Continuous assembly of a polymer on a metal–organic framework (CAP on MOF): a 30 nm thick polymeric gas separation membrane

Ke Xie a, Qiang Fu a, Chenglong Xu a, Hiep Lu a, Qinghu Zhao a, Roger Curtain b, Dunyin Gu a, Paul A. Webley *a and Greg G. Qiao *a
aDepartment of Chemical and Biomolecular Engineering, The University of Melbourne, Parkville, VIC 3010, Australia. E-mail:;
bAdvanced Microscopy Facility, Bio21 Institute, University of Melbourne, Australia

Received 29th June 2017 , Accepted 17th October 2017

First published on 17th October 2017

We have developed a bottom-up approach to fabricate an ultra-thin (∼30 nm), continuous and defect-free polymeric membrane on a rough micro-scale MOF layer. This polymer-on-MOF architecture exhibits a promising CO2/N2 separation performance with a CO2 permeance of >3000 GPU and a CO2/N2 selectivity of 34. To the best of our knowledge, this membrane has the best CO2/N2 separation performance compared to any other membrane reported in the open literature.

Broader context

Increasing CO2 emissions are believed to be responsible for the extreme climate events currently experienced, and the efficient removal of CO2 is therefore critical. Membrane separation of CO2 from various gas mixtures is potentially an energy-saving solution for these issues. For practical purposes, thin film composite membranes (TFCMs) are preferred to ordinary dense membranes due to their higher gas flux as well as lower capital cost. The polymeric gutter layer, an essential component in TFCMs, is known to be the major resistance to high gas flux. Herein, we reported our discovery that the use of a continuous MOF layer instead of a polymeric gutter layer can overcome this technical barrier. Consequently, we have developed a bottom-up approach to fabricate what is now the thinnest (∼30 nm) polymeric gas separation membrane on a rough micro-scale MOF layer, pre-grown on a substrate. This membrane exhibits an excellent gas separation performance for CO2 capture applications.

The emission of CO2 to the atmosphere from the use of fossil fuels has been linked to global warming and climate change. As estimated by the International Energy Agent (IEA), fossil fuels fulfilled 81% of the world energy demand in 2013.1 In order to limit the global temperature rise to below 2 °C, 95% of coal fired and 40% of gas fired power plants need to be equipped with carbon capture and storage (CCS) facilities.2 Among various CCS approaches, post-combustion CCS is the most urgent one since it can be directly retrofit to existing fossil fuel-fired power generators.3

Membrane technology has been considered as an economic and energy-saving alternative capture technology to solvent scrubbing for the mitigation of CO2 from post-combustion exhaust gas (10–14% CO2 with mostly N2).4 Permeance and selectivity are the two major criteria to evaluate the performance of any CO2 separation membrane system. Permeance represents how fast the CO2 passes through the membrane (usually expressed as the gas permeation unit, GPU, 1 GPU = 10−6 cm3(STP) cm−2 scmHg−1), while CO2/N2 selectivity indicates how effectively the CO2 is separated from other components in the mixture.5 Therefore, in a practical sense, increased gas permeance without the loss of selectivity has been recognized as an important consideration in achieving a competitive membrane process for CO2 separation. Based on recent economic (energy consumption) evaluations, membranes with high permeance (>1000 GPU) and good selectivity (between 20 and 120) are economically desired for post-combustion CO2 capture, which thus provides a “Target Area” in the selectivity vs. permeance diagram.3 Polymer based membranes, such as polyimide (PI), polysulfone (PSU), poly(ethyl glycol) (PEG) and their derivatives have been intensively investigated.4a,5 This is ascribed to their high separation selectivity, good membrane processability and relatively low cost.4,6 However, traditional dense polymeric membranes have shown insufficient capability for CO2 removal due to their overall low permeance of CO2 regardless of their permeability.

Thin film composite membrane (TFCM) systems have been shown to be one of the best ways to achieve high permeance.7 Most of the TFCMs consist of a porous substrate, a gutter layer and an ultra-thin (<100 nm) top layer. With such a configuration, the gutter layer has always been considered essential since it can provide a smooth surface and prevent the penetration of the dilute solution substance into the porous substrate. Polydimethylsiloxane (PDMS) has been widely used as a material for the gutter layer due to its good processability and high CO2 permeance (1500–3000 GPU) with a CO2/N2 selectivity of 6–11.7b,d

The performance of a TFCM can be understood by a well-established resistance model, wherein the gas permeance is inversely proportional to the overall resistance and the total resistance is the sum of all three layers (substrate, gutter and top layers).7a,b In our recent study, the state-of-art performance of a TFCM (CAP on PDMS membrane) can achieve a permeance of >1100 GPU with a CO2/N2 selectivity of 20–40, in which the PEG-based top layer was calculated to achieve ca. 3000 GPU with a CO2/N2 selectivity of 40–50.7b Based on economic evaluations, the gutter layer in this example will increase the CO2 capture cost by >100%.3 Thus, the gutter layer has become a major obstruction that hampers the further improvement of membrane performance. Indeed, with this membrane configuration, the gas permeance of current TFCMs can never be higher than that of the gutter layer itself. Therefore, a new design is needed for a new generation of TFCMs.

