Role of Solvent in the Oriented Growth of Conductive Ni‐CAT‐1 Metal‐Organic Framework at Solid–Liquid Interfaces

A controlled growth of two‐dimensional (2D) π‐conjugated metal‐organic frameworks (MOFs) on solid substrates can open exciting opportunities for the application of 2D MOFs as optoelectronic devices. Some factors like solvent composition and type of substrates are known to influence the properties of solution‐processed 2D MOF crystals; however, a mechanistic understanding of how interactions between solvent, substrate, and precursors affect heterogeneous nucleation has been limited. Here, it is reported that the structure of Ni‐catecholate (Ni‐CAT‐1) MOFs at a solid–liquid interface is controlled by solvent–substrate and solvent–MOF precursor interactions. Specifically, the structure of the MOF film can be controlled by varying the affinity of the solvent to the substrate. As a fraction of N,N‐dimethylformamide (DMF) in a binary solvent mixture of water and DMF increases, the arrangement of Ni‐CAT‐1 crystals varies from vertically aligned nanorods to the graphite substrate to less ordered nanorods with the lower initial nucleation number density of Ni‐CAT‐1 crystals on the surface.


Introduction
Two-dimensional (2D) -conjugated metal-organic frameworks (MOFs) [1] have attracted considerable attention as active materials for many applications including sensors, [2] energy storage, [3] electrocatalysis, [4] and spintronics [5] owing to their unique properties such as intrinsic porosity, [6] chemical and structural flexibility, [7] and high conductivity. [8]The integration of 2D -conjugated MOFs into devices [9] often requires a controlled growth of MOFs on solid substrates to precisely tune the performance of the devices, [10] which depends on the crystallinity, [11] orientation, [12] defect, [13] and heterostructure [14] of the MOF films.A comprehensive understanding of the physical principles underlying crystallization at the solid-liquid interface can give us insight into a controlled growth of 2D -conjugated MOFs on substrate in solution for achieving the desired performance of the devices.Unlike homogeneous nucleation, heterogeneous nucleation can be significantly influenced by threefold interfacial interactions of solvent-substrate, solvent-MOF precursors, and substrate-MOF precursors.For example, the nature of interfacial interactions between substrates and ligand molecules of MOFs directs how the interfacial layer assembles controlling subsequent MOF growth. [15]Although the role of solvent in homogeneous nucleation including the coordination between MOF precursor and solvent has been intensively investigated [16] and some reports have shown that solvent polarity [17] and composition [12a,18] can affect the arrangement of MOF crystals on substrates, the fundamental mechanism of a critical role of solvent in heterogeneous nucleation is still poorly understood.
In this work, we elucidate the role of solvent in heterogeneous nucleation and growth of 2D Ni-catecholate (Ni-CAT-1) at the solid-liquid interface.As a fraction of N,N-dimethylformamide (DMF) in a binary solvent mixture of water and DMF increases, the arrangement of Ni-CAT-1 crystals varies from wellaligned nanorods orthogonal to the solid substrate to less ordered nanorods with the lower initial nucleation number density of Ni-CAT-1 crystals on the surface.These observations can be attributed to interactions of solvent-solid substrate and solvent-MOF precursors.Our results provide insights into the role of solvent in the thermodynamics of heterogeneous nucleation of 2D -conjugated MOFs on solid substrates, which is essential for solution-processed crystal growth toward optoelectronic device applications.

