Research Article www.advmat.de Synthesis of 2D Porous Crystalline Materials in Simulated Microgravity Noemí Contreras-Pereda, David Rodríguez-San-Miguel, Carlos Franco, Semih Sevim, João Pedro Vale, Eduardo Solano, Wye-Khay Fong, Alessandra Del Giudice, Luciano Galantini, Raphael Pfattner,* Salvador Pané, Tiago Sotto Mayor,* Daniel Ruiz-Molina,* and Josep Puigmartí-Luis* 1. Introduction To date, crystallization studies conducted in space laboratories, which are prohibitively costly and unsuitable to most research laboratories, have Reactions are frequently conducted under shown the valuable effects of microgravity during crystal growth and mor- vigorous stirring to improve the mass phogenesis. Herein, an easy and highly efficient method is shown to achieve transport of reagents and to increase the space-like experimentation conditions on Earth employing custom-made reaction efficiency via turbulent flows. microfluidic devices to fabricate 2D porous crystalline molecular frameworks. However, turbulent conditions lead to a It is confirmed that experimentation under these simulated microgravity poor control of convective transport pro- cesses which may prompt the precipita- conditions has unprecedented effects on the orientation, compactness and tion of reaction products and significantly crack-free generation of 2D porous crystalline molecular frameworks as well affect the reaction performance. This as in their integration and crystal morphogenesis. It is believed that this scenario is especially critical during the work will provide a new “playground” to chemists, physicists, and materials synthesis of 2D and 3D functional crys- scientists that desire to process unprecedented 2D functional materials and tals with controlled sizes and shapes. For example, 3D crystalline porous mole- devices. cular frameworks such as metal–organic N. Contreras-Pereda, Dr. D. Ruiz-Molina J. P. Vale, Dr. T. S. Mayor Catalan Institute of Nanoscience and Nanotechnology (ICN2) SIMTECH Laboratory CSIC and BIST Transport Phenomena Research Centre Campus UAB Engineering Faculty of Porto University Bellaterra Porto 4200–465, Portugal Barcelona 08193, Spain E-mail: tiago.sottomayor@fe.up.pt E-mail: dani.ruiz@icn2.cat Dr. E. Solano Dr. D. Rodríguez-San-Miguel, Dr. C. Franco, S. Sevim ALBA Synchrotron Department of Chemistry and Applied Biosciences Carrer de la Llum 2–26 Institute for Chemical and Bioengineering Cerdanyola del Vallès, Barcelona 08290, Spain ETH Zurich Prof. W.-K. Fong Zurich 8093, Switzerland Discipline of Chemistry Dr. D. Rodríguez-San-Miguel, Prof. J. Puigmartí-Luis School of Environmental and Life Sciences Departament de Ciència dels Materials i Química Física University of Newcastle Institut de Química Teòrica i Computacional Callaghan, NSW 2308, Australia Barcelona 08028, Spain Dr. A. Del Giudice, Prof. L. Galantini E-mail: josep.puigmarti@ub.edu Department of Chemistry Sapienza University of Rome The ORCID identification number(s) for the author(s) of this article P. le A. Moro 5, Rome 00185, Italy can be found under https://doi.org/10.1002/adma.202101777. Dr. R. Pfattner Institut de Ciència de Materials de Barcelona © 2021 The Authors. Advanced Materials published by Wiley-VCH ICMAB-CSIC GmbH. This is an open access article under the terms of the C ­ reative Campus UAB, Bellaterra 08193, Spain Commons Attribution-NonCommercial License, which permits use, dis- E-mail: rpfattner@icmab.es tribution and reproduction in any medium, provided the original work is Prof. S. Pané properly cited and is not used for commercial purposes. Multi-Scale Robotics Lab ETH Zurich The copyright line for this article was changed on 21 June 2021 after Tannenstrasse 3, Zurich CH-8092, Switzerland original online publication. Prof. J. Puigmartí-Luis ICREA DOI: 10.1002/adma.202101777 Pg. Lluís Companys 23, Barcelona 08010, Spain Adv. Mater. 2021, 33, 2101777 2101777 (1 of 9) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbH www.advancedsciencenews.com www.advmat.de frameworks (MOFs) are often synthesized using turbulent thicknesses (from 200 to 500 µm) to create a microfluidic envi- mixing,[1] and/or solvothermal methods,[2,3] among others.[4] ronment that is 6 cm long and 1.5 cm wide (see Figure 1a and In these synthetic conditions, the formation of MOF crys- Section S3 in the Supporting Information for further details). tals depletes the surrounding monomer solution, thus cre- One of the surfaces has two machined inlet ports that allow for ating buoyancy-driven convection, which generates uneven the injection and complete filling of the microfluidic environ- growth rates and the precipitation (sedimentation) of crystals ment preventing the appearance of air bubbles (Figure 1b,c). in solution. Precipitation can significantly affect the reaction We believe that the environment created between the two sub- performance, for example, by leading to a non-controlled and strates can simulate space experimentation where the synthesis non-uniform growth rate of MOF crystals in solution. Other of Ni3(HITP)2 can be accomplished in the absence of buoyancy- methods that are particularly suitable for the generation of driven convection. This belief is clearly supported by numerical functional 2D porous crystalline molecular frameworks (e.g., simulations of the transport phenomena inside the microfluidic 2D MOFs) are surface-templated growth methods[5,6] or inter- device (see Section S4 in the Supporting Information for fur- facial synthetic approaches (e.g., liquid/air or liquid/liquid ther details). Two scenarios were considered in our numerical interfaces).