Metal–organic framework (MOFs) membranes have recently been developed for gas separation applications owing to their molecular sieving and/or Knudsen diffusion effects towards gas pairs with significant size and/or molecular weight differences (H2/CO2, H2/CH4etc.).8 These MOF membranes have shown a promising gas separation performance of H2 over other heavy gases. However, they are rarely applied in post-combustion CO2 separation because the kinetic diameters of CO2 (330 pm) and N2 (364 pm) are too close to separate by diffusion coefficients, and the CO2 concentration in post-combustion flue gas is too low (10–14%) for the adsorption selectivity of MOF membranes.8b,h,9 On the other hand, MOF membranes generally possess an extremely high CO2 permeability of >285[thin space (1/6-em)]000 Barrer,9a,b,10 which in turn means an extremely low CO2 resistance. Thus, MOF layers could potentially act as an attractive gutter layer. However, the ultra-thin top selectivity layers (on the gutter layer) are typically produced by dip coating, slot die coating or using doctor blades, which normally require a smooth flat surface for the gutter layer. These methods are not suitable for film casting on a rough MOF layer surface. Recently, our group has developed a novel nano-coating technology, namely the continuous assembly of a polymer (CAP).7b,11 Besides the flat PDMS layer, this bottom-up coating method also enables growth of an ultra-thin film from a rough substrate (i.e. MOF layer) with simultaneous crosslinking. Therefore, we saw this as an opportunity to fabricate a “CAP on MOF” membrane, where a MOF layer is covered by an ultra-thin (sub 100 nm), continuous and defect-free crosslinked polymeric layer. XPS, XRD and SEM measurements on the products confirmed the unique configuration of the “CAP on MOF” membrane as we proposed. The thickness of this CAP can be optimized to 30 nm without defects. To our knowledge, this is the thinnest defect-free polymeric membrane that can be used for gas separation. Because of this design, the optimized CO2/N2 separation performance achieved a permeance of 3000 GPU and a selectivity of 34, superior to all comparable membranes in this field. The permeance of this “CAP on MOF” membrane is 2.6 times higher than that of the CAP membrane using PDMS as a gutter layer in our previous study since the MOF layer is much more (>15 times) permeable than the PDMS layer.7b

A general approach to fabricate this CAP on MOF membrane via CAP technology is illustrated in Scheme 1. An amino functionalized MOF layer was prepared via in situ crystallization of NH2-MIL-53 crystals on a porous anodisc substrate (step i). Thereafter, the amino groups are functionalized by α-bromoisobutyryl bromide (step ii) which can serve as an initiator for the following CAP process. A PEG-based ultra-thin CAP layer was then formed on top of the MOF layer by immersing the functionalized substrates into aqueous solutions containing different concentrations of polyethylene glycol dimethacrylates (PEGDMA, macrocross-linker). The raw MOF membrane is denoted as Raw MOF/anodisc (RMA), and the resulting Polymer/MOF/anodisc (PMA) composite membranes are denoted as PMA-A, PMA-B, and PMA-C depending on the amount of PEGDMA used in each CAP film (PMA-A < PMA-B < PMA-C). NH2-MIL-53 was chosen as an example metal organic framework for the gutter layer in the composite membrane system because this MOF can form a porous continuous layer with plenty of amino groups that can be easily functionalized with an atom-transfer radical-polymerization initiator. In fact, our strategy is a general one and any porous material meeting these requirements will be applicable.

image file: c7ee02820b-s1.tif
Scheme 1 The bottom-up fabrication of the polymer/MOF architecture (PMA): (i) in situ formation of the MOF layer by hydrothermal treatment of anodisc in NH2-BDC solution. (ii) Initiator grafting of the MOF membrane by functionalizing the amino groups with BiBB. (iii) Coating of an ultra-thin polymer film on the MOF membrane by ARGET-ATRP. The chemical structure of the resulting CAP on MOF membrane can be found in Scheme S1 in the ESI.

