Unprecedented Topological Complexity in a Metal-Organic Framework Constructed from Simple Building Units A. Ken Inge,*,, Milan Köppen, Jie Su, Mark Feyand, Hongyi Xu, Xiaodong Zou, Michael O Keeffe,*, and Norbert Stock*, Institut für Anorganische Chemie, Christian-Albrechts-Universität zu Kiel, Max-Eyth-Str. 2, 24118 Kiel, Germany Berzelii Center EXSELENT on Porous Materials and Department of Materials and Environmental Chemistry, Stockholm University, Stockholm, S-106 91 Sweden School of Molecular Sciences, Arizona State University, Tempe, Arizona 85287-1604, U.S.A. *To whom correspondence should be addressed (Email: andrew.inge@mmk.su.se, mokeeffe@asu.edu, stock@ac.uni-kiel.de) Supporting Information S1
TABLE OF CONTENTS Figure S1. The asymmetric unit of CAU-17 Figure S2. Local environment around each bismuth cation Figure S3. The five independent channels in CAU-17 Figure S4. The 2-periodic htb net overlaid with symmetry elements and unit cell of CAU-17 Figure S5. Mismatch in the height of BTC 3- ligands in hypothetical links Figure S6. Misalignment of BTC 3- links in a hypothetical structure with smaller unit cell Table S1. Crystallographic data and refinement details on CAU-17 Structure determination and refinement details Rotation electron diffraction (RED) details Figure S7. Slices of the 3D reciprocal lattice of CAU-17 Figure S8. Unwarped images of the X-ray data Figure S9. The ab-plane of CAU-17 and the apparent smaller unit cell Figure S10. Observed and calculated X-ray powder diffraction patterns of CAU-17 Figure S11. Thermogravimetric analysis Figure S12. SEM micrograph of CAU-17 Synthesis motivation Table S2. Select bond lengths in CAU-17 References S3 S4 S5 S6 S7 S8 S9 S10 S11 S12 S12 S13 S14 S15 S16 S16 S17 S18 S2
Figure S1. The asymmetric unit of CAU-17. Thermal ellipsoids are drawn at a 50% probability level. Bismuth, oxygen, carbon and hydrogen atoms are colored purple, red, grey and white respectively. Hydrogen atoms are illustrated as spheres as they are described with isotropic displacement parameters. S3
Figure S2. Local environments around each independent Bi 3+ cation. Adjacent Bi 3+ cations are excluded for clarity. S4
Figure S3. The five symmetry-independent channels in CAU-17. Right and left handed helical rods are colored purple and yellow respectively. In both hexagonal channels neighboring rods are double linked by BTC 3- ions stabilized by π-π stacking forming double-walls around the channels. The rods in the triangular channel are single linked by BTC 3- ions resulting in a single wall. The two rectangular channels both have two single walls and two double walls formed by BTC 3- molecules. The RLRL rectangular channels have local 2-fold symmetry that is not expressed in the crystal symmetry. S5
Figure S4. The 2-periodic htb net overlaid with symmetry elements and unit cell edges of CAU-17. The purple and yellow nodes represent right and left handed rods in CAU-17 respectively. The five crystallographically independent channels are colored as in Figure S3. S6
Figure S5. Mismatch in the height of BTC 3- ligands. (a), (top) A triangular channel composed of three rods with handedness RRL. Such arrangements occur in the structure of CAU-17 and all BTC 3- ions connect the rods. (bottom) The purple and yellow nodes around the green triangle represent right and left handed rods respectively. The black edges of the triangle represent connected rods. (b), (top) A hypothetical triangular channel composed of three rods of the same handedness (RRR shown) leads to the misalignment of BTC 3- linkers between rods. The BTC 3- ions shown in blue in the foreground are aligned with the blue rod on the left side, but do not connect to the rod in green on the right side. Instead the green BTC 3- ions are in positions that connect with the green rod on the right side but not with the blue rods on the left side. The grey rod and BTC 3- ions in the background are properly connecting to the adjacent rods. (bottom) The red edge in the green triangle represents the misaligned BTC 3- link. (c), (top and bottom) The rods in the hexagonal channels in CAU-17 normally alternate in handedness and are double linked. (d), (top and bottom) Two adjacent rods of the same handedness in the hexagonal channel (LL is shown) would cause one of the two sets of BTC 3- linkers to be misaligned. The BTC 3- ions shown in blue in the foreground are aligned with the blue rod on the left side, but do not connect to the rod in green on the right side. Instead the green BTC 3- ions are in positions that connect with the green rod on the right side but not with the blue rods on the left side. The BTC 3- ions in the background are properly connecting to the adjacent rods. S7
