WorksPhysics and Astronomy
Reticulating Local Dynamics in Three-Dimensional Covalent Organic Frameworks
Abstract
In covalent organic frameworks (COFs), the geometry of the building units predetermines the structure of the framework, yet comparable design principles relating the COF linkers to the local motions that govern guest adsorption and diffusion as well as mechanical response have remained beyond reach. Here, we combine synchrotron X-ray pair distribution function (PDF) analysis with molecular dynamics (MD) simulations accelerated by machine learning interatomic potentials to resolve and compare the local dynamics of two three-dimensional imine-linked COFs, chosen to differ in the geometry and rigidity of one of their two building units. Both are constructed from the same 4-connected square planar linker, -(pyrene-1,3,6,8-tetrayl)tetraaniline (PyTTA), reticulated with the 12-connected nanographene hexakis{[3,5-bis(-formylphenyl)-4,6-dimethoxyphenyl]}hexabenzocoronene to give COF-412, {[(HBC-LA)(PyTTA)]}, and with the 6-connected triptycene 2,3,6,7,14,15-hexakis(4-formylphenyl)triptycene to give COF-564, {[(Trip-CHO)(PyTTA)]}. The PDF-MD approach resolves the amplitude, the time scale, and the spatial extent of the motion of all COF linkers, separating the limited libration of the fused aromatic cores from the larger reorientational tilts of the pendant aryl rings they carry and from the translation of whole linkers within the net. The simulations further reveal that exchanging the nanographene- for the triptycene-based linker tunes the equilibrium tilt and the positional probability density of PyTTA, raising the tilt amplitude of its linker partner, and, notably, strengthens the coupling between the translations of neighbouring linkers. This approach provides a blueprint for understanding local dynamical properties in COFs, and shows them to be selectable through the choice of building unit and amenable to design.
Introduction
Covalent organic frameworks (COFs) are a class of crystalline porous materials whose components are selected from a wide range of organic linkers and covalently joined into predetermined topologies, affording stable backbones and pore environments that are tunable in functionality and metrics. [1, 2, 3] The diversity of chemical building blocks and linkage geometries makes COFs an object of design for structural and functional control, with applications ranging from the selective capture of carbon dioxide [4, 5, 6, 7] and water[8, 9, 10] from air, to gas separation, [11, 12, 13] chemical sensing, [14, 15] and heterogeneous catalysis. [16, 17]
However, this control has been exercised almost entirely over the global static structures of COFs and has yet to be extended to the rational design of their local dynamic properties, which, as in all reticular materials, govern guest adsorption and diffusion as well as mechanical response. [18, 19] That such motion pervades COF materials is well documented. For instance, it has been shown that single crystals of dia COFs expand and contract reversibly as guests enter and leave their pores, [20] COFs may adapt their pore geometry to the guests they include, [21] imine linkages undergo pedal motion that influences framework flexibility, [22] and aromatic rotors mounted within porous organic frameworks spin at megahertz rates even at cryogenic temperatures. [23]
However, while reticular chemistry rests on the precise translation of molecular geometry into extended frameworks, no defined principles yet exist by which local structure may be translated into framework dynamics. With such principles in hand, COFs could desirably be made locally rigid to maintain pore shape and selectivity, or locally mobile to modulate the pore aperture, adapt to incoming guests, and facilitate their diffusion. Establishing such design guidelines requires an experimental probe sensitive both to the local and the extended order of COFs, information beyond the reach of powder X-ray diffraction (PXRD) methods ordinarily used to characterize COF systems. [24, 25]
We reasoned that the atomic pair distribution function (PDF) obtained from X-ray total scattering experiments could meet this need, for it converts the scattering of a material, Bragg and diffuse alike, into a real space map of interatomic distances spanning local, intermediate, and long range order in any solid regardless of crystallinity. [26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43] In conjunction with theoretical methods such as molecular dynamics (MD) simulations, now extended by machine learning interatomic potentials (MLIPs) to length and time scales beyond the reach of ab initio approaches, [44] the features of the PDF may further be traced quantitatively to the interatomic vectors that produce them and linked to the underlying dynamical behaviour. [45]
Here, we combine synchrotron X-ray PDF analysis with MLIP-accelerated MD simulations to resolve the local dynamics of selected three-dimensional (3D) COFs and point toward the design of COFs with tunable local dynamics. We demonstrate this approach on two 3D imine-linked COFs built from the same 4-connected square planar linker, 4, 4′, 4″, 4‴-(pyrene-1,3,6,8-tetrayl)tetraaniline (PyTTA), whose fused pyrene core is planar and rigid. In COF-412, [(HBC-LA₁₂)(PyTTA)₃], PyTTA is reticulated with the 12-connected hexagonal prismatic nanographene hexakis[3,5-bis(p-formylphenyl)-4,6-dimethoxyphenyl]hexabenzocoronene (HBC-LA₁₂), whose hexabenzocoronene core is likewise a rigid, planar sheet of fused rings and whose twelve connecting arms are carried on dimethoxyphenyl spacers, affording the (3,4,6)-connected cez topology and the highest porosity reported for COFs to date (Figure 1a). [24] Both building units of COF-412 therefore present extended aromatic surfaces of comparable rigidity. In COF-564, [(Trip-CHO)₂(PyTTA)₃], that rigidity is modified while PyTTA is left in place, the nanographene being replaced by the 6-connected trigonal prismatic 2,3,6,7,14,15-hexakis(4-formylphenyl)triptycene (Trip-CHO), whose three arene peripheral groups are fixed at 120° about a bridgehead axis, so that an extended planar surface gives way to a compact unit of 3D shape and the framework adopts the (4,6)-connected stp topology (Figure 1b). [46] Holding PyTTA constant while varying its linker partner isolates the influence of the second building unit on framework motion. With the ML-MD models of both COFs validated against the measured PDFs, we analyze the MD trajectories to resolve the local structure of each framework and to quantify the reorientational and translational motion of its linkers. We find that replacing the nanographene-based linker by the triptycene-based building unit in COF-564 influences both the tilt angle and the positional disorder of PyTTA and, importantly, allows the translations of neighbouring linkers to become correlated, a coupling that is impeded in COF-412. Taken together, these results establish how the local dynamics of COFs connect to the structural features encoded in the PDF, and how the dynamics of their building units may be understood and directed.