It should be noted that the configuration reported in this work is entirely distinct from the well-developed mixed-matrix membranes (MMMs) and the polymer-supported MOF membranes previously reported.12 In the MMMs, the MOF filler particles are distributed in a continuous polymeric matrix (Case A, Scheme 2). In CO2/N2 separation, the purpose of incorporation of the MOF in the polymer matrix (Case A) is to improve the gas permeability of the bulk membrane materials, at the expense of slightly reduced selectivity. For polymer-supported MOF membranes, the polymer serves as the porous mechanical support (Case D, Scheme 2) or binder (Case E, Scheme 2) of the MOF crystals, so they are essentially MOF membranes, and their separation properties come from the nature of the MOF. The present “CAP on MOF” membrane has a unique configuration which is different from our previously developed TFCM (Case B, Scheme 2) and all those membrane systems reported to date. In the “CAP on MOF” architecture, the ultra-thin polymer and micro-scale MOF are two respective layers, integrated by chemical bonding (Case C, Scheme 2). In this configuration, the MOF layer serves as the porous support while the CAP layer provides CO2/N2 separation properties. The schematics in Scheme 2 reveal the difference between these concepts.

image file: c7ee02820b-s2.tif
Scheme 2 Comparison of different types of MOF/polymer composite membranes: (A) dense (20–100 μm thick) Mixed Matrix Membranes (MMMs),12a,b,e (B) thin Film Composite Membranes (TFCMs),7b (C) CAP on MOF (this work), and (D) and (E) polymer supported MOF membranes.12c,d,f–h

Typical SEM images of RMA, PMA-A, PMA-B and PMA-C are shown in Fig. 1. We found that the RMA membrane has a rough but complete surface reminiscent of a layer of “rubble” consisting of micro-size MOF crystals with a size distribution of 1.5–2.5 μm. After the subsequent CAP process, the edges of the MOF crystals became smooth in all PMA cases, indicating the successful formation of a polymer coating. Moreover, the morphologies of the membranes strongly depended on the concentration of PEGDMA (macro-crosslinker) used in the CAP process. In the case of the PMA-A membrane, a thin polymer film was formed on the surface of the MOF crystal but still left gaps between the MOF crystals (Fig. 1b). This is indicative of insufficient PEGDMA to cover the whole MOF layer. By increasing the concentration of PEGDMA to 120 mM, the resulting polymer layer showed a continuous polymer phase covering the whole MOF layer, which is shown in the SEM image of PMA-B (Fig. 1c). Further increase of the concentration to 180 mM led to denser polymer layer formation on the MOF layer (PMA-C, Fig. 1d). The MOF crystals lost their intrinsic cubic shape and appear to be spherical. A cross-sectional view of RMA (Fig. 1e) shows that the MOF crystals are firmly packed with each other, forming a continuous rough MOF layer with a thickness of 1–2 μm. From the cross-section SEM images of the focused ion beam (FIB) fabricated PMA-B (Fig. 1f), we can find that the continuous MOF crystal is covered by another continuous polymeric layer. The high-resolution SEM and TEM measurements on the cross-section of PMA-B were also conducted (Fig. 1g and h). In these images, a continuous polymer layer (middle) with a thickness of 30 ± 2 nm shows a different contrast between the MOF layer (bottom) and the platinum deposition (top) that was pre-deposited on the surface of the CAP layer before imaging. This result provides direct evidence of the successful formation of an ultra-thin CAP coating.

image file: c7ee02820b-f1.tif
Fig. 1 Typical SEM and TEM images of RMA and PMAs. (a–d) The top view of RMA, PMA-A, PMA-B and PMA-C respectively. (e) The cross-section view of RMA. (f–h) The SEM (f and g) and TEM (g) cross-section view of PMA-B. The red dashed line indicates the boundary between Pt deposition and CAP on the MOF membrane. The sample in (f–h) is a lamellae fabricated by focused ion beam (FIB) with pre-deposition of platinum.

EDX mapping and AFM scratching analysis was conducted to confirm the thickness of the selective polymer layer, and the results are shown in Fig. 2. From the TEM image in Fig. 2a, a grey layer with a thickness of 25–35 nm was observed between the bright Pt deposition layer and the dark MOF layer. The regional EDX mapping on the boundary (highlighted in orange) is shown in Fig. 2b, in which a carbon-rich layer (red) is observed between the platinum-rich deposition layer (orange) and an aluminum-rich MOF layer (green). Moreover, the elemental intensity from EDX linear scanning (Fig. 2c) crossing this boundary also reveals a dramatic decrease of the aluminum signal and a simultaneous increase of the carbon signal. These elemental analyses, combined with SEM and TEM measurements (Fig. 1g, h and 2a) provide solid evidence for the successful formation of an ultrathin polymeric layer (ca. 30 nm) on top of the MOF layer.