Figure S6. Misalignment of BTC 3- links in a hypothetical structure with smaller unit cell. A hypothetical structure built of similar rods as those of CAU-17 packed in the htb net with only one independent hexagonal channel in the crystal structure would cause misalignment of BTC 3- molecules between rods. The non-connected rods are indicated as red edges. S8
Table S1. Crystallographic data and refinement details on CAU-17. identification code CAU-17 empirical formula [Bi 9 (C 9 H 3 O 6 ) 9 (H 2 O) 9 ] formula weight 434.11 temperature wavelength crystal system 150 K 0.6889 Å trigonal space group P3 (No. 147) unit cell dimensions a = 47.210 (2) Å c = 9.9682 (4) Å volume 19240 (2) Å 3 Z 6 density (calculated) 2.014 g/cm 3 absorption coefficient 11.070 mm -1 F(000) 10584 crystal size 0.40 0.02 0.02 mm 3 θ range for data collection 1.28 to 24.52 index ranges -10 h 56, -56 k 52, -10 l 11 reflections collected 46522 independent reflections 20983 [R(int) = 0.1621] absorption correction Multi-scan max. and min. transmission 0.339 and 1.000 data / restraints / parameters 20983 / 78 / 569 goodness-of-fit on F 2 1.043 final R indices [I>2σ(I)] R1 = 0.1061, wr2 = 0.2053 largest diff. peak and hole 3.983 and 3.367 e/å 3 S9
Structure determination and refinement details Single crystal X-ray diffraction data on a rod-shaped crystal of CAU-17 were collected on a Rigaku Saturn 724+ detector at 150 K using silicon double-crystal monochromated synchrotron radiation (λ = 0.6889 Å) at the Beamline I19, Diamond Light Source, Didcot, UK. Data reduction and absorption correction were applied using CrystalClear. Only the first 175 frames were included in the final data reduction and the exposure time was limited to one second per frame due to significant beam damage to the crystal throughout data acquisition. Beam damage had already occurred during the collection of these first 175 frames but the data where included in the data reduction and processing to allow for sufficient data completeness at the cost of higher R-values. Crystal structure determination of CAU-17 was initially impeded by the presence of twinning in all examined crystals. Rotation electron diffraction (RED) 1 on crushed fragments of CAU-17 confirmed the unit cell parameters (a = b = 47.21 Å, c = 9.97 Å, α = β = 90 and γ = 120 ) as determined by single crystal X-ray diffraction. The possibility of nonmerohedral twinning was eliminated suggesting that merohedral twinning was the likely reason for the complications. The crystal structure was solved by direct methods from single crystal X-ray diffraction data in the space group P3. All nine Bi 3+ ions in the asymmetric unit were located in the initial structure solution. The other framework atoms could not be located in the initial solution nor in difference Fourier maps after subsequent refinement without first applying a twin matrix corresponding to 2-fold along the c-axis. Upon realizing and applying the twin law, the positions of all nine BTC 3- anions and nine coordinated water molecules were immediately determined in the first subsequent difference Fourier map. Atoms were refined using a full-matrix least squares technique on F 2 with SHELXL. 2 All non-hydrogen atoms were refined with anisotropic displacement parameters. Methanol and water molecules are disordered in the pores of CAU-17 and confirmed by thermogravimetric analysis and elemental analysis. However, due to the merohedral twinning present in all examined crystals, SQUEEZE in the PLATON suite cannot be applied on the diffraction data by removing the unaccounted electron density from disordered guest species in the pores. Hence the R-values were not lowered in such a manner. A riding model was used to constrain the coordinates of hydrogen atoms bonded to aromatic carbon atoms of the BTC 3- linker. Anisotropic displacement parameters of carbon atoms and oxygen atoms on the same BTC 3- linker were constrained. Bond distance and bond angle restraints were applied to nine of the 27 carboxylate groups (see CIF). S10
