Bio-compatible MOFs for Storage and Separation of O2: A Molecular

Subscriber access provided by UNIV OF NEW ENGLAND ARMIDALE

Separations

Bio-compatible MOFs for Storage and Separation of O2: A Molecular Simulation Study Ezgi Gulcay, and Ilknur Erucar Ind. Eng. Chem. Res., Just Accepted Manuscript • DOI: 10.1021/acs.iecr.8b04084 • Publication Date (Web): 07 Feb 2019 Downloaded from http://pubs.acs.org on February 11, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

Bio-compatible MOFs for Storage and Separation of O2: A Molecular Simulation Study Ezgi Gulcaya and Ilknur Erucarb* a

Department of Mechanical Engineering, Faculty of Engineering, Ozyegin University, Cekmekoy, 34794, Istanbul, Turkey b Department of Natural and Mathematical Sciences, Faculty of Engineering, Ozyegin University, Cekmekoy, 34794, Istanbul, Turkey Submitted to Industrial & Engineering Chemistry Research Abstract

Metal Organic Frameworks (MOFs) are great candidates for capturing O2 due to their highly porous structures and tunable physical and chemical properties. In this study, we assessed the performance of 1525 bio-compatible MOFs which have endogenous linkers and non-toxic metal centers for adsorption-based and membrane-based O2 separation and also for highpressure O2 storage. We initially computed Henry’s constants of O2 and N2 at zero-coverage and 298 K by performing Grand Canonical Monte Carlo (GCMC) simulations and estimated infinite dilution adsorption selectivities for O2/N2 mixture. We performed binary mixture GCMC simulations for the top 15 candidates at various pressures and 298 K and compared mixture adsorption selectivities with those obtained from infinite dilution. We then estimated O2 working capacities of 315 bio-compatible MOFs obtained at 298 K and 140 bar for storage and 5 bar for release pressures. Our results showed that 15 bio-compatible MOFs outperform gravimetric O2 working capacities of the traditional adsorbent materials such as activated carbon and NaX, and some common MOFs such as NU-125 and UMCM-152 at 298 K. We finally calculated O2 and N2 permeabilities and membrane selectivities of 45 promising MOF candidates for O2/N2 separation. 17 bio-compatible MOF membranes were identified to exceed the Robeson’s upper bound established for polymers. This computational study will be useful to identify the promising bio-compatible MOFs for storage and separation of O2. The bio-MOF library constructed in this study will also guide both experimental and computational studies for design and development of bio-compatible MOFs for various medical applications. Keywords: molecular simulation; adsorption; metal organic frameworks; oxygen separation; oxygen storage. *

Corresponding author. Email: [email protected], Phone: +90 (216) 564-9297 1 ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1. Introduction High-pressure oxygen (O2) storage has various medical and aerospace applications such as photodynamic therapy in which high O2 loading efficiency is desired1 and control system of the spacecraft cabin O2 concentration.2 It is also required for the enrichment of air during the catalyst regeneration in industrial catalytic units.3 In order to produce pure O2, air separation is widely used which purifies O2 and nitrogen (N2).4 Cryogenic distillation, which requires high pressure (from 1320 kPa to 650 kPa) and ultra-low temperature (-185oC), is a traditional method to separate air mixtures based on the boiling point differences of components. However, this technology is complex, expensive and energy intensive. Adsorption-based and membrane-based gas separations have become alternative methods for O2/N2 separations. Comparison to cryogenic distillation, these two technologies require low energy and less complex infrastructures.5 Adsorption-based gas separation is based on the equilibrium adsorption of gas mixtures and separation performance is dependent on the identity of an adsorbent material. Both high working capacity and high selectivity are desired for a promising adsorbent material. Most adsorbent materials used for air separation are N2 selective or non-selective. For example, zeolites, Linde type A (LTA)-4A and LTA-5A are N2 selective due to more favorable electrostatic interactions of N2 with the zeolites’ atoms than those of O2.6 On the other hand, O2/N2 adsorption selectivity of carbon-based adsorbents such as activated carbons at 1 bar and 298 K is almost unity.7 Achieving high selectivity towards O2 with high O2 working capacity is highly required for an efficient adsorbent material for O2 separation. Membrane-based gas separation processes have also gained importance due to their lower energy consumption and lower operating costs compared to cryogenic distillation. Membranes do not require a regeneration process unlike adsorbents. Polymeric membranes have been widely used for air separation due to their low cost and ease of scale-up. Polymer membranes generally exhibit moderate O2/N2 selectivities (3.3-10.5), however they have low O2 permeabilities (0.05-370 Barrer).8 For example, Jeazet et al.9 investigated polysulfone (PSF) for O2 separation and reported the O2 permeability of PSF as 1.5 Barrer and selectivity of O2 over N2 as 5.9 at 3 bar and 303 K. Similarly, Rodrigues et al.10 fabricated pure polyurethane (PU) membrane, which exhibited low O2 permeability (2.8 Barrer), giving O2/N2 selectivity as 4 at 4 bar and 298 K. Polymer membranes generally have a trade-off between selectivity and gas permeability.11 To solve permeability-selectivity trade-off, mixed matrix membranes (MMMs), in which inorganic particles are embedded into a polymer 2 ACS Paragon Plus Environment

Page 2 of 31

Page 3 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


matrix have been developed. Various materials such as zeolites, silicas, carbon molecular sieves (CMS) and aerogels have been used as fillers in polymers. Recent studies showed that although incorporation of these fillers enhances O2 permeabilities of MMMs, O2/N2 selectivities do not change significantly.12 Performing the membrane-based gas separation with high O2 permeability and high O2 selectivity has been still required for industrial air separation applications. Metal-organic frameworks (MOFs), consisted of metal nodes linked by organic ligands, can be alternative adsorbent and/or membrane materials for O2 separation due to their permanent porosities (0.3-0.9), large surface areas (1000-10,000 m2/g) and versatile functionalities.13 Research on MOFs has been rapidly growing with potential applications including gas adsorption14-15, gas separation16, biomimetic catalysis17, sensing18, optical and luminescent applications19 and drug delivery20. To the best of our knowledge, there are only a few experimental studies on storage and separation of O2 with MOFs. Wang et al.21 synthesized a RPM3-Zn(Zn2(bpdc)2(bpee); bpdc = 4,4ꞌ-biphenyldicarboxylate; bpee = 1,2bipyridylethene) MOF and reported O2/N2 mixture selectivity as 5 at both 77 K and 87 K at 1 bar. O2 selectivity was attributed to the metal sites and gate-opening process of RPM3-Zn. In another study, Piscopo et al.22 synthesized two series of fluorine-containing UiO-66 (UiO for University of Oslo) and observed that O2 adsorption capacity of fluorine-containing UiO-66 series improved due to favorable oxygen-fluorine interactions. In a recent study, Gallagher et al.23 showed that MOFs with porphyrin ligand, which is a naturally occurring organic compound, such as PCN-224-Fe (PCN: Porous Coordination Network) and PCN-224-Co can be promising materials for biological O2 transport and storage due to enhanced interactions between O2 and framework atoms. MOFs have been also tested as fillers in polymer membranes. Zornoza et al.24 embedded Cu-BTC and ZIF-8 in a PSF matrix and O2 permeabilities and O2/N2 selectivities of MOF-based MMMs were enhanced compared to those of pure polymer membranes. The selectivity enhancement was more pronounced for ZIF-8 (3.4 Å) due to its smaller pore apertures than Cu-BTC (6 Å). Similarly, Bushell et al.25 fabricated ZIF-8-based PIM-1 (polymer of intrinsic microporosity) membranes. These MMMs surpassed the Robeson’s 2008 upper bound established for O2/N2 separation, when 43 vol% ZIF-8 was embedded into the pure PIM-1. Cu-BTC and MIL-101(Cr) were also used as fillers in polymers and O2 permeabilities of MMMs increased without any significant change in selectivities.9, 26

