Sol–Gel Based Structural Design of Macropores and Material Shapes of Metal–Organic Framework Gels

We have developed a general synthetic strategy to control macroporous structures and material shapes of metal–organic framework (MOF) gels via the sol–gel based structural control process. A series of 1,3,5-benzene tricarboxylic acid (BTC) based MOF gels, Cr-BTC and Zr-BTC, have been chosen as a proof of the concept.

Therefore, the advantage of sol-gel process, which realizes high level structural control for introducing (meso-)/macroscopic targeted structures, [4] has not yet been applied to MOF gels.
In 2019, we reported meso-/macroporous UiO-66-NH 2 , with the formula [Zr 6 O 4 (OH) 4 (BDC-NH 2 ) 6 ] (BDC-NH 2 = 2aminoterephthalic acid), monoliths via self-assembly-induced phase separation (SIPS) technique and subsequent solvothermal treatment. [5] Although commonly obtained MOF gels have low crystallinity, [3] we developed a strategy for higher degree of crystallization while maintaining macroscopic morphology of MOF gels by optimizing post-treatment conditions by taking the UiO-66-NH 2 system as a proof of the concept. [5] Therefore, establishing a synthetic route to control macroscopic morphology of MOF gels plays a pivotal role to realize multiscale structural control of the MOFs with targeted pore sizes, which facilitate the use of MOFs in applications in which efficient mass transfer of the guest molecule is important. However, since the first report on the UiO-66-NH 2 system, no reports applying the SIPS technique to other kinds of MOF gels have been reported.
Here, we report a synthetic route to control macroporous structures of 1,3,5-benzene tricarboxylic acid (BTC) based MOF Figure 1. Schematic of the synthesis of the macroporous Cr-BTC metalorganic framework gels, which consist of the amorphous metal-organic networks. (i) Gelation accompanied by the phase separation from homogeneous solution of metal salt, organic linker, solvent, and phase separation inducer. The macroporous robust Cr-BTC gel was prepared. Polytetrahydrofuran (PTHF, M n = 2 900), which is hydrophobic polymer, was used as phase separation inducer in a mixuture of ethanol and tetrahydrofuran (V ethanol , V tetrahydrofuran = 1:1). gels, Cr(III)-BTC gels (denoted as Cr-BTC gels) and Zr(IV)-BTC gels (denoted as Zr-BTC gels), by SIPS technique using a hydrophobic polymer as a phase separation inducer. In addition, we demonstrate the top-down processability of the obtained macroporous MOF gels toward macroporous microparticles by selecting Cr-BTC as representative.
Firstly, we discuss the formation and control of macroporous structures of Cr-BTC gels. The sol-gel transition behaviour of a solution of Cr(III) salt and BTC have been reported elsewhere. [3c,6] The typical chemical formula of Cr-BTC is Cr 3 XO(BTC) 2 ·nH 2 O (X = typically F, n ≈ 28), known as MIL-100(Cr), a crystalline MOF powder. [7] Figure 1 shows a schematic of synthesis of the macroporous Cr-BTC metalorganic gels. The suitable starting compositions and reaction conditions have been specified, in which transient macroscopic structure of the phase separation can be arrested by the sol-gel transition. Typically, Cr(NO 3 ) 3 ·9H 2 O (12 mmol) and BTC (8 mmol) were used as inorganic precursor and organic linker, respectively. We used a mixture of ethanol (EtOH) and tetrahydrofuran (THF) (13 mL, V EtOH :V THF = 1:1) as a solvent to ensure good solubility of BTC, and to maintain low internal pressure in a glass tube at gelation temperature, 60 °C. When we used non-alcoholic solvent such as N,N-dimethylformamide (DMF), gelation of Cr-BTC did not occur within 1 d. Polytetrahydrofuran (PTHF, M n = 2 900) was used as a phase separation inducer, with the idea that intrinsically hydrophilic Cr-BTC has low compatibility with hydrophobic polymers. This idea follows our previous report on Zr-BDC-NH 2 based sol-gel system, in which a hydrophobic polymer, poly(propylene glycol) (PPG), acted as a phase separation inducer in the course of polymerization. [5] The mixture of Cr(NO 3 )·9H 2 O, BTC, solvent, and the phase separation inducer was sealed in a glass tube, and transferred into the water bath at 60 °C. The macroporous opaque green gel was obtained (Figure 1) possibly as a result of the mismatched growth [3a] of Cr-BTC clusters. After washing with ethanol and n-hexane, the wet gels were dried at 40 °C, and then 150 °C in vacuo.