Figure 1. Synthesis of COF-412 and COF-564. a) Condensation of the twelve-connected, dodeca-benzaldehyde-functionalized nanographene HBC-LA₁₂ with the four-connected tetraamine PyTTA to obtain COF-412. b) Condensation of the six-connected aldehyde Trip-CHO with the four-connected tetraamine PyTTA to obtain COF-564. Hydrogen atoms are omitted in the crystal structures for clarity.
Results and Discussion
Both COFs were obtained by imine condensation, and their syntheses are summarized in Figure 1. COF-412 was prepared by condensing the dodeca-benzaldehyde-functionalized nanographene HBC-LA with PyTTA (Figure 1a), and COF-564 by condensing the hexa-benzaldehyde triptycene Trip-CHO with the same tetraamine (Figure 1b). By virtue of the differing geometry and connectivity of the polyaldehyde linkers, the two COFs adopt distinct topologies, as shown in Figure 1. In COF-412 the twelve arms of each nanographene linker extend within a hexagonal prismatic envelope and are met by square planar PyTTA units, leading to a cez net. In COF-564 the six arms of each triptycene linker project from a bridgehead axis into three directions, resulting in a stp net. [46] Additional details on the synthetic procedures and related characterization of the COFs are provided in Section S1 of the SI (see Figures S1–S4).
The local structure of COFs is challenging to access experimentally. [47, 48] Their synthesis often yields polycrystalline rather than single crystalline products, so that their structural characterization has been largely limited to structures modeled from PXRD data, which are poorly sensitive to the positional disorder of the linkers and the stacking of their layers. Moreover, the COF frameworks are composed entirely of light elements (C, N, O, B, H) and offer no absorbing center to hard X-ray spectroscopic probes, such as hard X-ray absorption spectroscopy (XAS),[49, 50, 51, 52, 53] that serve the study of local properties in metal-organic frameworks. [54, 55, 56] To obtain accurate local structure information, we collected X-ray PDF total scattering data on COF-412 and COF-564 at beamline 28-ID-1 of the National Synchrotron Light Source II at 290 K using radiation of wavelength Å. The resulting PDFs, shown as black curves in Figure 2a,Figure 2b, display a series of well-defined peaks that persist above 12 Å, indicating that both frameworks retain short and intermediate range order.

Figure 2. Experimental and MD-averaged theoretical PDFs of a) COF-412 and b) COF-564. The measured PDFs (black) and the PDFs computed as an ensemble average over the MLIP-MD trajectory (blue) are displayed in the left panels. The markers above each panel indicate the positions of selected maxima in the experimental PDFs. A visualization of the COF structures where these maxima are assigned to interatomic vectors drawn on the corresponding unit cells is displayed in the right panels. Hydrogen atoms are omitted for clarity.