3 ACS Paragon Plus Environment


These experimental studies showed that MOFs can be promising adsorbent and/or membrane materials for storage and separation of O2. However, the judicious choice of MOFs for each application is challenging due to the enormous number (~70,000) of available MOFs in the Cambridge Structural Database (CSD).27 For this reason, computational studies play a critical role to identify the best candidates prior to experiments. For example, DeCoste et al.28 studied O2 adsorption in 10,000 hypothetical MOFs by performing Grand Canonical Monte Carlo (GCMC) simulations. They reported that NU-125 (NU for Northwestern University) exhibited the highest O2 uptake capacity as 17.4 mol/kg at 298 K and 140 bar, which was also validated by experiments. In a recent study, Moghadam et al.29 computationally screened 2,932 existing MOFs for O2 uptake and reported that the maximum volumetric O2 working capacity (249 cm3 (STP) cm-3) and the maximum gravimetric O2 working capacity (20.4 mol O2/kg) are achieved by UMCM-152 (University of Michigan Crystalline Material) and DIDDOK, respectively. McIntyre et al.4 also performed GCMC simulations of IRMOFs-n (n=1-16), MOF-177 and UiO-66 and showed that IAST (Ideal Adsorption Solution Theory) selectivities of these MOFs for O2/N2 mixture are between 1-1.3 at 298 K. Wang et al.30 examined the effect of metal ions on the O2 adsorption performance of M3(BTC)2- type materials (M = Cr, Mn, Fe, Co, Ni and Cu; BTC = 1,3,5-benzenetricarboxylate acid) and found out that Ni3(BTC)2 shows 11 kJ mol-1 higher interaction energy toward O2 than that for N2, indicating that Ni3(BTC)2 can be a promising material for O2/N2 separation. Similarly, Parkes et al.31 computationally investigated M2(dobdc) (M = Cr, Mn, Fe) MOF series for pure gas and competitive gas adsorption of O2 and N2 and showed that unsaturated metal sites enhance O2 selectivity due to strong interactions between O2 and the metal centers of MOFs. These pioneering studies showed that MOFs with unsaturated metal sites and/or MOFs with functional units (porphyrin ligand and fluorine moieties) exhibit enhanced O2 adsorption. Considering the large number of synthesized MOFs in the literature, it is highly required to examine the potential of different MOFs which have various functional ligands for efficient storage and separation of O2. In this study, we first screened the recent MOF database (69,699 MOFs) and identified 1525 bio-compatible MOFs, which have endogenous organic linkers such as amino acids, porphyrins, nucleobases, proteins and cyclodextrins and metal centers.32 These endogenous linkers have different functional groups in their units which may enhance O2-framework interactions. The performance of MOFs with endogenous organic linkers has not been studied in the literature for storage and separation of O2. We examined the performance of bio-compatible MOFs for adsorption-based and membrane-based O2 4 ACS Paragon Plus Environment

Page 4 of 31

Page 5 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


separation and also for high-pressure O2 storage. We first compared our predictions with the available experimental and computational data for O2 and N2 adsorption in the literature. We then performed GCMC simulations for adsorption-based O2/N2 separation for 315 biocompatible MOFs at 298 K and infinite dilution. Mixture selectivities were then estimated and compared with those calculated at infinite dilution. High-pressure O2 storage is highly required for various medical and industrial applications including the treatment of respiratory insufficiency and the hyperbaric oxygen treatment for carbon monoxide poisoning and enrichment of air in the catalytic units.3 For this reason, we also estimated gravimetric and volumetric O2 working capacities of 315 bio-compatible MOFs at 298 K and 140 bar storage and 5 bar release pressures, considering the conventional pressures used in medical O2 tanks. For the best performing candidates, Molecular Dynamics (MD) simulations were performed to estimate O2 permeabilities and O2/N2 membrane selectivities by using initial loadings obtained from binary mixture (O2/N2:21/79) GCMC simulations at 298 K and 1 bar. Biocompatible MOF membranes which exhibited high O2 permeability and high selectivity towards O2 were finally identified. 2. Computational Methods 2.1. Selection of Bio-compatible MOFs Bio-compatible MOFs have endogenous linkers which are biologically friendly and non-toxic. We used the current MOF database (69,699 MOFs) and a search algorithm available in CSD27 with the keywords as follows: acetamide, acetate, adamantane, adenine, amino, aspartate, citrate, cyclodextrin, dicyanamide, formamide, formate, fumarate, gallate, glutamate, glutarate, glycine, guanine, maleato, malonate, metalloporphyrin, muconate, oxalate, penicillin, peptide, porphyrin, proline, succinate and thymine. These 28 different endogenous linkers were previously defined in the literature.33-34 After this search, we found 1472 MOFs and also added 9 MIL-series (Material of Institute Lavoisier), 10 MOF-74-series and 34 well-known MOFs such as ZIF-8, Cu-BTC and UiO-66 which were previously studied for biomedical applications in the literature.33,

35-38

MIL series are promising materials for

drug storage due to their large surface areas (ranging from 200 to 3000 m2/g).35-37 MOF-74 series are also used as drug carriers due to their non-toxic metal sites (Mg2+ and Zn2+ cations) and large pore sizes (ranged from 11 to 54 Å).38 At the end of this selection criteria, we have a total of 1525 MOFs with endogenous linkers. To examine these MOFs for biomedical applications, metal toxicity which is evaluated by the oral lethal dose parameter (LD50) should 5 ACS Paragon Plus Environment


be also considered. Cu, Fe, Mg, Mn, and Zn metals are commonly preferred based on LD50.33 For this reason, we also investigated these MOFs based on their metal sites. Among 1525 MOFs, the common metal sites are as follows: cobalt (Co in 165 MOFs), copper (Cu in 227 MOFs), iron (Fe in 76 MOFs), magnesium (Mg in 37 MOFs), manganese (Mn in 221 MOFs), zinc (Zn in 260 MOFs) and zirconium (Zr in 18 MOFs). However, a few number of MOFs (122) in this database may have toxicity in metals such as mercury (Hg), cadmium (Cd), lead (Pb) and ruthenium (Ru). The crystal structures of 1525 MOFs were taken from CSD27. We used the same modified crystal structures of bio-MOF-1, -100, -101 and -102 which were previously studied for ibuprofen, caffeine and urea adsorption.15 Prior to molecular simulations, we cleaned solvent molecules with the Python code developed by Moghadam et al.39 Largest cavity diameters (LCDs), pore limiting diameters (PLDs), densities, pore volumes (PVs) and surface areas (SAs) were calculated by Zeo++ software.40 For SA calculations, the trial number was set to 2000 and kinetic radius of N2 was used as 1.86 Å. Probe radius was used as 0 Å for initial screening and the trial number was set to 50,000 for PV calculations. PVs were defined as geometric pore volumes, all the volume of the unit cell that is not overlapping with the atoms of the framework.41 1525 MOFs have various PLDs (0.4-53.3 Å), accessible SAs (05537.3 m2/g) and PVs (0.03-7.9 cm3/g). Structural properties of 1525 MOFs were reported in the supplementary excel file in addition to the information about the type of endogenous linkers and metals in the frameworks. We also provided accessible pore volumes and the porosities using the helium probe (1.29 Å) and these data were also reported in the supplementary excel file. 2.2. Simulation Details GCMC and equilibrium MD simulations were used to study adsorption and diffusion of O2 and N2 in bio-compatible MOFs, respectively. All simulations were performed with the multipurpose code of RASPA simulation package42 at 298 K. Lennard-Jones 12-6 (LJ) and Coulomb potentials were used to model repulsion-dispersion forces and electrostatic interactions, respectively. The partial atomic charges of frameworks were estimated using charge equilibration method as implemented in RASPA to compute electrostatic interactions between gas molecules and the frameworks’ atoms. The electrostatic interactions were calculated by Ewald summation.43 Both O244 and N245 were modeled as a three-site linear molecule with two sites located at two atoms and the third one located at its center of mass 6 ACS Paragon Plus Environment

Page 6 of 31

Page 7 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(COM). The interaction parameters of O2 and N2 were given in Table S1. The potential parameters of MOFs were taken from the Universal Forcefield (UFF).46 This forcefield was selected based on the results of previous simulation studies that were in a good agreement with the experiments for various applications.20, 47 The cut-off radius was set to 13 Å for van der Waals terms. The simulation cell lengths were increased to at least 26 Å for each three dimensions. Periodic boundary conditions were applied, and rigid frameworks were assumed in all simulations. Four types of moves including translation, reinsertion, rotation and swap of the molecule were used in GCMC simulations. Identity change move was also considered in binary mixture GCMC simulations. The Lorentz-Berthelot mixing rules were employed for pair-wise interactions. Peng-Robinson equation of state was used to convert the pressure to the corresponding fugacity. GCMC simulations were carried out for a total of 104 cycles with the first 5×103 cycles for initialization and the last 5×103 cycles for taking ensemble averages. Henry’s constants of O2 and N2 species and isosteric heat of adsorption values were computed at zero-coverage (infinite-dilution) and 298 K by using the Widom particle insertion method.48 105 MC cycles were performed to compute Henry’s constants of O2 and N2 species and isosteric heat of adsorption values. Increasing MC cycles (106) in both single-component GCMC and infinite-dilution simulations did not change the adsorption properties. Considering inaccessible cavities of MOFs in GCMC simulations is important because without blocking inaccessible cavities, computed adsorption isotherms can overestimate the isotherms obtained from experimental measurements.49 Since the lowest PLD among the MOFs we studied in GCMC simulations is 3.72 Å, we did not identify the inaccessible cavities of MOFs and not consider artificially blocking of inaccessible cages in this study. For MD simulations, the initial loadings were taken from the binary mixture GCMC simulations computed at 1 bar and 298 K. Three types of moves including translation, reinsertion and rotation of molecules were considered for MD simulations. MD simulations within the NVT ensemble were performed with a step size of 1 fs up to a total 10 ns at 298 K. Nose-Hoover (NH) thermostat50-51 was used to keep the temperature constant. All MD simulations were carried out for a total of 1×107 cycles with the first 5×104 cycles for initialization and 1×105 cycles for equilibration. More details about these simulation methods can be found in the literature.48 In order to validate our computational methodology, we first compared our predictions with the available experimental and computational data of DeCoste et al.28 and Moghadam et al.29 for single component O2 adsorption. DeCoste et al.28 examined single-component O2 7 ACS Paragon Plus Environment