[7–9] Certainly, these methods can also suffer non- analyses: i) the conventionally used vapor-assisted conversion controlled convective transport processes (e.g., surface-tension- method,[15] where Ni3(HITP)2 is generated from a hypothetical driven convection (Marangoni convection)) and precipitation droplet of precursors when exposed to volatile alkaline gas and leading to fragmentation of the MOF films as well as to a poor atmospheric oxygen, and ii) the synthesis of Ni3(HITP)2 inside control over their orientation, compactness, roughness profile, our custom-made microfluidic device, where a controlled diffu- and/or morphology, features that greatly impact their function sion of the volatile reactive species in the static reaction media and performance.[7,8] can be easily achieved. The microfluidic device was assumed Recently, groundbreaking studies at the International Space to be filled with the dimethylsulfoxide (DMSO) solution of Station clearly confirmed the favorable effect of microgravity Ni3(HITP)2 precursors, and then the gases were allowed to dif- on the growth of crystalline matter.[10] It has been demonstrated fuse through the gas-liquid interface at the inlets. that the convection-free mass transport occurring in micro- Interestingly, we observed that in the droplet scenario, there gravity conditions can greatly favor the synthesis of materials are two counter-rotating vortices spanning most of the droplet with larger crystalline domain sizes, lower defect densities and diameter (see Figure 1d), where fluid elements move inward new morphologies.[11] However, the high costs and restricted along the droplet upper free surface (i.e., the gas-liquid inter- access to experimentation in space have hindered efforts to face) and outward near the substrate. In agreement with this, study and control the engineering of crystalline porous mole- there is an evident accumulation of the less dense solute in the cular frameworks under these synthetic conditions. Herein, we upper part of the droplet (Figure 1f), and a horizontal concen- show that microfluidic devices can simulate on Earth the effect tration front moving downward toward the substrate (i.e., in the of the microgravity in space experimentation. Then, we use this direction of gravity), see Figure 1f. In this scenario, the precur- microfluidic approach to control the synthesis and growth of a sors can mix and react in the upper part of the droplet, before prototypal conductive 2D MOF, i.e., Ni3(HITP)2 (where HITP the solute concentration front reaches the substrate. During this stands for 2, 3, 6, 7, 10, 11-hexaiminotriphenylene). We observe time, Ni3(HITP)2 powders may start to form in the upper part the formation of defect-free centimeter-sized thin films of of the droplet due to the oxidation reaction with atmospheric Ni3(HITP)2 that exhibit conductivities of up to 40 S cm–1 over oxygen, and the Ni3(HITP)2 precipitate will be pulled down by large distances at ambient conditions. Note that, thus far, similar gravity, adding to the already existing convective currents inside values outside of an inert atmosphere have only been achieved the droplet. In sharp contrast, and as shown in Figure 1e, the with pellets prepared under high pressures (≈1 GPa).[12] microfluidic device used in our investigations allows for the We chose Ni3(HITP)2 as an exemplar case for three main dramatic reduction of the convective mass transport effect reasons: i) Ni3(HITP)2 is a MOF whose growth has been diffi- during the reaction process. Indeed, much smaller/weaker cult to control as it quickly precipitates from the reaction media vortices are observed (and only near the device inlets), whose as black particles under alkaline media and an uncontrolled velocities (in the µm s−1 range) are ≈1000 times smaller (at atmospheric oxygen diffusion.[13] ii) Ni3(HITP)2 is a conductive 1 min) and ≈25 smaller (at 30 min) than those observed inside 2D MOF with extended π-conjugation, analogous to graphene, the droplet (see Figure 1d and Figure S5 (Supporting Informa- but whose conductivity properties have been derived mainly tion), respectively). In agreement with this, the concentration from pressed pellets and from polycrystalline Ni3(HITP)2 profile is also vastly different from that in the droplet scenario, powder but not in defect-free thin-films.[14] iii) Ni3(HITP)2 is with the microfluidic device generating a vertical concentration synthetized following a two-step vapor induced crystallization front moving horizontally along the substrate length (i.e., per- approach, first with volatile alkaline vapors and subsequently pendicular to the direction of gravity), see Figure 1g. The clearly with oxidant gases, that cannot be systematically controlled different synthetic conditions generated with our microfluidic employing common synthetic approaches.