Supporting Information. Perfluoroalkane Functionalization of NU-1000 via Solvent-Assisted Ligand Incorporation: Synthesis and CO2 Adsorption Studies

Transcription

1 Supporting Information Perfluoroalkane Functionalization of NU-1000 via Solvent-Assisted Ligand Incorporation: Synthesis and CO2 Adsorption Studies Pravas Deria, Joseph E. Mondloch, Emmanuel Tylianakis,, Pritha Ghosh, Wojciech Bury,,ζ Randall Q. Snurr, *, Joseph T. Hupp,*, and Omar K. Farha*, Departments of Chemistry and Chemical and Biological Engineering, Northwestern University, 2145 Sheridan Road, Evanston, Illinois 60208, United States Materials Science and Technology Department, University of Crete, P.O. Box 2208, Heraklion, Crete, Greece. ζ Department of Chemistry, Warsaw University of Technology, Noakowskiego 3, Warsaw, Poland. Table of Contents Contents S1. Materials. 2 S2. Instrumentation. 2 S3. Synthesis of SALI-n from NU-1000 (Zr 6 (µ 3 OH) 8 (OH) 8 (TBAPy) 2 ). 3 S4. Synthesis of SALI-1 from as Synthesized, NU-1000/BA (p-benzoic acid)) BA@NU1000 S5. Activation Procedure for NU-1000 and SALI-n 4 S6. 1 H and 19 F NMR of SALI-n. 4 7 S7. Expected SALI Stoichiometry. 7 S8. DRIFTS Data for NU-1000, SALI-n, SALI-1. 8 S9. SALI-7 and SALI-1 Structure Modeling 8-10 S10. Experimental PXRD Patterns for NU-1000, SALI-n and SALI-n Samples S11. Comparison of Simulated and Experimental Adsorption Isotherms S12. Experimental N 2 Adsorption Isotherms Analysis S13. Volumetric Experimental N 2 Isotherms and Pore Size Distribution S14. Pore Volume in SALI-n Considering the Volume Occupied by the Fluoroalkyl Chains S15. Experimental CO 2 Adsorption Isotherm Analysis and Comparison with Simulated CO 2 Adsorption Isotherms. Page Number S16. Experimental Q st Analysis and Comparison with Simulated Q st S17. Simulated CO 2 Density Distribution in SALI-7 at Various Pressures S18. CO 2 adsorption Selectivity in NU-1000 and SALI-9 for CO 2 /N 2 gas Mixture Calculated from IAST. S19. Experimental Water Adsorption Isotherms and Water Stability. 19 S20. References S1

