SUPPLEMENTARY INFORMATION

Transcription

1 DOI: 1.138/NCHEM.1457 Selectivity and Direct Visualization of Carbon Dioxide and Sulfur Dioxide in a Decorated Porous Host Sihai Yang, 1 * Junliang Sun, 2 Anibal J. Ramirez-Cuesta, 3 Samantha K. Callear, 3 William I.F. David, 3,4 Daniel Anderson, 1 Ruth Newby, 1 Alexander J. Blake, 1 Julia E. Parker, 5 Chiu C. Tang 5 and Martin Schröder 1 * [ 1 ] School of Chemistry, University of Nottingham, University Park, Nottingham, NG7 2RD (UK) Fax: Sihai.Yang@nottingham.ac.uk; M.Schroder@nottingham.ac.uk [ 2 ] College of Chemistry and Molecular Engineering, Peking University, Beijing, 1871 (China) [ 3 ] ISIS Facility, Rutherford Appleton Laboratory, Chilton, Oxfordshire, OX11 QX (UK) [ 4 ] Inorganic Chemistry Laboratory, University of Oxford, South Parks Road, OX1 3QR (UK) [ 5 ] Diamond Light Source, Harwell Science and Innovation NATURE CHEMISTRY 1

2 Index 1. Experimental Section 1.1 Physical Characterisation 1.2 Gas Adsorption Isotherms 1.3 Inelastic Neutron Scattering 1.4 DFT Modelling and Simulations 2. In Situ Synchrotron X-Ray Powder Diffraction Patterns 3. Transmission Electron Microscopic (TEM) Study 4. Additional View of Crystal Structures 5. TGA plot for NOTT-3-solvate 6. Synchrotron Powder Diffraction Studies of Solvated NOTT-3 7. Variable Temperature Powder Diffraction 8. Additional Gas Sorption Isotherm Plots for NOTT-3 9. Exposure of NOTT-3 to water 1. Additional Inelastic Neutron Scattering Spectra 11. Analysis and Derivation of the Isosteric Heat of Adsorption for 12. Calculation of Henry s Law selectivity for gas orption 13. Summary of the hydrogen bond interaction 14. Film of the Dynamics of the Crystal Lattice upon Inclusion in NOTT References NATURE CHEMISTRY 2

3 1. Experimental Section 1.1 Physical Characterisation. All reagents were used as received from commercial suppliers without further purification. Analyses for C, H and N were carried out on a CE-44 elemental analyzer (EAI Company). Thermal gravimetric analyses (TGA) were performed under N 2 flow (1 ml/min) with a heating rate of 2 C/min using a TA SDT-6 thermogravimetric analyzer (TA Company). IR spectra were recorded using a Nicolet Avatar 36 FT-IR spectrophotometer. High-resolution transmission electron microscopy (TEM) imaging was performed using a Jeol 21F transmission electron microscope using an accelerating voltage of 1 kv. TEM samples were prepared by casting several drops of a suspension of the NOTT-3-solvate complex in water onto copper-grid mounted lacy carbon film before drying under a stream of nitrogen. Variable temperature powder X-ray diffraction data (PXRD) were collected over the 2θ range 4-5 o on a Bruker Advance D8 diffractometer using Cu-Kα 1 radiation (λ = Å, 4 kv/4ma), and the temperature was controlled by an Oxford Cryosystems open-flow cryostat operating at K. 1.2 Gas Adsorption Isotherms., SO 2, CH 4,CO,N 2,O 2,H 2 and Ar sorption isotherms were recorded at 77 K (liquid nitrogen), 87 K (liquid argon) or K (temperature-programmed water bath from Hiden Company) on an IGA- 3 system at the University of Nottingham under ultra-high vacuum from a diaphragm and turbo pumping system. All gases used were ultra-pure research grade (99.999%) purchased from BOC or AIRLIQUIDE. The density of the desolvated NOTT-3 sample used in buoyancy corrections was 1.8 g cm -3 and was estimated from the crystallographic density of the desolvated sample derived from the PLATON/SOLV 1 results. In a typical gas orption experiment, ~1 mg of NOTT-3-solvate was loaded into the IGA, and degassed at 12 o C and high vacuum (1-1 bar) for 1 day to give fully desolvated NOTT Inelastic Neutron Scattering. INS spectra were recorded on the TOSCA spectrometer at the ISIS Facility at the Rutherford Appleton Laboratory (UK) for energy transfers between ~-2 and 5 mev. In this region TOSCA has a resolution of ~1% ΔE/E. The desolvated NOTT-3 sample was loaded into a cylindrical vanadium sample NATURE CHEMISTRY 3