Page 8 of 31

adsorption in Cu-BTC, NU-125 and UiO-66 at 310 K and various pressures (up to 140 bar).28 Cu-BTC with a refcode, FIQCEN was used in our simulations. Cu-BTC has Cu-Cu-dimers with a 3-dimensional network and mediocre pore sizes (13.7 Å×6.7 Å) and large surface area (2132.2 m2/g).52 NU-125 (with a refcode, REWNEO) has a large surface area (3334.8 m2/g) and high pore volume (1.3 cm3/g).53 UiO-66 (with a refcode, RUBTAK) is a zirconium MOF with 1,4-benzene dicarboxylate (BDC) linkers.54 Comparison with Cu-BTC and NU-125, UiO-66 has small pore sizes (8.6 Å×4.0 Å) and surface area (1073.1 m2/g). O2 uptakes in 10 different MOFs (ANUGIA, ANUGUM, BICDAU, DIDDOK, HIGRIA, HIHNUJ, ICALOP, KEFBEE, MOCKAR and WEBKOF) were also computed at 298 K, 5 bar and 140 bar to be consistent with the study of Moghadam et al.29 Predictions for the single-component N2 adsorption in Cu-BTC, IRMOF-1 and IRMOF-3 were also compared with the experimental data of Siberio-Perez et al.55 and computational data of Karra et al.56 IRMOF-1 and IRMOF-3 have cubic structures with large cavities (15 Å) and surface areas (~3500 m2/g). N2 uptakes in Cu-BTC, IRMOF-1 and IRMOF-3 were computed at 298 K up to 90 bar to be consistent with the literature. 2.3. Adsorbent and Membrane Evaluation Metrics Results obtained from MC simulations were used to predict infinite dilution adsorption 0

selectivities, S

(O2 / N2 )

by using the ratio of Henry’s constants, (KHenry) of each gas component;

S0 ( O 2 / N 2 ) 

K Henry,O2 K Henry, N 2

(2.1)

The isosteric heat of adsorption (Qst0) values at infinite dilution were calculated from:

Q 0 st  RT  ( U 0Total )

(2.2)

0 where, R is the ideal gas constant (kJ/mol  K), T is the temperature (K), U Total is the total

adsorption energy (kJ/mol) calculated at infinite dilution. In binary mixture GCMC simulations, the compositions of the bulk gas mixture were set to O2/N2:21/79 to mimic air composition. Then, mixture adsorption selectivity, Sads ( O2 / N 2 ) was estimated at 1 bar as the ratio of adsorbed amounts (x) of each gas component normalized by their bulk compositions (y):


Page 9 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Sads( O2 / N2 ) 

( x O2 / x N 2 )

(2.3)

( y O2 / y N 2 )

O2 working capacities ( WC O 2 ) also known as deliverable capacities were computed for 315 MOFs which have SAs > 0 m2/g and PLDs > 3.4 Å and at 298 K using the following equation:

WC O2  N ads  N des

(2.4)

where, N ads and N des are the gas uptakes computed at 140 bar (adsorption pressure) and 5 bar (desorption pressure), respectively. Equilibrium MD simulations were performed for the top performing 45 materials consisted of the best 30 candidates which gave the highest gravimetric and volumetric WC O 2 and the best 15 candidates which gave the highest S0 ( O 2 / N 2 ) . The mean square displacement (MSD) of O2 and N2 molecules in their binary mixture was computed as follows:  1 N   MSD(t)    r j (t)  r j (0)  N j1  

2

(2.5)

where N is the number of gas molecules and rj(t) is the position of the jth gas molecules at time t. Self-diffusivities of N2 ( DN2 ,self (cO

2

/ c N2 )

) and O2, ( DO2 ,self (cO

2

/ c N2 )

) were calculated from

the slope of MSD in the computation time limit from 1 to 5 ns. Diffusion selectivity ( Sdiff ( O 2 / N 2 ) ) was then estimated as the ratio of self-diffusivities of gas species at the adsorbed

initial loadings (CO / C N ) in the mixture: 2

2

Sdiff ( O2 / N2 ) 

D O2 ,self ( cO2 / cN 2 ) D N2 ,self ( cO2 / cN 2 )

(2.6)

To predict membrane-based gas separation performances of materials, both mixture adsorption and diffusion data were used. Keskin et al.57 showed a valid approximation to predict membrane selectivity, Smem ( O 2 / N 2 ) of MOFs for a desired gas separation as follows:

Smem ( O2 / N 2 )  Sads( O2 / N 2 )  Sdiff ( O2 / N 2 )


(2.7)


Page 10 of 31

Permeabilities of gas species in their binary mixture, PO 2 and PN2 (mol/m/s/Pa) were determined by using the equation that was suggested by Krishna et al.58:

PO2 

  DO2 ,self ( cO2 / cN2 )  cO2 f O2

, PN2 

  D N2 ,self ( cO2 / cN2 )  c N2 f N2

(2.8)

where,  is the fractional pore volume (porosity), c O 2 and c N 2 are the concentrations of the components at the upstream of the membrane (mol of gas/m3 of accessible pore volume of MOFs). These concentrations are the adsorbed O2 and N2 amounts obtained from binary mixture GCMC simulations at 298 K and 1 bar and defined in terms of accessible pore volume of MOFs. DO2 ,self (cO

2

/ c N2 )

and DN2 ,self (cO

2

/ c N2 )

are the self-diffusivities of gas species

(m2/s) and f O 2 and f N 2 are the bulk phase fugacities (Pa) of the components. To compute the accessible pore volume per unit cell of a MOF and the porosity in Equation 2.8, helium probe was used. Permeability results were converted to Barrer (1 Barrer=3.348×10-16 mol  m/(m2  s

 Pa)), the common gas permeability unit in the literature.11 3. Results and Discussion 3.1. Comparison of simulations with the available data Figure 1 shows comparisons with our simulation results with the available experimental and computational data for O2 and N2 uptakes in different types of MOFs. We first compared our predictions with the experimental measurements of Siberio-Perez et al.55 and the simulation results of Karra et al.56 for N2 uptake in Cu-BTC, IRMOF-1 and IRMOF-3 at 298 K. As shown in Figure 1(a), there is a good agreement between our predictions and experimental measurements for N2 uptake. For example, Siberio-Perez et al.55 measured 6.86 mol N2/kg IRMOF-1 at 90 bar and 298 K and we predicted N2 uptake in IRMOF-1 as 6.76 mol/kg under the same conditions. Similarly, Karra et al.56 measured N2 uptake in Cu-BTC (IRMOF-3) as 5.22 mol/kg (4.78 mol/kg) at 298 K and 42 bar, and our prediction for N2 uptake in Cu-BTC (IRMOF-3) was 5.02 mol/kg (4.95 mol/kg). O2 working capacities ( WC O 2 ) of 10 MOFs, ANUGIA, ANUGUM, BICDAU, DIDDOK, HIGRIA, HIHNUJ, ICALOP, KEFBEE, MOCKAR and WEBKOF were compared with the predictions of Moghadam et al.29 in Figure 1(b). It is important to note that Moghadam et al.29 used density derived electrostatic and chemical (DDEC) charge method for partial charges of MOFs’ atoms. 10 ACS Paragon Plus Environment

Page 11 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Although, in our simulations we used a quick charge equilibration method to assign partial charges of MOFs’ atoms, our results were consistent with the literature.29 This result may be attributed to weak coulombic interactions between gas molecules and the frameworks at high pressure (140 bar - 5 bar for adsorption and desorption, respectively). In Figure 1(c), a good agreement between our predictions and the data of DeCoste et al.28 was shown for the singlecomponent O2 adsorption in Cu-BTC, NU-125 and UiO-66. Simulations slightly underestimated O2 uptake in Cu-BTC at 303 K and 140 bar. Our predictions for O2 adsorption in NU-125 were in a better agreement with the experimental data than those reported in the literature.28 The good agreement between our predictions and experimental measurements motivated us to predict O2 and N2 uptakes in bio-compatible MOFs. 3.2. Adsorption-based O2/N2 separation After validating the accuracy of our computational methodology, we first computed KHenry values of O2 and N2 for 315 MOFs, which have SAs > 0 m2/g and PLDs > 3.4 Å, at infinite dilution and 298 K. KHenry values of O2 (N2) were ranged from 4.04×10-7 to 5.4×10-5 mol/kg/Pa (from 4.28×10-7 to 7.75×10-3 mol/kg/Pa). Figure 2(a) shows that as LCDs of MOFs increase, KHenry values of O2 and N2 generally decrease. The majority of MOFs gave similar KHenry values in magnitude for O2 and N2, indicating that equilibrium-based O2 separation from N2 is difficult. 21 MOFs exhibited higher KHenry values for N2 than those for O2. In these MOFs, as shown in Figure 2(b) Qst0 values for N2 (from 5.4 to 59.4 kJ/mol) were much higher than those for O2 (from 5.6 to 24.5 kJ/mol), indicating stronger adsorption of N2 molecules within these frameworks. This can be attributed to higher quadrupolar moment of N2 (almost 4 times) than that of O2.6 Figure 2(c) shows the relation between LCDs of MOFs and their adsorption selectivities calculated at infinite dilution and 298 K. S0 ( O 2 / N 2 ) of MOFs was in the range of 5×10-4 and 1.51. Since KHenry values of O2 and N2 were close to each other for 257 MOFs, their adsorption selectivities were found to be between 1 and 0.5. These MOFs may not be suitable candidates for adsorption-based O2/N2 separation. 37 MOFs which exhibited 1.0 < S0 ( O 2 / N 2 ) > 1.5 were identified as O2 selective adsorbents whereas 21 MOFs having

