Isolation of a structural intermediate during switching of degree of interpenetration in a metal–organic framework

A structural intermediate has been isolated for the first time during switching of interpenetration from twofold to threefold in the MOF [Co2(ndc)2(bpy)].

modelled to the best of our ability to make sense of the electron density maps. The remaining electron density within the channel was removed using the SQUEEZE routine of PLATON 24 .
The solvent-accessible void volumes were also calculated using PLATON.
Single-crystal to single-crystal conversion of 2fa to 3fa. A single crystal of 2fa was left at room temperature. The unit cell parameters for the same crystal were checked at regular intervals to monitor the progress of the transformation. Based on these parameters, the conversion appeared to be complete after almost one week and the single-crystal X-ray diffraction data analysis revealed that the system had converted to its triply-interpenetrated form (3fa).
Single-crystal to single-crystal conversion of 2fa′ to 3fa. A suitable single crystal of 2fa′ isolated from the bulk material was left at room temperature. The unit cell parameters of the same crystal were checked at regular intervals to monitor the progress of the transformation.
After 3 to 4 days the conversion appeared to be complete and single-crystal X-ray diffraction analysis revealed that the system had converted to its triply-interpenetrated form (3fa).
Single-crystal to single-crystal conversion of 2fa to 2fa′. Since the transformation from 2fa to 2fa′ is a time consuming process at room temperature, gentle heating was used to bring about the conversion. A suitable crystal of as-synthesized material was glued to the end of a glass fibre and SCD data were recorded at 100 K, yielding the structure of 2fa. The temperature was then ramped to 313 K at a rate of 120 K h -1 and kept constant while unit cell parameters were determined repeatedly. Once the unit cell parameters indicated apparentlycomplete conversion of 2fa to 2fa′, the temperature was then decreased back to 100 K. Single crystal X-ray diffraction data were again recorded, revealing a structure quite similar to that of 2fa′ isolated from the activation of bulk crystals. Thermogravimetric analysis. Thermal analyses were carried out using a TA Instruments Q500 thermogravimetric analyser. X-ray powder diffraction studies (XRPD). The diffractograms were recorded with a Bruker D2 PHASER equipped with a Lynxeye 1D detector and Ni-filtered copper Kα radiation (30 kV, 10 mA generator parameters; restricted by a 1.0 mm divergence slit and a 2.5° Soller collimator) with a 0.02° step width. The phase purity of the as-synthesized 2fa and the activated forms 2fa′ and 3fa was confirmed by comparison of experimental XRPD with the simulated XRPD patterns from the respective crystal structures. XRPD was used to study all the conversions under different conditions. Gas sorption measurement. An Intelligent Gravimetric Analyser (IGA-002 supplied by Hiden Analytical Ltd, Warrington, UK) 6 was used to record sorption isotherms at 298 K for samples of 2fa′ and 3fa. The instrument facilitates precise measurement of mass change and control of pressure and temperature, and is equipped with an advanced pressure rating allowing measurement up to 20 bar. The pressure is monitored using a pressure transducer with a range of 0-20 bar and buoyancy effects are corrected for. Temperature control is maintained to an accuracy of ca. 0.05 °C by a refrigerated recirculating bath. Data collection is controlled by Real-Time Processing computer software, 7 which continually analyses the equilibrium using least-squares regression to extrapolate a value of the asymptote.

Single-crystal X-ray diffraction (SCD)
Diffraction patterns for three single-crystal to single-crystal transformations: 2fa to 3fa (top), 2fa′ to 3fa (middle) and 2fa to 2fa′ (bottom).     (i) 1-x, 1-y, 1-z; (ii) 1-x, 2-y, -z; (iii) -x, 1-y, 2-z; (iv) 2-x, 2-y, 1-z.    The dihedral angles between metal plane and metal-carboxylate plane (θ) as well as metal plane and the aromatic plane of ndc (φ) in both 2fa and 2fa′ structures were measured. There are two crystallographically distinct ndc linkers with four different carboxylates and two different aromatic units in 2fa. As a result, a total of four θ and two φ values were determined. In the case of the bpy linker, similar measurements were made by measuring the torsion angle (δ) between the metal-metal axis and the metal-N linkages and torsion angle (ω) between metal-metal axis and aromatic part of bpy. Since bpy is disordered over three positions in 2fa (as shown above), torsion angles for each individual subunit were measured and the complete range reported. Similar meassurement were carried out for 2fa′.

Crystal Data and Structure Refinement:
Top: diagrammatic representation of key dihedral and torsion angles in 2fa and 2fa'. A: Co 2 ···Co 2 metal atom plane; B: Co 2 -carboxylate oxygen plane; C: aromatic plane of ndc; D: Co 2 cluster metal-metal axis; E: metal-nitrogen; F: N···N axis of bpy. Both crystallographic symmetry and geometrical disorders have been considered in measuring these parameters. Bottom: table of relevant angles in º.

Measurement of metal-metal distances between two paddlewheels:
In the case of ndc, the node···node distances between two paddlewheel units is taken as the distance between the Co-Co centroids of the two metal clusters bridged by the ndc ligand. The same protocol was used to determine the shortest distance between two paddlewheel nodes of different networks. In the case of bpy, the distance is measured between the two Co metal centers to which the ligand is attached. It should be noted that only the geometrically favourable Co centers are taken in to account while measuring the shortest distances between paddlewheel units of different networks.

Rietveld Refinement
The observed X-ray powder patterns were refined using the Rietveld 8 method. The program TOPAS was employed, using the corresponding published single-crystal X-ray structures as starting models. The resultant difference plots thus generated are given below.

NOTE:
For Rietveld refinements, the background was not subtracted for any of the three XRPD patterns. The XRPD patterns were originally collected on Bruker D2 phaser equipped with a Cu source. Owing to the presence of Co ions, each pattern contained a significant background signal, which was subtracted (for clarity) when comparing the XRPD patterns with simulated patterns obtained from single-crystal data.     Fig. S13: Observed (blue) and simulated (red) X-ray powder diffractograms (the latter obtained from Rietveld refinement) as well as the difference plot (grey) for 2fa.