2 S1. Materials. S1A. Reagents and Solvents: Acetone (Macron, 98%), chloroform (BDH, 99.8%), 1,4-dioxane (Aldrich, 99.8%, anhydrous), N,N -dimethylformamide (DMF) (Macron, 99.8%), diethylether (Aldrich, 99.0%), deuterated chloroform (d-cdcl 3 ) (Cambridge, 99.8%), deuterated dimethylsulfoxide (d 6 -DMSO) (Cambridge, 99%), deuterated sulfuric acid (D 2 SO 4 ) (Cambridge, 96 98% solution in D 2 O) were used as received without further purification. Perfluoroalkyl and fluoroalkyl carboxylic acids were purchased from Synquest lab and used as received. S1B. Synthesis of NU-1000 in DMF: Microcystalline NU-1000 was prepared via solvothermal method according to the published procedure except slight modification mg of anhydrous ZrCl 4 (0.30 mmol) and 2700 mg (22 mmol) of benzoic acid were mixed in 8 ml of DMF (in a 6-dram vial) and ultrasonically dissolved. The clear solution was incubated in an oven at 80⁰C for 1h. After cooling down to room temperature 40 mg (0.06 mmol) of H 4 TBAPy was added to this solution and the mixture was sonicated for 20 min. The yellow suspension was heated in an oven at 120⁰C for 48 h. After cooling down to 50 0 C the supernatant solution was decanted, yellow polycrystalline material was isolated by filtration (35 mg of activated material, 54% yield) and soaked in DMF for 8h at room temperature and finally washed with fresh DMF. This material was activated with HCl. S1C. Activation procedure for NU-1000: As synthesized NU-1000 was activated using a slightly modified method previously reported. 1-2 Approximately 40 mg of solvated ( wet ) material was soaked in 12 ml of DMF and 0.5 ml of 8 M aqueous HCl was added. This mixture was heated in an oven at 100⁰C for 24 h. After cooling to 50 0 C the supernatant solution was decanted and the material was soaked in DMF for 12h and subsequently washed twice with DMF to remove HCl impurities. The solid residue was then soaked in acetone for 12h, washed twice with acetone and soaked in acetone for additional 12h. NU-1000 was filtered, briefly dried on a filter paper and activated at 120⁰C under vacuum for 12h on the preparation station of ASAP 2020 instrument. Characteristic 1 H NMR, N 2 adsorption isotherms, and DRIFTS data are consistent with the removal of benzoic acid from the Zr 6 node and the incorporation of OH groups as reported in the previous publication. 1 S1D. Secondary structural element: It is important to note that 20 25% of the mesoporous channels of the NU-1000 contain a secondary structural element (for detail regarding residual electron density plots and N 2 adsorption simulations see reference S1). Modeling of the secondary element as [Zr 6 (µ 3 -O) 4 (µ 3 -OH) 4 ] 2 (TBAPy) 6 which connects to 12 Zr 6 nodes of the parent framework through six TBAPy ligands (see reference 1) provided. Bon et al. also recently detected a secondary element within the pores of a Zr-based MOF. 3 Finally, we cannot rigorously rule out the secondary framework being disordered about the six-fold axis running through the middle of a mesoporous channel. S2. Instrumentation. 1 H NMR spectra were recorded on a 500 MHz Varian INOVA spectrometer and referenced to the residual solvent peak. Powder X-ray diffraction measurements were carried out on a Bruker MX IµS microsource with Cu Kα radiation and an Apex II CCD detector. The samples were mounted in capillaries as powders, sealed with wax and placed on a goniometer head. The data were collected on an area detector with rotation frames over 180 in φ and at 2θ values of 12, 24, and 36 being exposed for 10 min at each frame. Overlapping sections of data were matched, and the resulting pattern was integrated using Bruker s APEX2 phase ID program. The powder patterns were treated for amorphous background scatter. Diffuse reflectance infrared spectra (DRIFTS) were recorded on a Nicolet 7600 FTIR spectrometer equipped with an MCT detector. The spectra were collected in a KBr mixture under N 2 purge (samples prepared in atmosphere); S2

3 KBr was utilized as the background. N 2 adsorption isotherms were collected on ASAP 2020 (Micromeritics). All pore size distributions were obtained using the BJH method on the desorption branch of the N 2 isotherms. CO 2 adsorption isotherms were collected on IGA-200 (Hiden Isochema) S3. Synthesis of SALI-n from NU-1000 (Zr 6 (µ 3 OH) 8 (OH) 8 (TBAPy) 2 ). A 60 mg portion of activated NU-1000 (0.027 mmol) was loaded in a 2 ml microwave vial (Biotage). Subsequently a 2.4 ml of 0.1M solution of fluoroalkane carboxylic acid (0.24 mmol) in DMF was added to the reaction vial, which was then sealed and heated at 60 0 C for h with occasional swirling. The supernatant of the reaction mixture was decanted and the MOF sample was soaked into fresh hot DMF which was then filtered, washed sequentially with DMF, acetone and ether (60, 40 and 30 ml each), and finally air dried. Scheme SI-1. Synthesis of SALI-n from NU S4. Synthesis of SALI-1 from as synthesized, NU-1000/BA. As-synthesized NU-1000 contains benzoate ligands (here after referred as NU-1000/BA) present on the Zr 6 nodes (4 benzoates are present per node) and the DRIFTS spectra 1 are identical to that of SALI-1 (Figure 2B). When NU-1000/BA (30 mg) was exposed to trifluoroacetic acid (0.11 mmol) in DMF (1.2 ml) at 60 0 C, benzoate ligands on the Zr 6 node were replace by trifluoroacetate, to produce SALI-1 via a ligand exchange process, which relies on a weaker benzoic acid (pka 4.2) being replaced by a stronger perfluoroalkyl carboxylic acid (pka 0.23). Formation of SALI-1 was confirmed via 1 H, 19 F NMR spectra (Figure SI-1). However, this ligand replacement process was not universal for respective SALI-n samples. Scheme SI-2. Synthesis of SALI-n from NU-1000/BA. S3

