Scalable and precise synthesis of two-dimensional metal organic framework nanosheets in a high shear annular microreactor

by Nicholas A. Jose, Hua Chun Zeng and Alexei A. Lapkin

Abstract

Despite their promise as next generation materials for adsorbents, membranes and sensors, two-dimensional metal organic frameworks (2D MOFs) are far from commercial adoption. Current synthesis methods are neither scalable nor precise enough to use at industrial scales. Furthermore, the characterization of 2D MOF nanostructure is problematic due to aggregation in post-processing. By accelerating precipitation kinetics and using the recently developed annular flow microreactor, we synthesized copper benzene-dicarboxylic acid nanosheets with significantly higher efficiency and precision than conventional batch methods. The reactor space-time yield was five orders of magnitude higher than the previously published batch methods. We used liquid cell transmission electron microscopy to reduce drying-induced aggregation and revealed a monodisperse particle size distribution. These developments are step-changes in the synthesis and analysis of 2D MOF structures, and may accelerate the commercialization of innovative 2D MOF technologies.

Introduction

Two-dimensional metal organic frameworks (2D MOFs) are a promising class of porous materials enabled by reticular chemistry [1], [2] for catalysis [3] separation [4], [5] and sensing [6]. They are characterized by high aspect ratios, abundant open metal sites, nanometer thinness and high tunability. They may also be used to form a variety of morphologies, from nanosheets to microporous clusters. For example, two-dimensional copper benzene dicarboxylic acid (CuBDC) was shown to have outstanding performance when used in composite membranes for gas purification and nanofiltration [4], [5], [7]. Its industrial implementation may reduce energy consumption and emissions in chemical manufacturing, given that 70% of input energy in most processes are spent on purification [8].

However, the lack of precise, scalable synthesis methods limits the wider use, development and commercialization of 2D MOFs. Current laboratory methods use liquid-phase, hydrothermal conditions in batch reactors, requiring elevated temperatures and reaction times up to 24 h. While convenient for gram-scale quantities, such methods are usually expensive, hazardous and imprecise when directly translated to ton-scale quantities.

Continuous synthesis methods are generally more space-time efficient and therefore more scalable. Although several continuous protocols have been reported for the synthesis of MOFs nanoparticles, such as ZIF-8 [9] and HKUST-1 [10], few continuous methods [11] for the precise synthesis of MOF nanosheets with controlled shape and size have been reported, to the best of the authors’ knowledge.

Efficient, continuous synthesis of MOF nanosheets is challenging due to the combination of three factors: 1) attainable mixing, 2) fast kinetics of particle growth, and 3) anisotropic growth. To achieve a narrow size distribution, reaction conditions must be tightly controlled by precise mixing. Microreactors can provide fast, continuous mixing, and have been used previously with other MOF systems. However, at high saturations reactor clogging and scaling can be problematic. To avoid this, reactor geometry, solvents, concentrations and flowrates must be optimized.

When designing a continuous process, multiple parameters must be selected, i.e. pressure, temperature, reactor geometry and flowrates. Knowing the rates of particle nucleation and growth kinetics is tremendously helpful; however, the kinetics of metal organic framework crystallization are not well-explored [12], [13], and are often complicated by the diversity of precursors, solvents and surfactants employed in different published methods.

Lastly, to form highly anisotropic structures, growth must be favored in one dimension. Most current techniques utilize interfaces between a liquid and solid [14], two immiscible liquids [4], or between a liquid and gas [15], at which precursor molecules may align and crystallize in two directions. Other methods include mixed solvent hydrothermal synthesis, in which solvent-material interactions encourage the growth of anisotropic shapes [16].

These methods are difficult to scale. In a conventional continuous reactor, convection disturbs the interface of immiscible fluids. Surfactants must be removed post-reaction, which increases operating and capital costs. High pressures and temperatures incur additional costs and introduce processing hazards.

High shear processing is an attractive alternative. Enhanced shear rates affect clogging and aggregation of nanoparticle suspensions, and can be used to produce highly anisotropic structures with graphene and other 2D materials [17], [18], [19]. It also provides rapid mixing. However, high shear processing requires uniform shear rate distributions, set to an optimal value at which both anisotropic crystallization is preferred and large aggregates are unstable [20]. “Top-down” shear-induced exfoliation of 2D materials is also limited in its control of particle size and yields. For example, in the synthesis of graphene via graphite exfoliation, yields were typically below 3% [18].