[7] approach enables the vapor-assisted synthesis of Ni3(HITP)2 in a reaction environment that guarantees the convection-free mass transport conditions achieved in space experimentation. 2. Results and Discussion Note that a microfluidic device as thick as the droplet also gen- erates horizontal concentration fronts inside reaction media, First, we engineered a microfluidic device that consists of two see Figures S6 and S7 (Supporting Information). Even though substrates sandwiched with a thin silicone film with variable other microfluidic devices have been used to synthesize MOF Adv. Mater. 2021, 33, 2101777 2101777 (2 of 9) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbH www.advancedsciencenews.com www.advmat.de Figure 1. a) Illustration showing an expanded view of the microfluidic device. b,c) Top and side view photographs of the microfluidic device, respec- tively. d,e) Velocity maps after 1 min of simulation time for the droplet and the microfluidic device, respectively. The images are zoomed in to allow the visualization of the flow details. Streamlines are also shown in the droplet velocity map to highlight the motion of the fluid elements. The velocity range in the droplet is ≈1000 times larger than that in the microfluidic device. f,g) Concentration maps after 30 min of simulation time for the droplet and the microfluidic device, respectively. The concentration maps are shown with ranges of 0.04–0.05 w/w and 0–0.05 w/w, respectively. h) Schematic illustration of the synthetic pathway followed to obtain Ni3(HITP)2. i) Photographs of the microfluidic environment at the different stages during the synthesis of Ni3(HITP)2. The initial colorless solution becomes pink upon exposure to ammonia vapors for 24 h, i.e., step 1. In step 2, the pink solution progressively becomes black from the inlets toward the center of the chamber upon exposure to air. After 72 h, a black thin film supported on quartz can be observed. crystals,[16–18] to date none of these microfluidic methods are growth of Ni3(HITP)2 by avoiding the movement of precipitates capable of generating vertical concentration fronts that move inside the microfluidic device. The microfluidic device was first horizontally along a substrate employing gases as reagents. filled with the solution of the reactants; next, the two-step vapor In order to experimentally confirm the controlled two-step induced crystallization approach of Ni3(HITP)2 (Figure 1h) was vapor induced crystallization of Ni3(HITP)2 inside our microflu- accomplished by controlling the diffusion of ammonia gas (step idic environment, we firstly assembled the microfluidic device 1) and air (step 2) inside the microfluidic device through its with two quartz slides as top and bottom walls (Figure 1a). This inlets (see the Supporting Information for further details). As configuration was preferred as it enables following the reac- shown in Figure 1i, the initially colorless solution composed by tion progress of Ni3(HITP)2 both optically and spectroscopi- a stoichiometric mixture of NiCl2 and HATP·6HCl turns pink cally. Anhydrous DMSO was chosen as the solvent of reaction. during step 1, i.e., when the microfluidic device is introduced in DMSO can solubilize all the reactants (NiCl2 and HATP·6HCl, a chamber full of ammonia vapor for 24 h. Subsequently, this where HATP stands for 2,3,6,7,10,11-hexaaminotriphenylene), pink solution turns black at the inlet ports in step 2, i.e., upon including intermediate species, thus preventing nucleation and exposure to air. After 72 h, the substrate is completely covered Adv. Mater. 2021, 33, 2101777 2101777 (3 of 9) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbH www.advancedsciencenews.com www.advmat.de Figure 2. a) Top-view SEM image of a Ni3(HITP)2@Quartz film. b) Cross-section SEM image of a Ni3(HITP)2@Quartz film. c) AFM image of a Ni3(HITP)2@Quartz film. d) 3D AFM image of the edge of a Ni3(HITP)2@Quartz film. e) 2D GIWAXS pattern of a Ni3(HITP)2@Quartz film. The inset shows a scheme of the orientation of the Ni3(HITP)2@Quartz film. f) Optical photograph of a Ni3(HITP)2@Au film. g) Top-view SEM image of a Ni3(HITP)2@Au film. h) Cross-section SEM image of a Ni3(HITP)2@Au film. i) 2D GIWAXS pattern of a Ni3(HITP)2@Au film. The inset shows a scheme of the orientation of the Ni3(HITP)2@Au film. j) Optical photograph of a Ni3(HITP)2@Au-SAM film. k) Top-view SEM image of Ni3(HITP)2@ Au-SAM film. l) Cross-section SEM image of a Ni3(HITP)2@Au-SAM film. m) 2D GIWAXS pattern of a Ni3(HITP)2@Au-SAM film. The inset shows a scheme of the orientation of the Ni3(HITP)2@Au-SAM film. with a black film that can readily be analyzed as Ni3(HITP)2, considerably right after the addition of the base. This rapid pro- see Figure 1i. The UV–vis spectra in Figure S11 (Supporting gress of the reaction in bulk, without the appearance of the two Information) clearly show these color changes. An intense distinct stages, showcases the lack of control over the oxygen absorption band at 1000 nm corresponding to the non-oxidized diffusion when conventional synthetic conditions are in place. complex between non-protonated HATP and the metal cation The black Ni3(HITP)2 films supported on quartz (here after is detected right after taking the device out of the ammonia referred to as Ni3(HITP)2@Quartz films) were initially charac- atmosphere. Then, when the device is exposed to air, the band terized by scanning electron microscopy (SEM), which revealed at 1000 nm starts to decrease while the characteristic band of a smooth, compact and crack-free continuous film with a Ni3(HITP)2 at 530 nm appears and increases as the film is homogeneous thickness of 100 nm (Figure 2a,b), that was formed.