4 container with an annealed copper vacuum seal and connected to a gas handling system. The sample was degassed at 1-7 mbar and 14 C for 1 day to remove any remaining trace guest water molecules. The temperature during data collection was controlled using a helium cryostat (7 ±.2 K). The loading of was performed at room temperature in order to ensure that was present in the gas phase when not orbed and also to ensure sufficient mobility of inside the crystalline structure of NOTT-3. The loading of H 2 was performed at 4-5 K in order ensure that H 2 was orbed into NOTT-3. Subsequently, the temperature was reduced to below 1 K in order to perform the scattering measurements with the minimum achievable thermal motion for or H 2. Background spectra (sample can plus NOTT-3) were subtracted to obtain the difference spectra. INS was used to study the binding interaction and structure dynamics in this case, because it has several unique advantages: INS spectroscopy is ultra-sensitive to the vibrations of hydrogen atoms, and hydrogen is ten times more visible than other elements due to its high neutron cross-section. The technique is not subject to any optical selection rules. All vibrations are active and, in principle, measurable. INS observations are not restricted to the centre of the Brillouin zone (gamma point) as is the case for optical techniques. INS spectra can be readily and accurately modelled: the intensities are proportional to the concentration of elements in the sample and their cross-sections, and the measured INS intensities relate straightforwardly to the associated displacements of the scattering atom. Treatment of background correction is also straightforward. Neutrons penetrate deeply into materials and pass readily through the walls of metal containers making neutrons ideal to measure bulk properties of this material. INS spectrometers cover the whole range of the molecular vibrational spectrum, -5 mev (-4 cm -1 ) INS data can be collected at below 1 K, where the thermal motion of the MOF material and orbed molecules can be significantly reduced. NATURE CHEMISTRY 4

5 1.4 DFT Modelling and Simulations. The vibrational properties of the NOTT-3 were calculated using a combination of density functional theory (DFT) and plane-wave pseudopotential methods as implemented in the CASTEP code, 2 using ultra-soft pseudopotentials with a plane-wave energy cutoff of 38 ev. Calculations were performed under the PBE approximation 3 for exchange and correlation. The unit cell used has a volume of Å 3 and contains 144 and 156 atoms for the bare and -loaded materials, respectively. The wave functions were sampled according to the Monkhorst-Pack scheme with a k-points mesh of spacing ~.5 Å -1. The normal modes of the solid were determined from dynamical matrices calculated using finite displacements, by numerical differentiation. The INS spectra was the calculated using the aclimax software. 4 The calculation of the bare material is computationally stable, in the calculation of the vibrational frequencies, all the frequencies are positive. In the calculation of the MOF material loaded with four molecules per unit cell (Fig. S1), corresponding to NOTT-3 1., 8 imaginary frequency modes were found. Following the methodology previously used in MOF modelling, 5 we calculated the potential energy surface (PES) along each of the negative frequency modes (corresponding to the rocking motion of the unit). The obtained PES (Fig. S1) confirms that the zero-point energy (ZPE) level (the ground state) lies at an energy above the local minimum calculated by DFT. This means that these PES, although corresponding to an unstable structure from a static classical point of view, are indeed stable structures from the quantum mechanical view. The location of SO 2 molecules in NOTT-3 has been optimised by DFT modelling based on the measured INS spectra. The DFT calculation was performed using the same settings as in the case of. Notably, a high symmetry configuration, which does not take into account the disorder of SO 2 molecules in the channel, was used. NATURE CHEMISTRY 5

6 Figure S1. The potential energy surface along the first vibrational eigenvector. At the centre it is clear that the potential would give an unstable configuration from a classical point of view. The calculated solution of the Schrödinger equation gives the energy levels at 14.2 and 81 cm -1. The ZPE is higher than the local maximum so that from a quantum point of view, the system is in a stable equilibrium. NATURE CHEMISTRY 6

7 2. In Situ Synchrotron X-Ray Powder Diffraction Patterns Intensity Observed Calculated Difference degrees a Intensity Observed Calculated Difference degrees b Figure S2. (a) PXRD patterns [observed (blue), calculated (red) and difference (grey)] for the Rietveld refinement of the as-synthesized NOTT-3-solvate [λ = (2) Å]; (b) higher angle data (2θ = o ) scaled up to show the quality of the fit between the observed and the calculated patterns. NATURE CHEMISTRY 7

8 Intensity Observed Calculated Difference degrees a Intensity Observed Calculated Difference degrees b Figure S3. (a) PXRD patterns [observed (blue), calculated (red) and difference (grey)] for the Rietveld refinement of the -loaded NOTT [λ = (2) Å]; (b) higher angle data (2θ = o ) scaled up to show the quality of fit between the observed and the calculated patterns. NATURE CHEMISTRY 8

9 Intensity Observed Calculated Difference degrees a Intensity Observed Calculated Difference degrees b Figure S4. (a) PXRD patterns [observed (blue), calculated (red) and difference (grey)] for the Rietveld refinement of the SO 2 -loaded NOTT-3 4SO 2 [λ = (2) Å]; (b) higher angle data (2θ = o ) scaled up to show the quality of fit between the observed and the calculated patterns. NATURE CHEMISTRY 9

10 regenerated NOTT-3 after SO 2 removal NOTT-3.SO 2 1. bar SO 2 NOTT-3.SO 2.5 bar SO 2 NOTT-3 NOTT-3-solv Intensity degrees a regenerated NOTT-3 after SO 2 removal NOTT-3 NOTT-3.SO 2.5 bar SO 2 NOTT-3-solv NOTT-3.SO 2 1. bar SO 2 Intensity degrees b Figure S5. (a) Comparison of the powder diffraction patterns for original, evacuated, SO 2 -loaded, and final desolvated samples at 273 K; (b) higher angle data (2θ = o ) has been scaled up to show the changes upon SO 2 inclusion. NOTT-3 retains crystallinity after removal of SO 2. NATURE CHEMISTRY 1