image file: c7ee02820b-f2.tif
Fig. 2 Elemental analysis on the polymeric film of PMA-B using TEM/EDX (a–c) and AFM (d–f) measurements. (a) The cross-sectional view of the TEM image for FIB fabricated PMA-B lamellae; the imaging condition was tuned for EDX. The 25–35 nm thick polymeric layer is indicated by the green dashed lines; the orange square indicates the region for EDX mapping; the red line and arrow indicate the range and direction for linear scanning. (b) The EDX mapping image in the orange region in (a). Elemental color: Al-grey; Pt-yellow; C-red. (c) The dependence of element intensity on the analysis position along the red line in (a). The green dashed lines in (c) refer to the boundaries of the polymeric layer in (a) (also marked by green dashed lines). (d and e) The 3D AFM topology images of PMA-B before (d) and after (e) scratching at exactly the same position. The “character peaks” are marked (1–5) for reference. (f) The SEM image (inset) of the AFM tip used for scratching and the corresponding EDX spectra on the site squared by white.

Owing to the difference in mechanical strength between the crystalline MOF layer and the amorphous polymer layer, the AFM tip was employed to scratch off the polymeric layer from the MOF layer (experimental details are available in Scheme S2, ESI). The corresponding 3D AFM images are illustrated in Fig. 2d (before the scratch) and Fig. 2e (after the scratch). There are several “character peaks” (marked with numbers 1–5), whose profiles are identical before and after scratching (Fig. 2d), indicating that the MOF crystal was barely damaged during the scratching process. By contrast, the flat-looking area of the membrane (Fig. 2d) becomes rough after scratching (Fig. 2e). This can be attributed to the exposure of rough MOF crystals. Additionally, we find that the depth of the “flat layer” is less than 50 nm. EDX spectra were also recorded to investigate the chemical composition of this thin layer. The SEM image and the corresponding EDX spectra for the indicated sites are shown in Fig. 2f. Not surprisingly, the EDX spectrum on the non-contact area (edge) of the Si-made AFM tip indicates negligible carbon and oxygen content. In contrast, the EDX scanning on the pinpoint (tip) shows 63.3 at% of carbon and considerable oxygen content, indicating that the layer scratched off is indeed a polymeric layer. Moreover, the Al is almost undetectable on the pinpoint, implying that the MOF crystals are barely damaged under the scratching force. This result is in good agreement with the observation in Fig. 2d and e. Therefore, the AFM scratching analysis provides further evidence of a thin polymeric layer coated on the rough MOF layer surface.

The XRD spectra of RMA and PMAs (Fig. S2a, ESI) exhibit several sharp diffraction peaks, implying the high crystallinity of their MOF layers. The XRD spectrum of RMA (black trace) is ascribed to the MOF NH2-MIL-53, in good agreement with that previously reported.8j,9b After the CAP process, a broad amorphous pattern overlaps the XRD spectra of RMA at ca. 2θ = 24°, which comes from the coated PEG layer.13 The higher PEG/NH2-MIL-53 ratio leads to intensive amorphous patterns. In addition, the high-resolution XPS spectra of C1s (Fig. S2b–d, ESI) show a significant difference between the RMA and PMAs. The C1s spectra of RMA can be fitted into two separated peaks which are assigned to C–C (285 eV) and C[double bond, length as m-dash]O (289 eV) owing to the benzene ring and carboxyl group, respectively. In the cases of PMA-B and PMA-C, a new peak ascribed to C–O (ca. 286.5 eV) is produced by the PEG coating on top of the MOF layer. Additionally, the intensity of the C–O signal increases as the concentration of PEGDMA increases (PMA-C > PMA-B), indicating the formation of a thicker PEG layer. The compositions of RMA, PMA-A, PMA-B and PMA-C were further investigated by thermogravimetric analysis (TGA) (Fig. S3 and Table S1, ESI). The TGA results indicate that RMA contained 10.8 wt% of MOF (NH2-MIL-53), and the polymer content gradually increases from 1.60 wt% for PMA-A to 2.88 wt% for PMA-B and 6.97 wt% for PMA-C. This result is consistent with the change of PEGDMA concentration used in the CAP process.

As one side of the MOF layer is sealed with a CAP coating and the other side is fixed on an anodisc, the pores of the MOF should be still accessible to gas molecules with part of the pores filled by the polymer. CO2 adsorption isotherms of the PMAs illustrated in Fig. S4 (ESI) demonstrated this expectation. Before the CAP coating, the RMA shows a higher CO2 uptake compared to the bare anodisc due to the formation of the MOF layer. The CO2 uptake decreases gradually as the thickness of the polymer coating increases. However, the values are still higher than the uptake of anodisc, implying the preservation of the porosity of the MOF in the PMAs. All of these characterizations, combined with the high-resolution SEM image confirm the successful formation of an ultra-thin polymer coating on the MOF crystal layer.