In this study we overcame these challenges and developed a more scalable technique for the precise synthesis of CuBDC. Using triethylamine, we accelerated the kinetics of CuBDC precipitation at ambient temperature and pressure. Using the recently developed annular microreactor [21], which provides rapid mixing and high, uniform shear, we synthesized pristine, monodisperse CuBDC nanosheets. To avoid the effects of aggregation caused by sample dry-down in conventional electron microscopy, we used liquid transmission electron microscopy to obtain a more accurate particle size distribution.

Materials and methods

Materials

Copper nitrate hexahydrate (99.5%), 1,4-benzene dicarboxylic acid and triethylamine were obtained from Merck. HPLC grade acetonitrile, dimethylformamide, ethanol and methanol were obtained from Tedia and Singapore Chemical Reagent Co.

Synthesis

CuBDC was synthesized in the annular microreactor previously described [19]. Three quartz tubes were installed in a “tube-in-tube” configuration to create two annular flow zones. Liquids were delivered to the reactor using a KDS Legato Dual Syringe Pump using disposable plastic 10 mL Terumo syringes. Compressed dried air was passed through a 200 µm filter (Swagelok) and delivered to the reactor using a Sierra SmartTrak C50L Mass Flow controller (20 L min−1 max, 2% accuracy).

Stainless steel tee connectors with 1/4″ and 1/16″ diameter compression fittings were purchased from Swagelok. Precision quartz capillary round tubes of the following dimensions: 0.30 mm ID × 0.4 mm OD × 100 mm L (T1), 0.50 mm ID × 0.7 mm OD × 100 mm L (T2), and 1 mm ID × 1.2 mm OD × 100 mm L and 300 mm L (T3) were purchased from VitroCom (supplied by ArteGlass Japan). Gas tight connections between stainless steel fittings and quartz tubes were made with graphite and Teflon 1/16″ compression ferrules. 0.8 mm ID and 0.5 mm ID ferrules were purchased from Restek, and 1/ 16″ PTFE ferrules were purchased from Swagelok. 0.8 mm ID PTFE tubing from ChemiKalie was used for delivery of liquid solutions to the reactor.

These fittings were mounted and aligned using a custom-built mounting base with precision-milled grooves and positioning clamps. The capillaries and fittings were installed in the clamps, and could be finely manipulated the positioning of the mounting bolts and positioning screws. Alignment of the capillaries was evaluated visually using a portable microscope (RSPro) or magnifying glass at various locations along the reactor length. The two reagent solutions, “A” and “B” were prepared as follows. A: Copper nitrate was dissolved in DMF and acetonitrile (1/1 by volume). B: TEA and BDC were dissolved in DMF and acetonitrile (1/1 by volume).

The molar ratios of Cu to BDC and TEA to BDC were varied from 0.5 to 4 and from 0.2 to 4 respectively. The concentration of BDC in “B” was varied from 0.015 to 0.24 M. Characterized material was synthesized with 0.06 M BDC in “B.” The solutions were sonicated for five min to completely dissolve the reagents. Solution A and Solution B were pumped simultaneously at 4 to 10 mL min−1 per reagent. Compressed dried air was simultaneously flowed at a high velocity at 0 to 3 L min−1 through the innermost tube. Unless otherwise stated, reactions were conducted at 6 mL min−1 per reagent and 1 L(STP) min−1 air. The resulting precipitates were centrifuged at 6000 rpm and rinsed with DMF three times.

Annular fluid flow characteristics were determined using the empirical model of Han et al. [22], which was previously validated in this reactor using high speed scanning fluorescence microscopy [21]. The viscosity and density of a 1/1 volumetric combination of DMF to acetonitrile (equivalent to a mole fraction of acetonitrile to DMF of 0.60) are approximately 0.55 mPa s−1 and 865 kg m−3 respectively [23] at ambient conditions (20–25 °C).

Two lower-shear methods were also evaluated for comparison. In one, solution A (Cu:BDC = 1:1) was combined with solution B (TEA:BDC = 1:1) dropwise over the course of a minute into a gently hand-shaken 30 mL vial. This was done to mimic a slow mixing process in which the shear rate and Re number are on the order of 1–100 (s−1 and dimensionless quantity respectively). In the second, the solutions were combined in the annular reactor without the addition of a gas flow, such that the reaction was conducted in laminar pipe flow, in which the Re number is ρυD/µ, where ρ is the liquid density (kg m−3 ), υ is the average liquid velocity (m s−1 ), D is the tube diameter (m) and µ is the liquid viscosity (Pa·s), and the average shear rate is 4υ/D. The flowrates tested and their calculated Re number, residence time, shear rate and characteristic mixing time are given in Table 1. The characteristic mixing time of the annular microreactor is calculated using the engulfment mixing model of Baldyga & Bourne (1989) [24], which was previously validated. The characteristic mixing time for laminar flow in a microchannel of 1 mm in diameter is approximated from the survey of mixers done by Falk & Commenge (2010) [25].