11 regenerated NOTT-3 after removal NOTT bar Intensity NOTT-3..5 bar NOTT-3 NOTT-3-solv degrees a regenerated NOTT-3 after removal NOTT-3 NOTT-3..5 bar NOTT-3-solv NOTT bar Intensity degrees b Figure S6. (a) Powder X-ray diffraction patterns for original, evacuated, -loaded, and final desolvated samples at 273 K; (b) higher angle data (2θ = o ) has been scaled up confirming that NOTT-3 retains crystallinity on removal of. NATURE CHEMISTRY 11

12 Table S1. Atomic positions for non-hydrogen atoms in NOTT-3-solvate. x y z B iso (Å 2 ) Al (5).3574(5).5.72(13) O (13) (12) O (9).28675(9).99869(15).51(12) O (1).37371(1).39436(12).51(12) C1.5928(15).3639(14).69453(19).7(13) C (1).43181(7).75722(1).7(13) C (14).7(13) C4.5471(1) (71).87716(8).7(13) C (11).7(13) O1w.918(23).8489(16).9477(28) 21.6(2) O2w.45734(19).73953(3).6123(32) 11.9(3) O3w.1341(22).29586(24).53782(37) 35.7(3) NATURE CHEMISTRY 12

13 Table S2. Atomic positions for NOTT X y z B iso (Å 2 ) Al1.3694(4) (4).5.93(4) O (13) (4) H1.8123(17) (4) O (22).37975(22).6276(28).82(4) O3.6511(22).2838(22).7532(28).82(4) C (19).3644(15).7222(15) 1.37(4) C2.5478(8).433(6).76456(13) 1.37(4) C (14) 1.37(4) H (7) 1.37(4) C4.5478(8).433(6).88199(13) 1.37(4) H4.5675(4).3842(6).925(6) 1.37(4) C (14) 1.37(4) C1_ (25) (12) 15.1(2) O1_1.9673(23) (5) O2_ (22) (24) 39.4(8) C1_ (2) O1_ (8) NATURE CHEMISTRY 13

14 Table S3. Atomic positions for NOTT-3 4SO 2. X y z B iso (Å 2 ) Al1.3668(4) (4).5.83(11) O1.7593(11) (11) H1.894(7).2737(13).6182(19).75(17) O2.8768(8).12169(8).1281(9).5(11) O (7).21365(7).2536(13).5(11) C (12).9466(13).79858(16).9(12) C (11).4384(7).75958(9).9(12) C (1).9(12) H (7) 1.9(14) C4.545(7).4296(5).87811(9).9(12) H (14) C (9).9(12) S (2) O1s (2) O2s (8) S (2) O3s (8) O4s (4) NATURE CHEMISTRY 14

15 3. Transmission Electron Microscopic (TEM) Study. A TEM image shows the crystals to have uniform morphology of ~1 μm plates (Fig. S7a,b), and a high resolution (HRTEM) image confirms the presence of extended crystalline planes (Fig. S7c). a b Figure S7. TEM images for NOTT-3-solvate. c NATURE CHEMISTRY 15

16 4. Additional View of Crystal Structures. Figure S8. View of the structure of NOTT-3-solvate along the a-axis. a Figure S9. View of the structure of NOTT-3 1. along the c-axis (a) and a-axis (b). The structure was obtained by DFT simulation. The orbed molecules in the pore channel are highlighted by the use of spacing filling style. b NATURE CHEMISTRY 16

17 a b c Figure S1. Detailed views of -OH and -CH groups binding in the pocket cavity of NOTT Views along (a) the a-axis, (b), the b-axis and (c) the c-axis. The moderate hydrogen bond between O(δ-) of and H(δ+) of -OH is highlighted in cyan, [O H = Å]. The weak cooperative hydrogen bond between O(δ-) of and H(δ+) from -CH is highlighted in purple, [O H = 3.29, 3.19Å with each occurring twice]. Therefore, each O(δ-) centre interacts with five different H(δ+) centres. a Figure S11. View of the structure of NOTT-3 4.SO 2 along the c-axis (a) and a-axis (b). The structure was obtained by Rietveld refinement of high resolution powder diffraction data collected for NOTT-3 4.SO 2. The orbed SO 2 molecules in the pore channel are highlighted by the use of spacing filling style. b NATURE CHEMISTRY 17

18 a b c Figure S12. Detailed views of -OH and -CH groups binding SO 2 in the pocket cavity of NOTT- 3 4.SO 2. Views along (a) the a-axis, (b), the b-axis and (c) the c-axis. The modest hydrogen bond between O(δ-) of SO 2 (I) and H(δ+) from -OH is highlighted in cyan, [O H = 2.376(13) Å]. The weak cooperative hydrogen bond between O(δ-) of SO 2 and H(δ+) from -CH is highlighted in purple, [O H = 2.86(14), 2.841(17), 3.111(16), 3.725(18) Å]. Therefore, each O(δ-) centre is interacting with five different H(δ+) centres. The bond distance between S(δ+) of SO 2 (I) and O(δ-) of SO 2 (II) is 3.34(7) Å and highlighted in blue. The S-O bond distances are 1.481(4) and 1.5(8) Å for SO 2 (I) and SO 2 (II), respectively. The <O-S- O angles are 117.5(11) and 19.1(9) for SO 2 (I) and SO 2 (II), respectively. NATURE CHEMISTRY 18