The gas separation performance of RMA, PMA-A, PMA-B and PMA-C was measured using a constant pressure variable volume (CPVV) apparatus, and the typical performance data is listed in Table 1 and plotted in Fig. 3. All data presented in this work were the average values collected from at least three membranes. As expected, the RMA presented a high permeance for both N2 (∼47[thin space (1/6-em)]000 GPU) and CO2 (∼45[thin space (1/6-em)]000 GPU). For PMA-A, it is non-selective for CO2 over N2 due to the defects of the CAP coating on the MOF layer (Fig. 1b). Remarkably, in the case of PMA-B the MOF layer is entirely sealed by the CAP layer (Fig. 1c) as indicated by a dramatic enhancement of CO2/N2 selectivity to 34 with an extremely high permeance of 3000 GPU. Noteworthily, this value is equal to the permeance of an ordinary PDMS gutter layer used in composite membranes,12a–g implying that the permeance of PMA-B is unachievable in these membrane systems. In addition, the defect-free MIL-53 or NH2-MIL-53 MOF membranes, no matter whether reported in previous literature or in this study, show a poor CO2/N2 selectivity of 0.7–1 using single gas measurements.9b,15 We thus can conclude that the overall separation performance of the PMAs mainly depends on the pristine ultra-thin CAP layer. Indeed, when subtracting the performance of RMA (45[thin space (1/6-em)]000 GPU with a selectivity of 1) we found that the net performance of the ultra-thin top layer is 3200 GPU with a selectivity of 36, which is in good agreement with previous studies on pristine PEG based membranes.7b,16 The apparent MOF's impact on overall performance is only 6.2% and 5.5% for CO2 permeance and CO2/N2 selectivity, respectively. If any defect exists in a gas separation membrane, the membrane is immediately compromised and its selectivity will be reduced significantly.4a,17 The fact that PMA-B presents a similar CO2/N2 selectivity as pristine bulk PEG membranes provides solid evidence for the formation of a defect-free, 30 nm thick top layer. Based on an economic analysis model, if the selectivity of a membrane is above 30, further increase of membrane selectivity has little benefit (see ESI Fig. S5 for details).3 As shown in Table S3 (ESI), a hypothetical membrane with the same CO2 permeance as PMA-B and a 1.6 times higher selectivity of 54 can only reduce the cost by 5%. Indeed, the PMA-B membrane has almost the lowest capture cost among all the currently documented non-facilitated transport membrane systems for post-combustion CO2 capture. The state-of-the-art, non-facilitated transport, high permeance membranes reported in the open literature are listed in Table 2 and plotted in Fig. 3.3,7b,d,e,9c,d,14 The top layer thicknesses of the reported TFCMs range from 45 to 500 nm, however their permeance is less than 1200 GPU due to the use of a polymeric gutter layer. Membrane PMA-B presents the highest CO2 permeance. For example, the top layer thickness of our previously reported “CAP on PDMS” membranes is 45 nm, which is comparable to this present “CAP on MOF” membrane.7b However, its CO2 permeance (1140 GPU) is far less than that of the PMA-B (3000 GPU) membrane despite the fact that the chemical components and the fabrication approaches for the top layers are similar. The lower CO2 permeance is due to the resistance of the PDMS gutter layer. By using the well-established resistance model,7b our early work reported that the top layer of the “CAP on PDMS” membrane has a CO2 permeance of 3170 GPU, which is close to the CO2 permeance of the top layer in PMA-B (3200 GPU). In fact the overall CO2 permeance of PMA-B is 3000 GPU, indicating that the MOF gutter layer has minimum resistance when we compare the “CAP on MOF” and “CAP on PDMS” membranes. The change from the PDMS gutter layer to the MOF gutter layer improves the CO2 permeance by 2.6 times.

Table 1 Summary of CO2/N2 separation performance of RMA and PMAs
Entry N2 permeance [GPU, STP]a CO2 permeance [GPU, STP]a CO2/N2 selectivitya
a Single gas performance was measured at 35 °C with a pressure difference of 100 kPa.
RMA ∼47[thin space (1/6-em)]000 ∼45[thin space (1/6-em)]000 ∼1
PMA-A ∼12[thin space (1/6-em)]000 ∼15[thin space (1/6-em)]000 ∼1.2
PMA-B 90 ± 8 3000 ± 320 34 ± 3
PMA-C 8.0 ± 0.4 310 ± 30 37 ± 2