4 Figure SI-1. SALI-1 from NU-1000/BA: 1 H NMR spectra of (A) as-synthesized NU-1000/BA and (B) SALI-1 from NU-1000/BA; the peaks for benzoate are denoted with (*). (C) 19 F NMR spectrum of SALI-1 sample shown in panel (B). Samples were decomposed in 10% D 2 SO 4 /DMSO-d 6. 1 H and 19 F signals of 2,5-bis(trifluoromethyl)-1,4-dibromobenzene, used as an internal standard, are denoted with (+). Note the complete disappearance in the 1 H NMR signal for the benzoate ligands in panel B. A complete functionalization considering the SALI-1 stoichiometry (Scheme SI-1,2) yields four carboxylates per Zr 6 node and the expected 19 F NMR signal intensity at 77 ppm (panel C) is 6. S5. Activation procedure for NU-1000 and SALI-n for gas sorption measurements. SALI-n and SALI-n samples were activated at 120⁰C under vacuum for 12h. Shown below are characterization data for NU-1000, SALI-n samples including 1 H, 19 F NMR, PXRD, N 2 adsorption measurements, and DRIFTS. The data are consistent with the removal of OH groups from the Zr 6 node and the incorporation of perfluoroalkyl carboxylate ligands. S6. 1 H and 19 F NMR of SALI-n. Functionalization of NU-1000 with perfluoroalkyl carboxylates was quantified by 1 H and 19 F NMR after decomposing the samples in a 10% D 2 SO 4 /DMSO-d 6 mixture; the 19 F signals of the perfluoroalkanes (1, 3, 7, and 9; Scheme 1) were integrated against the 1 H NMR signals of TBAPy ligand using an internal standard (2,5-dibromo-1,4-bis(trifluoromethyl)benzene (examples below). S4

5 Figure SI-2i. (A), (C) 1 H NMR and (B), (D) 19 F NMR spectra of SALI-1 and SALI-3 samples, respectively, derived from NU Samples were decomposed in 10% D 2 SO 4 /DMSO-d 6. 1 H and 19 F signals of 2,5-bis(trifluoromethyl)-1,4-dibromobenzene, used as an internal standard, are denoted with (+). A complete functionalization considering the SALI-n stoichiometry (Scheme SI-1) yields four carboxylates per Zr 6 node and the expected 19 F NMR signal intensities at 77 and 82 ppm (panel B, D) are 6. S5

6 Figure SI-2ii. (A), (C) 1 H NMR and (B), (D) 19 F NMR spectra of SALI-7 and SALI-9 samples, respectively, derived from NU Samples were decomposed in 10% D 2 SO 4 /DMSOd 6. 1 H and 19 F signals of 2,5-bis(trifluoromethyl)-1,4-dibromobenzene, used as an internal standard, are denoted with (+). A complete functionalization considering the SALI-n stoichiometry (Scheme SI-1) yields four carboxylates per Zr 6 node and the expected 19 F NMR signal intensities at 82 ppm (panel B, D) are 6. S6