Powder X-ray diffraction

Suspensions were diluted in ethanol and drop-cast onto a non-reflective silicon wafer (1 0 0) and dried at 80 °C for 30 min. The powder x-ray diffraction pattern was collected with a Brucker D8 Advance Powder Diffractometer using Cu Kα radiation (λ = 1.5418 Å) at 40 kV from a 2θ of 3° to 70° with a step size of 0.02° and a scanning rate of 1.25° min−1 .

Surface Area Analysis

Suspensions were thoroughly rinsed with MeOH and activated by Soxhlet extraction with methanol for three consecutive days as described in previous studies [4,26]. The activated solid was then dried at 80 °C and degassed under reduced pressure (< 10−2 Pa) at 160 °C for 12 h. N2 adsorption and desorption isotherms were collected on a Micromeritics ASAP2020 Surface Area and Porosity Analyzer with 23.9 mg of sample at 77 K. Surface area was calculated using Langmuir [27] and Brunauer Emmett-Teller [28] (BET) theories. The pore size distribution was determined by 2D Non-Local Density Functional Theory (2D-NLDFT) [29] over 5.5936 Å – 209.7892 Å.

Atomic force microscopy

Suspensions diluted in ethanol, sonicated for 1 min, dropped onto 1 mm quartz slides, and dried. Height and phase maps were taken with a Bruker Dimension ICON AFM with a silicon SPM probe (MikroMasch NSC15) while tapping in air. Maps were flattened using a 1st or 2nd order fit for each line scan and analyzed using Bruker Nanoscope Analysis.

Transmission electron microscopy

Suspensions were diluted in ethanol and dropped on holey carbon
200 mesh copper TEM grids (InLab Supplies) and dried at ambient
temperature. Images were taken with a JEOL 2100F FETEM at 200 kV.
Samples were imaged for brief time periods under low current density
to prevent the reduction of copper and radiation damage.

Liquid transmission electron microscopy

All liquid TEM work was started within an hour after synthesis. Post-synthesized samples were diluted 100x in DMF after synthesis and imaged using a Protochips Poseidon holder fitted with an E-chips liquid cell. The cell consists of two silicon chips, coated with a 50 nm thick silicon nitride membrane, with windows of approximately 500 × 50 µm. The spacer between the two chips, composed of 150 nm thick SU-8, was deposited in a modified flow configuration. A Tecnai F20 FETEM was used at 200 kV in bright field mode at an electron dose at approximately 80 e- nm−2 s −1 . Video was taken at 25 fps at 4 k × 4 k resolution using a Gatan OneView camera. Sample was imaged under static conditions. Particle sizes were measured in imageJ. The area-averaged side-length of at least 100 individual particles were measured to construct a reasonable particle size distribution.

Scanning electron microscopy

Suspensions were diluted in ethanol and drop-cast on silicon substrates, dried at ambient temperature and coated with Pd in a Cressington sputter coater for 70 s. Images were taken with a JEOL JSM-5600LV FESEM at 5 kV.

Fourier transform infrared spectroscopy

Suspensions were dried at 80 °C. Resulting powder was mixed with KBr and pressed into a pellet. FTIR spectra were measured on a Bruker Tensor II spectrometer.

Results and discussion

CuBDC precipitates via the reaction shown in Fig. 1, in which positively charged copper ions coordinate with negatively charged BDC ions, with two copper ions per cluster. Due to the lower polarity of BDC, a co-solvent system of dimethylformamide (DMF) and acetonitrile (CH3CN) at a 1/1 volumetric ratio was used to increase solubility. Triethylamine, which increases pH and the deprotonation of BDC ions, was added to accelerate the reaction such that a high yield could be obtained within seconds at ambient temperature. The initial reaction concentration of BDC was 0.03 M, while the molar ratios of Cu to BDC (Cu/BDC) and TEA to BDC (TEA/BDC) were varied from 0.5 to 4 and 0.2 to 4 respectively.

Synthesis was conducted in the recently developed high shear annular microreactor, which utilizes two-phase annular flow to generate films with high, uniform shear and rapid mixing. In this reactor, shown schematically in Fig. 2, three quartz capillary tubes are arranged in a staggered, coaxial configuration (i.e. “tube-in-tube”). While high velocity gas is pumped through the inner-most tube, liquid precursor streams containing copper nitrate (“Liquid 1”) and BDC (“Liquid 2”) were introduced via T-fittings to the outer tubes. This configuration generates high shear films that can combine with characteristic micromixing times below 1 ms. In this study, the flowrates were fixed at 1 L min−1 (air) and 6 mL min−1 (for each liquid reagent). The resulting suspension was a blue-green color and formed a viscous gel when centrifuged. It was rinsed with DMF before analysis.