The formation and coarsening of the interconnected macroporous structures were controlled by adopting PTHF amount as a variable. To apply SIPS technique to the sol-gel process, it is important to optimize the starting composition and synthetic parameters, in which transparent non-macroporous gels can be formed without a phase separation inducer. [5] Without PTHF, a transparent dark green gel was obtained, and the dried gel had a non-macroporous structure (Figure S1A), and with appropriate amounts of PTHF, macroporous structures were formed. Figure 2A and Figure S2 show scanning electron microscope (SEM) images of the Cr-BTC gels with different amount of PTHF, and formation and coarsening of the interconnected macroporous structure were observed with an increase of PTHF amount. At w PTHF = 1.3 g, in addition to the continuous macropores, isolated macropores [8] were observed within the skeletons in addition to the continuous macropores as a result of secondary phase separation ( Figure S2B). The mercury intrusion results show the macropore diameter was controlled from 1.3 μm to 3.0 μm in the range of starting composition ( Figure 2B-C, Table S1). When the macroporous Cr-BTC gel was prepared in the glass tube, the obtained macroporous Cr-BTC gel had a rod-like crack-free monolithic form ( Figure 2D). It is worth noting that this crack-free monolithic form can be preserved whenever the continuous macropores morphology was formed ( Figure S1 and Figure 2D). Introduction of continuous well-defined macroporous structures can reduce the local shrinkage stresses to allow isotropic shrinkage during the drying process, and the crackfree monolithic form of Cr-BTC gels was preserved.
The energy dispersive X-ray spectroscopy (EDS) results of the (non-)macroporous Cr-BTC gels show uniform distributions of Cr and BTC, and impurities were not observed ( Figure S3-6, Table S2-3). The Fourier transform infrared spectroscopy (FTIR)  (Figure S7). These results suggest that both of the non-macroporous and macroporous Cr-BTC gels had the similar chemical structures.
The X-ray diffraction (XRD) patterns of the samples show the diffraction peaks of the Cr-BTC gels (w PTHF = 0 g, 1.1 g, 1.2 g, and 1.3 g) had only negligible differences ( Figure S1B, Figure  S8). In addition, XRD patterns of the samples show only weak and broad peaks, which are similar to the previously reported Cr-BTC gels. [3c] Although the crystalline framework structure was undetectable, the nitrogen adsorption-desorption measurement of the Cr-BTC gels show high specific surface area about 653-661 m 2 g −1 calculated by the Brunauer-Emmett-Teller (BET) method, which suggests the presence of the microporous structure derived from the MOF network. (Figure 2E and Table  S1). To investigate structural difference between the conventional MIL-100(Cr) particles ( Figure S19) and the Cr-BTC gels, the micropore size distribution of the samples were estimated by the non-local density functional theory (NLDFT) (Figure 3). The pores derived from the two type of pore cages ( Figure S10) can be assigned to 1.2 nm and 1.5 nm in NLDFT pore size distributions by analysing the conventional MIL-100(Cr) particles with assumption of N 2 -carbon at 77 K on a slit pore model (Figure 3B). Comparing the results of NLDFT analysis between conventional MIL-100(Cr) particles and the as-dried Cr-BTC gels, both samples have micropores with two peaks at 1.2 nm and 1.5 nm (Figure 3B). These results imply the obtained Cr-BTC gels have the microporous structure derived from the local network of MOF crystallites. In addition, we consider that the space between the crystallites, which is derived from the low crystallinity, provided the pore distribution with the peak at 1.8 nm (Figure 3B).
The Barrett−Joyner−Halenda (BJH) pore size distributions and isotherms of the Cr-BTC gels indicate the presence of mesopores below 10 nm (Figure S11). We consider these mesopore size distributions are derived from the pores between the primary particles which constitute the skeletons (Figure S12).
In addition, to investigate the mechanical strength of macroporous Cr-BTC gels, uniaxial compression test was performed ( Figure 2F). The macroporous Cr-BTC gel (w PTHF = 1.2 g) was shaped into a cylinder with a diameter of 5 mm and a height of 5 mm. Young's modulus of the macroporous Cr-BTC gel calculated from the linear elastic region was 1.2 GPa. This result confirms the high mechanical stability of Cr-BTC gels.
The compatibility in a binary polymeric system can be estimated by the thermodynamic treatment known as the Flory-Huggins equation. [9] The Gibbs free energy change of mixing for a binary system, ΔG is expressed as [Eq. (1)]; ∆G ∝ RT[{(φ 1 /P 1 )lnφ 1 +(φ 2 /P 2 )lnφ 2 } + {χ 12 φ 1 φ 2 }] (1) Here, φ i and P i (i = 1,2) denote the volume fraction and the degree of polymerization of component i, respectively, and χ 12 shows the interaction parameter, and R and T are the gas constant and temperature, respectively. The former two terms in the bracket represent the entropic contribution, and the last term the enthalpic contribution. From the thermogravimetrydifferential thermal analysis (TG-DTA) results of the as-dried (non-)macroporous Cr-BTC gel (w PTHF = 0 g, 1.2 g) and PTHF (Figure S13), it is evident that PTHF was removed by the solvent exchange due to its weak attractive interaction with Cr-BTC. The volume fraction of macropores remains almost the same with the varied amount of PTHF ( Figure 2B). This tendency is the same in the Zr-BDC-NH 2 based sol-gel system, in which hydrophobic polymer PPG acted as a phase separation inducer. [5] In addition, Figure S14 shows that SEM images of the Cr-BTC gel prepared with PPG (M n = 4 000, w PPG = 1.3 g) as phase separation inducer instead of PTHF. Distinct phase separation was not induced by PPG in the Cr-BTC system, and only a fine macroporous structure was observed (Figure S14B and C). Considering the entropic contribution term in Eq. (1), higher molecular weight of the polymer induces phase separation strongly. While the molecular weight of PPG (M n = 4 000) is higher than that of PTHF (M n = 2 900), PTHF (M n = 2 900) induced phase separation more strongly than PPG (M n = 4 000). Therefore, enthalpic interaction between the hydrophilic Cr-BTC network and hydrophobic polymer is more important in inducing phase separation in the present system. Since PPG is more hydrophilic than PTHF, we consider PPG is not hydrophobic enough to induce the separation in the Cr-BTC based sol-gel system. In our previous report on Zr-BDC-NH 2 based sol-gel system, PPG (M n = 1 000) was used as phase separation inducer, and PPG induced macroscopic phase separation strongly. [5] We consider that appropriate choices of hydrophobicity and molecular weight of phase-separation inducer is a key to applying SIPS technique to other hydrophilic MOF gel-based sol-gel systems.