7 Figure SI-2iii. (A), (C) 1 H NMR and (B), (D) 19 F NMR spectra of SALI-1 and SALI-3 samples, respectively, derived from NU The residual the peaks for benzoate are denoted with (*), whereas residual solvents (DMF and ether) peak marked with (x). Samples were decomposed in 10% D 2 SO 4 /DMSO-d 6. A complete functionalization considering the SALIstoichiometry (Scheme SI-1) yields four carboxylates per Zr 6 node and the expected total (CH 2 ) signal intensities at ppm (panel A, C) are 12 and 8 respectively for SALI-1 and SALI-3. S7. Expected SALI stoichiometry. Considering 20 25% occupancy of secondary structural elements [Zr 6 (µ 3 -O) 4 (µ 3 -OH) 4 ] 2 (TBAPy) 6 in the mesoporous channels of the NU-1000 (as described in section S1D), SALI stoichiometry was estimated to be 3.2 carboxylates per Zr 6 nodes, whereas a 0% occupance of such secondary structural elements would provide 4 carboxylates per Zr 6 nodes. However, considering 20 25% occupancy of a secondary framework that is disordered about the six-fold axis running through the middle of a mesoporous channel, 4 carboxylates per Zr 6 nodes are expected as the SALI stiochiometry. Given that the SALI stoichiometry found for the majority of the fluoroalkyl incorporated NU-1000 is close to 4 carboxylates per Zr 6 nodes (3.4 for SALI-9 to a maximum of 4 for SALI-1), we think that secondary structural element is possibly a framework disordered about the six-fold axis running through the middle of a mesoporous channel. S7

8 S8. DRIFTS data for NU-1000, SALI-n and SALI-n. Figure SI-3. DRIFTS data of NU-1000, SALI-n and SALI-n samples were collected in a KBr mixture. Note that the peak at 3674 cm -1, assigned to the terminal OH groups on the Zr 6 node in parent NU-1000, completely disappears in SALI-3 and SALI-7 which underwent complete functionalization (based on the 1 H NMR data: 3.8 carboxylate ligands per Zr 6 node). On the contrary, in SALI-9, SALI-1 and SALI-3, the terminal OH stretch intensity was only partly reduced along with slight upshifts as a function of respective degree of carboxylate incorporation (3.4, 2 and 3 carboxylates were incorporated in SALI-9, SALI-1 and SALI-3 samples). We assign the lower energy peak at cm -1 to the H-bonded species on the Zr 6 nodes in the functionalized materials (Note that the corresponding peak (Figure 1B) in the fully activated SALI-1 sample disappears). S9. SALI-7 and SALI-1 Structure Modeling. Due to difficulties in obtaining crystals of suitable quality for single crystal analysis, we modeled the structures of SALI-7 and SALI-1 ab initio by applying a procedure based on molecular mechanics energy minimizations explained previously. 4 A description of the NU-1000 ab initio structure modeling and the crystallographic structure solution were presented in our prior publication. 1 For this work, the initial coordinates of NU-1000 were taken from the Crystallographic Information File reported in our prior publication. Starting with these coordinates, SALI-1 was built by adding the ab-initio optimized fluoroalkyl chains to the Zr 6 nodes in silico. Four C 7 F 15 COOH (7) or CF 3 C 3 H 6 COOH (1 ) chains were attached to each Zr 6 node. The new carboxylate ligands coordinate to each equatorial Zr IV as a bidentate ligand via two carboxylate oxygen atoms, and the resulting molecular formula is Zr 6 (µ 3 O) 4 (µ 3 OH) 4 (CO 2 ) 12. Similar to the UiO-66 node, this newly functionalized Zr 6 node provides a structure where the fluoroalkyl chains protrude into the mesoporous hexagonal channel of the NU Thereafter, the structures were subject to geometry optimization based on molecular mechanics calculations by modifying all the atomic positions. These calculations were performed with the Forcite S8