S0 ( O 2 / N 2 ) < 0.5 were identified as N2 selective adsorbents. Among these 21 MOFs, 11 MOFs whose LCDs are between 16-17 Å in Figure 2(c) were from CD-MOF (CD: cyclodextrin) series, having open 3-dimensional structures surrounded by free terminal hydroxyl groups and K+ ions. In all these MOFs, Qst0 values for N2 were much higher than those for O2, indicating stronger adsorption of N2 molecules due to the enhanced electrostatic interactions between N2 11 ACS Paragon Plus Environment


molecules and the free cations. Table 1 shows the top performing 15 MOFs ranked based on their S0 ( O 2 / N 2 ) . Among 315 MOFs, OREZES and XACZEH were identified as the top 2 candidates having the highest S0 ( O 2 / N 2 ) ,(1.4-1.5). These MOFs gave much higher Qst0 values for O2 adsorption (14.6-16.5 kJ/mol) than those for N2 adsorption (13.9-14.5 kJ/mol) which may be attributed to their open metal sites including Ca and Cu in OREZES and XACZEH, respectively. McIntyre et al.4 also reported that MOFs which have open metal sites can be great alternatives as adsorbents for O2 separation from air. CATDEH, CATDIL, CAYRIE and CAYROK have almost the same pore volumes (0.4 cm3/g) and pore sizes (~6.5Å× 5Å) due to their common Zn-porphyrin based structures. Thus, these MOFs exhibited almost the same

S0 ( O 2 / N 2 ) , ~1.3. SADZUV (peptide-based MOF with Cu sites) and DEPJIR02 (formate-based MOF with Zn) have the highest pore volume (~0.6 cm3/g) in these 15 MOFs, also they have similar S0 ( O 2 / N 2 ) , 1.2. FIFNUE, FIFNUE01, FIFPAM and FIFPAM01, which have all formate linkers exhibited similar performances due to their common structural properties, such as pore sizes (~5Å× 4Å MOFs) and pore volumes (~0.4 cm3/g). CAYSIE, which is a porphyrin-based MOF with saturated Co sites, gave the lowest adsorption selectivity of O2/N2 as 1.1 at infinite dilution in the top promising MOFs. In many computational screening studies, the adsorption-based gas separation performance of MOFs has been assessed using their infinite dilution adsorption selectivities.16 This approach is highly useful to identify the best candidates among thousands of materials in a reasonable computational time. However, mixture adsorption selectivities should be also considered for the top candidates to better understand their real separation performances for industrial applications. For this reason, we performed binary mixture (O2/N2 :21/79) GCMC simulations for the top 15 promising materials at five different pressures (1, 5, 10, 100 and 140 bar) and 298 K, and compared mixture adsorption selectivities with those obtained from infinite dilution conditions in Figure 3. Figure 3(a) shows the relation between S0 ( O 2 / N 2 ) and binary mixture adsorption selectivities at a wide range of pressures. Binary mixture adsorption selectivities for O2 over N2 were ranged from 1.1 to 1.5 for 15 MOFs at 1, 5 and 10 bar. Results showed that predictions of the infinite dilution simulations can give a precise description about the adsorption-based O2/N2 separation performance of MOFs up to 10 bar. Selectivity is not dependent on the bulk gas composition at low pressures. However, when the pressure increased up to 100 bar, mixture adsorption selectivities for O2/N2 separation differed (1.2-1.8) from those obtained at infinite dilution. This was attributed to packing 12 ACS Paragon Plus Environment

Page 12 of 31

Page 13 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


effects occurred at high pressures. Figure 3(b) shows the distribution of mixture adsorption selectivities of the top 15 MOFs computed from binary mixture GCMC simulations at 298 K and up to 140 bar. As shown in this figure, at low pressures adsorption selectivity is almost constant. While the pressure increased up to 10 bar, binary mixture adsorption selectivities showed a general tendency to increase with increasing pressure. Only for XACZEH, binary mixture adsorption selectivities followed a decreasing trend with increasing pressure. In order to better understand the adsorption mechanism at high pressures, we computed adsorption isotherms of O2 and N2 in two selected MOFs, WIWHUG and XACZEH which exhibited different trends in binary mixture selectivities. WIWHUG showed a sharp increase with increasing pressure, whereas XACZEH showed a sharp decrease with increasing pressure. Figure S1 (S2) shows adsorption isotherms (screenshots) of O2 and N2 in these MOFs at 298 K and up to 140 bar. As shown in Figure S1, O2 and N2 uptakes in WIWHUG were close to saturation at low pressures due to its smaller pore volume (0.3 cm3/g) than XACZEH (0.5 cm3/g). Since WIWHUG has narrow pore windows (4.4 Å×3.9 Å), small O2 molecules (3.46 Å) could find space at high pressures as shown in Figure S2. Therefore, O2/N2 selectivity increased. On the other hand, due to larger spherical-type pore apertures (10.6 Å× 7.2 Å) and higher pore volume of XACZEH, both N2 and O2 molecules were adsorbed in this MOF with increasing pressure. Since N2 adsorption in XACZEH was almost three times higher than O2 adsorption at high pressures, O2/N2 selectivity decreased. Overall, our results showed that adsorption-based O2/N2 separation using bio-compatible MOFs is challenging due to similar interaction energies of gas molecules with the frameworks’ atoms, but MOFs which have non-saturated metal sites can be used as adsorbents due to their higher adsorption selectivities for O2/N2 mixture. 3.2.1. O2 storage in 315 Bio-compatible MOFs After separation of O2 from N2, we tested the performance of MOFs for O2 storage. Since conventional compressed oxygen adsorption (desorption) pressure is 140 (5) bar for high-pressure tanks, we calculated O2 working capacities of 315 bio-compatible MOFs obtained at 140 bar storage and 5 bar desorption pressures at 298 K. Table 2 shows the best performing 15 MOFs ranked based on their gravimetric O2 working capacities. The best performing MOFs exhibited generally high pore volumes (2-8 cm3/g) and large pore apertures (9-54 Å). As shown in Table 2, MOSDIJ (PCN-332(Fe)) gave the highest gravimetric O2 working capacity (52.9 mol/kg) due to its highest pore volume (7.9 cm3/g) and the largest surface area (5668.9 m2/g). ADATEG and RAVXIX had similar working capacities for O2 13 ACS Paragon Plus Environment


(~30 mol/kg) due to their similar pore volumes (3.2 and 3.9 cm3/g, respectively). RAVWIW, RAVWOC, RAVWUI, RAVXAP and RAVXIX are in the family of MOF-74 series and exhibited high O2 working capacities (18-29 mol/kg) due to their large pore volumes and open-metal sites. SAPBIW (also known as bio-MOF-100) and TOCJEC (also known as bioMOF-102), which have adenine in their building units also exhibited high O2 working capacities (22-25 mol/kg) due to their high pore volumes (~3 cm3/g) and free DMA (dimethylammonium) cations which enhanced adsorbate-adsorbent interactions. BEWCUD, BIPSUQ, BORYOY, DOGBEI (also known as PCN-224-(Ni)) and MERLED have all porphyrin in their building units and showed similar O2 working capacities (20-23 mol/kg). Since MERLED has the highest pore volume and the largest pore size among these 5 MOFs, it gave the highest O2 working capacity (23 mol/kg). It is important to note that the top performing MOFs identified in this study outperformed the previously reported promising MOFs, including NU-125 (15.7 mol/kg),28 HKUST-1 (11.9 mol/kg)28 and UMCM-152 (19.6 mol/kg)29 and also traditional adsorbent materials such as activated carbon (8 mol/kg) and NaX zeolite (6.6 mol/kg) in terms of gravimetric O2 working capacity. We would like to add that the top performing MOFs identified for adsorption-based O2/N2 separation (given in Table 1) are not the same those identified for O2 storage as we expected. The pore geometry and/or pore sizes have a significant effect on O2/N2 adsorption selectivity as discussed in the literature.4 Since pore volume has a more pronounced effect on O2 storage, the best MOFs which have the highest O2 working capacities are not the best candidates for adsorbent-based O2/N2 separation. For adsorbent-based O2/N2 separation, MOFs which have mediocre pore volumes such as XACZEH can be selective adsorbents with moderate O2 working capacities. We also calculated volumetric O2 working capacities of 315 MOFs because volumetric capacities are required to determine the size of the tank. Figure S3 shows the relation between gravimetric and volumetric O2 working capacities of 315 MOFs together with their porosities. The MOFs with high gravimetric capacities generally exhibited low volumetric capacities. For example, MOSDIJ has the highest gravimetric O2 working capacity among 315 MOFs, but its volumetric deliverable capacity is moderate (141.8 cm3 (STP)/cm3) due to its very low density (0.12 g/cm3). Large surface area and high pore volume increased the gravimetric O2 working capacities, but low framework density decreased the volumetric O2 working capacities. Similar results were obtained in the literature.29 Table 3 shows the best performing 15 MOFs ranked based on their volumetric O2 working capacities. The best performing MOFs exhibited generally mediocre pore volumes (0.8-1.6 cm3/g) and pore 14 ACS Paragon Plus Environment