[19] In contrast to this controlled two-step vapor induced further confirmed by profilometry (Figure 2d). Atomic force crystallization of Ni3(HITP)2 inside the microfluidic device, per- microscopy (AFM) confirmed the low roughness of the film forming the same reaction in bulk conditions (e.g., in a vial) where a root mean square (RMS) roughness of 14.9 ± 0.5 nm results in a fast precipitation of black particles at the air/DMSO was measured (Figure 2c). The crystallinity and structure of interface characterized as Ni3(HITP)2 (see Section S5 in the the Ni3(HITP)2@Quartz films were assessed by 2D grazing- Supporting Information). In this case, monitoring the reaction incidence wide-angle X-ray scattering (GIWAXS). The reflec- by UV–vis spectroscopy (Figure S10, Supporting Information) tions corresponding to the Ni3(HITP)2 structure at qr = 3.4 nm–1 does not allow to identify the non-oxidized pink intermediate, (100), qr = 6.7 nm–1 (200), qr = 8.9 nm–1 (130), qr = 11.7 nm–1 since the band at 530 nm is present from the first spectrum (240), and qz ≈ 18 nm–1 (002) are clearly observed in Figure 2e, and the band at 1000 nm has already decreased its intensity where they appear with an anisotropic distribution.[19] That Adv. Mater. 2021, 33, 2101777 2101777 (4 of 9) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbH www.advancedsciencenews.com www.advmat.de is, plotting the intensities in an azimuthal representation some influence over the growth of Ni3(HITP)2 when a con- (Figure S12, Supporting Information) clearly shows that the vection-free mass transport condition is achieved. Note that (hk0) peaks have higher intensity in the in-plane direction, gold surfaces functionalized with chemical groups mimicking while the (002) peak is mainly localized in the out-of-plane those of linkers have been used to grow conductive 2D MOFs direction. These data suggest that the Ni3(HITP)2 structure is under convective mass transport processes and employing a preferentially oriented with its c-axis perpendicular to the sub- layer-by-layer approach.[20] Therefore, the amino groups of our strate. As shown in Figure S13 (Supporting Information), the SAM are expected to have strong directional interactions with layers of Ni3(HITP)2 extend through coordination bonds along the building blocks of Ni3(HITP)2, either by coordination to the a and b axes and are separated by a typical π–π stacking the nickel cations or through hydrogen bonds with the ligand. distance of 0.33 nm along the c-axis. Accordingly, and based Accordingly, a completely different microfluidic-based synthetic on the unit cell of Ni3(HITP)2, the GIWAXS data indicates environment was designed in these studies to perform the two- that the layers of the MOF in Ni3(HITP)2@Quartz films have step vapor induced crystallization of Ni3(HITP)2 films under grown parallel to the quartz surface. It is worth noting that if the convection-free mass transport condition achieved in our a quartz substrate is immersed in a vial containing the same microfluidic device when compared to the film grown onto bare reaction mixture (in the presence of buoyancy-driven convec- gold and quartz substrates. tion), Ni3(HITP)2 can be generated but the resulting coverage Even though the films grown on bare gold and on Au- of the substrate is not compact nor homogeneous. Addition- SAMs substrates looked similar to the naked eye (Figure 2f,j, ally, in this case, the Ni3(HITP)2 films produced do not have a respectively), the Ni3(HITP)2@Au-SAM films showed a very preferential orientation (Figure S14, Supporting Information). different morphology when observed by SEM. As shown in These results clearly highlight the importance of avoiding con- Figure 2k,l, the Ni3(HITP)2@Au-SAM films consist of densely vective mass transport during the synthesis of Ni3(HITP)2 thin packed nanoplatelets attached edge-on to the Au-SAM sub- films. Undoubtedly, with our microfluidic device, we allow the strate, while the thickness remains at ≈100 nm. This striking growth of Ni3(HITP)2 in a quiescent environment that ensures morphological change was also corroborated by AFM, which the generation of compact, smooth and crack-free Ni3(HITP)2 also allowed to assess the significant difference in roughness thin films. between Ni3(HITP)2@Au and Ni3(HITP)2@Au-SAM films Since Ni3(HITP)2 is a promising conductive material that (Figures S17 and S18: Supporting Information). From the AFM will have to be interfaced with electrodes in order to build func- images, the RMS of Ni3(HITP)2@Au films is 7.2 ± 0.5 nm while tional devices, we extended