The crystal structure, chemical composition, electronic structure and morphology were determined via powder x-ray diffraction (XRD), Fourier transform infrared spectroscopy (FTIR), scanning electron microscopy (SEM), transmission electron microscopy (TEM), and atomic force microscopy (AFM). The surface area and pore size distribution were characterized via nitrogen adsorption.

The molar ratios of Cu/BDC and TEA/BDC had significant effects on the purity of the precipitate, which was evaluated using XRD. At TEA/ BDC molar ratios exceeding 1, a copper hydroxide phase is more likely to form, and at TEA/BDC molar ratios below 1, pristine nanosheets are not formed. At Cu/BDC molar ratios above or below 1, which are nonstoichiometric, impurity phases also develop. When both Cu/BDC and TEA/BDC equal 1, pristine nanosheets are produced with few impurities, as seen in Fig. 3.

These findings are consistent with what is known about TEA in MOF synthesis [30–33]. It increases the rates of nucleation and growth, resulting in higher yields and smaller particle sizes. While TEA is not known to be a surface-capping agent, at excessive amounts it can alter the crystallinity and chemistry of the synthesized MOF.The FTIR spectrum and XRD patterns of the optimized Cu(BDC) show the characteristic peaks of DMF-intercalated CuBDC (DMFCuBDC) [35,36], seen in Fig. 4(a & b). The (Brunauer-Emmett-Teller) BET surface area was 324 m2 g−1 , which is consistent with previously reports [26]. The micropore size distribution in Fig. 4(d & e) shows a maximum peak at 6 and 6.4 Å. From 35 to 200 Å there is an additional increase in porosity, most likely arising from defects in the arrangement of nanosheets.Changing reactor flowrates and conditions (see conditions in Table 1) also produced noticeable changes in crystallinity, as measured by powder XRD (see Fig. 5a). The crystal structure of DMF-CuBDC has prominent reflections for the lattice planes (1 1 0) and − (201), which are highlighted in Fig. 5b, corresponding to crystallinity in the lateral and stacking directions respectively. For the DMF-CuBDC materials synthesized in the hand-mixed reactor (1–100 s−1 ) and in the pipe flow reactor (2.0 · 103 s −1 ), the peak intensity of the (1 1 0) reflection, relative to the ( − (201)) reflection intensity, is approximately 0.1 (see Fig. 5c). For the DMF-CuBDC synthesized in the annular flow microreactor, which possesses significantly higher shear rates (from 1.1·104 to 9.1·105 s −1 ) and more uniform shear rate distributions, the (1 1 0) peak intensities were less than 0.02. The intensities of the (0 2 0), ( − 1 1 1), (2 2 0) and (1 3 1) peak reflections at 12.3°, 13.6°, 20.5° and 24.8° respectively were also significantly decreased. No significant changes in peak intensities or peak widths were detected when shear rates were increased from 1.1 · 104 to 9.1 · 105 s −1 .When comparing crystallinity across the dimensionless Re number for each reactor condition, there appears to be a non-monotonic trend; however, this trend is misleading due to the significant differences between the hydrodynamics of pipe flow and of the two-phase annular flow. In this case, Re number only describes whether the flows are turbulent or laminar, and does not adequately describe other important parameters, such as mixing speed and uniformity.We attribute the dependence of crystallinity on increasing shear rate and hydrodynamic conditions to their influences on particle nucleation and growth [20]. When TEA is used as an accelerant for CuBDC precipitation nucleation and growth occur rapidly, instantaneously to the human eye. If the characteristic mixing time ™ is significantly less than the characteristic growth time (tg), nucleation and growth kinetics are segregated. One well-known embodiment of this segregation is known as the LaMer model [37]: supersaturation is consumed in the formation of nuclei above a critical monomer concentration, after which crystal growth occurs via diffusion and Oswald ripening. When tm ≥ tg, growth and nucleation compete, resulting in less nuclei and larger, less uniform crystals.

The increase of crystallinity in the lateral direction, corresponding to longer sheets of CuBDC, indicates that the characteristic mixing times of the hand-mixed reactor (~60 s) and capillary pipe flow reactor (~0.1) are on the order of the characteristic growth time, such that particles are larger and less uniform. The annular microreactor, which possesses characteristic micromixing times on the order of milliseconds (see Table 1), segregates nucleation and growth such that smaller, more uniform populations of nanosheets are formed. The dependence of CuBDC formation on reactor conditions further highlights the importance of the annular microreactor in controlling crystal properties.

