Communication pubs.acs.org/JACS
Pore Breathing of Metal−Organic Frameworks by Environmental Transmission Electron Microscopy Lucas R. Parent,†,‡,∥,⊥,#,¶ C. Huy Pham,†,§,¶ Joseph P. Patterson,†,∇ Michael S. Denny, Jr.,† Seth M. Cohen,† Nathan C. Gianneschi,*,†,‡,∥,⊥,# and Francesco Paesani*,†,‡,§ †
Department of Chemistry and Biochemistry, ‡Materials Science and Engineering, §San Diego Supercomputer Center, University of California, San Diego, 9500 Gilman Drive, La Jolla, California 92093, United States ∥ Department of Chemistry, ⊥Department of Materials Science & Engineering, #Department of Biomedical Engineering, Northwestern University, Evanston, Illinois 60208, United States ∇ Labratory of Materials and Interface Chemistry and Center of Multiscale Electron Microscopy, Department of Chemical Engineering and Chemistry, Eindhoven University of Technology, 5612 AZ Eindhoven, The Netherlands S Supporting Information *
molecular switches, memory devices, chemical sensors, gas storage materials, and drug delivery systems. MIL-53(M), with M = Al, Sc, Cr, Fe, and Ga, is a prototypical family of flexible MOFs, exhibiting a large, yet reversible, structural transition from large pore (LP) to narrow pore (NP) configurations upon hydration.13 These transitions are accompanied by an anisotropic contraction of the unit cell volume up to 32%.13 Besides water, MIL-53(M) materials also breathe upon adsorption of other molecular guests, including CO2, CH4, and aromatic compounds.23−25 Breathing structural transitions in MOFs induced by guest species are signaled by characteristic (multi)step profiles and large hysteresis in the corresponding adsorption isotherms.19 Adsorption measurements are usually supported by analysis of powder X-ray diffraction (PXRD) patterns, a bulk-average technique, combined with computer simulations.16,26 However, direct observations of breathing in MOF nanocrystals are scarce, with no in situ single-particle studies of breathing behavior at the lattice level.16,26,27 In this study, we used in situ environmental transmission electron microscopy (ETEM) in combination with molecular dynamics (MD) simulations to directly monitor and discern the breathing behavior at the lattice-level of single MIL-53(Cr) nanocrystals during water adsorption/desorption, controlled through temperature ramping. MIL-53(Cr) nanocrystal powders with submicron particle diameters were synthesized according to standard literature conditions using anhydrous organic solvents.13 After washing the sample, the dried powder was characterized in bulk by PXRD (Figure S1) to verify the expected MIL-53(Cr) structure prior to ETEM experiments. Because MOF crystals are known to be beam-sensitive materials,28−30 low-dose TEM beam conditions (∼5 e−/Å2 cumulative dose) were used during in situ ETEM experiments to mitigate potential damage (see Supporting Information for details). Upon insertion into the ETEM column, the MIL-53(Cr) nanocrystals were exposed to ultrahigh vacuum, with the substrate at ambient temperature (27 °C). An isolated nanocrystal, generating a strong single-
ABSTRACT: Metal−organic frameworks (MOFs) have emerged as a versatile platform for the rational design of multifunctional materials, combining large specific surface areas with flexible, periodic frameworks that can undergo reversible structural transitions, or “breathing”, upon temperature and pressure changes, and through gas adsorption/desorption processes. Although MOF breathing can be inferred from the analysis of adsorption isotherms, direct observation of the structural transitions has been lacking, and the underlying processes of framework reorganization in individual MOF nanocrystals is largely unknown. In this study, we describe the characterization and elucidation of these processes through the combination of in situ environmental transmission electron microscopy (ETEM) and computer simulations. This combined approach enables the direct monitoring of the breathing behavior of individual MIL-53(Cr) nanocrystals upon reversible water adsorption and temperature changes. The ability to characterize structural changes in single nanocrystals and extract lattice level information through in silico correlation provides fundamental insights into the relationship between pore size/shape and host− guest interactions.