19 5. TGA plot for NOTT-3-solvate. A TGA plot shows that the as-synthesised sample NOTT-3-solvate loses solvent rapidly between 3 and 1 C, with a plateau observed from 1-2 C indicating no further weight loss to give NOTT-3. The weight loss of 2. % from NOTT-3-solvate between 2 and 2 C corresponds to a loss of three water molecules per aluminium (calc. 2.6 wt%). Above 4 C NOTT-3 starts to decompose rapidly. 1 8 Weight loss (wt%) Temperature o C Figure S13. TGA plot for NOTT-3-solvate. NATURE CHEMISTRY 19

20 6. Synchrotron Powder Diffraction Studies of Solvated NOTT-3. To investigate the chemical stability of desolvated NOTT-3, an important feature for a capture material, PXRD patterns were collected for a range of NOTT-3 samples under variable chemical environments. After the collection of an original pattern for the as-synthesised sample, NOTT-3-solvate was degassed at 15 o C to generate desolvated NOTT-3. The desolvated material was then separated into ten portions, each of which was exposed to air for one month, or immersed in water, methanol, ethanol, CHCl 3, CH 3 CN, DMF, THF, benzene or toluene for one week. Comparison of the resultant PXRD patterns confirms the excellent stability of the desolvated NOTT-3 material under air, water and common organic solvents. ethanol methanol water exposed to air NOTT-3 toluene benzene THF DMF CH 3 CN NOTT-3-solv CHCl 3 Intensity degrees Figure S14. Comparison of powder diffraction patterns of NOTT-3 under different chemical environment [λ= (2) Å]. NATURE CHEMISTRY 2

21 Figure S15. Comparison of unit cell parameters of NOTT-3 under different chemical environments. NATURE CHEMISTRY 21

22 7. Variable Temperature Powder Diffraction. To investigate the possible framework phase change of NOTT-3-solvate as a function of temperature, variable temperature PXRD patterns were collected at K for NOTT-3-solvate (Fig. S16). Comparison of the PXRD patterns confirms that there is no framework phase transition over this temperature range and the framework of NOTT-3 remains intact after removal of the water molecules in the channels. The lattice parameters were refined via Le Bail methods, and results are summarised in Table S3. The overall change in the unit cell volume is less than 1. %, confirming the rigidity of the framework. 1 K 363 K 14 K 383 K 18 K 43 K 22 K 423 K 26 K 443 K 273 K 463 K 33 K 483 K 333 K Intensity Figure S16. Variable temperature PXRD patterns for NOTT-3-solvate (λ = Å). NATURE CHEMISTRY 22

23 Figure S17. Unit cell parameters for NOTT-3-solvate as a function of temperature. Table S4. Summary of Le Bail refinement results and unit cell parameters for NOTT-3-solvate. a, b (Å) c (Å) V (Å 3 ) Unit cell change (%) 1 K (3) (35) (13). 14 K (29) (33) (12).1 18 K (3) (34) (13) K (3) (34) (13) K (3) (35) (13) K (31) (37) (13) K (3) (36) (13) K (31) (38) (14) K (48) (61) (21) K (42) (54) (19) K (38) (47) (17) K (45) (51) (19) K (54) (59) (23) K (51) (6) (22) K (13) (15) (56) 1.4 NATURE CHEMISTRY 23

24 8. Additional Gas Sorption Isotherm Plots for NOTT Gas uptake (cc/g) H 2 77 K N 2 77 K Ar 87 K Gas uptake (mmol/g) Pressure (bar) Figure S18. H 2,N 2 and Ar sorption isotherms at 77 or 87 K for NOTT-3. No significant uptake was observed for these orption isotherms. NATURE CHEMISTRY 24

25 uptake (cc/g) Adsorption Desorption uptake (mmol/g) uptake (cc/g) Adsorption Desorption uptake (mmol/g) Pressure (bar) Pressure (bar) a b uptake (cc/g) Adsorption Desorption uptake (mmol/g) uptake (cc/g) Adsorption Desorption uptake (mmol/g) Pressure (bar) Pressure (bar) c d Figure S19. orption and desorption isotherms for NOTT-3 at (a) 273 K, (b) 283 K, (c) 293 K and (d) 33 K uptake (cc/g) Adsorption Desorption uptake (mmol/g) Pressure (bar) Figure S2. High pressure orption and desorption isotherms at 273 K for NOTT-3. NATURE CHEMISTRY 25