image file: c7ee02820b-f3.tif
Fig. 3 CO2/N2 selectivity versus permeance diagram comparing our PMA with the other state-of-art high permeance membranes reported in the literature.3,7b,d,e,9c,d,14 The facilitated transport membranes are excluded. The details can be found in Table 2. The target area is proposed by Merkel et al.3 PDMS is normally used as the gutter layer for the ordinary high permeance membranes which provides a smooth surface and prevents the penetration of top layer materials into the substrate.12a–g
Table 2 The comparison of the state-of-art high permeance membranes for CO2/N2 separation. The facilitated transport membranes are excluded
Types Symbols in Fig. 3 Membrane configuration Top layer thickness (nm) Gutter layer thickness (nm) CO2 permeance (GPU) CO2/N2 ideal selectivity Ref.
a The separation properties of these MOF membranes significantly depend on testing conditions. Their selectivities are below 1 (non-selective) when the upstream pressure is below 350 kPa, or the CO2 concentration is below 60 v/v%. The numbers in brackets are the highest selectivity acquired at high pressure (∼400 kPa) and high CO2 concentration (>88 v/v%) of CO2/N2 mixed gas.
TFCM image file: c7ee02820b-u1.tif PEG/PDMS/PAN 45 175 1140 22 7b
PEG-TMC/PDMS/PAN 125 230 1260 43 7b
image file: c7ee02820b-u2.tif PEG/PSS-PAH/α-Al2O3 50–200 Unknown 480 26 7e
image file: c7ee02820b-u3.tif SNP-Pebax®/PDMS/PAN 520 350 1160 20 14b
PPFPA-Pebax®/PDMS/PAN 300 255 1000 20 14c
PEG-Pebax®/PDMS/PAN Unknown Unknown 820 40 14d
ZIF-7-Pebax®/PTMSP/PAN 498 250 291 68 17
image file: c7ee02820b-u4.tif Polyactive®/PDMS/PAN 85 130 1200 53 7d and f
Unknown image file: c7ee02820b-u5.tif Polaris® Unknown Unknown 1000 50 3
MOF image file: c7ee02820b-u6.tif MOF-5/α-Al2O3 14[thin space (1/6-em)]000 N/A 1200 0.4 (70)a 9c
MOF-5/α-Al2O3 14[thin space (1/6-em)]000 N/A 482 0.8 (410)a 9d
Dense image file: c7ee02820b-u7.tif Cellulose acetate Unknown N/A 110 30 3
CAP on MOF PEG/NH2-MIL-53 30 N/A 3000 34 This work

Further increase of PEGDMA concentration resulted in a thicker top layer (PMA-C in Fig. 1d), which in turn led to a lower CO2 permeance of 310 GPU with a CO2/N2 selectivity of 37 (and permeance of 220 GPU with a selectivity of 39 for an even thicker polymer layer of PMA-D as listed in Table S2, ESI). The long-term stability of PMA-B was studied by exposing the membrane to lab conditions (room temperature, in air) and tracking its gas separation performance for >45 days. As shown in Fig. S6 (ESI), the CO2 permeance and CO2/N2 selectivity dropped ca. 30% and 40% respectively after 45 days. This can be ascribed to the plasticization and aging effects of the PEG based polymer membrane.7b PMA-B also shows high chemical and mechanical stability, which are demonstrated by mimicking the conditions in post-combustion CO2 capture (Fig. S7 and Table S4, ESI).

The gas separation performance of PMA-B and PMA-C was also investigated with a feed mixture of 10/90 mol% for CO2/N2 (Table S2, ESI). For the PMA-B, the CO2 permeance reduced to 2600 GPU and the selectivity reduced to 14, which is expected and is ascribed to the combined effects of diffusion competition, aging and plasticization (also known as the thin film effect).7a,b As expected, the PMA-C with a thicker polymer layer shows excellent resistance to those effects. These results suggest that optimized performance can be achieved by carefully adjusting the PEGDMA concentration during synthesis.


In conclusion, a novel CAP on MOF composite membrane consisting of a micro scale MOF layer and a nano-scale cross-linked polymer selective layer was successfully fabricated via CAP nanotechnology through a bottom-up approach. The CAP on MOF membrane with an extremely high CO2 permeance of 3000 GPU and a CO2/N2 selectivity of 34 was fabricated by using a PEGDMA concentration of 120 mM in the CAP process. This performance is well above the boundary of the well-known target area for post-combustion CO2 capture applications and is superior to most of the current existing membranes in the field. This method demonstrated an efficient general strategy to coat an ultra-thin, continuous and defect-free polymeric film on a macro-scale rough surface. The design can be further improved by using better polymer materials for the top layer as well as the introduction of additives in the top layer in future.

Conflicts of interest

There are no conflicts to declare.