9 module of Materials Studio 5 using an algorithm that is a cascade of the steepest descent, adjusted basis set Newton-Raphson, and quasi-newton methods. Bonded and non-bonded interactions were treated using the Universal Force Field (UFF). 6 A cutoff distance of 12.8 Å was used for the Lennard-Jones interactions. The long-range electrostatic interactions arising from the presence of partial atomic charges were modeled using a Coulombic term using the Ewald method. Partial atomic charges for the NU-1000 framework atoms were derived from the extended charge equilibration method (EQeq) using an in-house written code, 7 and the charges for the C 7 F 15 COOH (7) and CF 3 (CH 2 ) 3 COOH (1 ) atoms were taken from OPLS. 8 Figure SI-4. Atomistic representations of (A) SALI-7, and (B) C 7 F 15 COO functionalized Zr 6 node (only the phenyl-carboxylates of the 1,3,6,8-tetrakis(p-benzoic-acid)pyrene (TBAPy) linkers are shown for clarity; also hydrogen atoms were removed for clarity). S9

10 Figure SI-5. Comparison of simulated PXRD patterns (of the modeled structures) and experimental PXRD patterns for NU-1000 and SALI-7. S10. Experimental PXRD Patterns for NU-1000, SALI-n and SALI-n Samples. Figure SI-6. Experimental PXRD patterns for NU-1000, SALI-n and SALI-n samples. S11. Comparison of Simulated and Experimental Adsorption Isotherms. The adsorption of N 2 was investigated using grand canonical Monte Carlo (GCMC) simulations 9 performed with our in-house multi-purpose code RASPA. 10 We used an atomistic model for the NU-1000, SALI-7 and SALI-1 MOF structures, in which the framework atoms were kept fixed at the positions obtained from their respective optimized structures. We used the standard Lennard-Jones (LJ) + Coulomb potential to model the interactions between the framework and fluid atoms. The LJ parameters for the framework and perfluoroalkyl chain atoms were obtained from the UFF and OPLS force fields, respectively. Framework and perfluoroalkyl charges were obtained as described in Section S9. N 2 was modeled using the TraPPE potential S10

11 with charges placed on each atom ( e) and at the center of mass (0.8096e). 11 The Lorentz- Berthelot mixing rules were employed to calculate fluid/solid LJ parameters, and LJ interactions beyond 12.8 Å were neglected. The Ewald sum method was used to compute the electrostatic interactions. Up to 10 5 Monte Carlo cycles were performed, the first 50% of which were used for equilibration, and the remaining cycles were used to calculate the ensemble averages. To calculate the gas-phase fugacity we used the Peng-Robinson equation of state. 12 The geometric surface areas were calculated by rolling a Å-diameter sphere, which corresponds to a nitrogen molecule, across the surface of the material. 13 For adsorption of CO Monte Carlo cycles were performed, the first 50% of which were used for equilibration, and the remaining cycles were used to calculate the ensemble averages. CO 2 was modeled using the TraPPE potential with separate charges placed on each oxygen (- 0.35e) and on the center carbon (0.7e) using C-O bond lengths of 1.16 Å and an O-C-O angle of All other parameters were as described in the previous paragraph. Figure SI-7. (A) Experimental N 2 sorption isotherms for SALI-7 at 77 K. (B) Comparison of simulated (sim) and experimental (exp) N 2 adsorption isotherms for SALI-7 at 77 K. Note that the simulated isotherm overestimates the N 2 uptake possibly due to the fact that the experimental SALI-7 sample possess 20 25% secondary framework that was originally present in the NU starting material, 1 whereas the modeled structures contained no secondary framework. S12. Experimental N 2 Adsorption Isotherm Analysis. For all isotherm analyses, we ensured that the consistency criteria described by Rouquerol et al. 14 and Walton et al. 15 were satisfied. An example is given in Figure SI-8 and similar analyses were utilized for all isotherms reported herein. S11

