Mixed Matrix Membranes Based on Uio-66 Mofs in the Polymer of Intrinsic Microporosity PIM-1

Mixed Matrix Membranes based on UiO-66 MOFs in the Polymer of Intrinsic Microporosity PIM-1

Muhanned R. Khdhayyer1, Elisa Esposito2, Alessio Fuoco2, Marcello Monteleone2, Lidietta Giorno2, Johannes C. Jansen2*, Martin P. Attfield1, Peter M. Budd1*

1 School of Chemistry, University of Manchester, Manchester M13 9PL, UK. E-mail:

2 Institute on Membrane Technology, ITM-CNR, Via P. Bucci 17/C, 87036 Rende (CS), Italy. E-mail:

Abstract

This work presents a study of the gas transport properties for a novel class of mixed matrix membranes (MMMs) based on the polymer of intrinsic microporosity PIM-1 loaded with UiO-66 [Zr6O4(OH)4(O2CC6H4CO2)6] based metal-organic frameworks (MOFs). Three isoreticular MOFs were dispersed in the polymer matrix, standard UiO-66, UiO-66-NH2 (functionalized with an amino group) and UiO-66-(COOH)2 (functionalized with two carboxylic groups), in order to investigate the effect of the functionalization of the linker on the gas transport. The pure gas permeabilities of He, H2, O2, N2, CH4, CO2 were studied, for the as prepared membranes and after methanol treatment, focusing attention on the potential use of these membranes for CO2/CH4 separation. The pure gas transport of the MMMs was described on the basis of the Maxwell model. The predictions of the model are discussed and compared with the experimental permeability and selectivity of the MMMs and neat PIM-1. Mixed gas permeation tests were performed on a representative sample to investigate the actual separation performance with industrially relevant gas mixtures. These confirmed the good perspectives of these MMMs in applications like CO2 removal from biogas or from flue gas.

Keywords: Mixed Matrix Membranes, PIM-1, MOFs, UiO-66, gas transport, Maxwell Model.

1  Introduction

Mixed Matrix Membranes (MMMs) offer the opportunity to combine the benefits of easily processable polymeric materials with the excellent transport performance of fillers [1]. The design of these new materials for gas separation has the objective of producing innovative membranes with enhanced permeability and selectivity that exceed the Robeson upper bound limit. In the case of PIM-based MMMs, the aim is to combine the high gas permeability of the polymer and the good selectivity of the filler materials. The choice of the polymer and of the fillers is the most important parameter affecting the morphology and resulting performance of MMMs in gas separation [2]. In the present work, the polymer used is the polymer of intrinsic microporosity arising from the step-growth polymerization of 5,5,6,6-tetrahydroxy-3,3,3,3-tetramethyl-1,1-spirobisindane and tetrafluoroterephtalonitrile (PIM-1). This novel polymer is characterized by a high intrinsic microporosity. The polymers of intrinsic microporosity are recognized for their high intra-chain rigidity and high free volume with interconnected voids [3]. The free volume elements are formed as a direct consequence of the highly contorted shape and extreme rigidity of the polymer backbone structure [4]. The presence of high free volume makes this polymer a successful candidate for the development of highly permeable membranes [5]. PIM-1 has a gas separation performance that helped to define the 2008 upper bound [6]. It is expected that the addition of fillers like Metal-Organic Frameworks (MOFs) can improve further the performance of this polymer. For this reason, there is an active research field that is developing MMMs using PIM-1 as the polymer matrix [7–10]. MOFs are a new group of fillers and nanoporous materials made by linking organic units with metal inorganic complexes [11], in which the inorganic units form the vertices of a framework and the organic linkers form the porous structure with a definite size and shape [12]. MOFs present advantages compared to inorganic fillers, such as a better compatibility with the polymer due to their organic nature and a higher pore volume with a lower density [11,13]. The pre-defined dimension of the cages and eventually the functionalization of the ligands can increase the affinity of the MOFs for the organic polymer phase and for specific gases. This makes MOFs interesting for the gas separation application [14].

Bushell et al. [15] prepared MMMs based on the polymer matrix of PIM-1 and a zeolitic imidazolate framework, ZIF-8. They reported that an increase in MOF loading results in an increased permeability coefficient and selectivity, but there was an exception for the CO2/CH4 pair, where the selectivity is more or less independent of the ZIF-8 loading. Alentiev et al. [16] studied the gas transport of composite membrane material based on PIM-1 and MIL-101 (chromium terephthalate) MOFs. They demonstrated that the addition of MIL-101 leads to an increase in the permeability and diffusion coefficient for He, O2, N2 and CO2 gases in accordance with the behaviour observed by Bushell et al. for the PIM/ZIF-8 MMMs. Ma et al. [17] designed nanoporous membranes combining a task-specific ionic liquid and MIL-101(Cr) loaded in PIM-1. They found an improved selectivity, although their permeabilities are lower than those generally reported in the literature for PIM-1.

