platelets remains significant. Thus, the CO2 permeability decreases with increasing filler loading, but not as abruptly as in the case of GO or GO-PEG.

[0197] Figure 6 shows the effect of nanofiller loading on the permeability and selectivity.

[0198] Therefore, those skilled in the art can advantageously adjust the membrane properties, such as permeability or selectivity, by varying the nanofiller content.

[0199] Laboratory-scale gas permeation performance Hollow fiber membranes Both GO nanofillers and pGO nanofillers (GO Example 4) were dispersed in the SHPAA / PVA solution at two filler loadings of 0.2 wt% and 0.5 wt% in the selective layer. The total solids content in the casting solution (i.e., polymer + GO total) was kept low at 0.15 wt% (polymer basis), resulting in an ultrathin selective layer thickness of about 200 nm on the PPO hollow fibers. Mobile carriers were also added in the amounts shown below. Figure 4 shows the composite membrane of the present invention with neat polymer in the selective layer. The easy dip-coating procedure also ensures in-plane alignment of GO by shear alignment.

[0200] Both GO-based hollow fiber membranes and pGO-based hollow fiber membranes were developed, varying the size of GO by sonication time, resulting in GO3 / 6 / 9 and corresponding pGO3 / 6 / 9. The tests followed the protocol described above for flat sheet membranes. The neat polymer membrane with SHBPAA / PVA showed a CO2 permeability of 407 GPU and a CO2 / N2 selectivity of 32.2.

[0201] The addition of a small amount of pGO6 (optimized from Figures 7 and 8) at 0.2 wt% doubled the permeability up to 790 GPU while maintaining the selectivity at 31. Figures 7 and 8 demonstrate that there are both an optimal filler loading and an optimal size of the nanofiller.

[0202] Addition of mobile carriers Hollow fiber membranes with mobile carriers To increase the amount of reaction sites for CO2 to interact in the selective layer, low molecular weight CO2 affinity constitutive components were added. These diffuse through the membrane matrix and are called mobile carriers to improve permeation.

[0203] 0.2 wt% of pGO was dispersed in a polymer matrix containing 10 wt% of [Emim][OAc] or 20 wt% of ProK. The compositions were selected according to the optimal compositions detected experimentally and reported in the previous section. These resulting composite membranes with mobile carriers had increased performance for CO2 / N2 separation for the ProK-containing membranes, increasing the CO2 permeability up to 810 GPU as seen in Figure 9A. As seen in Figure 9B, even better results were demonstrated for the CO2 / CH4 separation performance. For these tests, a feed gas consisting of a 40 / 60 v / v CO2 / CH4 mixture was used to mimic a typical biogas composition. The membranes containing both mobile carriers were characterized by a significant increase in CO2 permeability while maintaining a CO2 / CH4 separation factor of approximately 20. At a feed pressure of 2 bar, the composite membrane containing 0.2 wt% of pGO6 with 20% of ProK peaked at a CO2 permeability of 825 GPU with a CO2 / CH4 separation factor of 20, while the neat polymer had a CO2 permeability of 497 GPU with a CO2 / CH4 separation factor of 21. The corresponding composite membrane containing 0.2 wt% of pGO6 without mobile carriers was limited to 727 GPU using the CO2 / CH4 feed gas mixture. A similar increase up to 782 GPU was observed in the CO2 permeability of the 10% [Emim][OAc]-containing membrane. The membranes containing mobile carriers showed increased CO2 permeability at a total upstream pressure of 2 bar using the CO2 / CH4 feed mixture due to the increased CO2 partial pressure and lower stage cut (higher feed flow).

[0204] The influence of pressure on the composite membrane performance was also investigated. The upstream pressure was increased from 2 bar to a maximum of 20 bar. The increase in the feed pressure resulted in further distinguishable separation performance. The composite membrane of the present invention is characterized by a carrier saturation phenomenon at a high partial pressure of CO2 in the feed. Since the availability of the fixed CO2 carrier (amine group) in the polymer matrix is limited, increasing the CO2 partial pressure in the feed gas leads to carrier saturation and thus reduces the CO2 permeability. Therefore, in all the systems discussed in this operation, as the pressure on the feed side increases, a decrease in the CO2 permeability is reflected, as seen in Figure 10.