To interpret these structural correlations, we computed the corresponding PDFs as ensemble averages over MD simulations of each COF, which made use of the universal model for atoms (UMA) [44] (see the Supporting Information, Section S2, for additional details on the employed MD procedure). UMA is a transferable MLIP trained across molecular and periodic quantum-mechanical datasets of wide chemical diversity, including organic molecules, molecular crystals, and reticular materials. It describes at near-DFT accuracy both the covalent bonding within the linkers and their non-covalent contacts, the two interactions that together set the local structure of an imine-linked COF, while granting access to length and time scales beyond the reach of ab initio methods when validated against high-fidelity experimental data such as synchrotron PDF measurements. [44, 45] In particular, 100 MD snapshots were extracted at equal intervals along each trajectory for both COFs, and the ensemble PDFs were computed as their converged average (Figures S5–S6). As displayed in Figure 2a,Figure 2b, the MD-averaged PDFs agree with the experimental data of both COFs in the positions of the main peaks, from the intramolecular bonds at low to the inter-linker correlations beyond 9 Å. The agreement extends to features that are diagnostic of the local geometry of the building units. For instance, the relative heights of the maxima at 1.39 and 2.46 Å, which measure the balance of nearest- and second-neighbour aromatic contacts, are reproduced for both COFs, as the intra-linker correlation at 3.81 Å and the widths of these peaks, which report on the spread of conformations sampled by the linkers. Above about 9 Å the measured PDFs display oscillations that decay weakly with . For COF-564 the computed PDF reproduces the positions of these maxima, though with reduced amplitudes (Figure 2b), whereas for COF-412 the computed peaks at these distances are less intense than the measured ones and their positions are slightly overestimated by the UMA-based MD model (Figure 2a). [45]
We next employed the MD models to assign the individual PDF peaks to specific interatomic vectors, mapping them onto the respective unit cells shown in the right panels of Figure 2. Assignments were made by counting bonds along the shortest covalent path in the structural model of each COF (Tables S1–S2). Below about 4.5 Å the PDFs of the two COFs are very similar, both being dominated by intra-linker correlations. The two sharp maxima at 1.39 and 2.46 Å arise from the nearest-neighbour aromatic C–C bond and from the second-neighbour (meta) C⋯C distance across the aromatic rings, as expected of materials built from phenylene and fused-ring units. The maximum at 2.89 Å arises from the para separation across a single six-membered ring and is more intense in COF-412, consistent with the larger fraction of fused rings in its linkers. The maximum at 3.81 Å is assigned to a three-bond path crossing a ring junction into the adjacent ring. The maxima between 4.3 and 6 Å arise from four- to seven-bond paths that walk outward along a single linker for both COFs. In COF-564, the peak centered around 4.99 Å is split into a main peak and a shoulder at higher , both of which can be assigned through the employed theoretical model. As shown in Figure 2b, the MD-averaged PDF of COF-564 displays a maximum at 4.79 Å which arises from four-bond C⋯C paths and is dominated by pairs connecting the pyrene core of PyTTA to its phenyl arms (41%) and the triptycene blades of Trip-CHO to their pendant aryl rings (31%). Conversely, the PDF shoulder peak at 5.14 Å arises from the corresponding five-bond paths, with nearly equal contributions from the blade–aryl pairs of Trip-CHO (34%) and the core–arm pairs of PyTTA (33%). The PDFs of the two COFs differ more markedly above 6 Å. The maximum at 6.66 Å in COF-412 arises mainly from six- to nine-bond paths within its larger nanographene linker, whereas the corresponding feature of COF-564 lies 0.10 Å lower, at 6.56 Å, and receives contribution from six-bond paths within PyTTA together with pairs bridging the imine linkage into the aryl rings of Trip-CHO. Beyond this distance, inter-linker correlations increasingly contribute to the PDFs of both frameworks. The maxima at approximately 7.95 and 10.9 Å are present in both PDFs and receive contributions from pairs that either bridge the imine linkage or span two linkers that are not directly connected (see Figure 2, right panels). Having established that the MD simulations reproduce the short-range order of both COFs and the
intermediate-range order of COF-564, we turned to investigating the dynamic properties of the COFs encoded in the MD trajectories, beginning with COF-412 (Figure 3). We treated each linker in two parts, the fused aromatic core and the covalently linked pendant aryl rings, and monitored the reorientation of each through the second-order orientational correlation function , where is the second Legendre polynomial, is the unit normal of the aromatic plane, and the average runs over time origins [55, 56] (Figure 3a). The behaviour of the two fused cores is similar. Both the pyrene core of PyTTA (L1) and the hexabenzocoronene core of HBC-LA (L2) exhibit values near unity throughout a 20 ps MD simulation time window, indicating that the fused cores of neither linker reorient on this time scale. The values of the linker pendant rings, by contrast, decay markedly and to different extents, so that within COF-412 the two linkers carry pendant groups of substantially different mobility.

Figure 3. Local dynamics of COF-412 obtained from the analysis of MLIP–MD trajectories. a) Second-order orientational correlation function of the ring normal of the fused cores and pendant aryl rings of the PyTTA (L1) and HBC-LA (L2) linkers. b) Distributions of the equilibrium tilt angle of the same four groups, with mean values indicated. c) Cross-correlation matrix of the centre-of-mass displacements of the six crystallographically distinct PyTTA units (L1-1 to L1-6) and the two HBC-LA units (L2-1 and L2-2). d,e) Positional probability densities of selected PyTTA (d) and HBC-LA (e) linkers. f) Distribution of the pore volume sampled over the MD trajectory. g) Visualization of the directions of motion of the linkers within the COF cell, where selected pore areas are highlighted.
The amplitudes of these motions follow the same ordering (Figure 3b). The tilt angle of the fused cores is confined to median values of 8.6° for L1 and 6.1° for L2, whereas the pendant rings exhibit ring equilibrium tilt median values of 38.2° and 18.0°, respectively. These analyses indicate that the fused cores librate about their equilibrium orientation while the pendant rings undergo large-amplitude reorientation. COF-412 may therefore be described as rigid in its linker cores and mobile at the joints that bridge PyTTA and HBC-LA.
To determine whether the building units of COF-412 translate independently or in concert, we computed the cross-correlation matrix of the linker centres of mass (Figure 3c). Its normalised elements, , where is the displacement of the centre of mass of linker from its mean position, equal unity for two linkers that move together and minus one for two linkers whose displacements are opposed. Because the conserved centre of mass of the simulation cell imposes a weak negative correlation on all pairs even for fully independent motion, the matrix is reported relative to this baseline, so that the displayed values vanish for uncoupled linkers (see Section S2 of the SI). The off-diagonal elements among the six PyTTA units (L1-1 to L1-6) and the two nanographene units (L2-1 and L2-2) are comparatively weak, indicating that the linkers of COF-412 translate largely independently of one another. Closer inspection nonetheless reveals that the displacements of the two nanographene units (L2-1 and L2-2) are on average opposed, as they exhibit the strongest off-diagonal element of the matrix (see darker blue regions in Figure 3c). We attribute this anticorrelation to the pore the two linkers share, for they face each other across a single aperture, so that the approach of one toward the pore centre is balanced by a displacement of the other (Figure 3g).