Other MOFs of particular interest are the UiO-66-isoreticulars, Zr-based MOFs with different functional linkers. One of the advantages of UiO-66 is that different linkers can be used to tailor the affinity with the polymer matrix, which is one of the factors that affect the gas transport properties of MMMs. Transport properties of various UiO-based MMMs have been reported. Nik et al. [18] prepared MMMs with different types of isoreticular-UiO-66 in 6FDA–ODA polyimide. The presence of –NH2 groups decreased the CO2 permeability and slightly increased the CO2/CH4 selectivity. Yang et al. [19] described the effect of ligand functionalization on the CO2/CH4 separation for the UiO-66 via a computational approach. They found CH4/CO2 adsorption selectivity in the following order of linker groups –(COOH)2 > –NH2 > UiO-66, in accordance with the results reported by Wu et al. [20]. Hu et al. [21] showed that the selectivity of the UiO-66-(COOH)2 can be tailored by exchanging the hydrogen of the carboxylic group with a monovalent cation. Recently, Smith et al. [22] described the use of Ti-exchanged UiO-66 to enhance the gas permeability of PIM-1.

This work presents the ideal gas transport properties of novel MMMs based on isoreticular UiO-66 in PIM-1 for the six gases He, H2, O2, N2, CO2, CH4 determined in a fixed-volume pressure increase instrument. The actual separation of the CO2/CH4 gas pair and the N2/CO2/O2 ternary mixture was also tested under mixed gas conditions to investigate the membrane performance in simulated industrial applications. Three different types of Zr-based isoreticular UiO-66 are used, with loadings up to 27 wt-%: UiO-66, UiO-NH2 and UiO-66-(COOH)2. The effect of the functional groups on the membrane performance is studied and the gas transport properties are described in terms of the Maxwell equation. This model was already successfully used to describe the transport properties of MMMs based on PIM-1 [23]. The membrane performance is compared with the state of art data and is found to exceed the Robeson upper bound in many cases.

2  Theoretical description of gas transport in mixed-matrix membranes

Several models can be applied to describe the effect of fillers in polymer matrix [24]. Among them, one of the simplest models that is usually used to describe the gas transport in mixed matrix membranes is the Maxwell model [2,25–27]:

1)

where the PMMM is the effective permeability of the mixed matrix membrane, Pc and Pd represent the gas permeabilities in the continuous and dispersed phase, respectively and Fd is the volume fraction of dispersed phase. Since it is remarkably accurate, this is the model that will be used in the present work. The Maxwell model is valid for systems containing spherical fillers with a relatively low loading [28]. The transport can occur in different ways depending on the type of nanoparticle dispersion. The incorporation of fillers into the polymer phase can significantly alter the transport properties of gases through the MMM. Several fundamentally different cases may occur.

First case

If the interaction between the polymer and filler particles is poor, the polymer chains do not completely adhere to the surface of the fillers, giving rise to the formation of undesirable channels between both phases. In this case, the permeability of the gas in the dispersed phase and around the particles is much higher than that of the continuous phase: Pd » Pc. Thus, becomes:

2)

The formation of these non-selective voids at the interface allows bypassing of the gas around the filler particles. At low filler loadings this results in a higher permeability but constant selectivity. At relatively high filler loadings, above the percolation threshold, continuous channels are formed across the membrane, which deteriorates the selectivity of the MMM, and thus the gas separation performance of the membrane [29].

Second case

Another phenomenon that can occur is the partial or complete blockage of the MOF cage by the polymer phase. In this situation, the permeability of the gas in the dispersed phase is much lower than that of the continuous phase: Pd « Pc. reduces to:

3)

Thus, the fillers behave as impermeable obstacles to the transport and in this case no positive effects of the addition of fillers on the transport properties of the membrane will be observed [7], although even impermeable fillers could affect the packing of the polymer matrix, and thus influence the permeability indirectly.

Third case

Generally, the permeability of the gases through the dispersed phase lies between the two extreme cases discussed above, and 0 « Pd « ∞. The fillers exhibit an intrinsic finite permeability. This permeability will be different from that of the polymer matrix and its behaviour must be described by the complete Maxwell equation . This requires that the MOFs are well distributed within the polymer matrix and adhere to the polymer without pore blockage. The three cases are schematically indicated in Figure 1.

Figure 1: Graphic representation of gas transport properties of the previous cases as a function of the filler concentration. The dashed lines delimit the range of permeabilities predicted by the Maxwell model. The points indicate an example where the filler in the MMM slightly increases the permeability of the polymer matrix.

Fourth case

This is a special example of the previous situation, in which the permeability of the continuous phase is exactly equal to that of the dispersed phase, Pc=Pd. In this case, simplifies to:

4)

This means that the presence of the nanoparticles has no influence on the transport properties of the membrane for a specific gas. In this case, the mixed matrix membrane will show an identical permeability to the neat polymer membrane for this gas, but since different gases are probably affected differently by the filler particles, the selectivity might change with respect to the neat polymer.

3  Experimental

3.1  Materials

All starting materials and solvents were purchased from Sigma-Aldrich, except for chromium nitrate nonahydrate (Cr (NO3)3.9H2O, Alfa, 98%) and sodium hydroxide (NaOH, Fisher Chemical). All chemicals were used as received. Single gases were supplied by Pirossigeno at a minimum purity of 99.9995%. Certified gas mixtures were supplied by Sapio at a purity of ±0.01% from the certified concentration (CO2/CH4 mixture with 47.89 mol.% CH4 and N2/CO2/O2 mixture with 10.10 mol.% CO2 and 10.02 mol.% O2).

3.2  PIM-1 synthesis

PIM-1 (Figure 2a) was synthesized by the method of Du et al. [4,30]. To a dry 500 mL three-necked round bottom flask equipped with a Dean-Stark trap, 5,5',6,6'-tetrahydroxy-3,3,3',3'-tetramethyl-1,1'-spirobisindane (TTSBI) (17 g, 50 mmol), tetrafluoroterephthalonitrile (TFTPN) (10 g, 50 mmol), anhydrous potassium carbonate (20.7 g, 150 mmol), dimethylacetamide DMAc (100 mL) and toluene (50 mL) were added under dry nitrogen gas. The reaction mixture was refluxed at 160 oC for 40 min. The product was then poured into methanol. The crude polymer was dissolved in chloroform and reprecipitated from methanol. The product was refluxed for six hours in deionized water and then dried at 100 oC for two days. Two batches were prepared following this procedure, ground into powder and blended together. The blend was dissolved in chloroform with stirring for 2 h at room temperature, then reprecipitated from methanol, collected by filtration and dried under vacuum at 100 °C for two days. For the PIM-1 blend, average molar masses from multi-detector Gel Permeation Chromatography (Viscotek 2001 instrument with two Polymer Laboratories mixed-B columns, chloroform as eluent) were Mw = 112,000 and Mn = 48,600 g mol-1, polydispersity Mw/Mn = 2.3.

Figure 2: a) Chemical structure of PIM-1; b) Porous structure of UiO-66 MOFs, tetrahedral cage (right) and octahedral cage (left); c) non-functionalized benzene-1,4-dicarboxylates and functionalized with (-NH2) and (-COOH)2 ligands

3.3  MOF Synthesis

In this work, the synthesis of UiO-66(Zr) and UiO-66(Zr)-NH2 were adapted from previous work by Katz et al. [33]. Carboxylic acid-functionalized UiO-66(Zr) was synthesized as described by Yang et al. [34]. These MOFs are a family of Zr-terephthalate based metal–organic frameworks, namely UiO-66. In UiO-66, each Zr6O4(OH)4 cluster is surrounded by maximally 12 terephthalate linkers resulting in large octahedral and smaller tetrahedral cages with diameters of 11 Å and 8 Å, respectively (Figure 2b) [31]. Triangular windows with a diameter of 6 Å guarantee the selective gas passage [32]. Each of these UiO-66 MOFs differs from the others in terms of the functionalization of the benzene-1,4-dicarboxylate linker, that allows a family to be obtained of isoreticular UiO-66. The UiO-66-isoreticulars used in the present work are UiO-66-NH2 and UiO-66-(COOH)2, functionalized with an amino group (-NH2) and with two carboxylic groups (-COOH)2, respectively.

3.3.1  UiO-66(Zr) Synthesis

Solutions of ZrCl4 (0.125 g, 0.5 mmol) dissolved in 5 mL DMF and 1 mL conc. HCl and terephthalic acid (0.123 g, 0.75 mmol) dissolved in 10 mL DMF were mixed and sonicated for 20 min at room temperature. The whole mixture was heated in a glass jar (30 mL) at 80 oC overnight. Upon cooling to room temperature, the solid product was filtered under vacuum, washed with DMF to remove unreacted terephthalic acid, washed with methanol and dried at 100 oC.

3.3.2  UiO-66(Zr)-NH2 Synthesis

Solutions of ZrCl4 (0.125 g, 0.5 mmol) dissolved in 5 mL DMF and 1 mL conc. HCl and 2-aminoterephthalic acid (0.134 g, 0.75 mmol) dissolved in 10 mL DMF were mixed and sonicated for 20 min. The whole mixture was heated in a glass jar (30 mL) at 80 oC overnight. Once cooled to room temperature, the solid product was separated under vacuum, washed with DMF, washed with methanol and dried at 100 oC.