[0205] Interestingly, the composite membranes with the filling of both 2D fillers of GO6 and pGO6 showed an increased resistance to the carrier saturation phenomenon, especially at pressures of 5 bar and 10 bar. Therefore, the corresponding CO2 permeability was maintained at 340 GPU and 450 GPU when compared with the neat polymer at 300 GPU at 5 bar, which is a typical operating pressure for the upgrading of biogas.

[0206] The composite membranes containing mobile carriers showed further resistance to the carrier saturation phenomenon, as expected, due to the increased availability of the CO2 carriers. This effect is still evident throughout the test pressure range for both the 10 wt% [Emim][OAc] filled membrane and the 20 wt% Pro-K filled membrane. These membranes showed CO2 permeabilities of 463 GPU and 468 GPU, respectively, and CO2 / CH4 separation factors of 24 and 25 at a feed pressure of 5 bar.

[0207] The selected membranes were scaled up and tested in the field.

[0208] Industrial tests of composite membranes with mobile carriers The hollow fiber composite membrane is fabricated using the same PPO support as described previously. The selective layer is applied by dip coating. The concentration of the coating solution is about 0.15 wt% total solids, and the amount of pGO is maintained at an optimized concentration of 0.2 wt% relative to the polymer content in the solution. The amount of mobile carrier is 10 wt% for [Emim][OAc] and 20 wt% for ProK.

[0209] 130 cm 2 ~200 cm 2 Three scaled-up modules each having a membrane area in the range of were assembled and tested. The material composition and mechanical aspects in the modules are summarized in Table 1.

Table 1

[0210] The gas test was carried out using flue gas from a stack (height: 105 m.) located close to the 5-stage cyclone preheater of the grey clinker production line at the Colacem Cement Plant in Gubbio (PG), Italy.

[0211] The sampling point was located 30 meters below the top of the stack. A hole was made in the side wall of the stack, and a vacuum pump was used to draw the flue gas from the stack into the membrane module. A 2μm ceramic filter, which is part of the gas sample probe (M&C model SP180H), was used to remove the particulate matter suspended in the flue gas. The temperature of the flue gas from the stack during the plant test was about 115°C. The composition of the dry flue gas during the plant test is summarized in Table 2.

Table 2

[0212] The flowchart of the in-field membrane permeation test is shown in Figure 1. Except for the removal of suspended particulate matter from flue gas by a filter, no further pretreatment was carried out on the feed gas. All gas transport pipes were covered by an electrically heated tube bundle to control the temperature desired for the test and mainly to avoid condensation of moisture. To control the supply pressure, one needle valve was placed on the non-permeate side. The gas composition of the feed gas was measured by the Multicomponent Analysis System ACF-NT of ABB S.p.A., while the gas composition of the non-permeate and permeate streams was measured by the HORIBA PG 350 SRM and using the TESTO model 350XL-350S gas analyzer. The flow rates of the three streams were measured by the TSI model 4143 and confirmed by several floating element flow meters selected according to the measured flow rates. Unless otherwise specified, all tests were carried out without vacuum or sweep gas. When sweep gas was used, the composition of the IP grade gas was 20.93% oxygen and 79.07% nitrogen.

[0213] Module 1: Influence of vacuum and stability test Module 1 showed a significant increase in CO2 flux by using vacuum on the permeate side of the membrane. The recorded flux was twice when compared with the use of sweep. The test was extended over a two-week period to simultaneously estimate the stability at the maximum performance of the membrane (Figure 11). Considering the lack of pretreatment of the stack gas still containing trace amounts of SO x and NOx, the long-term performance obtained was converted into significant material stability. In addition, the decrease in flux may be due to condensation of water in the pores and fluctuations in the real-time supply water content.