Page 14 of 31

Page 15 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


apertures (8-18 Å). As shown in Table 3, ADASOP gave the highest volumetric O2 working capacity (231.9 cm3 (STP)/cm3) due to its high porosity (0.8). Following ADASOP, HIKSIF (common name: Zr-PCN-221(Cu)), LEPKEZ and LEPKOJ gave similar working capacities (~225 cm3 (STP)/cm3) due to their large surface areas (~2800 m2/g) and high porosities (~0.7). ZEZFIV, DUPVER, CUBBEI, SAHYIK, ADUWON, PICZAE and FIQCEN (common name as Cu-BTC) gave similar volumetric O2 working capacities (212.0-218.8 cm3 (STP)/cm3) due to their similar porosities (0.7). Among these 7 MOFs, ZEZFIV gave the highest O2 working capacity because of its free DMA cations which enhance adsorbateadsorbent interactions. RAVVUH, ADUWIH and AGAXIP exhibited similar volumetric O2 working capacity (~210 cm3 (STP)/cm3). Establishing structure-performance relationships is important to understand materials’ behavior and to synthesize materials with desired properties. Therefore, we investigated the relations between gravimetric/volumetric O2 working capacities of MOFs and their physical and chemical structural properties including LCD, PLD, density, PV, SA, porosity (ϕ) and Qst0 in Figure S4. As shown in Figure S4, correlation coefficients (R2 > 0.8) were found to be very high for the relations between gravimetric working capacities and PV, SA and ϕ of MOFs. Since the framework density is an important parameter for volumetric capacities29, instead of PV, ϕ gave a high correlation coefficient (R2 =0.7) for the volumetric capacity. We observed weak correlations between volumetric capacities of MOFs and their LCDs, PLDs, PVs and Qst0 in Figure S4. Figures 4(a) and (b) show that there is an obvious linear relationship between gravimetric O2 working capacities of MOFs and their pore volumes and/or surface areas. O2 working capacities of MOFs increased from 0.8 mol/kg to 52.9 mol/kg, as the pore volumes (0.1-7.9 cm3/g) and surface areas (74.7-5668.9 m2/g) of MOFs increased. Figure 4(c) shows the relation between gravimetric O2 working capacities, infinite dilution adsorption selectivities and O2 heat of adsorption. As O2 working capacities of MOFs increased, their adsorption selectivities and heats of adsorption generally decreased. The highest O2 gravimetric capacity was obtained for MOFs with heats of adsorption (~7 kJ/mol), whose O2/N2 selectivity was around ~1 due to their large pore volumes. Figures 4(d) and (e) show relations between volumetric O2 working capacities of 315 MOFs (32.2-230 cm3 (STP)/cm3) and their porosities (0.3-0.9) and surface areas (74.7-5668.9 m2/g). No obvious correlation was found between volumetric capacities, adsorption selectivities and O2 heat of adsorption as shown in Figure 4(f). The highest O2 volumetric capacity was obtained for MOFs with heats of adsorption between 11-12 kJ/mol, whose O2/N2 selectivity was around ~1 15 ACS Paragon Plus Environment


due to their high porosities between 0.7-0.8. The high heats of adsorption for O2 (20-25 kJ/mol) was observed for MOFs with porosities between 0.4-0.6, resulting in low volumetric capacities. It is important to discuss structural stabilities of these promising 45 MOFs, which were ranked based on their S0 ( O 2 / N 2 ) , gravimetric and volumetric O2 working capacities. Porphyrin based materials such as, CAYSIE59, DOGBEI60 and MERLED (also known as PCN-600-Fe)61 have the highest thermal stabilities almost up to 350 oC. Similarly, 6 materials in MOF-74 series (RAVVUH, RAVWIW, RAVWOC, RAVWUI, RAVXAP and RAVXIX)62 and SAHYIK63 (common name:IRMOF-1) have high thermal stabilities up to 300 oC. These top performing materials have generally high thermal stabilities up to 200 oC except ADUWIH and ADUWON (family structures and also known as MOF-1 and MOF-2, respectively)64, BIPSUQ,65 BORYOY,66 DUPVER67, FIFNUE01,68 FIFPAM0168, LEPKEZ and LEPKOJ (family structures)69, PICZAE70 and XACZEH71 which have thermal stabilities up to almost 50 oC. Chemical and physical stabilities of these MOFs after solvent removal should be also investigated prior to industrial applications. It is also important to note that among the top performing materials, 35 MOFs have open metal sites except BEWCUD, bio-MOF-100 (SAPBIW), bio-MOF-102 (TOCJEC), CAYSIE, CUBBEI, IRMOF-1 (SAHYIK), MIXJOU, MOSDIJ and PICZAE. Developing more accurate theoretical methods (such as DFT-based partial atomic charges) is required to accurately capture the O2 adsorption behavior in these MOFs. In order to examine the effect of partial charge assignment method on O2 adsorption, we selected 6 MOFs with open-metal sites (RAVWAO, UXABOL, XACZEH, XIHSAJ, XIHSEN and YANBAR02) and reperformed our simulations for these MOFs using DDEC charges. Results (given in Table S2) showed that charge method do not significantly affect the gas uptakes and infinite dilution selectivities. However, investigation of a high number of MOFs with open metal sites is highly required to further test its accuracy for different types of MOFs in molecular simulations. 3.3. Membrane-based O2/N2 separation Equilibrium-based O2/N2 separation is challenging due to the similar interactions of O2 and N2 molecules with the atoms of the frameworks. Therefore, we also investigated kineticbased separation performances of the top 44 materials (one MOF, BEWCUD is common


Page 16 of 31

Page 17 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


among the MOFs with the highest gravimetric and volumetric working capacities). For the kinetic-based gas separation, both the knowledge of adsorption and diffusion of gas molecules are important. To examine gas transport in these materials, we performed binary mixture MD simulations using adsorbed O2 and N2 amounts obtained from binary mixture GCMC simulations at 298 K and 1 bar. Figure 5 shows O2 permeabilities (a range from 4.0×103 to 1.8×106 Barrer) and O2/N2 membrane selectivities (0.2-11) of 44 MOF membranes together with the Robeson’s 1991 and 2008 upper bounds. Since MOFs gave high O2 permeabilities, we extrapolated the Robeson’s upper bounds with a dash line to represent the high permeability region. Table 4 shows the top performing 17 materials which surpass the Robeson’s 2008 upper bound. Structural properties of these MOFs including linkers, metal centers and crystal types were given in Table S3. In all these MOFs, smaller O2 molecules (3.46 Å) diffused faster than N2 molecules (3.64 Å) through the pores of MOFs and diffusion selectivity favored O2. Since adsorption selectivities were almost unity in these MOFs, membrane selectivities were governed by diffusion selectivities. RAVWOC is the most promising membrane for O2/N2 separation among 44 MOFs due to its very high O2 permeability (1.8×106 Barrer) and high membrane selectivity (11). High membrane selectivity of RAVWOC was driven by high diffusion selectivity (11) towards O2 because O2 diffusion (4.9×10-3 cm2/s) was one order of magnitude higher than N2 diffusion (4.4×10-4 cm2/s) in this MOF. Following RAVWOC, AGAXIP showed high O2 permeability (8.9×105 Barrer) and moderate membrane selectivity (4.2). Similar to RAVWOC, membrane selectivity of AGAXIP was governed by diffusion selectivity towards O2 (4.2). Among these MOFs, BIPSUQ, BORYOY, CUBBEI, DOGBEI, HIKSIF, RAVXAP, RAVXIX, and SAPBIW (also known as bio-MOF-100) gave high O2 permeabilities (3.3×105 -1.6×106 Barrer) but low membrane selectivities (1.4-2.5). The high O2 permeabilities in these MOFs were attributed to their high pore volumes (0.6-2.9 cm3/g). Among these MOFs, RAVXAP and RAVXIX have isoreticular structures with RAVWOC. However, membrane selectivities of RAVXAP (1.4) and RAVXIX (2.3) were much lower than the membrane selectivity of RAVWOC (11.1). Self-diffusion coefficients of O2 and N2 had the same order of magnitude in these MOFs, resulting in low diffusion selectivities (1.3-2.3). For example, self-diffusion coefficients of O2 (1.3×10-3 cm2/s) and N2 (1×10-3 cm2/s) in RAVXAP were almost the same, indicating a nonselective diffusion. This may be attributed to slightly higher isosteric heat of adsorption value of O2 (11 kJ/mol) than that of N2 (10 kJ/mol) in this MOF. The strong binding of gas molecules retarded the transport of gas molecules.15 DEPJIR02 and WIWHUG exhibited similar performances due to their similar adsorption (1.2) and diffusion (1.6) selectivities. 17 ACS Paragon Plus Environment