12 Figure SI-8. An example of a (A) Rouquerol plot, and (B) BET plot for SALI-9 (N 2 isotherm Figure SI-9, red triangle). The R 2 value from the BET plot is and the y-intercept is S13. Volumetric Experimental N 2 Isotherms and Pore Size Distributions for NU- 1000, SALI-1, and SALI-n. Volumetric N 2 isotherms were calculated by using the carboxylic acid loading, determined by 1 H NMR and 19 F NMR, in Table 1 of the main text and Section S6. We assumed that one R-COO moiety was bound to each of the four equatorial Zr IV of the Zr 6 nodes and that two water molecules were removed (i.e., following the stoichiometry described in Scheme 1 of the main text and Scheme SI-1). We used the crystallographically predicted density of 0.49 cc/g for NU-1000 and estimated 0.56 cc/g for SALI-1, 0.64 cc/g for SALI-3, 0.81 cc/g for SALI-7, 0.85 cc/g for SALI-9, 0.54 cc/g for SALI-1, and 0.62 cc/g for SALI-3 based on simple molecular formula calculations. Figure SI-9. Gravimetric (A) and volumetric (B) isotherms for NU-1000, SALI-n and SALIn samples. S12

13 Figure SI-10. BJH pore size distributions from the desorption branch of the N 2 isotherms for NU-1000, SALI-n, and SALI-n samples. S14. Pore Volume in SALI-n Considering the Volume Occupied by the Fluoroalkyl Chains. As described in Table 1 in the main text, the pore volume systematically decreases in SALI-n relative to that of NU-1000 with increasing fluoroalkyl chain length. Utilizing the SALI stoichiometry for each SALI-n sample obtained from the 1 H NMR data (Table 1), we estimated that 0.18, 0.28, 0.42, and 0.46 g of fluoroalkyl carboxylates are present per gram of SALI-1, SALI-3, SALI-7 and SALI-9 materials, respectively; this corresponds to their liquid-phase volumes of 0.12, 0.17, 0.25, 0.27cc per g of their respective MOF samples. Note that the reductions in the respective pore volumes from the parent NU-1000, as measured from the N 2 isotherm at 77 K, are 0.4, 0.6, 0.8 and 0.8 cc/g. Additionally, we suspect that upon functionalization at the parent NU-1000 nodes, the smaller triangular pores might become less accessible by the N 2 probe molecules at 77 K. S15. Experimental CO 2 Adsorption Isotherm Analysis and Comparison with Simulated CO 2 Adsorption Isotherms. Experimental CO 2 isotherms for all MOF samples were measured gravimetrically at three temperatures (273 K, 283 K, and 298 K). Volumetric CO 2 isotherms were calculated using the density for each sample used in calculating the volumetric N 2 isotherm (Section S12). S13

14 Figure SI-11. Volumetric CO 2 adsorption isotherms for NU-1000, SALI-n, and SALI-n samples at 273 K. Figure SI-12. Gravimetric CO 2 sorption isotherms for (A) NU-1000 and (B) SALI-3 samples. Note the steeper uptake around 0.15 bar in the isotherms for SALI-3. S14

15 Figure SI-13. Comparison of simulated (Sim) and experimental (Exp) CO 2 adsorption isotherms for NU-1000, SALI-1, and SALI-7 at 293 K. Note that the simulated isotherms overestimate the CO 2 uptake similar to the simulated N 2 isotherms (Figure S7). The simulation results shown above treated the perfluoroalkane chains as rigid. To test whether chain flexibility could affect the results, we also simulated SALI-7 allowing for flexibility of the perfluoroalkane chains but holding the rest of the framework rigid. For these calculations, all force field and simulation parameters were identical to those described in sections S9 and S11, with the addition of bond bending, bond stretching, and torsional potentials taken from OPLS 8 for the perfluoroalkane chains. In addition to the traditional grand canonical Monte Carlo moves of insertion, deletion, reinsertion, translation, and rotation, we also included a hybrid Monte Carlo move 16 that allowed for a molecular dynamics trajectory of 5 steps, with a time step of 2 fs. As shown in Figure SI-14, the simulated isotherms and Q st values of the flexible and rigid SALI-7 structures are almost identical. Figure SI-14. Comparison between rigid and flexible models of SALI-7 for (A) CO 2 adsorption isotherms at 293 K and (B) CO 2 Q st as a function of loading. S16. Experimental Q st Analysis and Comparison with Simulated Q st. S15