[0214] Module 2 and 3: Influence of water content Water plays an important role in the facilitated transport membrane. Since the three modules 1, 2, and 3 have three different chemical compositions with respect to amine chemistry, the modules were subjected to changes in the water content in the feed. An external evaporator system (HOVACAL) was used to force additional water into the feed stream. The results of the permeation performance are shown in Figure 12.

[0215] Generally, all modules showed an increase in the CO2 flux and purity with an increase in the total water content in the feed gas. This demonstrates that water as a carrier plays a major role in the activation of the amine groups present in the main chain of the facilitated transport polymer chains. In addition, due to the increase in the total amount of CO2-reactive groups present in the selective layer, high fluxes were evident in both the ProK-containing module and the [Emim][OAc]-containing module. [Emim][OAc] also functions as a physical solvent for CO2, and thus, due to the physical adsorption of both CO2 and N2 in the mobile phase, the purity of the module containing [Emim][OAc] is lower than that of the other two modules.

[0216] Modules 2 and 3 were also tested for stability in the presence of SO x and NO x . Composite membranes containing mobile carriers (ProK and [Emim][OAc]) were exposed to SO x and NO x in simulated flue gas, and little change in the purity of CO2 in the permeate was observed. The results are presented in Figure 13.

[0217] Comparison with existing membranes tested at pilot scale The performance of membrane modules at a larger scale when compared to laboratory-scale tests has been quantified in the literature using the CO2 flux and selectivity / purity in the permeate. However, 1 GPU = 10 -6 cm 3 (STP) cm -2 s -1 cmHg -1 = 3.35x10 -10 mol m -2 s-1 Pa -1 ) There are problems associated with the estimation of the flux from the perspective of -1 . The transmittance (flux) in the GPU can only be calculated based on assumptions about the driving force in the steady state across the membrane. Using these assumptions, the fabricated membrane has the following transmittance in the GPU, as shown in Figure 3.

[0218] These estimations are based on the following assumptions - · Due to the smaller module size and the use of continuous vacuum, the partial pressure of CO2 can be ignored on the permeate side · A flat CO2 concentration profile on the supply side, which is justified by the low stage cut (<5%) and the shorter length of the module

Table 3

[0219] Therefore, a surprising aspect of the present invention is that a membrane containing a mobile carrier has an increased CO2 transmittance and an increased CO2 flux under industrially relevant conditions compared to the same module without a mobile carrier.

[0220] The benchmark performance of the fabricated membrane against other membranes tested in an industrial pilot can be carried out using Table 5 obtained from Int.J.Greenhouse Gas Control, 86(2019), pp.191 - 200 as shown below. (References to each membrane can be found in the said article).

Table 4

[0221] Conclusion A composite membrane containing a GO-based filler in an ultrathin selective layer was fabricated and tested. The GO-based fillers were found to benefit the composite membrane in increasing the CO2 separation performance depending on their lateral dimensions and loading. The pGO fillers derived from size-optimized GO nanosheets at a loading of 0.2 wt% form continuous CO2 permeation pathways along the CO2-philic pGO surfaces with reoriented water channels surrounding the 2D structures in the matrix. These composite membranes are characterized by a high CO2 permeability of 780 GPU and a corresponding CO2 / N2 separation factor of 30. Composite membranes with mobile carriers that reversibly react with CO2 were also developed as hollow fibers.

[0222] The mobile carriers ProK and [Emim][OAc] were found to particularly improve the separation performance of the immobilized site SHPAA / PVA membrane due to their high mobility and reversible interaction with CO2 to form bicarbonate / carbonate species and carbene-CO2 adducts, respectively. As a new concept, composite membranes were combined with mobile carriers, which resulted in membranes with a CO2 permeability of 825 GPU. These membranes were evaluated for both CO2 / N2 gas pairs and CO2 / CH4 gas pairs, and separation factors of 31 for CO2 / N2 and 20 for CO2 / CH4 were obtained. Due to the relative increase in the content of CO2-philic species and the enhancement by the addition of pGO, the composite membranes with mobile carriers were stable against supply pressures up to 20 bar and showed increased resistance to carrier saturation phenomena. This high stability and gas separation performance establish the commercial viability of the fabricated membranes for CO2 separation applications when combined with an easily scalable hollow fiber configuration.