CATDIL, CAYRIE, FIFNUE, FIFPAM and SADZUV gave similar mediocre O2 permeabilities (2.7×104 -9.6×104 Barrer) and membrane selectivities (2.3-3.1) due to their similar structural properties including pore volumes (0.05-0.14 cm3/g) and pore sizes (3.9-6.7 Å). Results showed that 17 bio-compatible MOFs outperformed traditional polymeric membranes in terms of O2 permeability and O2/N2 selectivity. However, it is important to note that our simulations do not give any information about the stabilities of these MOF membranes. The permeabilities and selectivities reported in this work were estimated by assuming defect-free and ideal membranes. For real synthesis, cracks and/or defects may be occurred at the interface which can affect the flux of gas molecules through the pores of MOF membranes. Additionally, industrial-scale synthesis of these MOFs requires large-scale production in which the cost and rapid availability of reactants, a high-yield synthesis procedure, and purity of the compounds should be considered.72 Among the top candidates, RAVWOC, RAVXAP and RAVXIX are isoreticular series of MOF-74, which is commercially produced.72 Non-toxic metal sources and/or bio-compatible linkers offer an advantage for environmentally-friendly MOF synthesis.73 However, future experimental work is required to address the challenges related to scalable synthesis of these MOFs. 4. Conclusion In this work, 1525 bio-compatible MOFs were initially identified, and their structural properties were estimated. Both adsorption-based and membrane-based separation of O2 from N2 using these MOFs were assessed at 298 K. O2 working capacities of 315 MOFs which have SAs > 0 m2/g and PLDs > 3.4 Å, were also computed at adsorption (140 bar) and desorption pressures (5 bar). Top 45 MOFs, which gave the best adsorption selectivities, the highest gravimetric and volumetric O2 working capacities at 298 K were identified. Results showed that XACZEH gave the highest adsorption selectivity (1.5) at infinite dilution and 298 K, due to its open Cu sites. Infinite dilution adsorption selectivities were also compared with the binary mixture adsorption selectivities. Results showed that infinite dilution adsorption selectivities can give quick and reasonable predictions about the adsorption-based O2/N2 separation performances of MOFs at low pressures, but mixture GCMC simulations should be performed for more realistic performance predictions of MOFs at high pressures. Among 315 bio-compatible MOFs, MOSDIJ and ADASOP exhibited the highest gravimetric (52.9 mol/kg) and volumetric (231.9 cm3 (STP)/cm3) O2 working capacities, respectively due to their high porosities and large surface areas. We finally performed binary-mixture MD simulations to investigate the membrane-based O2/N2 separation performances of the top 18 ACS Paragon Plus Environment

Page 18 of 31

Page 19 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


promising MOFs. Among these MOFs, RAVWOC gave the highest membrane selectivity (11) due to its high O2 permeability and high O2 selectivity. 17 MOFs outperformed polymeric membranes by surpassing the Robeson’s 2008 upper bound for O2/N2 separation. This computational study will be helpful for identifying the promising bio-compatible MOFs for air separation and O2 capture. The bio-compatible MOF library constructed in this work will also guide both experiments and computational studies, particularly in design and development of bio-compatible MOFs for various biomedical applications in future work. However, further studies are highly required to examine the stability and toxicity of linkers and metals in these materials for medical applications. Acknowledgment: The authors gratefully acknowledge the financial support provided by the Scientific and Technological Research Council of Turkey (TUBITAK) under the project no 217M675. Conflicts of interest: There are no conflicts of interest to declare. Supporting Information: Supporting data associated with this article can be found in online version: interaction parameters for adsorbate atoms used in this work; binary mixture adsorption isotherms for O2/N2 in WIWHUG and XACZEH; GCMC screenshots of O2 and N2 adsorption in WIWHUG and XACZEH at 298 K and at 1, 5, 10, 100 and 140 bar; gravimetric and volumetric working capacities of O2 for 315 MOFs with color mapped by porosity; R2 correlations between working capacities (both gravimetric and volumetric) and seven parameters; comparison between Qeq and DDEC charge methods for Henry constants, O2 working capacities and heats of adsorption of O2 and N2; crystal properties of the top 17 MOF membranes, which surpass the Robeson’s 2008 upper bound. 5. References (1) Gao, S.; Zheng, P.; Li, Z.; Feng, X.; Yan, W.; Chen, S.; Guo, W.; Liu, D.; Yang, X.; Wang, S., Biomimetic O2-Evolving Metal-Organic Framework Nanoplatform for Highly Efficient Photodynamic Therapy Against Hypoxic Tumor. Biomaterials 2018, 178, 83-94. (2) Wang, Y.; Helvensteijn, B.; Nizamidin, N.; Erion, A. M.; Steiner, L. A.; Mulloth, L. M.; Luna, B.; LeVan, M. D., High Pressure Excess Isotherms for Adsorption of Oxygen and Nitrogen in Zeolites. Langmuir 2011, 27 (17), 10648-10656. (3) Alezi, D.; Belmabkhout, Y.; Suyetin, M.; Bhatt, P. M.; Weseliński, Ł. J.; Solovyeva, V.; Adil, K.; Spanopoulos, I.; Trikalitis, P. N.; Emwas, A.-H., MOF Crystal Chemistry Paving the Way to Gas Storage Needs: Aluminum-Based soc-MOF for CH4, O2, and CO2 Storage. J. Am. Chem. Soc. 2015, 137 (41), 13308-13318.



(4) McIntyre, S.; Shan, B.; Wang, R.; Zhong, C.; Liu, J.; Mu, B., Monte Carlo Simulations to Examine the Role of Pore Structure on Ambient Air Separation in Metal-Organic Frameworks. Ind. Eng. Chem. Res. 2018, 57 (28), 9240-9253. (5) Hashim, S.; Mohamed, A.; Bhatia, S., Oxygen Separation from Air Using Ceramic-Based Membrane Technology for Sustainable Fuel Production and Power Generation. Renewable Sustainable Energy Rev. 2011, 15 (2), 1284-1293. (6) Krishna, R., Methodologies for Evaluation of Metal-Organic Frameworks in Separation Applications. RSC Adv. 2015, 5 (64), 52269-52295. (7) Delavar, M.; Nabian, N., An Investigation on the Oxygen and Nitrogen Separation From Air Using Carbonaceous Adsorbents. J. Eng. Sci. Technol. 2015, 10 (11), 1394-1403. (8) Murali, R. S.; Sankarshana, T.; Sridhar, S., Air Separation by Polymer-Based Membrane Technology. Sep. Purif. Rev. 2013, 42 (2), 130-186. (9) Jeazet, H. B. T.; Staudt, C.; Janiak, C., A Method for Increasing Permeability in O 2/N2 Separation with Mixed-Matrix Membranes Made of Water-Stable MIL-101 and Polysulfone. Chem. Commun. 2012, 48 (15), 2140-2142. (10) Rodrigues, M. A.; de Souza Ribeiro, J.; de Souza Costa, E.; de Miranda, J. L.; Ferraz, H. C., Nanostructured Membranes Containing UiO-66(Zr) and MIL-101(Cr) for O2/N2 and CO2/N2 Separation. Sep. Purif. Technol. 2018, 192, 491-500. (11) Robeson, L. M., The Upper Bound Revisited. J. Membr. Sci. 2008, 320 (1-2), 390-400. (12) Himma, N. F.; Wardani, A. K.; Prasetya, N.; Aryanti, P. T.; Wenten, I. G., Recent Progress and Challenges in Membrane-Based O2/N2 Separation. Rev. Chem. Eng. 2018, 1-35. (13) Eddaoudi, M.; Li, H.; Yaghi, O., Highly Porous and Stable Metal-Organic Frameworks: Structure Design and Sorption Properties. J. Am. Chem. Soc. 2000, 122 (7), 1391-1397. (14) Atci, E.; Erucar, I.; Keskin, S., Adsorption and Transport of CH4, CO2, H2 Mixtures in a Bio-MOF Material From Molecular Simulations. J. Phys. Chem. C 2011, 115 (14), 68336840. (15) Erucar, I.; Keskin, S., Efficient Storage of Drug and Cosmetic Molecules in Biocompatible Metal Organic Frameworks: A Molecular Simulation Study. Ind. Eng. Chem. Res. 2016, 55 (7), 1929-1939. (16) Altintas, C.; Erucar, I.; Keskin, S., High-Throughput Computational Screening of the Metal Organic Framework Database for CH4/H2 Separations. ACS Appl. Mater. Interfaces 2018, 10 (4), 3668-3679. (17) Vilela, S. M.; Horcajada, P., MOFs as Supports of Enzymes in Biocatalysis. MetalOrganic Frameworks: Applications in Separations and Catalysis 2018, 447-476. (18) Wang, H.-S., Metal-Organic Frameworks for Biosensing and Bioimaging Applications. Coord. Chem. Rev. 2017, 349, 139-155. (19) Miller, S. E.; Teplensky, M. H.; Moghadam, P. Z.; Fairen-Jimenez, D., Metal-Organic Frameworks as Biosensors for Luminescence-Based Detection and Imaging. Interface Focus 2016, 6 (4), 1-14. (20) Erucar, I.; Keskin, S., Computational Investigation of Metal Organic Frameworks for Storage and Delivery of Anticancer Drugs. J. Mater. Chem. B 2017, 5 (35), 7342-7351. (21) Wang, C.-Y.; Wang, L.; Belnick, A.; Wang, H.; Li, J.; Lueking, A. D., Oxygen-Selective Adsorption in RPM3-Zn Metal Organic Framework. Chem. Eng. Sci. 2017, 165, 122-130. (22) Piscopo, C. G.; Trapani, F.; Polyzoidis, A.; Schwarzer, M.; Pace, A.; Loebbecke, S., Positive Effect of the Fluorine Moiety on the Oxygen Storage Capacity of UiO-66 MetalOrganic Frameworks. New J. Chem. 2016, 40 (10), 8220-8224. (23) Gallagher, A. T.; Kelty, M. L.; Park, J. G.; Anderson, J. S.; Mason, J. A.; Walsh, J. P.; Collins, S. L.; Harris, T. D., Dioxygen Binding at a Four-Coordinate Cobaltous Porphyrin Site in a Metal-Organic Framework: Structural, EPR, and O2 Adsorption Analysis. Inorg. Chem. Front. 2016, 3 (4), 536-540. 20 ACS Paragon Plus Environment

