High‐Throughput Electron Diffraction Reveals a Hidden Novel Metal–Organic Framework for Electrocatalysis

Abstract Metal‐organic frameworks (MOFs) are known for their versatile combination of inorganic building units and organic linkers, which offers immense opportunities in a wide range of applications. However, many MOFs are typically synthesized as multiphasic polycrystalline powders, which are challenging for studies by X‐ray diffraction. Therefore, developing new structural characterization techniques is highly desired in order to accelerate discoveries of new materials. Here, we report a high‐throughput approach for structural analysis of MOF nano‐ and sub‐microcrystals by three‐dimensional electron diffraction (3DED). A new zeolitic‐imidazolate framework (ZIF), denoted ZIF‐EC1, was first discovered in a trace amount during the study of a known ZIF‐CO3‐1 material by 3DED. The structures of both ZIFs were solved and refined using 3DED data. ZIF‐EC1 has a dense 3D framework structure, which is built by linking mono‐ and bi‐nuclear Zn clusters and 2‐methylimidazolates (mIm−). With a composition of Zn3(mIm)5(OH), ZIF‐EC1 exhibits high N and Zn densities. We show that the N‐doped carbon material derived from ZIF‐EC1 is a promising electrocatalyst for oxygen reduction reaction (ORR). The discovery of this new MOF and its conversion to an efficient electrocatalyst highlights the power of 3DED in developing new materials and their applications.

Electrochemical evaluation. All electrochemical tests were carried out on CHI 660E electrochemical workstation at 30 °C. Oxygen reduction reaction (ORR) performances were evaluated by a three-electrode system. The working electrode was a rotating disk electrode (RDE) with a diameter of 5 mm (0.196 cm 2 ). The counter electrode was a graphite rod and the reference electrode was a saturated Ag/AgCl electrode. The cyclic voltammetry (CV) test was carried out in 0.1 M KOH saturated with O2 (scan rate is 50 mV s −1 ). The linear sweep voltammetry (LSV) test was carried out in 0.1 M KOH saturated with O2, and the speed range was 400 to 2025 rpm (scan rate was 5 mV s −1 ). Current-time chronoamperometric responses were measured at 0.66 V (relative to RHE). The catalyst ink for ORR tests was prepared as the following: 4 mg catalyst was dispersed in a mixed solution containing 660 μL isopropanol, 330 μL deionized water and 10 μL Nafion aqueous solution (5 wt%). The mixed solution was treated by ultrasound for 1 h to form a homogeneous suspension. Then 20 µL catalyst solution was taken from the pipette and uniformly dripped onto the rotating disc electrode (RDE) (catalyst loading was about 0.4 mg cm −2 ). After natural drying, a homogeneous film was formed.

Section 2. Synthesis of ZIF-EC1
Synthesis of ZIF-EC1 and ZIF-CO3-1 mixture. In a typical synthesis, 300 μL of a 440 mM aqueous solution of HmIM, 60 μL of a 36 mg mL -1 aqueous solution of BSA and 1140 μL of deionized water were mixed in a 2 ml plastic centrifuge tube for 1 minute. Then, the solution was added to 500 μL of an 80 mM aqueous solution of Zn(OAc)2•2(H2O). The mixture was left under S3 static conditions at RT for 24 h. White powder was then harvested by centrifugation followed by washing by deionized water for at least six times.
Synthesis of pure ZIF-EC1. In a typical synthesis of pure ZIF-EC1, 0.125 mL of a 3.84 M aqueous solution of HmIM was mixed with 1.875 mL of deionized water. 1 mL of 0.24 M aqueous solution of Zn(OAc)2•2(H2O) was added into the above solution under vigorous stirring condition. The mixture was kept under stirring at RT for at least 4 h. White powder was then harvested by centrifugation followed by washing by deionized water for at least six times.
Synthesis of ZIF-8. In a typical synthesis of pure ZIF-8, 100 mg HmIm was dissolved in 10 mL of methanol to form solution A. 140 mg Zn(NO3)2·6H2O was dissolved in another 10 mL methanol to form solution B. Then, solution B was poured into solution A under magnetic stirring for 10 min and aged for 24 h at room temperature. The product was collected by centrifugation and washed with methanol. It was then dried overnight in a freeze dryer for characterization.

Synthesis of ZIF-1.
In a typical synthesis of pure ZIF-1, 100 mg Zn(NO3)2·6H2O and 150 mg HIm are dissolved in a glass bottle containing 18 mL DMF. After vigorous stirring, the vial was capped and heated in a thermostat at 85°C for 24 h. After the reaction is completed, it is naturally cooled to room temperature. The product was collected by centrifugation and washed with dichloromethane. It was then dried overnight in a freeze dryer for characterization.