Rotation Electron Diffraction (RED) details One of the advantages of using the RED technique is that smaller crystals could be studied in the TEM. Although the data was not sufficient for space group determination and structure solution of CAU-17, it provides evidences to support that nonmerohedral twinning did not exist in the crystal. Powders of the as-made CAU-17 were crushed, dispersed in absolute ethanol and treated by sonication for 2 minutes. A droplet of the suspension was then transferred onto a TEM copper grid. The TEM specimen was observed on a JEOL JEM-2100 microscopy operated at 200kV using a single-tilt tomography sample holder. The RED data sets were collected using the software package RED-data collection. 1 Due to the electron beam sensitivity of CAU-17 crystals, an average of 100 frames could be collected covering a tilt range up to 31 with an interval of 0.5 from a total of 32 studied crystals. The selected area electron diffraction (SAED) frames were combined to reconstruct the 3D reciprocal lattice using the software RED-data processing, which performed shift correction, peak search, unit cell determination, indexation of the reflections and intensity extraction. Since the CAU-17 crystals are electron beam sensitive, only small portions of 3D data could be obtained. By using limited information, the unit cell of CAU-17 could be confirmed. Figure S7a shows the hk0 slice of the 3D reciprocal lattice of CAU-17, the obtained apparent unit cell parameters attained from this slice alone was a = b = 27 Å, γ = 120 while the c, a and β parameters were unknown due to limited information. From a different crystal, the 0kl slice of the 3D reciprocal lattice was obtained (Figure S7b). The apparent unit cell deduced from the data suggests b = 47 Å, c = 10 Å, α = 90 and the parameters a, β and γ were unknown due to limited data. The hk0 slice suggests a smaller value for the unit cell parameter b (and also a) by a factor of 3. This was also observed in the X-ray data (Figure S8). This is due to the apparent reflection condition of h - k = 3n for hk0 reflections which is not a reflection condition of any space group we are aware of, except in rhombohedral lattice h - k = 3n applies to all hkl reflections, not just hk0 reflection as in CAU-17. The unusual reflection conditions and the discrepancies in the cell parameters a and b are best explained by viewing the structure along [001] (Figure S9). In this projection the unit cell appears to have the smaller values of the unit cell parameters a = b = 27 Å as outlined in blue. This is due to the fact that the handedness of the helical rods cannot be distinguished without information contained along the c-axis. Left and right-handed helices that run along that c-axis look identical viewed from a projection of the ab-plane. When left and right-handed helices cannot be distinguished all of the hexagonal channels appear identical to one another. Hence it appears that all hexagonal channels are centered around the corners of the unit cell and therefore only in the hk0 plane do the unit cell parameters appear to be a = b = 27 Å. All other reflections (hk-2, hk-1, hk1, hk2, etc.) indicate the true longer unit cell lengths a = b = 47 Å as they contain information regarding the handedness of the helices. S11