Page 20 of 31

Page 21 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(24) Zornoza, B.; Seoane, B.; Zamaro, J. M.; Téllez, C.; Coronas, J., Combination of MOFs and Zeolites for Mixed‐Matrix Membranes. ChemPhysChem 2011, 12 (15), 2781-2785. (25) Bushell, A. F.; Attfield, M. P.; Mason, C. R.; Budd, P. M.; Yampolskii, Y.; Starannikova, L.; Rebrov, A.; Bazzarelli, F.; Bernardo, P.; Jansen, J. C., Gas Permeation Parameters of Mixed Matrix Membranes Based on the Polymer of Intrinsic Microporosity PIM-1 and the Zeolitic Imidazolate Framework ZIF-8. J. Membr. Sci. 2013, 427, 48-62. (26) Duan, C.; Jie, X.; Liu, D.; Cao, Y.; Yuan, Q., Post-Treatment Effect on Gas Separation Property of Mixed Matrix Membranes Containing Metal Organic Frameworks. J. Membr. Sci. 2014, 466, 92-102. (27) Allen, F. H., The Cambridge Structural Database: A Quarter of a Million Crystal Structures and Rising. Acta Crystallogr., Sect. B: Struct. Sci. 2002, 58 (3), 380-388. (28) DeCoste, J. B.; Weston, M. H.; Fuller, P. E.; Tovar, T. M.; Peterson, G. W.; LeVan, M. D.; Farha, O. K., Metal-Organic Frameworks for Oxygen Storage. Angew. Chem. 2014, 126 (51), 14316-14319. (29) Moghadam, P. Z.; Islamoglu, T.; Goswami, S.; Exley, J.; Fantham, M.; Kaminski, C. F.; Snurr, R. Q.; Farha, O. K.; Fairen-Jimenez, D., Computer-Aided Discovery of a MetalOrganic Framework with Superior Oxygen Uptake. Nat. Commun. 2018, 9 (1), 1378. (30) Wang, Y.; Yang, J.; Li, Z.; Zhang, Z.; Li, J.; Yang, Q.; Zhong, C., Computational Study of Oxygen Adsorption in Metal-Organic Frameworks with Exposed Cation Sites: Effect of Framework Metal Ions. RSC Adv. 2015, 5 (42), 33432-33437. (31) Parkes, M. V.; Greathouse, J. A.; Hart, D. B.; Gallis, D. F. S.; Nenoff, T. M., Ab Initio Molecular Dynamics Determination of Competitive O2 vs. N2 Adsorption at Open Metal Sites of M2(dobdc). Phys. Chem. Chem. Phys. 2016, 18 (16), 11528-11538. (32) McKinlay, A. C.; Morris, R. E.; Horcajada, P.; Férey, G.; Gref, R.; Couvreur, P.; Serre, C., BioMOFs: Metal-Organic Frameworks for Biological and Medical Applications. Angew. Chem., Int. Ed. 2010, 49 (36), 6260-6266. (33) Horcajada, P.; Gref, R.; Baati, T.; Allan, P. K.; Maurin, G.; Couvreur, P.; Ferey, G.; Morris, R. E.; Serre, C., Metal-Organic Frameworks in Biomedicine. Chem. Rev. 2011, 112 (2), 1232-1268. (34) Cai, H.; Huang, Y.-L.; Li, D., Biological Metal-Organic Frameworks: Structures, HostGuest Chemistry and Bio-Applications. Coord. Chem. Rev. 2018, DOI: 10.1016/j.ccr.2017.12.003. (35) Horcajada, P.; Serre, C.; Vallet-Regí, M.; Sebban, M.; Taulelle, F.; Férey, G., MetalOrganic Frameworks as Efficient Materials for Drug Delivery. Angew. Chem. 2006, 118 (36), 6120-6124. (36) Horcajada, P.; Serre, C.; Maurin, G.; Ramsahye, N. A.; Balas, F.; Vallet-Regi, M.; Sebban, M.; Taulelle, F.; Férey, G., Flexible Porous Metal-Organic Frameworks for a Controlled Drug Delivery. J. Am. Chem. Soc. 2008, 130 (21), 6774-6780. (37) Horcajada, P.; Chalati, T.; Serre, C.; Gillet, B.; Sebrie, C.; Baati, T.; Eubank, J. F.; Heurtaux, D.; Clayette, P.; Kreuz, C., Porous Metal-Organic Framework Nanoscale Carriers as a Potential Platform for Drug Delivery and Imaging. Nat. Mater. 2010, 9 (2), 172. (38) Hu, Q.; Yu, J.; Liu, M.; Liu, A.; Dou, Z.; Yang, Y., A Low Cytotoxic Cationic MetalOrganic Framework Carrier for Controllable Drug Release. J. Med. Chem. 2014, 57 (13), 5679-5685. (39) Moghadam, P. Z.; Li, A.; Wiggin, S. B.; Tao, A.; Maloney, A. G.; Wood, P. A.; Ward, S. C.; Fairen-Jimenez, D., Development of a Cambridge Structural Database Subset: A Collection of Metal-Organic Frameworks for Past, Present, and Future. Chem. Mater. 2017, 29 (7), 2618-2625.



(40) Willems, T. F.; Rycroft, C. H.; Kazi, M.; Meza, J. C.; Haranczyk, M., Algorithms and Tools for High-Throughput Geometry-Based Analysis of Crystalline Porous Materials. Microporous Mesoporous Mater. 2012, 149 (1), 134-141. (41) Ongari, D.; Boyd, P. G.; Barthel, S.; Witman, M.; Haranczyk, M.; Smit, B., Accurate Characterization of the Pore Volume in Microporous Crystalline Materials. Langmuir 2017, 33 (51), 14529-14538. (42) Dubbeldam, D.; Calero, S.; Ellis, D. E.; Snurr, R. Q., RASPA: Molecular Simulation Software for Adsorption and Diffusion in Flexible Nanoporous Materials. Mol. Simul. 2016, 42 (2), 81-101. (43) Ewald, P. P., Die Berechnung Optischer und Elektrostatischer Gitterpotentiale. Ann. Phys. 1921, 369 (3), 253-287. (44) Mellot, C.; Lignieres, J., Monte Carlo Simulations of N2 and O2 Adsorption in Silicalite and CaLSX Zeolites. Mol. Simul. 1997, 18 (6), 349-365. (45) Makrodimitris, K.; Papadopoulos, G. K.; Theodorou, D. N., Prediction of Permeation Properties of CO2 and N2 Through Silicalite via Molecular Simulations. J. Phys. Chem. B 2001, 105 (4), 777-788. (46) Rappé, A. K.; Casewit, C. J.; Colwell, K.; Goddard Iii, W.; Skiff, W., UFF, a Full Periodic Table Force Field for Molecular Mechanics and Molecular Dynamics Simulations. J. Am. Chem. Soc. 1992, 114 (25), 10024-10035. (47) Keskin, S.; Liu, J.; Rankin, R. B.; Johnson, J. K.; Sholl, D. S., Progress, Opportunities, and Challenges for Applying Atomically Detailed Modeling to Molecular Adsorption and Transport in Metal-Organic Framework Materials. Ind. Eng. Chem. Res. 2008, 48 (5), 23552371. (48) Frenkel, D.; Smit, B., Understanding Molecular Simulation: From Algorithms to Applications. Academic press: 2001. (49) Zhang, K.; Nalaparaju, A.; Chen, Y.; Jiang, J., Crucial Role of Blocking Inaccessible Cages in the Simulation of Gas Adsorption in a Paddle-Wheel Metal-Organic Framework. RSC Adv. 2013, 3 (36), 16152-16158. (50) Nosé, S., A Unified Formulation of the Constant Temperature Molecular Dynamics Methods. J. Chem. Phys. 1984, 81 (1), 511-519. (51) Hoover, W. G., Canonical Dynamics: Equilibrium Phase-Space Distributions. Phys. Rev. A 1985, 31 (3), 1695-1697. (52) Chui, S. S.-Y.; Lo, S. M.-F.; Charmant, J. P.; Orpen, A. G.; Williams, I. D., A Chemically Functionalizable Nanoporous Material [Cu3(TMA)2(H2O)3]n. Science 1999, 283 (5405), 1148-1150. (53) Wilmer, C. E.; Farha, O. K.; Yildirim, T.; Eryazici, I.; Krungleviciute, V.; Sarjeant, A. A.; Snurr, R. Q.; Hupp, J. T., Gram-Scale, High-Yield Synthesis of a Robust Metal-Organic Framework for Storing Methane and Other Gases. Energy Environ. Sci. 2013, 6 (4), 11581163. (54) Cavka, J. H.; Jakobsen, S.; Olsbye, U.; Guillou, N.; Lamberti, C.; Bordiga, S.; Lillerud, K. P., A New Zirconium Inorganic Building Brick Forming Metal Organic Frameworks with Exceptional Stability. J. Am. Chem. Soc. 2008, 130 (42), 13850-13851. (55) Siberio-Pérez, D. Y.; Wong-Foy, A. G.; Yaghi, O. M.; Matzger, A. J., Raman Spectroscopic Investigation of CH4 and N2 Adsorption in Metal-Organic Frameworks. Chem. Mater. 2007, 19 (15), 3681-3685. (56) Karra, J. R.; Walton, K. S., Molecular Simulations and Experimental Studies of CO2, CO, and N2 Adsorption in Metal-Organic Frameworks. J. Phys. Chem. C 2010, 114 (37), 15735-15740.


