Efficient Photoluminescence via Metal-Ligand Alteration in a New MOFs Family

Transcription

1 Supporting Information for: Efficient Photoluminescence via Metal-Ligand Alteration in a New MOFs Family Dorina F. Sava Gallis, Lauren E. S. Rohwer, Mark A. Rodriguez, and Tina M. Nenoff Nanoscale Sciences Department, Sandia National Laboratories, Albuquerque, NM 87185, USA. Microsystems Integration Department, Sandia National Laboratories, Albuquerque, NM 87185, USA. Materials Characterization Department, Sandia National Laboratories, Albuquerque, NM 87185, USA. S1

2 Table of Contents Section S1: Powder X-Ray Diffraction (PXRD) measurements Section S2: Thermogravimetric (TGA) analyses Section S3: SEM-EDS analysis Section S4: X-ray single-crystal data collection and determination Section S5: References S3 S9 S14 S15 S25 S2

3 Section S1: Powder X-Ray Diffraction (PXRD) measurements Powder X-Ray Diffraction (PXRD) measurements were performed on a Siemens Kristalloflex D500 diffractometer, CuKα radiation (λ = Å). Figure S1. Calculated (red) and experimental (blue) PXRD pattern for compound 1. S3

4 Figure S2. Calculated (red) and experimental (blue) PXRD pattern for (a) as synthesized and (b) thermally treated at 150 C for 10 hours compound 2. (a) As-synthesized (b) Experimental Calculated Intensity (a.u.) After thermal treatment at 150ºC, 10 hours theta/ 0 S4

5 Figure S3. Calculated (red) and experimental (blue) PXRD pattern for compound 3. S5

6 Figure S4. Calculated (red) and experimental as synthesized (magenta) and after thermal treatment at 150 C for 10 hours (blue) PXRD pattern for compound 4. S6

7 Figure S5. Experimental (blue) PXRD pattern for compound 5. Table S1. PXRD peak determination for compound 5. 2-Theta d(å) BG Height H% Area A% FWHM S7

8 In order to gather additional structural information about compound 5, additional analyses were performed. Microelemental analyses confirmed an equimolar metal to ligand ratio, as well as the anticipated guest exchange (DMA/MeOH with EtOH/H 2 O). Elemental analyses were conducted by the Galbraith Laboratories, Inc. on as synthesized and EtOH-exchanged compound 5. Suggested formula (as synthesized): Eu(TCBB)(DMA) 4.5 (MeOH); Found: C, 60.99; H, 5.64; N, 5.03; Eu, Calculated: C, 61.8; H, 5.79; N, 5.07; Eu, Suggested formula (EtOH exchanged): Eu(TCBB)(EtOH) 2 (H 2 O) 4 Found: C, 60.68; H 4.16; N <0.5; Eu, 15.4 Calculated: C, 59.8; H, 4.7; Eu, 15.4 FT-IR spectra of TCBB ligand (red) and compound 5 (blue). The data was recorded on a Thermo Scientific Nicolet is10 FT-IR Spectrometer. Compound 5 TCBB ligand S8

9 Section S2: Thermogravimetric analyses The structural thermal stability was investigated by TGA-DSC. Measurements used a SDTQ600 TA instrument. Samples were heated at 10 C/min to 800 C under nitrogen flow. Figure S6. TGA (red trace) and DSC (blue trace) for compound 1. S9

10 Figure S7. TGA (red trace) and DSC (blue trace) for compound 2. S10

11 Figure S8. TGA (red trace) and DSC (blue trace) for compound 3. Figure S. TGA (red trace) and DSC (blue trace) for compound 3. S11

12 Figure S9. TGA (red trace) and DSC (blue trace) for compound 4. S12

13 Figure S10. TGA (red trace) and DSC (blue trace) for compound 5. S13

14 Section S3: SEM-EDS analysis Scanning Electron Microscopy (SEM) images were captured on a FEI NovaNano SEM 230, and Energy Dispersive Spectroscopy (EDS) analyses were collected on a EDAX Genesis Apex 2 with an Apollo SDD detector. Iodine loading in compound 5 An activated sample of 5 (sample exchanged in EtOH for 24 hours and degassed at room temperature for 18 hours on a Micromeritics ASAP 2020 surface area and porosity analyzer) was sealed with excess I 2 in an adsorption chamber and placed in a convection oven at 75 C. The iodine loaded sample was recovered, allowed to cool in an inert environment and the mass change recorded. Figure S11. SEM-EDS analyses of I 5. S14