[0223] Specifically, a surprising aspect of the present invention is that membranes containing mobile carriers have an increased CO2 permeability and an increased CO2 flux under industrially relevant conditions compared to the same module without mobile carriers.

Claims

1. A composite membrane suitable for separating a gas from a gas mixture, comprising a selective layer coated on a support, wherein the selective layer comprises a) a polymer matrix containing an amine polymer, b) a graphene oxide nanofiller having an average lateral dimension of 1000 nm or less, c) a mobile carrier selected from an ionic liquid or an amino acid salt, and the amount of the graphene oxide nanofiller in the selective layer is 1 wt% or less.

2. The polymer matrix comprises a polymer containing a repeating unit of formula (I), 【Chemical 1】 wherein R 1 and R 2 are each independently selected from hydrogen or a C 1 -C 10 hydrocarbyl group, and the integer m is from 0 to 6. The composite membrane according to claim 1.

3. The polymer matrix comprises polyallylamine having a repeating unit of formula (II), [Chemical 2] In the formula, R is C 1 ~C 10 The composite membrane according to any one of claims 1 to 2, which is a hydrocarbyl group.

4. The composite membrane according to any one of claims 1 to 3, wherein the graphene oxide is physically or chemically modified.

5. The composite membrane according to any one of claims 1 to 4, wherein the graphene oxide is porous and / or contains a polymer grafted thereto.

6. The composite membrane according to any one of claims 1 to 5, wherein the graphene oxide nanofiller has an average lateral dimension in the range of 10 to 1000 nm.

7. The composite membrane according to any one of claims 1 to 6, wherein the support is a flat sheet or in the form of one or more hollow fibers.

8. The composite membrane according to any one of claims 1 to 7, wherein the nanofiller is present in the selective layer in an amount of less than 1 wt%.

9. The composite membrane according to any one of claims 1 to 8, wherein the amount of the mobile carrier in the membrane is in the range of 1.0 to 40 wt%.

10. The composite membrane according to any one of claims 1 to 9, wherein the mobile carrier comprises an ionic liquid in which the cation is selected from 1-alkyl-3-methylimidazolium, 1-alkylpyridinium, fluorosulfonyl-trifluoromethanesulfonylimide (FTFSI) N-methyl-N-alkylpyrrolidinium, or a salt of a naturally occurring amino acid.

11. The composite membrane according to any one of claims 1 to 10, wherein the selective layer has a thickness of 20 nm to 100 μm.

12. The composite membrane according to any one of claims 1 to 11, wherein the support is made of polyethersulfone (PES), polytetrafluoroethylene (PTFE), polypropylene, sulfonated polysulfone, polyvinylidene fluoride (PVDF), polyacrylonitrile (PAN) and related block copolymers, cellulose-based such as cellulose acetate (CA), polyimide, polyetherimide (PEI), aliphatic polyamide, polyetheretherketone (PEEK), polyphenylene oxide (PPO), and polysulfone (PSf).

13. The composite membrane according to any one of claims 1 to 12, wherein the support is porous.

14. A process for forming the composite membrane according to any one of claims 1 to 13, comprising: (I) a) forming an aqueous solution comprising a polymer matrix containing an amine polymer, b) a graphene oxide nanofiller, and c) a mobile carrier selected from an ionic liquid or an amino acid salt; and (II)casting the aqueous solution onto a support such as a hollow fiber support.

15. The support is a flat sheet support, and the casting process uses a bar roller to apply the selective layer, or the support is a hollow fiber, and the casting process involves dip coating, according to the process of claim 14.

16. The process according to claim 14 or 15, wherein the support is treated with a pore filler before casting the aqueous solution in step (II).

17. A process for separating a gas from a gas mixture, comprising contacting the gas mixture with the membrane according to any one of claims 1 to 13.

18. Use of the membrane according to any one of claims 1 to 13 in the separation of a gas from a gas mixture.

19. Use of the membrane according to claim 18 in the separation of carbon dioxide from a mixture containing carbon dioxide.

Images