M
etal−organic frameworks (MOFs) are an attractive class of porous materials characterized by large surface areas and programmable structures, which have applications in a variety of settings.1−12 MOFs are three-dimensional networks in which metal ions or clusters are connected to multidentate ligands. Unlike nanoporous zeolites, several MOFs are characterized by flexible, periodic frameworks that can undergo reversible structural deformations, known as “breathing” effect, upon the application of external stimuli (e.g., guest adsorption, temperature and pressure changes, light irradiation, and exposure to electric fields).13−21 Furthermore, the breathing in MOFs can be modulated by chemical functionalization of the framework.22 The ability to engineer the breathing behavior of MOFs at the molecular level thus holds promise for the development of multifunctional materials for applications as © 2017 American Chemical Society
Received: June 26, 2017 Published: September 23, 2017 13973
DOI: 10.1021/jacs.7b06585 J. Am. Chem. Soc. 2017, 139, 13973−13976
Communication
Journal of the American Chemical Society
that previously determined by PXRD for MIL-53(Cr)-ht(“empty”).13 In the as-synthesized form, free terephthalic acid (H2bdc) molecules are expected in the pore channels,13 but under the ultrahigh vacuum of the microscope column and the initial e− beam irradiation, the crystal pores were largely emptied of residual H2bdc, solvent, or water molecules (confirmed by additional control TEM experiments in UHvacuum, Figure S6). This “emptying” transition could not be captured, as it likely occurred upon initial pump-down to UHvacuum in the ETEM column (prior to the first acquired diffraction pattern in Figure 1a). Using the ETEM gas manifold system, we introduced H2O vapor into the microscope column at ∼3 mbar, with directed flow onto the specimen. The spatial resolution was slightly degraded by background scattering caused by the gas molecules in the column, but we note that no changes in either the overall nanoparticle morphology (TEM image) or its crystal structure (diffraction pattern) were observed after ∼25 min under H2O vapor at 27 °C (Figure 1b). Following the previous report,13 the MIL-53(Cr) sample was rapidly heated to, and held at, 300 °C in situ while maintaining ∼3 mbar H2O vapor within the ETEM. An image and diffraction pattern of the nanoparticle was acquired after ∼15 min at this condition (Figure 1c). Comparing to Figure 1b, it is evident that the nanocrystal has experienced a dramatic structural transformation, reflected by the changes in its diffraction pattern, while maintaining overall integrity. A qualitative change in the overall particle dimensions is evident (the width has contracted and the height has extended, Table S2) in the TEM micrographs, with no detectable particle degradation. Despite the significant spot location shifts, the pattern remains single-crystalline, indicating that the nanostructure transition is consistent throughout the entire crystal domain. However, the cause of the shift is unclear from these ETEM data alone: either the lattice nanostructure has changed because of breathing, or possibly the lattice orientation has simply changed, or some combination of the two. Furthermore, without correlating with MD simulations, ETEM diffraction cannot determine the molecular adsorption, or other, processes, that gave rise to the nanostructure transitions. Seeking insight into the breathing behavior of MIL-53(Cr) at the lattice level, MD simulations were independently carried out at 27 and 300 °C for a variable number of water molecules (Nw) adsorbed per unit cell (Nw = 0−10, 12, 14, 16, 18, 20, 23−28, 30, 34, 40, 45, and 50) following a reported procedure,31,32 which generated a library of modeled lattice structures (see Supporting Information for details). The MD structures were used in STEM_CELL software to simulate electron diffraction patterns (Figure S5).33 As noted above, the [322] zone axis gave the best fit to the experimental ETEM patterns of the selected nanoparticle for an empty MIL-53(Cr)
crystalline diffraction pattern, was selected for the subsequent in situ breathing ETEM experiment (Figure 1 and Figures S2, S3, S4).