This dynamic behaviour affects the positional probability densities of both COF linkers, which are shown in Figure 3d,Figure 3e. In both probability density projections the fused cores appear as sharp, atomically resolved features, whereas the pendant rings are smeared into broad lobes, reflecting the greater reorientational freedom of the pendant aryl groups.[45] Despite this local mobility, the pores of COF-412 exhibit a narrowly distributed volume, as shown in Figure 3f which displays the MD-derived pore volume analysis (refer to Section S2 of the SI for additional details). The width of the resulting distribution amounts to a few percent of the mean volume, indicating that, within the model, the large local librations of the linker pendant aromatic rings occur without significantly altering the metrics of the pore they line. We note that these values refer to the simulated COF, so that the absolute pore volume of COF-412 may carry a corresponding offset.
To determine how the second building unit influences the COF local dynamics, we applied the same analysis to COF-564, in which PyTTA is retained while the extended planar nanographene linker is replaced by the compact triptycene-based building unit, so that any difference in the dynamics may be attributed to this substitution. In triptycene, the three arene wings are held about a bridgehead axis rather than fused into a single plane as in the nanographene. We reasoned that this exchange would replace a broad aromatic platform, which anchors twelve closely spaced arms to one rigid sheet whose plane defines their equilibrium orientation, by a compact node bearing six arms, so that each arm would be less crowded by its neighbours and less constrained by the surface on which it is mounted, while the building unit itself remained rigid and directionally defined. The pendant aryl arms of the partner should therefore sample larger tilt amplitudes than those borne on the nanographene sheet, and the motion of PyTTA, although the linker itself is unchanged, should be altered at least in part.
As shown in Figure 4, the pattern of motion found in COF-564 exhibits differences compared to that of COF-412. The fused cores of both linkers again show almost no reorientation, with the values of the pyrene core of PyTTA (L3) and of the triptycene core of Trip-CHO (L4) remaining above 0.95 over 20 ps, and their pendant rings decaying to approximately 0.50 and 0.58, respectively (Figure 4a). The corresponding equilibrium tilt amplitudes display median values of and for the fused cores of PyTTA and Trip-CHO, and of and for their pendant rings (Figure 4b). Notably, the mobility of PyTTA is itself altered by the substitution as the tilt of its fused core rises by from the median value of found in COF-412 while that of its pendant rings falls by from a median of . Although small, these shifts are in line with our hypothesis that a building unit retains only part of its dynamic signature when its linker partner is changed. We attribute the freer libration of the pyrene sheet to the loss of the rigid and broad aromatic platform of HBC-月 opposite it. The mobility of the Trip-CHO linker partner changes more significantly. The tilt of the pendant rings of the triptycene-based linker rises by , from for the arms of the nanographene to for those of the triptycene, and that of its fused core by , from to , as anticipated for arms borne on a compact node rather than closely spaced about one broad sheet. Together, these changes bring the pendant rings of the two linkers to comparable mobility, differing by in COF-564, whereas in COF-412 they differ by . Part of the dynamic signature of a building unit is therefore carried with it from one COF into another, and part is set by the partner with which it is reticulated.

Figure 4. Local dynamics of COF-564 obtained from the analysis of MLIP-MD trajectories. a) Second-order orientational correlation function of the ring normal of the fused cores and pendant aryl rings of the PyTTA (L1) and Trip-CHO (L4) linkers. b) Distributions of the equilibrium tilt angle of the same four groups, with mean values indicated. c) Cross-correlation matrix of the centre-of-mass displacements of the six crystallographically distinct PyTTA units (L3-1 to L3-6) and the four Trip-CHO units (L4-1 and L4-2). d,e) Positional probability densities of selected PyTTA (d) and Trip-CHO (e) linkers. f) Distribution of the three pore volumes sampled over the MD trajectory. g) Visualization of the directions of motion of the linkers within the COF cell, where selected pore areas are highlighted.
The translational behaviour changes more markedly. The cross-correlation matrix of COF-564 (Figure 4c) exhibits appreciable off-diagonal elements among the six PyTTA units (L3-1 to L3-6) and the four triptycene units (L4-1 to L4-4). In particular, the PyTTA–PyTTA elements remain weak and mildly anticorrelated, much as in COF-412. Interestingly, several pairs among the triptycene units are markedly negatively and positively correlated, while also the PyTTA–triptycene matrix elements display correlations. This pattern may be understood by extending the picture drawn for the cez topology: triptycene-based linkers that bound a common pore displace in mutually opposed fashion as one approaches the pore centre, as found for the two nanographene cores in COF-412, whereas triptycene-based units on the same side of a pore are carried along together, indicating that the motions of neighbouring linkers in the COF-564 are coupled to a degree not seen in COF-412. The probability densities of L3 and L4 are shown in Figure 4d,Figure 4e and localize the fused cores and delocalize the pendant rings of both linker.