our synthetic microfluidic approach the RMS for Ni3(HITP)2@Au-SAM films is 24.1 ± 0.6 nm. More to gold substrates. The use of gold was twofold: i) On the one interestingly, 2D GIWAXS measurements showed that this hand, it provides a conductive surface to analyze the electronic morphological change arises from a change in the orientation properties of the as-prepared Ni3(HITP)2 thin films. On the of the Ni3(HITP)2 crystals generated. As seen in Figure 2m and other hand, ii) it facilitates the rational surface engineering of Figure S19 (Supporting Information), the (hk0) peaks acquired the microfluidic environment. Gold can be readily functional- from the 2D GIWAXS patterns of Ni3(HITP)2@Au-SAM films ized with thiols, forming self-assembled monolayers (SAMs), are more intense in the out-of-plane direction, which is in a feature that can be exploited to trigger a controlled orienta- agreement with a preferential orientation of the Ni3(HITP)2@ tion of Ni3(HITP)2 films during their growth in the simulated Au-SAM films with the c-axis parallel to the substrate. These microgravity conditions achieved with our microfluidic device. results are in sharp contrast to earlier observations where As shown in Figure 2f, assembling the microfluidic device Ni3(HITP)2@Au films were grown with a preferred orientation with two glass slides sputtered with gold as top and bottom walls of the c-axis perpendicular to the substrate. HR-TEM images and performing the synthesis of Ni3(HITP)2 as indicated above, of the material removed from Ni3(HITP)2@Au-SAM films yielded a large centimeter-scale Ni3(HITP)2 film, hereafter called reveal hexagonal patterns with periodicities of 2 nm, matching Ni3(HITP)2@Au film. The morphology of the Ni3(HITP)2@Au the pores of the Ni3(HITP)2 structure. Additionally, SAED pat- films is again compact, smooth and crack-free with a thickness of terns also show the (100) and (200) reflections of Ni3(HITP)2 ≈100 nm (Figure 2g,h). HR-TEM of the material removed from crystals as for Ni3(HITP)2@Au films (Figure S20, Supporting Ni3(HITP)2@Au films shows the high quality of the Ni3(HITP)2 Information). Altogether, these results show the versatility of crystals generated, in which the interlayer distance of 0.33 nm our approach for the preparation of compact and crack-free can be observed directly from the images, while selected area Ni3(HITP)2 thin films both in conductive and insulating sub- electron diffraction (SAED) patterns show peaks matching the strates, i.e., on quartz and/or gold substrates, respectively. Strik- (100) and (200) crystallographic planes (Figure S15, Supporting ingly, it was also demonstrated that the simulated microgravity Information). Finally, the 2D GIWAXS images are very similar conditions achieved with our microfluidic device can trigger to those reported for Ni3(HITP)2@Quartz films, with the (hk0) the growth of compact, smooth and crack-free Ni3(HITP)2 thin peaks displaying a higher intensity in the in-plane direction, indi- films with controlled preferred orientations, either perpen- cating that Ni3(HITP)2@Au films are also preferentially oriented dicular or parallel to the substrate, avoiding time-consuming, with the c-axis perpendicular to the substrate (Figure 2i and labor-intensive layer-by-layer approaches where fluid flow con- Figure S16: Supporting Information). ditions are not controlled. Next, we decided to functionalize the gold surface with an Since the Ni3(HITP)2 films have the same crystallographic amino-terminated SAM (see Section S7 in the Supporting orientation on quartz and gold, a microfluidic device assembled Information for further details) to study if the nature of the with quartz slides sputtered with gold electrodes was consid- substrate with the presence of anchoring groups can have ered as the best case scenario to study the electrical anisotropy Adv. Mater. 2021, 33, 2101777 2101777 (5 of 9) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbH www.advancedsciencenews.com www.advmat.de Figure 3. a) Ni3(HITP)2@Electrodes film (dark region in the center) grown on top of pre-patterned gold electrodes deposited by sputtering on a quartz substrate. b,c) Schematic views of the area highlighted in green in (A) showing the contact geometries employed to measure the electrical anisotropy of a Ni3(HITP)2@Electrodes film. b) Shows the contact used for in-plane conductivity, while (c) shows the gold contact thermally evaporated on a polyimide substrate and subsequentially pressed on top of the sample to measure the out-of-plane conductivity. d) Ni3(HITP)2@Electrodes film (dark region in the center) grown on top of pre-patterned gold electrodes deposited by sputtering on a quartz substrate, analogously to the scheme in (a). e) Lock-in AC V–I measurements of a Ni3(HITP)2@Electrodes film measured in-plane (red, geometry (i)) and vertical directions (blue, geometry (ii)). of the Ni3(HITP)2 films generated (Figure 3). Accordingly, A-B mode was used to efficiently filter DC contact voltages quartz substrates with pre-patterned gold electrodes were used (see Figure S23 (Supporting Information) for calibration of the to grow Ni3(HITP)2@Electrodes films with their c-axis