The authors appreciate the Bio21 Advanced Microscopy Facility for the assistance with material characterization.

Notes and references

  1. Carbon Capture and Storage: The solution for deep emissions reductions, International Energy Agent.
  2. .
  3. T. C. Merkel, H. Lin, X. Wei and R. Baker, J. Membr. Sci., 2010, 359, 126–139 CrossRef CAS .
  4. (a) T.-S. Chung, L. Y. Jiang, Y. Li and S. Kulprathipanja, Prog. Polym. Sci., 2007, 32, 483–507 CrossRef CAS ; (b) M. Rezakazemi, A. E. Amooghin, M. M. Montazer-Rahmati, A. F. Ismail and T. Matsuura, Prog. Polym. Sci., 2014, 39, 817–861 CrossRef CAS .
  5. S. Wang, X. Li, H. Wu, Z. Tian, Q. Xin, G. He, D. Peng, S. Chen, Y. Yin and Z. Jiang, Energy Environ. Sci., 2016, 9, 1863–1890 CAS .
  6. (a) D. Bastani, N. Esmaeili and M. Asadollahi, J. Ind. Eng. Chem., 2013, 19, 375–393 CrossRef CAS ; (b) P. Bernardo, E. Drioli and G. Golemme, Ind. Eng. Chem. Res., 2009, 48, 4638–4663 CrossRef CAS .
  7. (a) Q. Fu, A. Halim, J. Kim, J. M. Scofield, P. A. Gurr, S. E. Kentish and G. G. Qiao, J. Mater. Chem. A, 2013, 1, 13769–13778 RSC ; (b) Q. Fu, J. Kim, P. A. Gurr, J. M. Scofield, S. E. Kentish and G. G. Qiao, Energy Environ. Sci., 2016, 9, 434–440 RSC ; (c) A. Halim, Q. Fu, Q. Yong, P. A. Gurr, S. E. Kentish and G. G. Qiao, J. Mater. Chem. A, 2014, 2, 4999–5009 RSC ; (d) W. Yave, H. Huth, A. Car and C. Schick, Energy Environ. Sci., 2011, 4, 4656–4661 RSC ; (e) A. M. Balachandra, G. L. Baker and M. L. Bruening, J. Membr. Sci., 2003, 227, 1–14 CrossRef CAS ; (f) W. Yave, A. Car, J. Wind and K.-V. Peinemann, Nanotechnology, 2010, 21, 395301 CrossRef PubMed .
  8. (a) A. J. Brown, N. A. Brunelli, K. Eum, F. Rashidi, J. Johnson, W. J. Koros, C. W. Jones and S. Nair, Science, 2014, 345, 72–75 CrossRef CAS PubMed ; (b) H. Bux, F. Liang, Y. Li, J. Cravillon, M. Wiebcke and J. r. Caro, J. Am. Chem. Soc., 2009, 131, 16000–16001 CrossRef CAS PubMed ; (c) D. Nagaraju, D. G. Bhagat, R. Banerjee and U. K. Kharul, J. Mater. Chem. A, 2013, 1, 8828–8835 RSC ; (d) O. Shekhah, H. Wang, D. Zacher, R. A. Fischer and C. Wöll, Angew. Chem., Int. Ed., 2009, 48, 5038–5041 CrossRef CAS PubMed ; (e) Y. Sun, F. Yang, Q. Wei, N. Wang, X. Qin, S. Zhang, B. Wang, Z. Nie, S. Ji, H. Yan and J.-R. Li, Adv. Mater., 2016, 28, 2374–2381 CrossRef CAS PubMed ; (f) A. Huang, Q. Liu, N. Wang and J. Caro, J. Mater. Chem. A, 2014, 2, 8246–8251 RSC ; (g) Z. Y. Gu and X. P. Yan, Angew. Chem., Int. Ed., 2010, 49, 1477–1480 CrossRef CAS PubMed ; (h) C. Scherb, A. Schödel and T. Bein, Angew. Chem., 2008, 120, 5861–5863 CrossRef ; (i) J. Yao, D. Dong, D. Li, L. He, G. Xu and H. Wang, Chem. Commun., 2011, 47, 2559–2561 RSC ; (j) T. T. Tan, M. R. Reithofer, E. Y. Chen, A. G. Menon, T. A. Hor, J. Xu and J. M. Chin, J. Am. Chem. Soc., 2013, 135, 16272–16275 CrossRef CAS PubMed .