Where COF-412 exhibits one main kind of pore (see Figure 3g, purple regions), the six-connected triptycene node of COF-564 results in three kinds of pores, of approximately 19, 9.7, and 4.8 × Å, each with a narrow distribution of volume (Figure 4f), which tile the lattice in a repeating pattern (Figure 4g). The exchange of one building unit therefore alters not only the amplitude of local motion but also the number and size of the pores whose walls that motion modulates. The narrow spread of all pore volume distributions indicates retention of the pore metrics in both COFs, so that local mobility is accommodated without pore deformation.
Conclusion
In summary, we have combined synchrotron X-ray PDF analysis with MD simulations accelerated by MLIPs to resolve the local structure and dynamics of two 3D imine-linked COFs built from a common linker. The combined PDF measurements and MD simulations establish a consistent picture of motion. Fused aromatic cores librate within median values of at 290 K and show reduced reorientational ability, while the pendant aryl rings they carry sample tilt angles three to more than four times larger. Exchanging the nanographene-based HBC-LA linker of COF-412 for the triptycene-based Trip-CHO linker of COF-564 (i) influences the dynamics of the shared PyTTA linker, (ii) raises the tilt amplitude of the pendant rings of the Trip-CHO partner by at 290 K, and (iii) increases the coupling between the translational motions of neighbouring linkers. Two comparative design principles follow. First, stiffness and mobility occupy separate and predictable positions in these frameworks, rigidity residing in the fused cores and motion at the pendant rings and at the joints between building units, so that each can be placed by the choice of linker without compromising the pore metrics. Second, the mobility of a linker is a property of the reticulated pair rather than of the linker alone. Indeed, a building unit carries only part of its dynamic signature from one COF to the other, so that pairing a given linker with a broad aromatic sheet or with a compact node offers a direct handle on its dynamical properties.
We anticipate that the employed method will aid the screening of candidate combinations of COF building units before synthesis, with the measured PDFs serving as the experimental benchmark for theoretical validation. Extension of this approach to guest-loaded COFs would then allow the local dynamics to be followed in realistic operating conditions, and COFs to be designed in which the intra- and inter-linker dynamics are tailored to function.
Data Availability
Experimental PDF data will be made available on the Zenodo platform.
Supporting Information
General experimental methods and materials; syntheses and characterization of COF-412 and COF-564; total scattering data collection and PDF processing; MLIP-MD protocol and convergence tests; additional PDF results.
Acknowledgments
F.T. acknowledges funding from the European Union’s Horizon Europe research and innovation program under the Marie Skłodowska-Curie grant agreement No. 101201413 (AQUAFRAME). Effort of S.J.L.B. and C.M. on PDF analysis, modeling and writing was supported by U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences (DOE-BES) under contract No. DE-SC0024141. This research used beamline 28-ID-1 of the National Synchrotron Light Source II, a U.S. DOE Office of Science User Facility operated for the DOE Office of Science by Brookhaven National Laboratory under Contract No. DE-SC0012704. Claude-Opus5.0 was used for grammar and sentence structure checking during the review and editing of this manuscript.
References
- [1]Côté, A. P.; Benin, A. I.; Ockwig, N. W.; O’Keeffe, M.; Matzger, A. J.; Yaghi, O. M. Porous, Crystalline, Covalent Organic Frameworks. Science 2005, 310, 1166–1170.pubmed.ncbi.nlm.nih.gov/16293756
- [2]Han, S. S.; Furukawa, H.; Yaghi, O. M.; Goddard, W. A. I. Covalent Organic Frameworks as Exceptional Hydrogen Storage Materials. J. Am. Chem. Soc. 2008, 130, 11580–11581.DOI
- [3]Diercks, C. S.; Yaghi, O. M. The Atom, the Molecule, and the Covalent Organic Framework. Science 2017, 355, eaal1585.DOI
- [4]Lyu, H.; Li, H.; Hanikel, N.; Wang, K.; Yaghi, O. M. Covalent Organic Frameworks for Carbon Dioxide Capture from Air. J. Am. Chem. Soc. 2022, 144, 12989–12995.DOI
- [5]Lyu, H.; Li, H.; Hanikel, N.; Wang, K.; Yaghi, O. M. Covalent Organic Frameworks for Carbon Dioxide Capture from Air. J. Am. Chem. Soc. 2022, 144, 12989–12995.DOI
- [6]Alenazi, M. H.; Helal, A.; Khan, M. Y.; Khalil, A.; Khan, A.; Usman, M.; Zahir, M. H. Covalent Organic Frameworks (COFs) for CO₂ Utilizations. Carbon Capture Sci. Technol. 2025, 14, 100365.DOI
- [7]Wei, W.; Wang, Z.; Xu, C.; Li, Z.; Bai, H.; Liu, Z. Advances in the Structural Design of Covalent Organic Frameworks for Direct Air Capture of CO₂. Ind. Eng. Chem. Res. 2025, 64, 24276–24299.DOI
- [8]Nguyen, H. L.; Darù, A.; Chheda, S.; Alawadhi, A. H.; Neumann, S. E.; Wang, L.; Bai, X.; Alawad, M. O.; Borgs, C.; Chayes, J. T.; Sauer, J.; Gagliardi, L.; Yaghi, O. M. Pinpointing the Onset of Water Harvesting in Reticular Frameworks from Structure. ACS Cent. Sci. 2025, 11, 665–671.DOI
- [9]He, Y.; Ran, J.; Gao, X.; Ding, J.; Templeton, M. R.; Peng, C.; Chu, W. Covalent Organic Frameworks Enable Efficient Atmospheric Water Harvesting in Arid Climates. Environ. Sci.: Water Res. Technol. 2025, 11, 2810–2824.DOI
- [10]Sun, C.; Zhu, Y.; Shao, P.; Chen, L.; Huang, X.; Zhao, S.; Ma, D.; Jing, X.; Wang, B.; Feng, X. 2D Covalent Organic Framework for Water Harvesting with Fast Kinetics and Low Regeneration Temperature. Angew. Chem. Int. Ed. 2023, 62, e202217103.