26 Pore volume, dv/d(d) (cm 3.1nm -1 g -1 ) dv/d(d) Pore volume Pore width (.1nm) Cumulative pore volume (cm 3 g -1 ) Figure S21. Pore size distribution (PSD) plot and cumulative pore volume for NOTT-3. Data were calculated the orption isotherm at 273 K using DFT/Monte Carlo methods Gas uptakes (cc/g) O 2 H 2 Ar SO 2 CH 4 CO N Gas uptakes (mmol/g) Pressure (bar) Figure S22. Comparison of the gas orption isotherms for NOTT-3 at 273 K and.15 bar. NOTT-3 exhibits highly selective uptake for and SO 2 compared with CH 4,CO,N 2,H 2,O 2 and Ar. The selectivities, calculated from the ratio of isotherm uptakes at.15 bar are 88, 99, 148, 197, 85, and 16 for CH 4,CO,N 2,H 2,O 2, and Ar, respectively. The SO 2 selectivities, calculated from the ratio of isotherm uptakes at.15 bar are 25, 278, 418, 557, 239, and 451 for CH 4,CO,N 2,H 2,O 2 and Ar, respectively. NATURE CHEMISTRY 26

27 SO 2 uptake (cc/g) Adsorption Desorption SO 2 uptake (mmol/g) SO 2 uptake (cc/g) Adsorption Desorption SO 2 uptake (mmol/g) Pressure (bar) Pressure (bar) a b SO 2 uptake (cc/g) Adsorption Desorption SO 2 uptake (mmol/g) SO 2 uptake (cc/g) Adsorption Desorption SO 2 uptake (mmol/g) Pressure (bar) Pressure (bar) c d Figure S23. SO 2 orption and desorption isotherms for NOTT-3 at (a) 273 K, (b) 283 K, (c) 293 K and (d) 33 K. NATURE CHEMISTRY 27

28 Gas uptakes (cc/g) des O 2 H 2 Ar SO 2 SO 2 des CH 4 CO N Gas uptakes (mmol/g) Pressure (bar) a Gas uptakes (cc/g) des SO 2 SO 2 des O 2 H 2 Ar CH 4 CO N Gas uptakes (mmol/g) Pressure (bar) b Gas uptakes (cc/g) O 2 H 2 Ar CH 4 CO N 2 SO Gas uptakes (mmol/g) des SO 2 des Pressure (bar) c Figure S24. Comparison of the gas orption isotherms for NOTT-3 at 283 K (a), 293 K (b), 33 K (c) and 1. bar. NOTT-3 exhibits highly selective uptake for and SO 2 compared with CH 4,CO,N 2,H 2, O 2,andAr. NATURE CHEMISTRY 28

29 9. Exposure of NOTT-3 to water. The NOTT-3-solvate material was loaded into an IGA and degassed at 12 o C and 1-1 bar for 24 h to give the fully desolvated NOTT-3 material. A orption isotherm was then measured at 273 K and up to 1. bar (noted as first cycle). The desolvated sample was then exposed to high temperature (9-1 o C) water vapour for 1 h as a humidity treatment. The hydrated sample was then loaded into IGA and degassed again at 12 o C and 1-1 bar for 24 h to give desolvated NOTT-3 material. A second orption isotherm was then measured at 273 K and up to 1. bar (noted as the second cycle). The same humidity treatment (hydration), degassing (de-hydration), and orption were repeated twice more (noted as third and fourth cycles). Comparisons of these four orption isotherms confirm that there is no apparent loss of uptake capacity and that the pore surface can be fully regenerated, showing that the framework has good stability upon exposure to the above humidity cycle uptake (cc/g) First cycle Second cycle Thrid cycle Fourth cycle uptake (mmol/g) Pressure (bar) Figure S25. orption isotherms at 273 K for NOTT-3 upon cyclic hydration and desolvation process. NATURE CHEMISTRY 29

30 1. Additional Inelastic Neutron Scattering Spectra. Neutron Energy Loss/cm S(Q, )/Arb. Units Bare NOTT-3.25 H 2 /Al loading.5 H 2 /Al loading Neutron Energy Loss/meV Figure S26. Comparison of INS spectra for bare NOTT-3,.25 H 2 /Al and.5 H 2 /Al loaded-nott-3. NATURE CHEMISTRY 3

31 Neutron Energy Loss/cm Neutron Energy Loss/cm S(Q, )/Arb. Units.5 S(Q, )/Arb. Units Neutron Energy Loss/meV Neutron Energy Loss/meV a b Figure S27. Difference INS spectra plot for forward (red) and back (black) scattering between bare and.25h 2 /Al-loaded NOTT-3. The broad hump shows recoil of hydrogen; a very small and poorly defined peak at 1meV suggests a very weak interaction. A detailed view of the low energy transfers has been scaled up in Figure b. Neutron Energy Loss/cm Neutron Energy Loss/cm S(Q, )/Arb. Units S(Q, )/Arb. Units Neutron Energy Loss/meV Neutron Energy Loss/meV a b Figure S28. Difference INS spectra plot for forward (red) and back (black) scattering between bare and.5 H 2 /Al-loaded NOTT-3. The broad hump shows recoil of hydrogen. The peak at 1meV is still broad and very weak, but can be seen to have increased in intensity slightly. Furthermore, an additional peak at ~15 mev can be observed. This peak is very close to the rotational line of hydrogen (14.7 mev), further indicating that the interactions between the hydrogen and the NOTT-3 framework are very weak. A detailed view of the low energy transfers has been scaled up in Figure b. NATURE CHEMISTRY 31