per- method). As shown in Figure 3e, the V–I plots obtained for pendicular to the substrate (Figure 3a,d). This electrode con- both geometries are highly linear, which indicates good ohmic figuration, represented as geometry (i) in Figure 3b, allowed contacts[22] and negligible Fermi level pinning[23] along all direc- the direct characterization of the in-plane conductivity of tions of the Ni3(HITP)2@Electrodes films. Moreover, the phase Ni3(HITP)2@Electrodes films right after their generation during the measurements remains close to zero, showcasing inside our microfluidic device. However, the out-of-plane con- negligible capacitive coupling in both geometries (Figure S24, ductivity properties of Ni3(HITP)2@Electrodes films required Supporting Information). Linear fits of the V–I plots provided a second set of electrodes on top of the Ni3(HITP)2@Elec- high conductivities of 39.7 ± 0.2 S cm–1. Electrical conductivity trodes films. Taking into account that the thermal evaporation measurements at different temperatures indicate an in-plane or the sputtering of top electrode contacts onto the generated activation energy (Ea) of 64 ± 1 meV. A clear Arrhenius-type Ni3(HITP)2@Electrodes films will facilitate the intercalation of thermally activated transport with a very small thermal hyster- metal atoms into the film altering its electrical properties and esis was obtained for in-plane measurements under ambient eventually leading to electrical shortcuts, we conducted a dif- pressure (see Figures S26 and S27 in the Supporting Informa- ferent approach.[21] First, 50 nm of gold were thermally evapo- tion). Interestingly, such high conductivities close to 40 S cm–1 rated on polyimide substrates and these were mechanically in triphenylene-based MOFs are normally only obtained under transferred on top of the 100 nm-thick Ni3(HITP)2@Electrodes inert atmosphere or vacuum rather than in ambient condi- films, resulting in the contact geometry (ii) shown schemati- tions (see Table S2 in the Supporting Information), as water cally in Figure 3c. These geometries allowed for the charac- uptake diminishes the conductivity.[24] Furthermore, note that, terization of two relevant charge transport channels in the as- at ambient conditions, only pelletization at high pressures synthesized Ni3(HITP)2@Electrodes films. While geometry (i) (≈1 GPa)[12] has proved efficient to generate similar conductivi- allowed for the study of the in-plane charge transport (i.e., σ||), ties. Again, these results clearly showcase the good packing of i.e., along the Ni3(HITP)2@Electrodes films, geometry (ii) gives Ni3(HITP)2 obtained with our microfluidic approach. On the access to out-of-plane charge transport (σ┴), i.e., perpendicular other hand, the out-of-plane conductivity obtained from geom- to the Ni3(HITP)2@Electrodes films. Performing this out-of- etry (ii) is 0.75 ± 0.05 mS cm–1 with a corresponding activation plane measurement through the extremely thin Ni3(HITP)2@ energy of Ea = 121 ± 13 meV (Figure 3e and Figure S26: Sup- Electrodes films is challenging using standard direct current porting Information, respectively), which leads to a high anisot- (DC) measurement techniques because of the non-negligible ropy factor of more than 5 × 104. Similar high anisotropy factors contribution of contact resistances, so a Lock-in amplifier in have only been reported for graphite so far.[25–27] To ensure an Adv. Mater. 2021, 33, 2101777 2101777 (6 of 9) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbH www.advancedsciencenews.com www.advmat.de Figure 4. a,b) Illustrations showing expanded views of the microfluidic device. c) Top view photograph of the PEEK clamp (top clamp), showing the square sealing groove defining the confined reaction space. d) Side view photograph of the assembled microfluidic device, showing one of the ports for the connectors. e) Scheme of the structure of COF-TAPB-BTCA overlaid on a spacefill model of the structure. f–h) Optical photographs of COF-TAPB- BTCA films grown on 1 cm × 1 cm quartz (001), graphene and silicon (111) substrates, respectively. i) Side view SEM image of a thin film of COF-TAPB- BTCA grown on quartz (001). j) SEM image of COF-TAPB-BTCA synthesized onto a quartz surface immersed in the reaction mixture prepared in a vial. optimal contact and the quality of the data obtained with the microfluidic device consists of a piece with a cavity in which geometry (ii), we conducted additional out-of-plane V–I meas- the commercial surface is placed (see Figure 4a (top clamp) urements using polyimide substrates that incorporated gold and Figure 4c). Then, the bottom clamp is clamped pressing contacts with different areas. Remarkably, same values on the the surface against a 200 µm-thick sealing groove to create the vertical conductivity were obtained (0.78 ± 0.05 mS cm–1), thus microfluidic environment presented in Figure 1g (Figure 4a–d confirming the validity of our measurements (Figure S27, Sup- and Section S9 in the Supporting Information). Note that the porting Information). dimensions achieved with this new microfluidic device enable To assess the generality of our approach, we also investigated to keep the reduced convection benefits of the previous larger the effect of the simulated microgravity conditions during the version (see Section S4 in the Supporting Information). When synthesis of COF-TAPB-BTCA, a well-known porous and crys- carrying out the synthesis of COF-TAPB-BTCA inside the new talline 2D covalent organic framework (COF) resulting from microfluidic device and using a quartz surface, an extremely the Schiff-base reaction between 1,3,5-tris(4-aminophenyl)ben- smooth thin film was obtained 24 h after loading the reac- zene (TAPB) and 1,3,5-benzenetricarbaldehyde (BTCA), see tion mixture in the cell (Figure 4f). See for example the SEM Figure 4e.