Results and Discussion
The role of solvent affinity to the substrate in directing heterogeneous nucleation and growth of Ni-CAT-1 MOF was studied by systematically changing the composition of the mixed water/DMF solvent.In these experiments, Ni salt and 2,3,6,7,10,11-Hexahydroxytriphenylene (HHTP) were dissolved in a solvent in the presence of highly oriented pyrolytic graphite (HOPG) and the mixture was heated to 85˚C leading to direct growth of Ni-CAT-1 on HOPG.12b,15] When the solvent was changed to a binary mixture of water and DMF, the orientation of Ni-CAT-1 nanorods on HOPG was altered whereas no significant changes in the morphology of Ni-CAT-1 crystals were found under all reaction conditions.The increase in DMF content from 10 vol% to 20 vol% in a binary solvent introduced the disorder in nanorod orientation.The average deviation from the vertical orientation of Ni-CAT-1 nanorods with respect to HOPG also increased with the increase in DMF concentration (Figure 1b,c).
The orientation of the Ni-CAT-1 nanorods on HOPG was further investigated by using X-ray diffraction (XRD) analysis (Figure 2a).The out-of-plane scan of Ni-CAT-1 on HOPG synthesized in pure water exhibited a very small peak corresponding to the (100) plane and a sharp and intense peak for the (002) plane, indicating that (00l) planes are aligned parallel to the HOPG surface. [15]As the fraction of DMF in the binary solvent increased, the relative intensity of two characteristic peaks corresponding to (100) and (002) planes (I (100) /I (002) ) increased, implying that preferential alignment of (00l) planes on the HOPG surface weakens in the presence of DMF. [17]These observations can be explained by the disruption of water structure at the HOPG interface in the presence of DMF.In pure water, the water molecules form a hydrophobic, double-layer structure on a graphitic surface. [19]Weak interfacial binding between water and surface facilitates epitaxial nucleation of (00l) plane of Ni-CAT-1 through - interaction between the triphenylene linker molecules with the HOPG surface, followed by the formation of vertically aligned Ni-CAT-1 nanorods on HOPG.By contrast, in a binary solvent mixture of water and DMF, the double-layer structure of water at the HOPG interface could be disrupted due to stronger interactions of DMF molecules with the graphite surface.Density functional theory simulations predict that the difference in adsorption energies of DMF and water molecules is about 0.1 eV owing to strong interactions between the nitrogen of DMF and the surface carbon atom (Figure 2b,c).Furthermore, the adsorption of DMF reduces the number of available sites for HHTP adsorption onto the HOPG surface and disrupts formation of the epitaxial layer.We confirmed that Ni-CAT-1 nanorods do not grow on HOPG surface in pure DMF (Figure S1, Supporting Information).The result supports our hypothesis that DMF interferes with epitaxial growth of Ni-CAT-1 MOFs on HOPG.
To gain insight into the role of solvent in the crystal orientation, we conducted a quantitative analysis of the orientation of Ni-CAT-1 crystals on the graphitic surface at the early nucleation stage by high-resolution transmission electron microscopy (HRTEM).The HRTEM image of Ni-CAT-1 crystals synthesized in pure water (Figure 3a) clearly showed crystal domains with a hexagonal arrangement of pores and the spacing d (100) of 1.84 nm indicating that (00l) basal planes of Ni-CAT-1 are parallel to the graphitic surface.The fast-Fourier transform (FFT) pattern (Figure 3a) obtained from the entire region of the HRTEM image exhibited a scattered diffraction ring showing that the crystal domains are not aligned with each other on the HOPG surface.The orientation distribution of the (100) plane of Ni-CAT-1 crystals on the HOPG surface was quantified by calculating the signal intensity of each point around the circumference of the diffraction ring on the FFT image.The relationship of diffraction intensity versus the angle with the horizontal line produced a profile with multiple peaks (17.9˚apart) in Figure 3a, indicating that (100) planes are misaligned by 17.9˚with each other on the graphitic surface.Similarly, the HRTEM and FFT images of Ni-CAT-1 crystals synthesized in a binary solvent with 10% DMF (Figure 3b) showed a honeycomb arrangement of pores with d (100) = 1.84 nm and d (110) = 0.98 nm indicating that (00l) planes of Ni-CAT-1 are parallel to the graphitic surface like those in pure water.The diffraction intensity plot exhibited line graphs with peaks (17.7˚apart) for the (100) plane (Figure 3b) suggesting that (100) planes are misaligned by 17.7˚with each other on the graphitic surface.By contrast, the hexagonal array of pores was not observed in the HRTEM image of Ni-CAT-1 crystals synthesized in a binary solvent with 20% DMF (Figure 3c).Instead, the FFT image (Figure 3c) showed a continuous diffraction ring with d (004) = 0.34 nm from Ni-CAT-1 implying (00l) planes of some crystal domains are perpendicular to the graphitic surface and these domains do not have a specific orientation with respect to the graphitic surface (Figure 3c).A scattered diffraction ring with d (100) = 0.21 nm came from the graphitic layer.
We note that DMF does not participate in the coordination with the Ni node in the Ni-CAT-1 structure under our experimental condition.In pure water, the Ni node coordinates HHTP ligands and water molecules to complete octahedral coordination in the Ni-CAT-1 structure. [20]Unlike the reported Fe-HHTP structure that was synthesized in a binary solvent of DMF (50%) and water and DMF molecules are coordinated with octahedral Fe nodes, [21] no DMF is incorporated in Ni-CAT-1 nanorods synthesized in the binary solvents due to the absence of characteristic peaks of DMF by the Fourier-transform infrared spectra (Figure S2a, Supporting Information).The lack of DMF peaks in the spectra suggests that the Ni node does not coordinate DMF molecules at least in the final crystal structure.Moreover, identical powder XRD patterns of three Ni-CAT-1 powders in Figure S2b in Supporting Information indicate that structural features such as pore size (1.88 nm), interlayer spacing (0.32 nm), and the AB stacking orientation of Ni-CAT-1 crystals synthesized in binary solvents are consistent with those synthesized in pure water.In contrast to the Ni-based 2D MOFs in which the Ni node can adopt octahedral coordination in pure water and square-planar coordination in a binary solvent of water and 20% DMF, [22] the Ni node in the Ni-CAT-1 structure under our experimental condition appears to form octahedral coordination both in pure water and in binary solvents of DMF and water.The porosity of Ni-CAT-1 powders synthesized in binary solvents is comparable to those synthesized in pure water (Figure S2c, Supporting Information).When the fraction of DMF in the binary solvent is higher than 50%, however, no characteristic XRD peaks of Ni-CAT-1 were observed showing that Ni-CAT-1 crystallization was significantly inhibited by DMF (Figure S2b, Supporting Information).
To investigate the effect of solvent on heterogeneous nucleation of the Ni-CAT-1, the early growth stages of Ni-CAT-1 crystals formed in these solvents were evaluated by atomic force microscopy.As shown in Figure 4a, the Ni-CAT-1 crystals were densely packed on HOPG within 1 minute of reaction time when synthesized in pure water.As the fraction of DMF in the binary solvent increased, the number density of Ni-CAT-1 crystals decreased (Figure 4b,c).As shown in Figure 4d, the crystal number density on HOPG was decreased from 3885.3 ± 226.7 μm −2 (pure water) to 2994.6 ± 176 μm −2 (10% DMF in binary solvent) and 2064 ± 192 μm −2 (20% DMF in binary solvent).The decrease in the number density of Ni-CAT-1 crystals in the presence of DMF can be explained by three aspects.First, given that the solubility of HHTP in DMF is higher than that in pure water, [23] the addition of DMF will lower the supersaturation of MOF leading to a reduced nucleation rate of Ni-CAT-1.Second, the association between HHTP and Ni ions for Ni-CAT-1 nucleation be- comes more kinetically unfavorable in the presence of DMF due to the stronger DMF-HHTP interaction than water-HHTP interaction, which results in more stable DMF-HHTP complexes. [24]hird, the binding affinity of solvent with graphite surface is higher for DMF than for water with the adsorption energies equal to −0.26 eV and −0.156 eV, respectively (Figure 2b,c).The higher binding energy between DMF and graphite leads to a higher partition of DMF on the HOPG surface inhibiting the nucleation of Ni-CAT-1 in these DMF-rich regions.Therefore, the total nucleation number density of Ni-CAT-1 was decreased in the presence of DMF.