32 11. Analysis and Derivation of the Isosteric Heat of Adsorption for. To estimate the isosteric enthalpies (ΔH) for orption, all isotherms at K were fitted to the van t Hoff equation (1): d l n ( p ) d ( 1 / T ) H R (1) where p is pressure, T is the temperature, R is the real gas constant. Selected linear fitting plots at.5, 1., 1.5 and 2. mmol g -1 are shown in Figure S25. All linear fittings show R 2 above.999, indicating consistency of the isotherm data Equation y = a + b*x Adj. R-Square Value Standard Error.5 Intercept Slope Equation y = a + b*x Adj. R-Square Value Standard Error 1 Intercept Slope lnp (ln pa) Experiment at.5 mmol/g Linear fitting /T (K -1 ) lnp (ln pa) Experiment at 1. mmol/g Linear fitting /T (K -1 ) lnp (ln pa) Equation y = a + b*x Adj. R-Square Value Experiment at 1.5 mmol/g Linear fitting Standard Error 1.5 Intercept Slope /T (K -1 ) lnp (ln pa) Equation y = a + b*x Adj. R-Square Experiment at 2. mmol/g Linear fitting Value Standard Error 2. Intercept Slope /T (K -1 ) Figure S29. Linear fitting of Van t Hoff plots for the orption isotherms at.5, 1., 1.5 and 2. mmol g -1 loadings. NATURE CHEMISTRY 32

33 12. Calculation of Henry s Law selectivity for gas orption. To estimate the selectivity of and SO 2 over other gases at zero surface coverage, all low pressure isotherm data at 273 K were fitted using a linear virial-type expression (2) employed previously to model gas sorption in MOFs. For the isotherms with overall low uptakes, where a good linear fitting cannot be obtained at low pressure, the non-linear virial type expression (3) was employed to achieve reasonable virial fitting with inclusion of data at relatively high pressure. 6 l n ( n 1 / p ) A A n. (2) 2 ln( n / p ) A A n A n... (3) 1 2 where p is the pressure expressed, n is the amount orbed, A i are virial coefficients, and i represent the number of coefficients required to adequately describe the isotherms with low uptakes. The results of the fitting for all isotherms give R 2 greater than.99 and the Henry constants for each component were extracted from the virial coefficients (Tables S5). The Henry s constant (K H ) can be extracted from the values of the virial coefficients A using expression (4). K H exp( A ) (4) The Henry s Law selectivity for component i( or SO 2 ) over other gas component j(co, CH 4,N 2,O 2, Ar or H 2 ) was estimated based on the ratio of their Henry s constants (equation 5). The results are listed in Table S6. The selectivity data from virial fittings and Henry s Law analysis are confined to the zero surface coverage situations. S K / K ij Hi Hj (5) NATURE CHEMISTRY 33

34 ln(n/p) ln(mol/g Pa) Equation Experiment Linear Fit loading (mol/g) a y = a + b*x Adj. R-Square Value Standard Error D Intercept D Slope ln(n/p) ln(mol/g Pa) Equation Experiment Linear Fit SO 2 loading (mol/g) b y=a+b*x Adj. R-Square Value Standard Error P Intercept P Slope Figure S3. Linear virial fitting plots for the orption isotherms for (a) and (b) SO 2 for NOTT-3 at 273 K. lnp (ln mbar) n mmol/g a n mmol/g b lnp (ln mbar) lnp (ln mbar) lnp (ln mbar) n mmol/g c n mmol/g d NATURE CHEMISTRY 34

35 lnp (ln mbar) n mmol/g e Figure S31. Non-linear virial fitting plots for the orption isotherms for (a) CO, (b) CH 4, (c) O 2, (d) N 2 and (e) Ar for NOTT-3 at 273 K. ln(n/p) ln(mol/g Pa) Equation Experiment Linear fit y = a + b*x Adj. R-Square.9938 Value Standard Error D Intercept D Slope ln(n/p) ln(mol/g Pa) Equation Experiment Linear fit y = a + b*x Adj. R-Square.9967 Value Standard Error D Intercept D Slope loading (mol/g) a SO 2 loading (mol/g) b Figure S32. Linear virial fitting plots for the orption isotherms for (a) and (b) SO 2 for NOTT-3 at 283 K. NATURE CHEMISTRY 35

36 lnp (ln mbar) n mmol/g a n mmol/g b lnp (ln mbar) lnp (ln mbar) n mmol/g c n mmol/g d lnp (ln mbar) lnp (ln mbar) n mmol/g e Figure S33. Non-linear virial fitting plots for the orption isotherms for (a) CO, (b) CH 4, (c) O 2, (d) N 2 and (e) Ar for NOTT-3 at 283 K. NATURE CHEMISTRY 36

37 ln(n/p) ln(mol/g Pa) Equation Experiment Linear fit y=a+b*x Adj. R-Squar Value Standard Erro D Intercept D Slope ln(n/p) ln(mol/g Pa) Equation D Linear Fit of D y=a+b*x Adj. R-Squa Value Standard Err D Intercept D Slope loading (mol/g) a SO 2 loading (mol/g) b Figure S34. Linear virial fitting plots for the orption isotherms for (a) and (b) SO 2 for NOTT-3 at 293 K. lnp (ln mbar) n mmol/g a n mmol/g b lnp (ln mbar) lnp (ln mbar) lnp (ln mbar) n mmol/g c n mmol/g d NATURE CHEMISTRY 37