16 The CO 2 isotherms for NU-1000, SALI-n and SALI-n samples show Langmuir type behavior. Gravimetric isotherms were measured for these samples at three temperatures (273 K, 283 K, and 293 K; e.g. Figure SI-12) and they were subsequently fit with single or dual site Langmuir models. (1) Equation of single site Langmuir (SSL) model is: N = n sat, 1 *b 1 *P/(1 + b 1 *P) (2) Equation of dual site Langmuir (DSL) model is: N = n sat, 1 *b 1 *P/(1 + b 1 *P) + n sat, 2 *b 2 *P/(1 + b 2 *P) where N is the total gravimetric uptake of CO 2 (mmol/g) at pressure P; n sat, i and b i are saturation loading and Langmuir affinity parameter for site i, respectively. For NU-1000 no discernible bend in the isotherm was observed, hence the SSL model was used to fit the isotherms. In contrast, for all other SALI-derived MOF samples, steeper uptake was observed in their respective isotherms (e.g. Figure SI-12) and the DSL model was used for their respective isotherm fitting. Figure SI-15. Representative SSL and DSL fits for CO 2 adsorption isotherms at 273 K for (A) NU-1000, and (B) SALI-3 samples. Table 1. SSL and DSL parameters. MOF Model Ads. Site Parameters 273 K 283 K 293 K Qst 0 (kj/mol) n sat, (0.3) 27.5(0.5) 25.2(0.6) NU-1000 SSL Site 1 17 b 1 9.9(0.1)E-7 9.0(0.2)E-7 7.6(0.2)E-7 n sat, (0.01) 0.72(0.02) 0.76(0.02) Site 1 b 1 8.0(0.3)E-5 5.0(0.2)E-5 3.0(0.1)E-5 SALI-3 DSL 28 n sat, (0.3) 21.5(0.7) 23.3(1.3) Site 2 b 2 8.1(0.2)E-7 5.4(0.2)E-7 3.6(0.2)E-7 The errors in the parameters are listed in the parenthesis; R 2 fit = S16

17 These SSL or DSL fitting parameters were used to deduce a form where P is expressed as a function of N from Eq-1 or Eq-2 and then Q st values were calculated using the Clausius- Clapeyron equation: (3) Clausius-Clapeyron equation: (ln P) N = (Q st /R)(1/T) + Const Figure SI-16. Exemplary CO 2 adsorption isosteres for (A) NU-1000 and (B) SALI-3 samples. Figure SI-17. Representative Q st plots calculated from experimental isotherm data for NU-1000 (blue squares) and SALI-3 (purple triangles). S17

18 Figure SI-18. Comparison of simulated (sim) and experimental (exp) CO 2 Q st plots for NU- 1000, SALI-1, and SALI-7. Note the general good agreement between the experimental and simulated Q st plots. S17. Simulated CO2 Density Distributions in SALI-7 at Various Pressures. Figure SI-19. CO 2 density distributions in SALI-7 from GCMC simulations at (A) 0.01, (B) 0.10, (C) 0.50, and (D) 1.00 bar (T = 293 K). S18. CO 2 Adsorption Selectivity in NU-1000 and SALI-9 for CO 2 /N 2 Gas Mixtures Calculated from Ideal Adsorbed Solution Theory. Ideal Adsorbed Solution Theory (IAST) was applied to estimate the selectivity for CO 2 /N 2 mixed-gas adsorption in two representing MOF structures, NU-1000 and SALI-9. Pure component isotherms of N 2 and CO 2 from experiment were used to calculate the selectivity of CO 2 over N 2 for a gas phase composition of CO 2 /N 2 = 10: S18