15 Section S4: X-ray single-crystal data collection and determination The X-ray intensities were measured using a Bruker-APEX/CCD diffractometer (Mo Kα,λ = Å). Indexing and frame integration was performed using the APEX-II software suite. S1 Absorption correction was performed using face-indexing (numerical method) also within the APEX-II software. The structures were solved and refined using SHELXS-97 contained in SHELXTL v6.10 packages. S2 The hydrogen atoms were placed in geometrically calculated positions and included in the refinement process using a riding model. Since many MOFs are inherently low-density materials, it is not uncommon to have significantly reduced X-ray scattering, especially at the higher theta angles. The main focus of the reported MOF structures in this manuscript is the establishment of the crystalline framework and connectivity of the metal nodes via the linker molecules. The main drawback in regard to the missing high-angle data for a structure typically has to do with reduced confidence in bond length values in particular for the shortest of these bonds (e.g. C-C ~1.4 Angstroms). This concern has been addressed in our datasets through the use of constraints/restraints. Since the linker molecules most susceptible to refinement error are well-known, S3 these molecules can be constrained in a reasonable fashion (i.e. equal C-C distance in 6 membered rings) so as to improve the overall refinement stability. While this is not ideal, it is not an uncommon tactic for other large molecules with low density (e.g. proteins). S4 The restraining of the linker bond lengths/angles facilitates the refinement of a reasonable structure in the absence of meaningful high angle data. Determinations of the theta max cutoff for each dataset were made based on usefulness of the observed scattered intensities above the background level. Further increase of the analysis to larger theta max values only served to decrease the quality of the refinement and subsequently move the R value higher. S15

16 Crystal data for compound 1, (SMOF-2) The BTB ligand illustrated some difficulty in refinement of a few carbon atoms. Since the BTB molecule is fairly well defined as containing 4 benzene rings, the carbon atoms within the problematic ring (C57 > C62) were constrained to exhibit the expected bond lengths (and angles). In addition, this troublesome ring was assumed to be planar, and was constrained to be flat (using the FLAT command). Also, the bond distances between the O and C atoms at the termination of the BTB ligand were constrained (i.e. O13 and O14 to C63). The FLAT command was extended from the C57 > C62 ring to include the C63, O13 and O14 atoms. Note that the most difficult ring to refine with stability was terminated by the monodentate bonding (where the terminal or non-bound oxygen was the O14). The O14 is an OH group but due to the low resolution of the data it was difficult to locate the hydrogen atom. Hence, the hydrogen atom was left off of the O14 oxygen in the refinement, but was accounted for in the formula of the final refinement. The problematic nature of the C57 > C62 ring and its subsequent terminal oxygen branches suggests that this particular ring was subject to some disorder due to the absence of a more rigid bidentate bonding configuration. Elsewhere in the structure some additional bond length and FLAT constraints were added to improve the bidentate bonding of the O9-C36-O10 bonding. The above listed constraints added considerable stability to the refinement. In addition to the OH group, there were two terminal water molecules in the structure that coordinate to the In metal center. These water molecules are associated with O15 and O16 oxygen atoms. The hydrogen atoms present on these water molecules were not included in the structure model, but were added to the total formula in the final refinement cycle, just as in the case of the O14 OH hydrogen. There was significant scatter observed within the large channels/voids in the structure; this suggested the presence of solvent in the void space. Solvent density was detected in at least three separate crystallographic locations. We list these three center-of-density locations but acknowledge that there are likely more sites necessary to completely elucidate the modeled solvent density. The fractional coordinates for the sites are as follows: Site 1 at ( ), Site 2 at ( ), and Site 3 S16

17 at ( ). Attempts to refine the highly disordered solvent were unproductive. Therefore, the program Platon/Squeeze S5 was employed to model the solvent presence. Thermogravimetric analysis (TGA) suggested that the solvent loss was from diethylformamide (DEF) which was employed in the crystallization process. The application of the Squeeze routine reported solvent accessible void space of Å 3 containing 6073 electrons/cell. Assuming all the density in the solvent voids was from DEF molecules, this roughly translated to approximately 108 DEF molecules per cell. Association of sites 1, 2 and 3 with DEF molecules would account for 54 of the 108 DEF molecules (or half of the solvent density predicted by the electron count determined by Squeeze). The remaining solvent sites proved more difficult to locate due a lack of clearly located electron density nodes. This suggests that the solvent is highly disordered at these alternative sites, resulting in smearing of the electron density. Even so, the Squeeze process was still able to model this density and account for the total DEF presence. The estimated weight of DEF present within the structure (based on the electron density of DEF detected via Squeeze) was 31 wt%. This is in good agreement with TGA analysis. Refinement of the structure after solvent modeling resulted in a significant drop in residual error. While the final structure factors do not reflect the now absent DEF solvent, the final refinement included the additional DEF chemical species within the chemical formula and the reported crystal data reflect the presence of 6 DEF molecules per formula unit for compound 1 reported in this manuscript. S17