38 lnp (ln mbar) n mmol/g e Figure S35. Non-linear virial fitting plots for the orption isotherms for (a) CO, (b) CH 4, (c) O 2, (d) N 2 and (e) Ar for NOTT-3 at 293 K Equation y=a+b*x Adj. R-Squar Value Standard Erro D Intercept D Slope Equation y=a+b*x Adj. R-Squar.9763 Value Standard Erro D Intercept D Slope ln(n/p) ln(mol/g Pa) Experiment Linear fit ln(n/p) ln(mol/g Pa) Experiment Linear fit loading (mol/g) SO 2 loading (mol/g) a b Figure S36. Linear virial fitting plots for the orption isotherms for (a) and (b) SO 2 for NOTT-3 at 33 K. NATURE CHEMISTRY 38

39 lnp (ln mbar) n mmol/g a n mmol/g b lnp (ln mbar) lnp (ln mbar) n mmol/g c n mmol/g d lnp (ln mbar) lnp (ln mbar) n mmol/g e Figure S37. Non-linear virial fitting plots for the orption isotherms for (a) CO, (b) CH 4, (c) O 2, (d) N 2 and (e) Ar for NOTT-3 at 33 K. NATURE CHEMISTRY 39

40 Table S5. Virial fitting results and Henry s constants K H for, SO 2, CH 4,N 2,H 2,O 2 and Ar in NOTT-3 from isotherm data at 273 K. SO 2 CO CH 4 N 2 H 2 O 2 Ar A /ln(mol g -1 Pa -1 ) (8) 6.22(4) -2.27(3) -2.34(3) -2.81(1) -2.4(2) -2.57(2) K n.a. * H / mol g -1 Pa (57) 58(167) 1.57(5) 1.46(4) 9.14(1) 1.98(2) 1.16(2) x1-7 x1-9 x1-9 x1-1 x1-9 x1-9 Fitting R >.999 >.999 >.999 >.999 >.999 Residual error * The uptake of H 2 isotherm at 273 K (below.1 wt%) is too low to obtain a reasonable virial fitting curve, and the Henry constant is therefore considered to be approximately zero. Table S6. Virial fitting results and Henry s constants K H for, SO 2, CH 4,N 2,H 2,O 2 and Ar in NOTT-3 from isotherm data at 283 K. SO 2 CO CH 4 N 2 H 2 O 2 Ar A /ln(mol g -1 Pa -1 ) (5) 5.75(5) -2.38(2) -2.29(2) -2.84(1) -2.11(2) -2.7(3) K n.a. * H / mol g -1 Pa (24) 314(123) 1.41(3)x1 1.54(3) 8.88(9) 1.85(4) 1.3(4) x x1-9 x1-1 x1-9 x1-9 Fitting R >.999 >.999 >.999 >.999 >.999 Residual error * The uptake of H 2 isotherm at 283 K (below.1 wt%) is too low to obtain a reasonable virial fitting curve, and the Henry constant is therefore considered to be approximately zero. NATURE CHEMISTRY 4

41 Table S7. Virial fitting results and Henry s constants K H for, SO 2, CH 4,N 2,H 2,O 2 and Ar in NOTT-3 from isotherm data at 293 K. SO 2 CO CH 4 N 2 H 2 O 2 Ar A /ln(mol g -1 Pa -1 ) (3) 5.46(126) -2.37(2) -2.23(3) -2.78(3) (2) -2.69(2) K n.a. * H / mol g -1 Pa (6) 235(168) 1.42(2) 1.63(4) 9.42(22) 2.1(5) 1.3(2) x1-7 x1-9 x1-9 x1-1 x1-9 x1-9 Fitting R >.999 >.999 >.999 >.999 >.999 Residual error * The uptake of H 2 isotherm at 283 K (below.1 wt%) is too low to obtain a reasonable virial fitting curve, and the Henry constant is therefore considered to be approximately zero. Table S8. Virial fitting results and Henry s constants K H for, SO 2, CH 4,N 2,H 2,O 2 and Ar in NOTT-3 from isotherm data at 33 K. SO 2 CO CH 4 N 2 H 2 O 2 Ar A /ln(mol g -1 Pa -1 ) (1) 4.3(1) -2.27(2) -2.15(2) -2.73(1) -2.1(1) -2.42(2) K n.a. * H / mol g -1 Pa (8) 73.7(465) 1.57(2) 1.77(4) 9.89(1) 2.4(1) 1.35(2) x1-8 x1-9 x1-9 x1-1 x1-9 x1-9 Fitting R >.999 >.999 >.999 >.999 >.999 Residual error * The uptake of H 2 isotherm at 283 K (below.1 wt%) is too low to obtain a reasonable virial fitting curve, and the Henry constant is therefore considered to be approximately zero. NATURE CHEMISTRY 41