19 Figure SI-20. (A) Experimental CO 2 and N 2 adsorption isotherms measured for NU-1000 and SALI-9 at 293 K. (B) Calculated (IAST) CO 2 adsorption selectivity of NU-1000 and SALI-9 for CO 2 /N 2 : 10:90 gas mixtures at 293 K. S19. Experimental Water Adsorption Isotherms and Water Stability. Water adsorption isotherms of the NU-1000, and SALI-n samples were recorded at 298 K at 1 bar using helium as carrier gas (Figure SI-21). Compared to the parent NU-1000, all the SALI-n samples show systematic lower gravimetric uptake with increasing CF n chain length. These water adsorption data indicate only very mild enhancement in hydrophobicity upon perfluoroalkane functionalization. Congruent with their pore volumes (Table 1 in the main text), all the SALI-n samples show systematically lower water uptake with increasing chain length and. Figure SI-21. Water isotherms (left) of NU-1000 and SALI-n samples. Experimental condition: T = 298 K, P = 1 bar; He carrier gas. PXRD patterns (right) for SALI-n samples before and after measuring the water isotherms (water) showing their water stability. S20. References. S19

20 (1) Mondloch, J. E.; Bury, W.; Fairen-Jimenez, D.; Kwon, S.; DeMarco, E. J.; Weston, M. H.; Sarjeant, A. A.; Nguyen, S. T.; Stair, P. C.; Snurr, R. Q.; Farha, O. K.; Hupp, J. T. J. Am. Chem. Soc. 2013, 135, (2) Feng, D.; Gu, Z.-Y.; Li, J.-R.; Jiang, H.-L.; Wei, Z.; Zhou, H.-C. Angew. Chem. Int. Ed. 2012, 51, (3) Bon, V.; Senkovska, I.; Baburin, I. A.; Kaskel, S. Cryst. Growth Des. 2013, 13, (4) (a) Fairen-Jimenez, D.; Colón, Y. J.; O.K., F.; Bae, Y. S.; Hupp, J. T.; Snurr, R. Q. Chem. Commun. 2012, 48, 10496; (b) Farha, O. K.; Yazaydın, A. Ö.; Eryazici, I.; Malliakas, C. D.; Hauser, B. G.; Kanatzidis, M. G.; Nguyen, S. T.; Snurr, R. Q.; Hupp, J. T. Nature Chem. 2010, 2, 944; (c) Strutt, N. L.; Fairen-Jimenez, D.; Iehl, J.; Lalonde, M. B.; Snurr, R. Q.; Farha, O. K.; Hupp, J. T.; Stoddart, J. F. J. Am. Chem. Soc. 2012, 134, (5) Accelrys Software Inc: San Diego, CA 92121, USA. (6) Rappé, A. K.; Casewit, C. J.; Colwell, K. S.; Goddard, W. A.; Skiff, W. M. J. Am. Chem. Soc. 1992, 114, (7) Wilmer, C. E.; Kim, K. C.; Snurr, R. Q. J. Phys. Chem. Lett. 2012, 3, (8) Watkins, K.; Jorgensen, W. L. J. Phys. Chem. A 2001, 105, (9) Frenkel, D.; Smit, B. Understanding Molecular Simulations: From Algorithms to Applications.; 2nd ed.; Academic Press: San Diego, CA, USA, (10) Dubbeldam, D.; Calero, S.; Ellis, D. E.; Snurr, R. Q.; RASPA, 1.0; Northwestern University: Evanston, IL, USA, (11) Potoff, J. J.; Siepmann, J. I. AlChE J. 2001, 47, (12) Reid, R. C.; Pausnitz, J. M.; Poling, B. E. The properties of gases & liquids; 4th ed.; McGraw-Hill Companies: New York, USA, (13) Düren, T.; Millange, F.; Férey, G.; Walton, K. S.; Snurr, R. Q. J. Phys. Chem. C 2007, 111, (14) Rouquerol, J.; Llewellyn, P.; Rouquerol, F. Stud. Surf. Sci. Catal. 2007, 160, 49. (15) Walton, K. S.; Snurr, R. Q. J. Am. Chem. Soc. 2007, 129, (16) Chempath, S.; Clark, L. A.; Snurr, R. Q. J. Chem. Phys. 2003, 118, (17) Myers, A. L.; Prausnitz, J. M. AlChE J. 1965, 11, 121. S20