Synthesis of ZIF-95.
In a typical synthesis of pure ZIF-95, 237.5 mg Zn(NO3)2·6H2O and 1221 mg HcbIm were added to a beaker containing 55 ml DMF and 5 ml water. After stirring for 2 h, the mixed solution was added to a stainless steel autoclave lined with Teflon, and then heated in an air circulating oven at 120 °C for 3 days. The product was collected by centrifugation, washed with methanol and DMF, and then dried overnight in a freeze dryer for characterization.
Synthesis of N doped carbon from ZIF materials. The as-prepared ZIF-EC1, ZIF-1, ZIF-8 and ZIF-95 materials were annealed at 900 °C for 2 h at a ramp rate of 5 °C min −1 in flowing Ar in a tube furnace. Finally, the N doped carbon electrocatalysts were obtained.

Section 3. Structural analysis by cRED
The samples for high throughput cRED investigations were crushed in a mortar and dispersed in deionized water. A droplet was then taken from the suspension, transferred to a copper grid covered with lacey carbon, and dried in air at room temperature. cRED data were collected on a JEOL JEM2100 microscope operated at 200 kV (Cs 1.0 mm, point resolution 0.23 nm). TEM images were recorded with a Gatan Orius 833 CCD camera (resolution 2048 x 2048 pixels, pixel size 7.4 μm). cRED data collection was controlled by using the data acquisition software Instamatic [1] , and electron diffraction (ED) frames were recorded with a Timepix hybrid detector QTPX-262k (512 x 512 pixels, pixel size 55 μm, max 120 frames/second, Amsterdam Sci. Ins.). A single-tilt holder was used for the data collection, which could tilt from -60° to +60° in the TEM. The area used for cRED data collection was about 1.0 μm in diameter, defined using a selectedarea aperture. To minimize electron beam damage on the crystals, a low electron dose and high rotation speed were applied. The high throughput cRED method can benefit for virtualization and identification of each individual nanocrystals. As shown in Figure 1b, ~30 nanocrystals can be analyzed in an area of 35 ×35 μm 2 . By taking advantage of continuous rotation, it takes less than 5 minutes to collect a complete cRED dataset. For ZIF-CO3-1, a typical cRED dataset covered a crystal rotation angle of 100.2° and took 3.7 min to collect. For ZIF-EC1, the rotation range was 117.5° and the collection time was 4.3 min.
The obtained cRED data were analyzed by using REDprocessing software package [2] . Two sets of unit cells and space groups were determined from 11 nanocrystals in phase mixture, indicating it contains two different structures. For the major phase, one typical unit cell was determined as a = 10.57 Å, b = 12.40 Å, c = 4.65 Å, α = 90.8°, β = 90.9°, and γ = 91.6°. The intensity distribution of reflections in the 3D reciprocal lattice indicates that the crystal is orthorhombic with a Laue class of mmm ( Figures S2a-d). The unit cell angles α, β, and γ are near 90°, which also confirms the orthorhombic crystal system. The reflection conditions were deduced from the 2D slice cuts as 0kl: k = 2n; h0l: h = 2n; h00, h = 2n; 0k0: k = 2n, which corresponds to two possible space groups of Pba2 (No. 32), and Pbam (No. 55). For the minor phase, the unit cell parameters were determined to be a = 13.65 Å, b = 14.36 Å, c = 14.30 Å, α = 90.3°, β = 117.5°, and γ = 90.6°. The intensity distribution of reflections in the 3D reciprocal lattice indicates that the crystal is monoclinic with a Laue class of 2/m ( Figures S2e-f). The reflection conditions was deduced as 0k0: k=2n; 00l: l=2n; h0l: l=2n, which corresponds to the space group: P21/c (No. 14). The details of data collection and unit cell determination are summarized in Table S2.
After changing the synthetic conditions, we applied the high throughput cRED method on the product obtained using the optimized synthesis condition for ZIF-EC1. In total, nine datasets were collected from different crystals, and all of them gave the structure of ZIF-EC1. The sample purity was further confirmed by Pawley fitting of the PXRD pattern ( Figure S6). With improved crystallinity, cRED dataset with higher resolution and higher completeness were obtained ( Figure  S7). The intensities of the reflections were extracted from the cRED data using the X-ray Detector Software (XDS) [3] . To obtain a higher data completeness, nine cRED datasets were merged to a resolution of 0.78 Å and a completeness of 89.5%. Structure solution and refinement were conducted on each of the nine cRED datasets as well as on the merged dataset by using the SHELX software package 5 . In a typical refinement of the merged dataset, distance (DFIX) and planarity (FLAT) restraints were applied to the 2-methylimidazolate linkers to maintain a reasonable geometry. EADP was applied on two carbon atoms on the imidazolate group. In addition, EXTI was used in the final refinement, which converged with the agreement values R1=0.1811 for 4103 Fo > 4σ(Fo) and 0.1984 for all 5116 data for 302 parameters. The refinement results show a good agreement among the different nanocrystals (Table S3).