Conclusion
In summary, we have shown that the architecture of Ni-CAT-1 MOFs at a solid-liquid interface is controlled by solventsubstrate and solvent-HHTP interactions.These interactions dictate the orientation and nucleation number density of Ni-CAT-1 crystals on the HOPG substrate.As a fraction of DMF in a binary solvent mixture of water and DMF increases, the arrangement of Ni-CAT-1 crystals changes from well-aligned nanorods perpendicular to the HOPG substrate to less ordered nanorods due to a decrease in - interaction between HHTP and the HOPG surface in the presence of DMF.Strong DMF-HOPG and DMF-HHTP interactions hinder heterogeneous nucleation of Ni-CAT-1, leading to a decrease in the nucleation number density of Ni-CAT-1 crystals on HOPG.Our results provide insights into the role of solvents in directing the thermodynamics and kinetics of heterogeneous nucleation of 2D -conjugated MOFs on solid substrates, which is essential for solution-processed crystal growth for optoelectronic applications.

Figure 1 .
Figure 1.Top and tilted view SEM images of Ni-CAT-1 nanorods on HOPG synthesized in different solvent compositions.a) Highly oriented nanorods in pure water.b) Vertically protruded nanorods with different tilt angles formed in binary solvent of water and 10% DMF.c) More disorganized nanorods in the binary solvent of water and 20% DMF.

Figure 2 .
Figure 2. a) The out-of-plane XRD patterns and relative peak intensities, I (100) /I (002) , of Ni-CAT-1 on HOPG synthesized in pure water, binary solvent of water and 10% DMF, and binary solvent of water and 20% DMF.Simulated b) DMF and c) water molecule on graphite surface in side view (left) and top view (right) and calculated adsorption energy.C atoms are shown in brown, O in red, N in gray, and H in light gray.The unit cell is shown as a black parallelogram.

Figure 3 .
Figure 3. Structure of Ni-CAT-1 crystals synthesized in different solvent compositions on graphitic surface at the early growth stage.HRTEM image (left), FFT pattern (middle), and diffraction intensity profile (right) of Ni-CAT-1 synthesized in a) pure water (Inset: HRTEM image showing hexagonal arrangement of pores), b) binary solvent of water and 10% DMF, and c) binary solvent of water and 20% DMF.

Figure 4 .
Figure 4. Atomic force microscopy images of Ni-CAT-1 crystals on HOPG at the early growth stage synthesized in a) pure water, b) binary solvent of water and 10% DMF, and c) binary solvent of water and 20% DMF.d) The number density of Ni-CAT-1 crystals on HOPG as a function of DMF fraction in solvent measured from images in (a-c).