Figure S7. Slices of the 3D reciprocal lattice of CAU-17. (a), hk0 slice of the RED reconstruction taken from a typical CAU-17 crystal. The drawn reciprocal cell and the reciprocal cell axes are those of the apparent unit cell from this projection where a and b appear to be 27 Å. (b), 0kl slice of RED reconstruction taken from a different CAU-17 crystal where the length of the b-axis is seen correctly as 47 Å. Multiple scattering causes additional reflections that do not obey hk0: h - k = 3n to appear along 0k0. Figure S8. Unwarped images of the X-ray data. The hk0, hk1 and 0kl slices of reciprocal space are shown. Figures are drawn to scale with respect to one another. Comparison of the hk0 layer with that of the hk1 layer shows that two-thirds of the hk0 reflections (those that do not follow the condition hk0: h - k = 3n) have no clearly observable intensities. S12
Figure S9. The ab-plane of CAU-17 and the apparent smaller unit cell. The smaller apparent unit cell with a = b = 27 Å as determined from the hk0 plane from RED data is outlined in blue and the correct larger unit cell is outlined in black. S13
Figure S10. Observed and calculated X-ray powder diffraction patterns of CAU-17. Discrepancies in intensities are attributed to the disordered guest species in the pores that have not been accounted for in the crystal structure model. X-ray powder diffraction data were collected on a 0.5 mm capillary of CAU-17 with a STOE Stadi P with Cu K α1 radiation with a MYTHEN detector. S14
Figure S11. Thermogravimetric analysis of CAU-17. The first two steps until 220 C combined are attribute to the loss of three water molecules and one solvent methanol molecule per Bi 3+ cation (calc. 17.14% obs. 15.56%). The framework collapses and the BTC 3- are lost at 350 C (calc. 36.46% obs. 35.39%). S15
Figure S12. SEM micrograph of CAU-17. Scanning electron microscopy was performed on a Philips XL20 FEG microscope. Synthesis motivation The synthesis experiments of the title compound were inspired by two aspects of previous reports on bismuth carboxylates. Firstly, energy dispersive in-situ X-ray diffraction performed on dense bismuth 1,3,5-benzenetricarboxylates (BTC 3- ) during hydrothermal synthesis revealed the presence of shortlived crystalline intermediate phases whose structures provided insight towards the crystallization mechanisms. 3 This work has encouraged us to continue pursuing the prospect of discovering novel crystalline structures that are otherwise often overlooked due to the narrow time span in which they form under solvothermal synthesis conditions. Secondly, CAU-7 ([Bi(BTB)], BTB = 1,3,5- benzenetrisbenzoate), the first reported permanently porous Bi-MOF, forms in methanol under short reaction times (~20 min). 4 The title compound was synthesized using methanol as a solvent and BTC 3- as the linker with relatively short synthesis time. S16
Table S2. Select bond lengths in CAU-17. Atom 1 Atom 2 Dist.(Å) Atom 1 Atom 2 Dist.(Å) Atom 1 Atom 2 Dist.(Å) Bi1 O75 2.20(3) Bi5 O53 2.24(3) Bi9 O06 2.21(3) O72 2.36(3) O56 2.26(3) O03 2.30(3) O73 2.36(3) O42 2.36(3) OW9 2.44(5) O74 2.38(3) O55 2.39(3) O52 2.48(4) OW1 2.40(6) OW5 2.46(5) O04 2.52(3) O71 2.57(3) O41 2.56(4) O51 2.52(3) O76 2.64(3) O15 2.61(4) O05 2.58(3) O86 2.73(4) O54 2.62(3) O34 2.74(5) O44 3.02(2) O63 2.94(3) O23 2.87(2) Bi2 O82 2.23(4) Bi6 O64 2.29(3) O85 2.31(4) O61 2.29(3) O01 2.36(3) O31 2.39(5) O02 2.40(3) OW6 2.45(4) O86 2.50(3) O32 2.55(4) OW2 2.50(3) O62 2.58(3) O81 2.52(3) O63 2.62(3) O45 2.69(4) O55 2.80(3) O76 2.98(4) O14 3.09(4) Bi3 O46 2.29(3) Bi7 O33 2.23(4) O43 2.29(3) O36 2.24(4) OW3 2.30(5) O65 2.41(3) O84 2.47(4) O66 2.46(3) O83 2.47(4) O34 2.52(3) O44 2.49(3) OW7 2.54(2) O45 2.54(3) O35 2.64(3) O71 2.76(3) O22 3.00(4) O81 2.95(4) O05 3.01(4) Bi4 O16 2.19(4) Bi8 O22 2.18(3) O13 2.29(4) O21 2.24(3) O25 2.35(3) O24 2.27(3) O26 2.41(2) O11 2.33(5) OW4 2.41(3) O23 2.53(3) O14 2.47(3) OW8 2.55(5) O15 2.59(3) O12 2.56(4) O62 2.66(5) O04 2.71(2) O54 2.84(4) O35 2.88(5) S17
References (1) Wan, W.; Sun, J.; Su, J.; Hovmöller, S.; Zou, X. J. Appl. Crystallogr. 2013, 46, 1863. (2) Sheldrick, G. Acta Crystallogr. Sect. A: Found. Crystallogr. 2008, 64, 112. (3) Feyand, M.; Köppen, M.; Friedrichs, G.; Stock, N. Chem. Eur. J. 2013, 19, 12537. (4) Feyand, M.; Mugnaioli, E.; Vermoortele, F.; Bueken, B.; Dieterich, J. M.; Reimer, T.; Kolb, U.; De Vos, D.; Stock, N. Angew. Chem. Int. Ed. 2012, 51, 10373. S18