Section 4. Energy calculation of ZIF-EC1
Atomic Simulation Environment (ASE) [4] was used to handle the simulation and the QUANTUM ESPRESSO program package [5] to perform electronic structure calculations. The electronic wavefunctions were expanded in plane waves up to a cutoff energy of 800 eV, while the electron density is represented on a grid with an energy cutoff of 8000 eV after carrying out the convergence tests. Core electrons were approximated using ultrasoft pseudopotentials [6] . The ground state energies of ZIF-EC1, ZIF-CO3-1 and ZIF-8 bulk structures were calculated using PBE exchangecorrelation functional with dispersion correction [7] . The Brillouin zone were converged and sampled with (3  3  3) Monkhorst-Pack k-points.

Section 5. Electrochemical analysis of ZIF-EC1 and its derivatives
The potentials corresponding to the reversible hydrogen electrode (RHE) electrode were calculated with the following equation: Eq (1) The electron transfer numbers (n) were calculated with Koutecky-Levich (K-L) equation: where is the measured current density; is the diffusion current density; is the kinetic current density; is the rotation speed in rpm; can be confirmed by Koutecky-Levich (K-L) equation: where n is the transfer number; F is the The LSV curve of ZIF-EC1 in 0.1 M KOH solution saturated by O2 is shown in Figure  S9. The E1/2 of ZIF-EC1 was measured to be 0.757 V, indicating a moderate ORR activity. The high density of surface Zn-Nx structure could attribute to the good ORR activity. However, MOF materials are limited by their conductivities when being used as electrocatalysts for ORR. ZIF-EC1 possesses a high density of activity sites. NC-ZIF-EC1 derived from ZIF-EC1 has the highest N loading compared to those of NC-ZIF-1, NC-ZIF-8, and NC-ZIF-95 (Figures S15, S16 and Table S5). The ID/IG of NC-ZIF-EC1, NC-ZIF-1, NC-ZIF-8, and NC-ZIF-95 is 1.08, 1.35, 1.04, 1.08, respectively ( Figure S17). PXRD shows that all samples exhibit two similar broad peaks assigned to partial graphitized carbon. The results shown in Figure 5 indicates that NC-ZIF-EC1 exhibits the best ORR activity among all carbon based materials derived from ZIFs, which is also compatible to that of commercial Pt/C. A Tafel slope of 96.9 mV dec −1 is observed for NC-ZIF-EC1, which is considerably lower than that of NC-ZIF-1 (135.5 mV dec −1 ), NC-ZIF-8 (103.4 mV dec −1 ), and NC-ZIF-95 (215.8 mV dec −1 ) (Figure 5c). This indicates a favorable reaction kinetics for NC-ZIF-EC1 compared to NC-ZIF-1, NC-ZIF-8, and NC-ZIF-95.   Figure S4. N2 adsorption-desorption isotherm of ZIF-EC1. ZIF-EC1 is nonporous to N2. The porosity showed in the P/Po range > 0.9 is caused by interparticle voids. The BET surface area is 19.7 m 2 g -1 . Figure S5. Comparison of the observed PXRD pattern of the phase mixture with the simulated PXRD patterns of ZIF-CO3-1 and ZIF-EC1. All peaks in the observed pattern can be indexed using the two phases. The PXRD patterns were simulated from the corresponding structural models using a pseudo-Voigt peak shape function. Figure S6. Pawley fit of the experimental PXRD pattern (λ = 1.5406 Å) of ZIF-EC1, which shows a good agreement indicating the sample is pure ZIF-EC1. Figure S7. Reciprocal lattices reconstructed from cRED data. (a-d) 2D slice cuts of (a) hk0, (b) 0kl, and (c) h0l planes from the reconstructed 3D reciprocal lattice (d) of a ZIF-EC1 crystal (inset in d) in the phase mixture. (e-h) 2D slice cuts of (e) hk0, (f) 0kl, and (g) h0l planes from the reconstructed 3D reciprocal lattice (h) of a ZIF-EC1 crystal (inset in h) in the pure ZIF-EC1 sample. S10 Figure S8. Comparison of observed and simulated PXRD patterns of ZIF-EC1 ( = 1.5406 Å). The PXRD pattern were simulated using a pseudo-Voigt peak shape function, with the preferred orientation simulated using a weighted March-type correction.      Table S5.    Table S1. Details of cRED data collection on different nanocrystals in the phase mixture, and the corresponding unit cell parameters and space groups deduced from the cRED data.