18 Table S2. Crystal data and structure refinement for compound 1. Empirical formula C63 H40 Eu In O16, 6(C5 H11 O N) Formula weight Temperature Wavelength 193 K Å Crystal system, space group Trigonal, R-3 Unit cell dimensions a = (6) Å alpha = 90 b = (6) Å beta = 90 c = (5) Å gamma = 120 Volume 60330(14) Å 3 Z, Calculated density 18, Mg/m 3 F(000) Crystal size 0.46 x 0.36 x 0.25 mm Theta range for data collection 1.24 to Reflections collected / unique / 6776 R indices R1 = , wr2 = Largest diff. peak and hole and eå -3 S18

19 Crystal data for compound 3, (SMOF-5) The ligand illustrated some difficulty in refinement of carbon atoms. Since this molecule is fairly well defined as containing 3 benzene rings, connected via the central C3N3 ring, all the rings in both ligands were restrained to have the planar aromatic-type bonding. In addition, oxalate bridging ligands were also constrained to have the expected C-C, O-C bond lengths. This added considerable stability to the refinement without causing a significant increase in the overall residual error. There were two OH groups present in the lattice due to monodentate coordination of the linker molecule. These OH groups were associated with O14 and O18 oxygen atoms. The hydrogen atoms were left off of the O14 and O18 oxygen atoms in the refinement, but were accounted for in the formula of the final refinement. There was significant scatter observed within the large channels/voids in the structure; this suggested the presence of solvent in the void space. Solvent density was detected in at least 8 separate crystallographic locations. We list these eight center-of-density locations but acknowledge that there are likely more sites necessary to completely elucidate the modeled solvent density. The fractional coordinates for the sites are as follows: Site 1 at (0, 0.75, ), Site 2 at (0.542, 0.678, 0.259), Site 3 at (0.871, 0.786, 0.460), Site 4 at (0.737, 0.934, 0.187), Site 5 at (0.25, 0.873, 0.5), site 6 at (0, 0, 0), Site 7 at (0.145, 0.957, 0.429), Site 8 at (0.102, 0.977, 0.75). Attempts to refine the highly disordered solvent were unproductive. Therefore, the program Platon/Squeeze S3 was employed to model the solvent presence. Thermogravimetric analysis (TGA) suggested that the solvent loss was from dimethylformamide (DMF) which was employed in the crystallization process. The application of the Squeeze routine reported solvent accessible void space of Å 3 containing 4827 electrons/cell. Assuming all the density in the solvent voids was from DMF molecules, this roughly translated to approximately 120 DMF molecules per cell. Association of sites listed above with DMF molecules would account for 88 of the 120 DMF molecules (or about 70 percent of the solvent density predicted by the electron count determined by Squeeze). The remaining solvent sites proved more difficult to located due a lack of clearly located electron density nodes. This suggests that the S19

20 solvent is highly disordered, resulting in smearing of the electron density. Even so, the Squeeze process was still able to model this density and account for the total DMF presence. The estimated weight of DMF present within the structure (based on the electron density of DMF detected via Squeeze) was 27 wt%. This is in good agreement with TGA analysis. Refinement of the structure after solvent modeling resulted in a significant drop in residual error. While the final structure factors do not reflect the now absent DMF solvent, the final refinement included the additional DMF chemical species within the chemical formula and the reported crystal data reflect the presence of 15 DMF molecules per formula unit for compound s 3 structure. S20

21 Table S3. Crystal data and structure refinement for compound 3. Empirical formula C108 H64 In6 N12 O48, 15(C3 H7 N O) Formula weight Temperature Wavelength Crystal system, space group 193 K Å Orthorhombic, I b c a Unit cell dimensions a = (18) Å alpha = 90 b = Å beta = 90 c = (19) Å gamma = 90 Volume 37501(29) Å 3 Z, Calculated density 8, Mg/m 3 F(000) Crystal size 0.18 x 0.15 x 0.12 mm Theta range for data collection 1.19 to Reflections collected / unique 3611 / 2204 R indices R1 = , wr2 = Largest diff. peak and hole and eå -3 S21