- [11]Alom, J.; Rahman, M. A.; Hossein, M.; Ban, S.; Íñigo Martínez-Visus; Ahmed, M. B.; Keskin, S.; Téllez, C.; Zornoza, B. Covalent Organic Framework (COF) Based Mixed Matrix Membranes as an Alternative Solution for Gas Separation. J. Membr. Sci. 2026, 753, 125620.DOI
- [12]Jiang, Y.; Ren, J.; Yuan, Y.; Keshta, B. E.; Zhang, Y.; Guan, X.; He, Y.; Zhao, D.; Chen, B. Covalent Organic Frameworks for Gas Storage and Separation. Chem. Soc. Rev. 2026, 55, 4522–4584.DOI
- [13]Aksu, G. O.; Keskin, S. The COF Space: Materials Features, Gas Adsorption, and Separation Performances Assessed by Machine Learning. ACS Mater. Lett. 2025, 7, 954–960.DOI
- [14]Yue, Y.; Ji, D.; Liu, Y.; Wei, D. Chemical Sensors Based on Covalent Organic Frameworks. Chem. Eur. J. 2024, 30, e202302474.DOI
- [15]Liu, X.; Huang, D.; Lai, C.; Zeng, G.; Qin, L.; Wang, H.; Yi, H.; Li, B.; Liu, S.; Zhang, M.; Deng, R.; Fu, Y.; Li, L.; Xue, W.; Chen, S. Recent Advances in Covalent Organic Frameworks (COFs) as a Smart Sensing Material. Chem. Soc. Rev. 2019, 48, 5266–5302.DOI
- [16]Guo, J.; Jiang, D. Covalent Organic Frameworks for Heterogeneous Catalysis: Principle, Current Status, and Challenges. ACS Cent. Sci. 2020, 6, 869–879.DOI
- [17]Shirani, M. A.; Afshari, M.; Romero-Muñiz, L.; Platero-Prats, A. E.; Dinari, M.; Zamora, F. Designing Covalent Organic Frameworks for Organic Catalysis: Bridging Reticular and Molecular Chemistry. J. Mater. Chem. A 2025, 13, 19172–19192.DOI
- [18]Chattopadhyay, M. K.; Mondal, U.; Datta, S.; Banerjee, P. Recent Advances in Reticular Frameworks and Allied Composite Materials for Ultrasensitive Optical Detection of Biogenic and Volatile Amines in Food Spoilage Monitoring. Dalton Trans. 2026, 55, 3602–3627.
- [19]Ji, C.; Kang, C.; Patra, B. C.; Zhao, D. Flexible Covalent Organic Frameworks: Design, Synthesis, and Applications. CCS Chemistry 2024, 6, 856–881.DOI
- [20]Yao, A.; Xu, H.; Shao, K.; Sun, C.; Qin, C.; Wang, X.; Su, Z. Guest-induced structural transformation of single-crystal 3D covalent organic framework at room and high temperatures. Nat. Commun. 2025, 16, 1385.DOI
- [21]Liu, S.; Wei, L.; Zeng, T.; Jiang, W.; Qiu, Y.; Yao, X.; Wang, Q.; Zhao, Y.; Zhang, Y.-B. Single-Crystal Dynamic Covalent Organic Frameworks for Adaptive Guest Alignments. J. Am. Chem. Soc. 2024, 146, 34053–34063.DOI
- [22]Chi, H.; Liu, Y.; Li, Z.; Chen, W.; He, Y. Direct visual observation of pedal motion-dependent flexibility of single covalent organic frameworks. Nat. Commun. 2023, 14, 5061.DOI
- [23]Comotti, A.; Bracco, S.; Ben, T.; Qiu, S.; Sozzani, P. Molecular Rotors in Porous Organic Frameworks. Angew. Chem. Int. Ed. 2014, 53, 1043–1047.
- [24]Mirzaei, S.; Mirzaei, M. S.; Gao, M.-Y.; Gándara, F.; Wang, K.; Tavani, F.; Zhu, H.; Proserpio, D. M.; Yaghi, O. M. Single-Crystalline Twelve-Connected Nanographene-Based Covalent Organic Frameworks. J. Am. Chem. Soc. 2026, 148, 25241–25248.DOI
- [25]Mirzaei, S.; Mirzaei, M. S.; Wang, W.; Nguyen, H. L.; Wang, K.; Gándara, F.; Gao, M.-Y.; Khosropour, A.; Naning, B. L.; Zhang, Y.-B.; Abbaspourrad, A.; Yaghi, O. M. Single-Crystalline Covalent Organic Frameworks Directed by π-stacking. 2026; https://chemrxiv.org/doi/pdf/10.26434/chemrxiv.10001788/v1.DOI
- [26]Terban, M. W.; Billinge, S. J. L. Structural Analysis of Molecular Materials Using the Pair Distribution Function. Chem. Rev. 2022, 122, 1208–1272.DOI
- [27]Billinge, S. J. L.; Kanatzidis, M. G. Beyond Crystallography: the Study of Disorder, Nanocrystallinity and Crystallographically Challenged Materials with Pair Distribution Functions. Chem. Commun. 2004, 40, 749–760.