[28] In sharp contrast to Ni3(HITP)2, COF-TAPB-BTCA image of the COF-TAPB-BTCA film generated in Figure 4i or is now generated via covalent bonds instead of ionic interac- Figure S29 (Supporting Information). Further, we characterized tions. Briefly, we initially chose acetone as a solvent since it the morphology of the COF-TAPB-BTCA films by AFM (Figure S30, completely dissolves TAPB and BTCA, and further, it avoids the Supporting Information). The data collected shows that defect- fast precipitation of COF-TAPB-BTCA inside the microfluidic free centimeter-sized COF-TAPB-BTCA films as thin as 60 nm device. Accordingly, we designed a new solvent-resistant micro- and with an extremely low roughness (RMS = 2.4 nm) can fluidic device that can also accommodate a large variety of com- be prepared with our microfluidic device. Indeed, annealed mercially and technologically relevant surfaces with a standard films[29] show the characteristic GIWAXS pattern of COF- 1 cm × 1 cm size (e.g., silicon (111) and/or graphene). This new TAPB-BTCA with peaks at 4.1, 7.5, and 8.5 nm–1 corresponding Adv. Mater. 2021, 33, 2101777 2101777 (7 of 9) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbH www.advancedsciencenews.com www.advmat.de to the (100), (200), and (110) reflections, respectively. Addition- microCrysFact (ERC-2015-STG No. 677020), the Horizon 2020 FETOPEN ally, in Figure S31 (Supporting Information), it can be seen that project SPRINT (No. 801464), the Swiss National Science Foundation (project no. 200021_181988), grant MAT 2015-70615-R from the Spanish these peaks are in plane and that there is a broad band in the Government funds and by the European Regional Development Fund out-of-plane direction at q = 18 nm–1 corresponding to the (001) (ERDF). The ICN2 is funded by the CERCA programme/Generalitat reflection. This data indicates that the COF-TAPB-BTCA films de Catalunya. The ICN2 is supported by the Severo Ochoa Centres of obtained on the quartz slides and inside our microfluidic device Excellence programme, funded by the Spanish Research Agency (AEI, are preferentially oriented with their c-axis perpendicular to the grant no. SEV-2017-0706). N.C.-P. acknowledges support from “laCaixa” substrate. With this solvent-resistant microfluidic device, we Foundation (fellowship ID 100010434). The fellowship code is LCF/BQ/ also carried out the synthesis of COF-TAPB-BTCA onto com- ES17/11600012. N.C.-P. also acknowledges the financial support of COST action MP1407. R.P. acknowledges support from the Marie Curie Cofund, mercially available silicon (111) and single-layer graphene. The Beatriu de Pinós Fellowship AGAUR 2017 BP 00064. The GIWAXS COF-TAPB-BTCA films generated showed no differences to the experiments were conducted on the NCD-SWEET beamline of the ALBA COF-TAPB-BTCA films grown onto quartz (see Figure 4g,h and synchrotron, Spain and on the SAXS/WAXS beamline of the Australian Figures S32–S34: Supporting Information). Remarkably, COF- Synchrotron, ANSTO, Australia. GIWAXS experiments of Ni3(HITP)2 TAPB-BTCA samples grown on quartz substrates immersed and COF-TAPB-BTCA were performed at the NCD-SWEET beamline at inside a vial containing the same reaction mixture used for the ALBA Synchrotron with the collaboration of ALBA staff. The authors acknowledge the support of the ANSTO, in providing the facilities for microfluidic synthesis but where buoyancy-driven convection is the GIWAXS experiments used for the characterization of COF-TAPB- not suppressed, exhibited a completely different morphology. BTCA in this work. As shown in Figure 4j, mixing COF-TAPB-BTCA monomers inside a vial and allowing them to react for 72 h only yields spheres of several hundreds of nanometers. Note that this mor- phology is typical and well-known in the literature for this type Conflict of Interest of material when a poor control of the convective transport pro- The authors declare no conflict of interest. cesses dominates.[30] Therefore, these results clearly confirm that simulated microgravity conditions are not only ideal to control the orientation, compactness and crack-free generation of 2D porous crystalline molecular frameworks, but they can Data Availability Statement also change their crystal morphogenesis. The data that support the findings of this study are available from the corresponding author upon reasonable request. 