Fig. 3. Powder XRD patterns of material synthesized at (a) increasing ratios of Cu/BDC where TEA/BDC = 1, and (b) increasing ratio of TEA/BDC, where Cu/ BDC = 1. Impurity peaks (‘o’) associated with Cu2(OH)3(BDC)0.5 and Cu2(OH)3NO3 [34] are assigned to the following reflections: 8.3°, 15.4°, 26.1°, 27.8°, 29.9°; Representative CuBDC-DMF peaks (‘*’) and reflections are assigned at 10.2° (1 1 0), 16.9° ( − 201), and 34.2° − (4 0 2)) [35]. The peak (‘x’) at 26.1° represents an unknown amorphous impurity.

The increased shear of the reactor also affects other growth mechanisms such as aggregation. In our previous application of high shear to layered double hydroxides, we found that high shear can facilitate oriented attachment, and that crystallinity has a non-monotonic dependence on shear rate. In this work, a non-monotonic dependence was not observed, indicating that oriented assembly is not a dominant growth mechanism for nanosheets in our reaction conditions.

Conventional techniques of nanostructure analysis were hampered by the aggregation of particles during sample drying. As seen in TEM and SEM images in Fig. 6 (a & b), the CuBDC nanoparticles tended to attach to each other. AFM was slightly more effective in obtaining heights and thicknesses of individual particles. Single particles could adhere to the mica substrate, but aggregation persisted, as seen in Fig. 6 (c & d). AFM analysis of individual particles showed high aspect ratio particles. The smallest measured particles possessed a side length of approximately 37 nm and thickness of < 1 nm (seen in Fig. 6d).

To avoid drying-induced aggregation, we used liquid transmission electron microscopy (LTEM), which involves encapsulating a liquid sample between silicon nitride membranes that are transparent to the electron beam, thus enabling nanoscale resolution of nanoparticles in a fluid environment. In a suspension of DMF, particles were significantly more dispersed and easier to analyze, though at a lower resolution than conventional TEM. As seen in Fig. 6 (e & f), dispersed single particles could be clearly resolved with LTEM. The average particle side length, 33 ± 4 nm, was relatively uniform considering the resolution of LTEM, and comparable to the minimum size measured in AFM.

Relatively few studies of MOFs have been conducted using LTEM [38,39]. The further use of LTEM to study MOF nanoparticles may be beneficial in similar studies when sample drying and the high vacuum environment of TEM affects aggregation and crystalline structure. The dynamics of aggregation itself may also be further studied; for example, we noticed that the CuBDC nanosheets tended to cluster into squarelike configurations, which suggested that oriented attachment may be an alternate crystallization mechanism. It is also interesting to note that good contrast was achieved with thin organic particles. The use of perpendicular to the (1 1 0) and ( − 201) lattice planes, shown in green and purple respectively. Legend: red = oxygen, brown = carbon, white = hydrogen, dark blue = copper, light blue = nitrogen. (c) Relative peak intensity of the 110 reflection across different shear rates, (d) Re number and reactor types. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) LTEM may decrease the sensitivity of MOFs to high energy electron beams due to the dissipation of heat and charge through solvent. Furthermore, because MOF structures are known to be very sensitive to solvents, LTEM provides an interesting avenue to evaluate how solvents affect MOF properties in situ.

Fig. 4. FTIR spectrum, XRD pattern and surface area and pore size distribution analysis (BET). (a) XRD pattern showing characteristic reflections of DMF-CuBDC: (1 1 0) 10.2°, ( − 201) 16.9°, (1 3 1) 24.5°, (2 2 2) 27.9°, ( − 3 1 2) 29.2°, ( − (4 0 2)) 34.2°, ( − 4 2 2) 36.4° and (2 4 2) 41.4°. (b) FTIR transmission spectra showing characteristic peaks of DMF-CuBDC: (1387 cm−1 ) symmetric stretching of the BDC carboxylate group (“1”), (1576 cm−1 ) asymmetric stretching of the BDC carboxylate group (“2”), and (1662 cm−1 ) DMF coordinated with a Cu2+ center. (“3”). (c) N2 isotherm adsorption and desorption curve. (d & e) Incremental surface areas vs. pore size.