- [28]Griffiths, J.; Suzana, A. F.; Wu, L.; Marks, S. D.; Esposito, V.; Boutet, S.; Evans, P. G.; Mitchell, J. F.; Dean, M. P. M.; Keen, D. A.; Robinson, I.; Billinge, S. J. L.; Bozin, E. S. Resolving Length-Scale-Dependent Transient Disorder through an Ultrafast Phase Transition. Nat. Mat. 2024, 23, 1041–1047.DOI
- [29]Kløve, M.; Sommer, S.; Iversen, B. B.; Hammer, B.; Dononelli, W. A Machine-Learning-Based Approach for Solving Atomic Structures of Nanomaterials Combining Pair Distribution Functions with Density Functional Theory. Adv. Mater. 2023, 35, 2208220.
- [30]Young, C. A.; Goodwin, A. L. Applications of Pair Distribution Function Methods to Contemporary Problems in Materials Chemistry. J. Mater. Chem. 2011, 21, 6464–6476.DOI
- [31]Zhu, H.; Huang, Y.; Ren, J.; Zhang, B.; Ke, Y.; Jen, A. K.-Y.; Zhang, Q.; Wang, X.-L.; Liu, Q. Bridging Structural Inhomogeneity to Functionality: Pair Distribution Function Methods for Functional Materials Development. Adv. Sci 2021, 8, 2003534.DOI
- [32]Romero-Muñiz, I.; Mavrandonakis, A.; Albacete, P.; Vega, A.; Briois, V.; Zamora, F.; Platero-Prats, A. E. Unveiling the Local Structure of Palladium Loaded into Imine-Linked Layered Covalent Organic Frameworks for Cross-Coupling Catalysis. Angew. Chem. Int. Ed. 2020, 59, 13013–13020.
- [33]Cliffe, M. J.; Wan, W.; Zou, X.; Chater, P. A.; Kleppe, A. K.; Tucker, M. G.; Wilhelm, H.; Funnell, N. P.; Coudert, F.-X.; Goodwin, A. L. Correlated Defect Nanoregions in a Metal–Organic Framework. Nat. Commun. 2014, 5, 4176.DOI
- [34]Romero-Muñiz, I.; Loukopoulos, E.; Xiong, Y.; Zamora, F.; Platero-Prats, A. E. Exploring Porous Structures without Crystals: Advancements with Pair Distribution Function in Metal- and Covalent Organic Frameworks. Chem. Soc. Rev. 2024, 53, 11772–11803.
- [35]Bennett, T. D.; Cheetham, A. K. Amorphous Metal-Organic Frameworks. Acc. Chem. Res. 2014, 47, 1555–1562.DOI
- [36]Bennett, T. D.; Goodwin, A. L.; Dove, M. T.; Keen, D. A.; Tucker, M. G.; Barney, E. R.; Soper, A. K.; Bithell, E. G.; Tan, J.-C.; Cheetham, A. K. Structure and Properties of an Amorphous Metal–Organic Framework. Phys. Rev. Lett. 2010, 104, 115503.DOI
- [37]Cao, S.; Bennett, T. D.; Keen, D. A.; Goodwin, A. L.; Cheetham, A. K. Amorphization of the Prototypical Zeolitic Imidazolate Framework ZIF-8 by Ball-Milling. Chem. Commun. 2012, 48, 7805–7807.DOI
- [38]Gaillac, R.; Pullumbi, P.; Beyer, K. A.; Chapman, K. W.; Keen, D. A.; Bennett, T. D.; Coud- ert, F.-X. Liquid Metal–Organic Frameworks. Nat. Mat. 2017, 16, 1149–1154.
- [39]Keen, D. A.; Bennett, T. D. Structural Investigations of Amorphous Metal–Organic Frameworks Formed via Different Routes. Phys. Chem. Chem. Phys. 2018, 20, 7857–7861.DOI
- [40]Terban, M. W.; Banerjee, D.; Ghose, S.; Medasani, B.; Shukla, A.; Legg, B. A.; Zhou, Y.; Zhu, Z.; Sushko, M. L.; De Yoreo, J. J.; Liu, J.; Thallapally, P. K.; Billinge, S. J. L. Early Stage Structural Development of Prototypical Zeolitic Imidazolate Framework (ZIF) in Solution. Nanoscale 2018, 10, 4291–4300.DOI
- [41]Molina, M. A.; Manjón-Sanz, A.; Sánchez-Sánchez, M. On the contribution of Pair Distribution Function (PDF) to the characterization of nanocrystalline MOFs: The case of M-MOF-74. Micropor. Mesopor. Mat. 2021, 319, 110973.DOI
- [42]Sapnik, A. F.; Bechis, I.; Bumstead, A. M.; Johnson, T.; Chater, P. A.; Keen, D. A.; Jelfs, K. E.; Bennett, T. D. Multivariate Analysis of Disorder in Metal–Organic Frameworks. Nat. Commun. 2022, 13, 2173.DOI
- [43]Sapnik, A. F.; Sun, C.; Laulainen, J. E. M.; Johnstone, D. N.; Brydson, R.; Johnson, T.; Midgley, P. A.; Bennett, T. D.; Collins, S. M. Mapping Nanocrystalline Disorder within an Amorphous Metal–Organic Framework. Commun. Chem. 2023, 6, 92.DOI
- [44]Wood, B. M. et al. UMA: A Family of Universal Models for Atoms. 2026; https://arxiv.org/abs/2506.23971.