  9. (a) S. R. Venna and M. A. Carreon, J. Am. Chem. Soc., 2009, 132, 76–78 CrossRef PubMed ; (b) F. Zhang, X. Zou, X. Gao, S. Fan, F. Sun, H. Ren and G. Zhu, Adv. Funct. Mater., 2012, 22, 3583–3590 CrossRef CAS ; (c) Z. Zhao, X. Ma, A. Kasik, Z. Li and Y. Lin, Ind. Eng. Chem. Res., 2012, 52, 1102–1108 CrossRef ; (d) Z. Rui, J. B. James, A. Kasik and Y. Lin, AIChE J., 2016, 62, 3836–3841 CrossRef CAS .
  10. S. Qiu, M. Xue and G. Zhu, Chem. Soc. Rev., 2014, 43, 6116–6140 RSC .
  11. D. Mertz, C. J. Ochs, Z. Zhu, L. Lee, S. N. Guntari, G. K. Such, T. K. Goh, L. A. Connal, A. Blencowe and G. G. Qiao, Chem. Commun., 2011, 47, 12601–12603 RSC .
  12. (a) S. Kanehashi, G. Q. Chen, C. A. Scholes, B. Ozcelik, C. Hua, L. Ciddor, P. D. Southon, D. M. D’Alessandro and S. E. Kentish, J. Membr. Sci., 2015, 482, 49–55 CrossRef CAS ; (b) T. Rodenas, M. van Dalen, E. García-Pérez, P. Serra-Crespo, B. Zornoza, F. Kapteijn and J. Gascon, Adv. Funct. Mater., 2014, 24, 249–256 CrossRef CAS ; (c) J. Hou, P. D. Sutrisna, Y. Zhang and V. Chen, Angew. Chem., 2016, 128, 4015–4019 CrossRef ; (d) E. Barankova, X. Tan, L. F. Villalobos, E. Litwiller and K.-V. Peinemann, Angew. Chem., Int. Ed., 2017, 56, 2965–2968 CrossRef CAS PubMed ; (e) J. E. Bachman and J. R. Long, Energy Environ. Sci., 2016, 9, 2031–2036 RSC ; (f) M. S. Denny and S. M. Cohen, Angew. Chem., Int. Ed., 2015, 54, 9029–9032 CrossRef CAS PubMed ; (g) R. Zhang, S. Ji, N. Wang, L. Wang, G. Zhang and J. R. Li, Angew. Chem., Int. Ed., 2014, 53, 9775–9779 CrossRef CAS PubMed ; (h) Y. Zhang, X. Feng, H. Li, Y. Chen, J. Zhao, S. Wang, L. Wang and B. Wang, Angew. Chem., 2015, 127, 4333–4337 CrossRef .
  13. K. Xie, Q. Fu, Y. He, J. Kim, S. Goh, E. Nam, G. Qiao and P. Webley, Chem. Commun., 2015, 51, 15566–15569 RSC .
  14. (a) J. Kim, Q. Fu, K. Xie, J. M. Scofield, S. E. Kentish and G. G. Qiao, J. Membr. Sci., 2016, 515, 54–62 CrossRef CAS ; (b) J. M. Scofield, P. A. Gurr, J. Kim, Q. Fu, S. E. Kentish and G. G. Qiao, J. Membr. Sci., 2016, 499, 191–200 CrossRef CAS ; (c) S. Tan, Q. Fu, J. M. Scofield, J. Kim, P. A. Gurr, K. Ladewig, A. Blencowe and G. G. Qiao, J. Mater. Chem. A, 2015, 3, 14876–14886 RSC ; (d) J. Lillepärg, P. Georgopanos, T. Emmler and S. Shishatskiy, RSC Adv., 2016, 6, 11763–11772 RSC ; (e) X. Yu, Z. Wang, Z. Wei, S. Yuan, J. Zhao, J. Wang and S. Wang, J. Membr. Sci., 2010, 362, 265–278 CrossRef CAS ; (f) H. Yin, J. Wang, Z. Xie, J. Yang, J. Bai, J. Lu, Y. Zhang, D. Yin and J. Y. Lin, Chem. Commun., 2014, 50, 3699–3701 RSC .
  15. Y. Hu, X. Dong, J. Nan, W. Jin, X. Ren, N. Xu and Y. M. Lee, Chem. Commun., 2011, 47, 737–739 RSC .
  16. H. Lin and B. D. Freeman, J. Membr. Sci., 2004, 239, 105–117 CrossRef CAS .
  17. T. Li, Y. Pan, K.-V. Peinemann and Z. Lai, J. Membr. Sci., 2013, 425, 235–242 CrossRef .


Electronic supplementary information (ESI) available: Experimental, supplemental SEM images, TGA results, gas adsorption measurements, and membrane stability studies. See DOI: 10.1039/c7ee02820b
These authors contributed equally.

This journal is © The Royal Society of Chemistry 2018