Fig. 5. Comparison of CuBDC crystallinity across different flowrates. (a) Overlaid XRD patterns of DMF-CuBDC synthesized at various shear rates (in s −1). (b) Stick-and-ball model of the crystal structure of DMF-CuBDC viewed perpendicular to the (1 1 0) and ( − 201) lattice planes, shown in green and purple respectively. Legend: red = oxygen, brown = carbon, white = hydrogen, dark blue = copper, light blue = nitrogen. (c) Relative peak intensity of the 110 reflection across different shear rates, (d) Re number and reactor types. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

This synthesis method is significantly more efficient and scalable than the previously reported hydrothermal methods by Carson et al. [36] and Rodenas et al. [4]. Those methods, which have yields of 10–80%, require sealed batch reactors with 8 to 87 mL of suspension over 24–36 h, at elevated temperatures from 40 to 110 °C. In the annular reactor at the optimal reagent concentrations a yield of 78.8% could be achieved at 20 °C after purification. The space–time-yield (mass/time/reactor volume), a measure of the efficiency of the process, is 6.58 g L−1 s −1 —which is 5 orders of magnitude higher than the previous methods. For single reactors, this translates to dry mass equivalent productivities of 1.08 · 10−3 g s−1 (this work), 2.37 · 10- 7 g s−1 (Rodenas et al.) and 6.13 · 10-6 g s−1 (Carson et al.). Compared to the efficiency of the continuous method published by Wang et al. [11] for 2D MF-ZrBTB—a reported 385 kg m−3 day−1 (0.0045 g L-1 s −1 and 2 · 10-4 g s−1 ) at 130 °C – our method is 1500 times more efficient in terms of space time yield and produces a more precise particle size distribution at a lower operating temperature.

These improvements, illustrated in Fig. 7, are due to the continuous nature of the microreactor, the use of a high shear rate to generate delaminated MOF nanosheets, and the use of TEA as an accelerant. This method may be extended to a range of 2D-MOF or covalent organic framework (COF) structures and opens new avenues to large-scale manufacture of 2D-MOF technologies.

Conclusions

We have shown several key advancements in the synthesis and analysis of two dimensional metal organic frameworks, which are essential for their future implementation and commercialization as ad[1]sorbents, separation agents, drug delivery agents and sensors. Using high shear annular microreactor technology we developed an ambient temperature process that is up to 105 times more efficient in terms of the process space–time-yield. This method not only makes laboratory synthesis easier but may also lead to more efficient and economically feasible industrial-scale manufacturing processes. Liquid transmission electron microscopy was shown to be an important technique in accurately characterizing particle size distributions. Because few studies of 2D MOFs (and MOFs in general) have been conducted with LTEM, the results of this study suggest further usage of LTEM for the in situ ana[1]lysis of particle size, morphology and aggregation dynamics.

Funding sources

This project was supported by the National Research Foundation (NRF), Prime Minister’s Office, Singapore, under its Campus for Research Excellence and Technological Enterprise (CREATE) program as a part of the Cambridge Centre for Advanced Research and Education in Singapore Ltd (CARES). Funding for the liquid TEM holder was provided by the Engineering and Physical Sciences Research Council’s Henry Royce Institute (EP/P024947/1).

Declaration of Competing Interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: The authors declare competing financial interests: a patent application has been filed for the reactor described (Singapore patent app. no. 10201801303T).

Fig. 7. Laboratory-scale annular reactor and its efficiency. (a) Reactor and 500 mL of DMF-CuBDC suspension synthesized in 30 min. (b) Comparison of mass-space velocity and synthesis temperature with previously reported methods.

Acknowledgements

We thank Dr. Zhan Guowu for his assistance in synthesis, Wang Yu-Xiang’s assistance with surface area analysis, and Dr. Giorgio Divitini and Dr. Caterina Ducati for their assistance in liquid TEM studies.

Competing Interests

Elements of this work – the reactor design and its use to synthesize 2D MOFs – are included in patent application PCT/GB2019/050406.

References

[1] O.M. Yaghi, M. O’Keeffe, N.W. Ockwig, H.K. Chae, M. Eddaoudi, J. Kim, Reticular synthesis and the design of new materials, Nature 423 (2003) 705–714.

[2] O.M. Yaghi, Reticular chemistry: molecular precision in infinite 2D and 3D, Mol. Front. J. 03 (2019) 66–83.

[3] N.T.S. Phan, T.T. Nguyen, K.D. Nguyen, A.X.T. Vo, An open metal site metal-organic framework Cu(BDC) as a promising heterogeneous catalyst for the modified Friedlander reaction, Appl. Catal. a-Gen. 464 (2013) 128–135.

[4] T. Rodenas, I. Luz, G. Prieto, B. Seoane, H. Miro, A. Corma, F. Kapteijn, F.X.L.I. Xamena, J. Gascon, Metal-organic framework nanosheets in polymer composite materials for gas separation, Nat. Mater. 14 (2015) 48–55.