- [45]Tavani, F.; Mirzaei, S.; Yin, J.; Lin, Y.-H.; Myers, C.; Lin, C.-H.; Abeykoon, A. M.; Billinge, S. J. L.; Yaghi, O. M. Local Structure and Dynamics of Three-Dimensional Covalent Organic Frameworks. J. Am. Chem. Soc. 2026,arxiv.org/abs/2608.19106
- [46]Mirzaei, S.; Mirzaei, M. S.; Wang, K.; Tavani, F.; Yin, J.; Proserpio, D. M.; Yaghi, O. M. Single-Crystalline Triptycene-Based Covalent Organic Frameworks. 2026; https://chemrxiv.org/doi/full/10.26434/chemrxiv.15004178/v1.DOI
- [47]Pütz, A. M.; Terban, M. W.; Bette, S.; Haase, F.; Dinnebier, R. E.; Lotsch, B. V. Total Scattering Reveals the Hidden Stacking Disorder in a 2D Covalent Organic Framework. Chem. Sci. 2020, 11, 12647–12654.DOI
- [48]Romero-Muñiz, I.; Albacete, P.; Platero-Prats, A. E.; Zamora, F. Layered Copper-Metallated Covalent Organic Frameworks for Huisgen Reactions. ACS Applied Materials & Interfaces 2021, 13, 54106–54112.
- [49]Tavani, F.; Capocasa, G.; Martini, A.; Sessa, F.; Di Stefano, S.; Lanzalunga, O.; D’Angelo, P. Direct Structural and Mechanistic Insights into Fast Bimolecular Chemical Reactions in Solution through a Coupled XAS/UV–Vis Multivariate Statistical Analysis. Dalt. Trans. 2021, 50, 131–142.DOI
- [50]Frateloreto, F.; Tavani, F.; Di Berto Mancini, M.; Del Giudice, D.; Capocasa, G.; Kieffer, I.; Lanzalunga, O.; Di Stefano, S.; D’Angelo, P. Following a Silent Metal Ion: A Combined X-ray Absorption and Nuclear Magnetic Resonance Spectroscopic Study of the Zn2+ Cation Dissipative Translocation between Two Different Ligands. J. Phys. Chem. Lett. 2022, 13, 5522–5529.DOI
- [51]Tavani, F.; Capocasa, G.; Martini, A.; Sessa, F.; Di Stefano, S.; Lanzalunga, O.; D’Angelo, P. Activation of C–H bonds by a Nonheme Iron(iv)–Oxo Complex: Mechanistic Evidence through a Coupled EDXAS/UV-Vis Multivariate Analysis. Phys. Chem. Chem. Phys. 2021, 23, 1188–1196.DOI
- [52]Del Giudice, D.; Tavani, F.; Di Berto Mancini, M.; Frateloreto, F.; Busato, M.; Oliveira De Souza, D.; Cenesi, F.; Lanzalunga, O.; Di Stefano, S.; D’Angelo, P. Two Faces of the Same Coin: Coupling X-Ray Absorption and NMR Spectroscopies to Investigate the Exchange Reaction Between Prototypical Cu Coordination Complexes. Chem. Eur. J. 2022, 28, e202103825.DOI
- [53]Tavani, F.; Busato, M.; Veclani, D.; Braglia, L.; Mauri, S.; Torelli, P.; D’Angelo, P. Investigating the High-Temperature Water/MgCl2 Interface through Ambient Pressure Soft X-ray Ab- sorption Spectroscopy. ACS Appl. Mater. Interf. 2023, 15, 26166–26174.
- [54]Tavani, F.; Tofoni, A.; D’Angelo, P. Exploring the Methane to Methanol Oxidation over Iron and Copper Sites in Metal–Organic Frameworks. Catalysts 2023, 13.DOI
- [55]Tavani, F.; Tofoni, A.; Pietropaoli, E.; Stoian, D. C.; van Beek, W.; Marshall, K.; Pettiti, I.; Latini, A.; D’Angelo, P. Decoding the Water Harvesting Mechanism of MIL-100(Fe) Across Short- and Long-Range Length Scales. J. Am. Chem. Soc. 2025, 147, 40507–40518.DOI
- [56]Tavani, F.; Tofoni, A.; Vandone, M.; Busato, M.; Braglia, L.; Torelli, P.; Stanzione, M. G.; Armstrong, A. R.; Morris, R. E.; Colombo, V.; D’Angelo, P. A Combined Soft X-ray and Theoretical Investigation Discloses the Water Harvesting Behaviour of Mg-MOF-74 at the Crystal Surface. Chem. Sci. 2025, 16, 9462–9471.