42 Table S9. Summary of gas orption selectivity data obtained by two methods: (i) a the ratio of slopes of initial orption isotherm plot; (ii) b Henry s Law analysis at 273 K. Selectivity ratio Isotherm plot slop Henry s Law analysis /CO /CH /N /H 2 >1 5 >1 5 /O /Ar SO 2 /CO 315 >1 5 SO 2 /CH >1 5 SO 2 /N >1 5 SO 2 /H 2 >1 5 >1 5 SO 2 /O >1 5 SO 2 /Ar 4974 >1 5 a This method represents the selectivity at low pressure region (5-35 mbar) and is close to the situation from the direct comparison of gas uptakes. b This method represents the extreme selectivity at zero surface coverage of a given material, and therefore is higher than the values from method (i). Selectivity data obtained from method (i) are reported in the main text. Table S1. Summary of gas orption selectivity data obtained by two methods: (i) a the ratio of slopes of initial orption isotherm plot; (ii) b Henry s Law analysis at 283 K. Selectivity ratio Isotherm plot slop Henry s Law analysis /CO /CH /N /H 2 >1 5 >1 5 /O /Ar SO 2 /CO 3586 >1 5 SO 2 /CH >1 5 SO 2 /N >1 5 SO 2 /H 2 >1 5 >1 5 SO 2 /O >1 5 SO 2 /Ar 4831 >1 5 a This method represents the selectivity at low pressure region (5-35 mbar) and is close to the situation from the direct comparison of gas uptakes. b This method represents the extreme selectivity at zero surface coverage of a given material, and therefore is higher than the values from method (i). NATURE CHEMISTRY 42

43 Table S11. Summary of gas orption selectivity data obtained by two methods: (i) a the ratio of slopes of initial orption isotherm plot; (ii) b Henry s Law analysis at 293 K. Selectivity ratio Isotherm plot slop Henry s Law analysis /CO /CH /N /H 2 >1 5 >1 5 /O /Ar SO 2 /854 >1 5 SO 2 /CH >1 5 SO 2 /N >1 5 SO 2 /H 2 >1 5 >1 5 SO 2 /O >1 5 SO 2 /Ar 47 >1 5 a This method represents the selectivity at low pressure region (5-35 mbar) and is close to the situation from the direct comparison of gas uptakes. b This method represents the extreme selectivity at zero surface coverage of a given material, and therefore is higher than the values from method (i). Table S12. Summary of gas orption selectivity data obtained by two methods: (i) a the ratio of slopes of initial orption isotherm plot; (ii) b Henry s Law analysis at 33 K. Selectivity ratio Isotherm plot slop Henry s Law analysis /CO /CH /N /H 2 >1 5 >1 5 /O /Ar SO 2 /CO 1586 >1 5 SO 2 /CH >1 5 SO 2 /N >1 5 SO 2 /H 2 >1 5 >1 5 SO 2 /O >1 5 SO 2 /Ar 22 >1 5 a This method represents the selectivity at low pressure region (5-35 mbar) and is close to the situation from the direct comparison of gas uptakes. b This method represents the extreme selectivity at zero surface coverage of a given material, and therefore is higher than the values from method (i). NATURE CHEMISTRY 43

44 13. Summary of the hydrogen bond interaction. A hydrogen bond system is conventionally represented as a linear A H B arrangement of a hydrogen donor (A H) and an acceptor (B). Relevant properties of the different strengths of hydrogen bonds are given in Table C1. In this system, the hydrogen bond length H O is around 2.3 Å (Figure 4d), and therefore it can be classed as a moderate-to-weak hydrogen bond. The four C-H supramolecular contacts are likely to be of lower energies. Based on this analysis, we view the observed value of Q st as entirely reasonable and consistent with the likely hydrogen bond energies. In addition, there is the possibility of electrostatic Al(III)/XO 2 interactions. The high uptakes also reflect the relatively narrow pore size of the host which provides strong overlap potentials. Table S13. Properties of strong, moderate and weak H-bonds. 7,8 Properties Strong H-bonds Moderate H-bonds Weak H-bonds Bond energy (kj mol -1 ) <1 Bond nature mostly covalent mostly electrostatic electrostatic Bond linearity, A H B always linear mostly linear sometimes linear Bond length A H (Å) 1.2 to 1.5 ca 1. ca 1. Bond length H B (Å) 1.2 to to to 3.2 Bond length A B (Å) 2.2 to to to Film of the Dynamics of the Crystal Lattice upon Inclusion in NOTT-3. A separate file has been deposited to show the dynamics for these highlighted bonds. References: 1 Spek, A. L. Structure validation in chemical crystallography. Acta Crystallogr. Sect. D 65, , (29). 2 Clark, S. J. et al. First principles methods using CASTEP. Z. Kristall. 22, , (25). 3 Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized gradient approximation made simple. Phys. Rev. Lett. 77, , (1996). 4 Ramirez-Cuesta, A. J. aclimax 4..1, The new version of the software for analyzing and interpreting INS spectra. Comput. Phys. Commun. 157, , (24). 5 Liu, Y. et al. Reversible structural transition in MIL-53 with large temperature hysteresis. J. Am. Chem. Soc. 13, (28). 6 Chen, B. et al. Surface interactions and quantum kinetic molecular sieving for H 2 and D 2 orption on a mixed metal-organic framework material. J.Am.Chem.Soc.13, (28). 7 Jeffrey, G.A. An Introduction to Hydrogen Bonding, Oxford University Press, Oxford (1997). 8 Desiraju, G. R., Steiner, T. The Weak Hydrogen Bond in Structural Chemistry and Biology, IUCr Monographs on Crystallography 9. Oxford University Press, Oxford (1999). NATURE CHEMISTRY 44