Secondly, we discuss the formation and control of macroporous structures of Zr-BTC gels. The sol-gel transition behaviour of a solution of Zr(IV) salt and BTC has been reported. [3d] The typical chemical fomula of Zr-BTC is Zr 6 O 4 (OH) 4 BTC 2 (HCOO) 6 , known as MOF-808, a crystalline MOF powder. [10] As mentioned above, macroporous structures of two kinds of MOF gels, Zr-BDC-NH 2 and Cr-BTC, have been successfully controlled by using hydrophobic polymer as phase separation inducer. In the Zr-BTC based sol-gel system, we also applied a hydrophobic polymer as phase separation inducer. The suitable starting compositions and reaction conditions have been specified. We summarized the starting compositions of synthesis of macroporous MOF gels via SIPS technique in Figure  S15. In Zr-BTC system, we used PPG as hydrophobic polymer ( Figure S15). Without PPG, transparent non-macroporous Zr-BTC gels have been obtained via the sol-gel process from the mixture of metal salt ZrOCl 2 ·8H 2 O, organic linker BTC, and solvent N,N-dimethylformamide (DMF) and formic acid (HCOOH) (Figure S16). In the presence of hydrophobic polymer PPG (M n = 1 000), however, formation of the interconnected macroporous structures was observed (Figure S17A-C). The details of synthetic conditions and discussion of the synthesis of (non-)macroporous Zr-BTC can be found in the Electronic Supplementary Information.
Furthermore, by taking the advantage of the shape of the bulky monolithic gels, we demonstrate here the top-down processability toward macroporous microparticles. In applications in which mass transfer rate of the guest molecule has significant impact, the material shapes are also important depending on the specific applications and set-up. 11 Figure 4A shows a schematic of the synthesis of the macroporous Cr-BTC microparticles. Simply, the obtained as-dried gels were processed with top-down crushing and classifying processes into large sized microparticles (45-100 μm) by using stainless sieves. Figure 4B-D show the SEM images of the processed macroporous Cr-BTC microparticles. The as-dried macroporous Cr-BTC microparticles had well-defined macroporous structure, which did not collapse during the crushing process. This result shows the high formability of the macroporous monolithic MOF gels by the top-down processing.
These results demonstrate a general synthetic strategy to control macroporous structures of MOF gels. Macroporous structures of hydrophilic MOF gels can be controlled by employing specific hydrophobic polymer which can induce phase separation. In addition, the obtained macroporous structures tend not to collapse during the top-down crushing process, and this result provides opportunities to prepare macroporous MOF gels with different material shapes from crack-free monoliths to microparticles. Furthermore, the obtained MOF gels have similar pore size distributions with that of conventional MOF particles in the size range derived from the framework structures of MOFs.
We have previously demonstrated that reorganization from amorphous-like macroporous MOF gels to crystalline hierarchical MOFs can be controlled without deteriorating macroscopic morphology by the post-treatment in the Zr-BDC-NH 2 gels system. [5] In addition, by optimizing this reorganization process, the mesopores derived from the space between the primary particles can be controlled with target pore size. We believe the methodology of controlling macropores and material shapes of MOF gels, which have been generalized in this study, will be a platform to connect reticular chemistry and bulk materials [12] through further investigation of crystallization and/or functionalization conditions of MOF gels into the materials with controlled multimodal pore structures. In the future, we aim to develop a novel class of model porous materials which have well-defined hierarchical structures and reticular structures toward investigation for the further applications. This work was supported by a Grant-in-Aid for JSPS Fellow (19J22552 for Y.H.) and Grant-in-Aid for Scientific Research (B) (18H02056 for K.N. and K.K.) both from JSPS, Japan. The authors also thank Prof. Shuhei Furukawa (Kyoto University) for valuable discussion. The authors are grateful to Dr. Kei Morisato (Kyoto University) for technical support for the mercury porosimetry analysis. The authors also thank Prof. Hiroshi Kitagawa (Kyoto University) for his support.

Conflicts of interest
There are no conflicts to declare.