[5] A. Sabetghadam, B. Seoane, D. Keskin, N. Duim, T. Rodenas, S. Shahid, S. Sorribas, C. Le Guillouzer, G. Clet, C. Tellez, M. Daturi, J. Coronas, F. Kapteijn, J. Gascon, Metal organic framework crystals in mixed-matrix membranes: impact of the filler morphology on the gas separation performance, Adv. Funct. Mater. 26 (2016) 3154–3163.

[6] F.F. Su, S. Zhang, H.F. Ji, H. Zhao, J.Y. Tian, C.S. Liu, Z.H. Zhang, S.M. Fang, X.L. Zhu, M. Du, Two-dimensional zirconium-based metal organic framework nanosheet composites embedded with au nanoclusters: a highly sensitive electrochemical aptasensor toward detecting cocaine, ACS Sensors 2 (2017) 998–1005.

[7] H.X. Ang, L. Hong, Polycationic polymer-regulated assembling of 2D MOF nanosheets for high-performance nanofiltration, ACS Appl. Mater. Inter. 9 (2017) 28079–28088.

[8] S.B. Adler, Earl; Bryan, Paul; Robinson, Sharon; Watson, Jack, Vision 2020: 2000 Separations Roadmap, in: C.f.W.R. Technologies (Ed.), American Institute of Chemical Engineers, New York, 2000.

[9] S. Watanabe, S. Ohsaki, T. Hanafusa, K. Takada, H. Tanaka, K. Mae, M.T. Miyahara, Synthesis of zeolitic imidazolate framework-8 particles of controlled sizes, shapes, and gate adsorption characteristics using a central collision-type microreactor, Chem. Eng. J. 313 (2017) 724–733.

[10] M. Faustini, J. Kim, G.Y. Jeong, J.Y. Kim, H.R. Moon, W.S. Ahn, D.P. Kim, Microfluidic approach toward continuous and ultrafast synthesis of metal organic framework crystals and hetero structures in confined microdroplets, J. Am. Chem. Soc. 135 (2013) 14619–14626.

[11] Y. Wang, L. Li, L. Yan, X. Gu, P. Dai, D. Liu, J.G. Bell, G. Zhao, X. Zhao, K.M. Thomas, Bottom-Up fabrication of ultrathin 2D Zr metal-organic framework nanosheets through a facile continuous microdroplet flow reaction, Chem. Mater. 30 (2018) 3048–3059.

[12] A. Dighe, R. Nemade, M. Singh, Modeling and simulation of crystallization of metal organic frameworks, Processes 7 (2019) 527.

[13] R. El Osta, M. Feyand, N. Stock, F. Millange, R.I. Walton, Crystallisation kinetics of metal organic frameworks from in situ time-resolved X-ray diffraction, Powder Diffr. 28 (2013) S256–S275.

[14] H. Li, J. Hou, T.D. Bennett, J. Liu, Y. Zhang, Templated growth of vertically aligned 2D metal–organic framework nanosheets, J. Mater. Chem. A 7 (2019) 5811–5818.

[15] R. Makiura, O. Konovalov, Interfacial growth of large-area single-layer metal-organic framework nanosheets, Sci. Rep. 3 (2013) 2506.

[16] M. Mazaj, T. Birsa Čelič, G. Mali, M. Rangus, V. Kaučič, N. Zabukovec Logar, Control of the crystallization process and structure dimensionality of Mg–benzene–1,3,5-tricarboxylates by tuning solvent composition, Cryst. Growth Des. 13 (2013) 3825–3834.

[17] E. Varrla, C. Backes, K.R. Paton, A. Harvey, Z. Gholamvand, J. McCauley, J.N. Coleman, Large-scale production of size-controlled MoS2 nanosheets by shear exfoliation, Chem. Mater. 27 (2015) 1129–1139.

[18] K.R. Paton, E. Varrla, C. Backes, R.J. Smith, U. Khan, A. O’Neill, C. Boland, M. Lotya, O.M. Istrate, P. King, T. Higgins, S. Barwich, P. May, P. Puczkarski, I. Ahmed, M. Moebius, H. Pettersson, E. Long, J. Coelho, S.E. O’Brien, E.K. McGuire, B.M. Sanchez, G.S. Duesberg, N. McEvoy, T.J. Pennycook, C. Downing, A. Crossley, V. Nicolosi, J.N. Coleman, Scalable production of large quantities of defect-free few layer graphene by shear exfoliation in liquids, Nat. Mater. 13 (2014) 624–630.

[19] E. Varrla, K.R. Paton, C. Backes, A. Harvey, R.J. Smith, J. McCauley, J.N. Coleman, Turbulence-assisted shear exfoliation of graphene using household detergent and a kitchen blender, Nanoscale 6 (2014) 11810–11819.