22 Crystal data for compound 4, (SMOF-7) Special consideration was given to the O7 atom which appeared to be coordinated by a diethylformamide (DEF) molecule. This molecule showed significant disorder, but one orientation appeared to be the most consistent. Bond lengths and angles for a DEF molecule were employed to constrain this DEF molecule to coordinate to the metal site. Restraints regarding isotropic temperature parameters for this DEF molecule were also employed. It is important to bear in mind that this orientation is one of many possible orientations. There was a second oxygen atom, the O8 oxygen site, that might possibly be associated with a second coordinated DEF, but all attempts to model such a molecule to attach to the O8 atom were unsuccessful. It appeared that there was not sufficient clearly defined electron density to warrant placement of a DEF molecule off of the O8 atom. So in the case of O8 it was assumed that this terminal oxygen was a water molecule. The hydrogen atoms were left off of the O8 water as well as the DEF molecule coordinated via O7. However, these hydrogen atoms were included in the final refinement to generate proper crystallographic output. There was significant scatter observed within the large channels/voids in the structure; this suggested the presence of solvent in the void space. Solvent density was detected in at least 2 separate crystallographic locations: Site 1 at (0.598, 0.130, 0.516), Site 2 at (0.730, 0.126, 0.697). These were determined via significant Q peaks present in the Difference-Fourier maps. The program Platon/Squeeze S3 was employed to model the solvent presence. Thermogravimetric analysis (TGA) suggested that the mass loss was from the DEF solvent which was employed in the crystallization process. The application of the Squeeze routine reported solvent accessible void space of Å 3 containing 1115 electrons/cell. Assuming all the density in the solvent voids was from DEF molecules, this translated to approximately 20 DEF molecules per cell as uncoordinated solvent. Association of sites listed above with DEF molecules would account for 16 of the 20 DEF solvent molecules (or about 75 percent of the solvent density predicted by the electron count determined by Squeeze). The remaining solvent sites in the channels proved more difficult to locate due a lack of clearly located electron density nodes. This suggests that the solvent is highly S22

23 disordered, resulting in smearing of the electron density. Even so, the Squeeze process was still able to model this density and account for the total DEF presence via the electron count. The estimated weight of DEF present within the structure (based on the electron density of DEF detected via Squeeze) was 26.5 %. The inclusion of the coordinated solvent (O7) and the O8 water molecules resulted in an increase to 43.9 weight percent. This is in good agreement with the TGA analysis. Refinement of the structure after solvent modeling resulted in a significant drop in residual error. While the final structure factors do not reflect the now absent DEF solvent molecules, the final refinement included the additional DEF chemical species within the chemical formula and the reported crystal data reflect the presence of 7 DEF molecules per formula unit for the compound 4 structure. Of these DEF molecules, 2 of them are due to the O7 coordinated DEF and there are 5 additional DEF solvent molecules present as uncoordinated solvent. S23

24 Table S4. Crystal data and structure refinement for compound 4. Empirical formula C48 H30 N6 O12 Eu2, 7(C5 H11 N O), 2(H2O) Formula weight Temperature Wavelength 193 K Å Crystal system, space group Monoclinic, C 2/c Unit cell dimensions a = (8) Å alpha = 90 b = (3) Å beta = (3) c = (5) Å gamma = 90 Volume 8515(4) Å 3 Z, Calculated density 4, Mg/m 3 F(000) 3976 Crystal size 0.15 x 0.10 x 0.02 mm Theta range for data collection 1.29 to Reflections collected / unique 2548 / 1807 R indices R1 = , wr2 = Largest diff. peak and hole and eå -3 S24

25 Section S5: References S1. Bruker AXS, Inc. APEX2 ver Madison, Wisconsin, USA S2. (a) Sheldrick, G. M. Acta Cryst. 2008, A64, ; (b) Bruker AXS, Inc. XSHELL Version 4.01 and SHELXTL Version Madison, WI S3. Furukawa, H.; Ko, N.; Go, Y. B.; Aratani, N.; Choi, S. B.; Choi, E.; Yazaydin, A. O.; Snurr, R. Q.; O Keeffe, M.; Kim, J.; Yaghi, O. M. Science 2010, 329, 424. S4. (a) Chao, J. A.; Williamson, J. R. Structure 2004, 12, ; (b) Von Dreele, R. B. ; Stephens, P. W.; Smith, G. D.; Blessing, R. H. Acta Cryst. 2000, D56, ; (c) Prabhakaran, P.; Kale, S.S.; Puranik, V.G.; Rajamohanan, P. R.; Chetina, O.; Howard, J.A. K.; Hofmann, H.-J.; Sanjayan, G.J. J. Am. Chem. Soc. 2008, 130, S5. Spek, A. L. Platon, A Multipurpose Crystallographic Tool, Utrecht University, Utrecht, The Netherlands S25