Figure 1. In situ ETEM images (top row) and diffraction patterns (middle row) of one MIL-53(Cr) nanocrystal at four different environmental conditions during breathing; (a) 27 °C and UHvacuum, (b) 27 °C and water vapor (“precalcination”), (c) 300 °C and water vapor (“calcinated”), and (d) 27 °C and water vapor (“postcalcination”). The bottom row illustrates the lengths (a, b, c) and angles (α, β, γ) in the ETEM diffraction patterns (Table 1).
By comparing the experimental diffraction pattern (Figure 1A and Table 1) to simulated electron diffraction patterns Table 1. Measured Lengths and Angles between the Diffraction Spots in the Four Experimental MIL-53(Cr) ETEM Diffraction Patterns in Figure 1a ETEM condition 27 °C vacuum 27 °C H2O vapor 300 °C H2O vapor 27 °C H2O vapor (cooled)
a (nm−1)
b (nm−1)
c (nm−1)
α (deg)
β (deg)
γ (deg)
1.66 1.66 1.82 1.68
1.72 1.72 1.79 1.67
1.16 1.16 1.67 1.05
67.3 67.3 63.2 72.2
72.4 72.4 62.1 71.2
40.4 40.4 54.6 36.3
Length and angle measurements have an uncertainty error of ±0.03 nm−1 and ±0.4 deg, respectively. a
generated by independent theoretical modeling of the MIL53(Cr) lattice (Figure S5, Table 2), we determined that the selected nanocrystal had the MIL-53(Cr)-ht-“empty” (ht = high-temperature in the original report)13 crystal structure on the [322] zone axis, with a mismatch error of ∼3% (Table S1). Our simulations of the MIL-53(Cr)-“empty” structure (27 °C, vacuum) found a lattice with the Imma space group, matching
Table 2. Lengths and Angles between the Diffraction Spots in the Three Simulated Diffraction Patterns Obtained from MD Simulations of MIL-53(Cr) on the [322] Zone Axis (Figure S5) at Different Environmental Conditions (Mimicking ETEM Conditions in Figure 1, Table 1), and the Number of Water Molecules Adsorbed Per Unit Cell and Respective Crystal Space Groups for Each Lattice Structure MD conditions
a (nm−1)
b (nm−1)
c (nm−1)
α (deg)
β (deg)
γ (deg)
# H2O per unit cell
space group
27 °C vacuum 27 °C H2O vapor 300 °C H2O vapor 27 °C H2O vapor
1.66 − 1.87 1.64
1.77 − 1.82 1.61
1.13 − 1.68 1.11
65.74 − 64.27 71.85
75.97 − 61.51 68.48
38.3 − 54.22 39.67
0 − 1 25
Imma − P21 P1
13974
DOI: 10.1021/jacs.7b06585 J. Am. Chem. Soc. 2017, 139, 13973−13976
Communication
Journal of the American Chemical Society
the initial step that activates the breathing effect, providing a means for including up to 25 H2 O molecules when subsequently cooled to 27 °C. As shown by the MD simulations, this activation step involves hydrogen bonding between the first adsorbed water molecules and the bridging μOH groups of the framework. By anchoring to the framework, these individual H2O molecules become effective hydrogen bonding sites that facilitate the adsorption of additional water molecules when the material is cooled, and they remain when other water molecules are removed from the pores during subsequent heating. By monitoring the lattice structure of a single MOF nanocrystal in situ and correlating the results with MD simulations, we precisely determined the extent of water adsorption/desorption during the breathing cycle. The combination of ETEM measurements and computer simulations is shown to provide molecular-level insights into waterframework interactions and their effects on overall crystal structures. We propose that the same integrated ETEM-MD approach is directly applicable for the study of other similar environmentally induced lattice transformations in MOFs or other mesoporous nanomaterials.
lattice, and this orientation was maintained for the simulated diffraction patterns’ parameters listed in Table 2 for variable amounts of water (unit cell parameters in Table S3). By comparing the simulated (Table 2) and experimental ETEM (Table 1) diffraction patterns for the 300 °C in H2O vapor condition (“calcinated” MIL-53(Cr)-cal), assuming no orientation change ([322] zone axis), we find an excellent agreement (