[20] N. Jose, A. Lapkin, Chapter 2 – Influence of hydrodynamics on wet syntheses of nanomaterials, in: V.A. Sadykov (Ed.), Advanced Nanomaterials for Catalysis and Energy, Elsevier, 2019, pp. 29–59.

[21] N.A. Jose, H.C. Zeng, A.A. Lapkin, Hydrodynamic assembly of two dimensional layered double hydroxide nanostructures, Nat. Commun. 9 (2018) 4913.

[22] Y. Han, H. Kanno, Y.J. Ahn, N. Shikazono, Measurement of liquid film thickness in micro tube annular flow, Int. J. Multiphas. Flow 73 (2015) 264–274.

[23] Acetonitrile-N,N-Dimethylformamide (DMF) Mixture Viscosity: Datasheet from “Dortmund Data Bank (DDB) – Thermophysical Properties Edition 2014” in SpringerMaterials, Springer-Verlag Berlin Heidelberg & DDBST GmbH, Oldenburg, Germany.

[24] J. Baldyga, J.R. Bourne, Simplification of micromixing calculations. 1. Derivation and application of new model, Chem. Eng. J. Bioch. Eng. 42 (1989) 83–92.

[25] L. Falk, J.M. Commenge, Performance comparison of micromixers, Chem. Eng. Sci. 65 (2010) 405–411.

[26] Y.D. Cheng, X.R. Wang, C.K. Jia, Y.X. Wang, L.Z. Zhai, Q. Wang, D. Zhao, Ultrathin mixed matrix membranes containing two-dimensional metal-organic framework nanosheets for efficient CO2/CH4 separation, J. Membr. Sci. 539 (2017) 213–223.

[27] I. Langmuir, The adsorption of gases on plane surfaces of glass, mica and platinum, J. Am. Chem. Soc. 40 (1918) 1361–1403.

[28] S. Brunauer, P.H. Emmett, E. Teller, Adsorption of gases in multimolecular layers, J. Am. Chem. Soc. 60 (1938) 309–319.

[29] J. Jagiello, J.P. Olivier, A simple two-dimensional NLDFT model of gas adsorption in finite carbon pores. Application to pore structure analysis, J. Phys. Chem. C 113 (2009) 19382–19385.

[30] H. Lu, S. Zhu, Interfacial synthesis of free-standing metal-organic framework membranes, Eur. J. Inorg. Chem. 2013 (2013) 1294–1300.

[31] N.A.H.M. Nordin, A.F. Ismail, A. Mustafa, P.S. Goh, D. Rana, T. Matsuura, Aqueous room temperature synthesis of zeolitic imidazole framework 8 (ZIF 8) with various concentrations of triethylamine, RSC Adv. 4 (2014) 33292–3300.

[32] C.V. McGuire, R.S. Forgan, The surface chemistry of metal–organic frameworks, Chem. Commun. 51 (2015) 5199–5217.

[33] P. Long, Q. Zhao, J. Dong, J. Li, Synthesis of metal-organic frameworks from the system metal/L-glutamic acid/TEA/H2O, J. Coord. Chem. 62 (2009) 1959–1963.

[34] S.P. Newman, W. Jones, Comparative study of some layered hydroxide salts containing exchangeable interlayer anions, J. Solid State Chem. 148 (1999) 26–40.

[35] C.G. Carson, G. Brunnello, S.G. Lee, S.S. Jang, R.A. Gerhardt, R. Tannenbaum, Structure solution from powder diffraction of copper 1,4 benzenedicarboxylate, Eur. J. Inorg. Chem. 2014 (2014) 2140–2145.

[36] C.G. Carson, K. Hardcastle, J. Schwartz, X.T. Liu, C. Hoffmann, R.A. Gerhardt, R. Tannenbaum, Synthesis and structure characterization of copper terephthalate metal-organic frameworks, Eur. J. Inorg. Chem. (2009) 2338–2343.

[37] V.K. Lamer, R.H. Dinegar, Theory, production and mechanism of formation of monodispersed hydrosols, J. Am. Chem. Soc. 72 (1950) 4847–4854.

[38] J.P. Patterson, P. Abellan, M.S. Denny, C. Park, N.D. Browning, S.M. Cohen, J.E. Evans, N.C. Gianneschi, Observing the growth of metal-organic frameworks by in situ liquid cell transmission electron microscopy, J. Am. Chem. Soc. 137 (2015) 7322–7328.

[39] K.M. Vailonis, K. Gnanasekaran, X.B. Powers, N.C. Gianneschi, D.M. Jenkins, Elucidating the growth of metal-organic nanotubes combining isoreticular synthesis with liquid-cell transmission electron microscopy, J. Am. Chem. Soc. 141 (2019) 10177–10182.

This article is cited by

Science Direct