3. Conclusions Keywords We have shown an easy way to achieve space-like experimen- tation conditions on Earth employing a custom-made micro- 2D porous crystalline materials, covalent organic frameworks, metal–organic frameworks, microfluidic technologies, simulated fluidic device for the fabrication of 2D crystalline molecular microgravity frameworks. The results presented throughout this work con- firm that experimentation under simulated microgravity con- Received: March 5, 2021 ditions has unprecedented effects on the controlled growth of Revised: April 10, 2021 2D crystalline molecular frameworks. Homogeneous and large Published online: June 4, 2021 thin films formed by either metal–organic or covalent bonds could be obtained in a wide variety of substrates with a high control on their crystalline orientation. We have proved that these simulated microgravity conditions can open new routes [1] A. R. Millward, O. M. Yaghi, J. Am. Chem. Soc. 2005, 127, 17998. to straightforwardly study properties such as the anisotropy [2] M. Latroche, S. Surblé, C. Serre, C. Mellot-Draznieks, P. L. Llewellyn, of conductivity in 2D conductive crystalline molecular frame- J.-H. Lee, J.-S. Chang, S. H. Jhung, G. Férey, Angew. Chem., Int. Ed. 2006, 45, 8227. works. Accordingly, we believe that this work will provide a new [3] M. Kandiah, M. H. Nilsen, S. Usseglio, S. Jakobsen, U. Olsbye, “playground” to chemists, physicists and materials scientists M. Tilset, C. Larabi, E. A. Quadrelli, F. Bonino, K. P. Lillerud, Chem. that want to generate unprecedented 2D functional materials Mater. 2010, 22, 6632. and devices. [4] S. H. Jhung, J.-H. Lee, J. W. Yoon, C. Serre, G. Férey, J.-S. Chang, Adv. Mater. 2007, 19, 121. [5] Z.-G. Gu, S.-C. Chen, W.-Q. Fu, Q. Zheng, J. Zhang, ACS Appl. Mater. Interfaces 2017, 9, 7259. Supporting Information [6] B. Sen, J. C. C. Santos, R. Haldar, Q. Zhang, T. Hashem, P. Qin, Supporting Information is available from the Wiley Online Library or Y. Li, F. Kirschhöfer, G. Brenner-Weiss, H. Gliemann, L. Heinke, from the author. C. Barner-Kowollik, A. Knebel, C. Wöll, Nanoscale 2020, 12, 24419. [7] K. Yuan, T. Song, X. Zhu, B. Li, B. Han, L. Zheng, J. Li, X. Zhang, W. Hu, Small 2019, 15, 1804845. [8] V. Rubio-Giménez, M. Galbiati, J. Castells-Gil, N. Almora-Barrios, Acknowledgements J. Navarro-Sánchez, G. Escorcia-Ariza, M. Mattera, T. Arnold, N.C.-P., D.R.-S.-M., C.F. contributed equally to this work. This work J. Rawle, S. Tatay, E. Coronado, C. Martí-Gastaldo, Adv. Mater. 2018, was supported by the European Research Council Starting Grant 30, 1704291. Adv. Mater. 2021, 33, 2101777 2101777 (8 of 9) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbH www.advancedsciencenews.com www.advmat.de [9] T. Kambe, R. Sakamoto, K. Hoshiko, K. Takada, M. Miyachi, [20] M.-S. Yao, X.-J. Lv, Z.-H. Fu, W.-H. Li, W.-H. Deng, G.-D. Wu, G. Xu, J.-H. Ryu, S. Sasaki, J. Kim, K. Nakazato, M. Takata, H. Nishihara, Angew. Chem., Int. Ed. 2017, 56, 16510. J. Am. Chem. Soc. 2013, 135, 2462. [21] Y. Liu, J. Guo, E. Zhu, L. Liao, S.-J. Lee, M. Ding, I. Shakir, [10] A. McPherson, L. J. DeLucas, npj Microgravity 2015, 1, 15010. V. Gambin, Y. Huang, X. Duan, Nature 2018, 557, 696. [11] R. L. Kroes, D. Reiss, J. Cryst. Growth 1984, 69, 414. [22] M. Farmanbar, G. Brocks, Adv. Electron. Mater. 2016, 2, [12] T. Chen, J.-H. Dou, L. Yang, C. Sun, N. J. Libretto, G. Skorupskii, 1500405. J. T. Miller, M. Dincă, J. Am. Chem. Soc. 2020, 142, 12367. [23] Y. Liu, P. Stradins, S.-H. Wei, Sci. Adv. 2016, 2, 1600069. [13] R. W. Day, D. K. Bediako, M. Rezaee, L. R. Parent, G. Skorupskii, [24] V. Rubio-Giménez, N. Almora-Barrios, G. Escorcia-Ariza, M. Q. Arguilla, C. H. Hendon, I. Stassen, N. C. Gianneschi, P. Kim, M. Galbiati, M. Sessolo, S. Tatay, C. Martí-Gastaldo, Angew. Chem., M. Dincă, ACS Cent. Sci. 2019, 5, 1959. Int. Ed. 2018, 57, 15086. [14] L. S. Xie, G. Skorupskii, M. Dincă, Chem. Rev. 2020, 120, 8536. [25] K. S. Krishnan, N. Ganguli, Nature 1939, 144, 667. [15] A. Mähringer, A. C. Jakowetz, J. M. Rotter, B. J. Bohn, J. K. Stolarczyk, [26] A. K. Dutta, Phys. Rev. 1953, 90, 187. J. Feldmann, T. Bein, D. D. Medina, ACS Nano 2019, 13, 6711. [27] O. J. Guentert, C. A. Klein, Appl. Phys. Lett. 1963, 2, 125. [16] C. Hu, Y. Bai, M. Hou, Y. Wang, L. Wang, X. Cao, C.-W. Chan, [28] A. de la Peña Ruigómez, D. Rodríguez-San-Miguel, H. Sun, W. Li, J. Ge, K. Ren, Sci. Adv. 2020, 6, eaax5785. K. C. Stylianou, M. Cavallini, D. Gentili, F. Liscio, S. Milita, [17] M. Faustini, J. Kim, G.-Y. Jeong, J. Y. Kim, H. R. Moon, W.-S. Ahn, O. M. Roscioni, M. L. Ruiz-González, C. Carbonell, D. Maspoch, D.-P. Kim, J. Am. Chem. Soc. 2013, 135, 14619. R. Mas-Ballesté, J. L. Segura, F. Zamora, Chem. – Eur. J. 2015, 21, [18] A. Sorrenti, L. Jones, S. Sevim, X. Cao, A. J. DeMello, 10666. C. Martí-Gastaldo, J. Puigmartí-Luis, J. Am. Chem. Soc. 2020, 142, [29] B. J. Smith, A. C. Overholts, N. Hwang, W. R. Dichtel, Chem. 9372. Commun. 2016, 52, 3690. [19] D. Sheberla, L. Sun, M. A. Blood-Forsythe, S. Er, C. R. Wade, [30] D. Rodríguez-San-Miguel, A. Yazdi, V. Guillerm, J. Pérez-Carvajal, C. K. Brozek, A. Aspuru-Guzik, M. Dincă, J. Am. Chem. Soc. 2014, V. Puntes, D. Maspoch, F. Zamora, Chem. – Eur. J. 2017, 23, 136, 8859. 8623. Adv. Mater. 2021, 33, 2101777 2101777 (9 of 9) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbH