Article pubs.acs.org/JACS
Opening of an Accessible Microporosity in an Otherwise Nonporous Metal−Organic Framework by Polymeric Guests Benjamin Le Ouay,†,‡ Susumu Kitagawa,†,§ and Takashi Uemura*,†,‡ †
Department of Synthetic Chemistry and Biological Chemistry, Graduate School of Engineering, Kyoto University, Katsura, Nishikyo-ku, Kyoto 615-8510, Japan ‡ CREST, Japan Science and Technology Agency (JST), 4-1-8 Honcho, Kawaguchi, Saitama 332-0012, Japan § Institute for Integrated Cell-Material Sciences (iCeMS), Kyoto University, Yoshida, Sakyo-ku, Kyoto 606-8501, Japan S Supporting Information *
ABSTRACT: The development of highly porous metal−organic frameworks (MOFs) is greatly sought after, due to their wide range of applications. As an alternative to the development of new structures, we propose to obtain new stable configurations for flexible MOFs by insertion of polymeric guests. The guests prevent the otherwise spontaneous closing of the host frameworks and result in stable opened forms. Introduced at a fraction of the maximal capacity, polymer chains cause an opening of the occupied nanochannels, and because of the MOF reticular stiffness, this opening is propagated to the neighboring nanochannels that become accessible for adsorption. Composites were obtained by in situ polymerization of vinyl monomers in the nanochannels of an otherwise nonporous MOF, resulting in homogeneously loaded materials with a significant increase of porosity (SBET = 920 m2/ g). In addition, by limiting the accessible configurations for the framework and forbidding the formation of a reactive intermediate, the polymeric guest prevented the thermal degradation of the host MOF even at very low loading (as low as 3 wt %) and increased its stability domain by more than 200 °C.
■
INTRODUCTION Metal−organic frameworks (MOFs), also known as porous coordination polymers (PCPs), constitute a very promising category of materials.1−3 Their success is mainly due to the association of a very high porosity with a unique coordination environment, finding applications in numerous fields.4−6 The main aim of this research field is the design of new task-specific structures, driven by the high versatility in terms of linkers, metal nodes, and topology.7−9 However, among this vast diversity, obtainable structures often present a low porosity and are deemed of low interest. As such, without a rational design, structures are likely to present very small pores or catenated subnetworks that block the porosity.10,11 The porosity can also be blocked by guest solvent molecules that contribute to the stabilization of the structure, causing a collapse upon removal.12 As an alternative to the design of intrinsically porous MOFs, we propose in this article the insertion of a guest in a flexible MOF as a way to induce a porosity in the material, which becomes accessible for other species. In order to obtain materials that can still adsorb secondary species, preincluded primary guests were introduced at a fraction of the maximal host’s capacity. The amount of guest introduced could be controlled continuously and allowed to adjust the porosity and flexible characteristics. Whereas the dynamic opening of a structure upon introduction of a guest is well-known as the “pore-breathing effect”,13−15 the structural changes are usually associated with an increased stability for the adsorbates.6,16 As a result, the new pore volume is automatically filled during the © 2017 American Chemical Society
opening and cannot be accessed by other species. So far, only a few studies focused on the impact of one molecule on the opening and the coadsorption of another molecule.17−24 Here, we propose the realization of guest@MOF composites as a way to control the adsorption properties of the flexible host frameworks. Preparation of composites by inclusion of molecular species, polymers, or nanoparticles is a common approach to bring new properties to MOFs.25−27 In most cases, the introduced guest has an intrinsic functionality (e.g., catalytic activity,28 luminescence29,30) and is used to bring directly this functionality to the composite. Here, we followed a different approach and introduced relatively inert guests. Thus, the new properties of the composites did not originate directly from the guest reactivity. Instead, they emerged from the structural modifications caused by the presence of the guests and from the cooperative effects that propagated the deformation through the ensemble of the framework. To increase the materials’ stability over time, primary guests were selected with a very low vapor pressure, so that an activation treatment is not susceptible to displace them. We especially focused on the introduction of polymeric guests inside the MOF.31−33 These composites presented a very high stability, as the vapor pressure of polymers is practically nil below their decomposition temperature, and threaded chains are very difficult to extract from the host. In situ polymerization Received: March 10, 2017 Published: May 16, 2017 7886
DOI: 10.1021/jacs.7b02402 J. Am. Chem. Soc. 2017, 139, 7886−7892
Article
Journal of the American Chemical Society
porous structure, as was observed in powder X-ray diffraction (PXRD) measurements (Figure 3A). This open form is manifested by the apparition of diffraction peaks below 7° as the unit cell expands with the accommodation of guest molecules.46 The pore opening is also manifested by a dramatic color change from purple to dark green, likely due to a change in the Co coordination environment during the phase transformation (Figure 3B,C).46 The excess amount of St outside host crystals was selectively removed by gentle evacuation under controlled pressure (0.3 kPa), then the polymerization was performed at 80 °C. By changing the reaction time, it was possible to control the monomer conversion and hence the polymer content (Figure S1 and Table S1). Longer polymerization time resulted in higher polymer content. After reaction, the unreacted monomers were removed by washing and evacuation under vacuum, yielding composites of 1 with polystyrene (PSt). Scanning electron microscopy (SEM) indicated that composites were constituted of microcrystals (ca. 5−10 μm), with no observable free PSt (Figure S2). Absence of PSt chains outside 1 was further confirmed by differential scanning calorimetry (DSC), where no glass transition for free PSt could be detected (Figure S3). We denote such composite materials as 1-PSt(X%), with X% being the loading of PSt in 1, as a mass fraction of the composite. Note that the initial amount of St in 1-St was 27 wt %. After polymerization, PSt was immobilized in the nanochannels of 1 and could not be extracted by washing the composites with solvents. Gel permeation chromatography of PSt collected after decomposition of the host MOF reveals a molecular weight (Mn) around 8000 Da, with a polydispersity index of 1.8 (Figure S4), as is usual for vinyl polymerization inside MOFs.47−49 PSt chains of this molecular weight contain ca. 75 monomeric units, corresponding to a length of ca. 20 nm when fully extended.50 One polymeric chain is thus susceptible to expand through 15 adjacent unit cells inside one nanochannel. Figure 3A shows the PXRD patterns of activated 1-PSt composites with different loading ratio. For a very low loading amount of PSt (1-PSt(1%)), the quantity introduced is not sufficient to lead to the structural transformation of 1. In contrast, the characteristic peaks of 1NP could not be detected for loading rates of 3 wt % and above. These activated 1-PSt composites presented a diffraction peak around 7°, as was also observed in 1 containing St or other solvating molecules.46 Composites also presented the dark green color, characteristic of an open form (Figure 3D). As the ideal maximum loading of PSt in 1 could be 27 wt %, these results manifest the complete conversion into an open structure of 1 upon partial loading with PSt. The opening of the framework 1 was attributed to the homogeneous distribution of the polymer chains inside pores that effectively prevented the contraction of the network into its closed form 1NP. We also demonstrated the essential role of in situ polymerization to prepare composites, by attempting to insert preformed PSt from a solution (Figures S3 and S5). In this case, nonporous 1NP was observed exclusively with the existence of PSt outside the pores, indicating the impossibility of direct introduction for this polymer. Depending on the quantity of polystyrene in the pores, 1-PSt composites presented different PXRD patterns, revealing different conformation at equilibrium for the framework 1 (Figure 3A). The Pawley refinement of the diffraction pattern for 1-PSt(16%) revealed a triclinic unit cell with a volume of 2333 Å3 (Figure S7 and Table S2). This cell volume is higher
proved to be an efficient strategy to obtain composites with a homogeneous partial loading among all the MOF crystals. Because of the relative stiffness of the framework, the pore opening caused by a guest is propagated to the adjacent channels, even if they are themselves empty (Figure 1). The
Figure 1. Realization of porous MOF−polymer composites by in situ polymerization.
resulting composite adopts thus a conformation at equilibrium that presents open pores. Since nanochannels are only partially occupied, the porosity could be accessed for adsorption of gas molecules. The phase at equilibrium, as well as its adsorption capacity, could be adjusted by controlling the nature and the concentration of guest in the composite. In addition, by modifying the accessible conformations and the mobility of the framework, the introduction of a polymeric guest increased significantly the thermal stability of the host. While polymers inside MOFs have already been used to add new functional properties to pre-existing frameworks (e.g., conductivity,34−40 selective ion transport,41,42 gas separation39,43,44), it is to our knowledge the first time they are used to modify the stable crystalline states of the host and to generate porosity by this mean.
■
RESULTS AND DISCUSSION In this study, we used [Co2(NDC)2(Bipy)]n (1; NDC = 2,6naphthalenedicarboxylate; Bipy = 4,4′-bipyridine).45,46 This compound and its dynamic behavior has been extensively described by Aggarwal et al.,46 and we refer to this previous study for further information. The synthetic procedure yields a two-fold interpenetrated structure, with each subnetwork being constituted of stacked Co(NDC) square grids with Bipy acting as pillars. This arrangement defines 1D nanochannels that contain residual solvent molecules. As indicated in Figure 2, this
Figure 2. Structural changes in configuration and interpenetration of 1, as described by Aggarwal et al.46
two-fold phase is flexible and is converted into a “narrow pore” form (1NP) upon removal of the solvent with a gentle vacuum activation. Because of the small pore diameter, 1NP presents practically no adsorption for N2 at 77 K. However, solvent molecules cause a structural change into an “open” two-fold form (1-Solvent) and can readily be adsorbed. We used this property to introduce high quantities of neat monomer in the pores, before initiating an in situ radical polymerization, following a “ship-in-a-bottle” approach.32 To realize the composites, crystals of activated 1NP were immersed in neat styrene (St), yielding 1-St with a typical open 7887
DOI: 10.1021/jacs.7b02402 J. Am. Chem. Soc. 2017, 139, 7886−7892
Article
Journal of the American Chemical Society
Figure 3. (A) PXRD diagrams of the 1-PSt composite series. The peaks below 7° characteristic of open phases are highlighted in green. The quadruplet between 16 and 19° indicative of 1NP is highlighted in red. (B−D) Photographs of 1NP (B), 1-St (C), and 1-PSt(16%) (D).
than that for 1 opened by inclusion of DMF (2144 Å3).46 Composites with an intermediate loading of PSt (3−10 wt %) presented a different crystalline phase with a smaller cell volume (2217 Å3, Figure S8 and Table S2) than that of the highly loaded 1-PSt(16%). This value is, however, significantly higher than the one reported for 1NP (V = 1640 Å3 per two-fold cage).46 This reveals the existence of a stable state for 1 with an intermediate open pore structure during the loading of PSt. In addition to the inclusion of PSt into 1, synthesis of other vinyl polymers, poly(methyl methacrylate) (PMMA) and polyacrylonitrile (PAN), was performed in 1 to yield 1-PMMA(18%) and 1-PAN(13%) (Figure S1), also showing the conversion into open phases upon inclusion of these polymers (Figures S9 and S10). Whereas the opening of 1 by polymers prepared in situ was a general trend, the exact structures of the opened states (and the resulting PXRD diagrams, Figure S6) were dependent on the nature of the guest polymers and on their concentrations (Table S2). Unlike polymers prepared in situ, introduction of solvent molecules in quantities below the maximal capacity resulted in PXRD diagrams where characteristic peaks of both 1-Solvent (open phase) and 1NP (closed phase) could be detected (Figure S11). This was indicative of the partition of 1 crystals between two distinct populations, implying the heterogeneous loading of the small molecules among crystals. The coexistence of an open and a closed phase upon insertion of small molecules is a common phenomenon observed during the adsorption of gases,51,52 with only a few counter-examples.24,53 This manifests the very high potential of polymeric guests to control the structure and the properties of MOFs. First, polymers allow homogeneous composites to be obtained with well-dispersed guests, instead of heterogeneous crystal population. In addition, polymer chains are stable inside MOFs even under high
vacuum. Such composites can thus be activated and considered as new materials for gas adsorption. The porosity of the composites was establish by measuring their gas adsorption. N2 adsorption isotherms (77 K) for the 1PSt series are shown in Figure 4A. For this gas and at this
Figure 4. N2 adsorption isotherms (77 K) of the 1-PSt composites. Closed symbols denote the adsorption, and empty symbols denote the desorption.
temperature, 1NP presented almost no adsorption capacity (