Page 22 of 31

Page 23 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(57) Keskin, S.; Sholl, D. S., Efficient Methods for Screening of Metal Organic Framework Membranes for Gas Separations Using Atomically Detailed Models. Langmuir 2009, 25 (19), 11786-11795. (58) Krishna, R.; van Baten, J. M, In Silico Screening of Zeolite Membranes for CO2 Capture. J. Membr. Sci. 2010, 360 (1-2), 323-333. (59) Lin, K. J., SMTP-1: The First Functionalized Metalloporphyrin Molecular Sieves with Large Channels. Angew. Chem., Int. Ed. 1999, 38 (18), 2730-2732. (60) Feng, D.; Chung, W.-C.; Wei, Z.; Gu, Z.-Y.; Jiang, H.-L.; Chen, Y.-P.; Darensbourg, D. J.; Zhou, H.-C., Construction of Ultrastable Porphyrin Zr Metal-Organic Frameworks Through Linker Elimination. J. Am. Chem. Soc. 2013, 135 (45), 17105-17110. (61) Wang, K.; Feng, D.; Liu, T.-F.; Su, J.; Yuan, S.; Chen, Y.-P.; Bosch, M.; Zou, X.; Zhou, H.-C., A Series of Highly Stable Mesoporous Metalloporphyrin Fe-MOFs. J. Am. Chem. Soc. 2014, 136 (40), 13983-13986. (62) Deng, H.; Grunder, S.; Cordova, K. E.; Valente, C.; Furukawa, H.; Hmadeh, M.; Gándara, F.; Whalley, A. C.; Liu, Z.; Asahina, S., Large-Pore Apertures in a Series of MetalOrganic Frameworks. Science 2012, 336 (6084), 1018-1023. (63) Li, H.; Eddaoudi, M.; O'Keeffe, M.; Yaghi, O. M., Design and Synthesis of an Exceptionally Stable and Highly Porous Metal-Organic Framework. Nature 1999, 402 (6759), 276. (64) Wang, Y.; Yang, J.; Liu, Y. Y.; Ma, J. F., Controllable Syntheses of Porous MetalOrganic Frameworks: Encapsulation of LnIII Cations for Tunable Luminescence and Small Drug Molecules for Efficient Delivery. Chem.-Eur. J. 2013, 19 (43), 14591-14599. (65) Gao, W.-Y.; Wojtas, L.; Ma, S., A Porous Metal-Metalloporphyrin Framework Featuring High-Density Active Sites for Chemical Fixation of CO2 Under Ambient Conditions. Chem. Commun. 2014, 50 (40), 5316-5318. (66) Williams, D. E.; Rietman, J. A.; Maier, J. M.; Tan, R.; Greytak, A. B.; Smith, M. D.; Krause, J. A.; Shustova, N. B., Energy Transfer on Demand: Photoswitch-Directed Behavior of Metal-Porphyrin Frameworks. J. Am. Chem. Soc. 2014, 136 (34), 11886-11889. (67) Cai, H.; Li, M.; Lin, X. R.; Chen, W.; Chen, G. H.; Huang, X. C.; Li, D., Spatial, Hysteretic, and Adaptive Host-Guest Chemistry in a Metal-Organic Framework with Open Watson-Crick Sites. Angew. Chem., Int. Ed. 2015, 54 (36), 10454-10459. (68) Wang, Y.; Cao, R.; Bi, W.; Li, X.; Yuan, D.; Sun, D., Four Novel Porous Frameworks Constructed by Formate Ligand. Microporous Mesoporous Mater. 2006, 91 (1), 215-220. (69) He, W.-L.; Yang, X.-L.; Zhao, M.; Wu, C.-D., Suspending Ionic Single-Atom Catalysts in Porphyrinic Frameworks for Highly Efficient Aerobic Oxidation at Room Temperature. J. Catal. 2018, 358, 43-49. (70) Zou, C.; Zhang, T.; Xie, M.-H.; Yan, L.; Kong, G.-Q.; Yang, X.-L.; Ma, A.; Wu, C.-D., Four Metalloporphyrinic Frameworks as Heterogeneous Catalysts for Selective Oxidation and Aldol Reaction. Inorg. Chem. 2013, 52 (7), 3620-3626. (71) García-Terán, J. P.; Castillo, O.; Luque, A.; García-Couceiro, U.; Román, P.; Lezama, L., An Unusual 3D Coordination Polymer Based on Bridging Interactions of the Nucleobase Adenine. Inorg. Chem. 2004, 43 (15), 4549-4551. (72) Firmino, A. D.; Mendes, R. F.; Tomé, J. P.; Almeida Paz, F. A., Synthesis of MOFs at the Industrial Scale. Metal‐Organic Frameworks: Applications in Separations and Catalysis 2018, 57-80. (73) Julien, P. A.; Mottillo, C.; Friščić, T., Metal-Organic Frameworks Meet Scalable and Sustainable Synthesis. Green Chem. 2017, 19 (12), 2729-2747.



TOC / Graphical Abstract

Table 1. Top performing 15 MOFs identified for S0 ( O 2 / N 2 ) at 298 K together with their calculated physical properties. MOFs

LCD - PLD (Å)

PV (cm3/g)

S0 ( O 2 / N 2 )

XACZEH

10.61 – 7.24

0.50

1.51

OREZES

6.25 – 4.66

0.22

1.44

CATDEH

6.43 – 4.83

0.40

1.29

CAYROK

6.46 – 4.90

0.39

1.28

MIXJOU

5.04 – 3.82

0.48

1.25

CAYRIE

6.67 – 5.10

0.41

1.23

CATDIL

6.68 – 5.04

0.40

1.23

SADZUV

5.80 – 4.08

0.63

1.21

DEPJIR02

4.39 – 3.81

0.58

1.21

FIFPAM

4.70 – 3.85

0.40

1.19

WIWHUG

4.44 – 3.92

0.30

1.18

FIFPAM01

4.72 – 3.78

0.41

1.18

FIFNUE01

4.77 – 3.93

0.41

1.18

FIFNUE

4.78 – 3.96

0.41

1.16

CAYSIE

7.01 – 4.99

0.39

1.14


Page 24 of 31

Page 25 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Table 2. Top performing 15 MOFs identified for gravimetric WC O 2 obtained at 140 bar adsorption and 5 bar desorption pressures at 298 K together with their calculated physical properties and S0 ( O 2 / N 2 ) . MOF

LCD – PLD (Å)

PV (cm3/g)

S0 ( O 2 / N 2 )

WC O 2 (mol/kg)

MOSDIJ

48.55 – 29.10

7.96

1.03

52.85

ADATEG

26.34 – 13.79

3.18

1.03

32.99

RAVXIX

53.58 – 53.26

3.85

1.01

29.42

TOCJEC

31.40 – 26.28

3.21

1.02

24.72

RAVXAP

34.86 – 34.36

2.95

1.06

24.40

MERLED

26.46 – 25.19

2.44

1.00

22.70

RAVWUI

36.79 – 36.43

2.54

1.04

21.63

SAPBIW

20.23 – 14.72

2.64

1.03

21.61

BIPSUQ

18.61 – 16.93

1.92

1.00

20.78

BORYOY

18.18 – 14.41

2.08

0.98

20.69

RAVWIW

30.70 – 30.15

2.28

1.00

20.25

BEWCUD

10.59 – 8.54

1.61

1.01

19.87

ADUROI

25.67 – 17.11

2.08

1.01

19.71

DOGBEI

26.04 – 14.54

1.98

0.99

19.66

RAVWOC

28.22 – 27.56

2.10

0.99

17.97

Table 3. Top performing 15 MOFs identified for volumetric WC O 2 obtained at 140 bar adsorption and 5 bar desorption pressures at 298 K together with their calculated physical properties and S0 ( O 2 / N 2 ) . MOF

LCD – PLD (Å)

PV (cm3/g)

S0 ( O 2 / N 2 )

WC O 2 (cm3 (STP)/cm3)

ADASOP

18.04 – 6.69

1.22

1.00

231.85

HIKSIF

16.57 – 4.85

0.97

0.97

226.17

LEPKEZ

10.10 – 5.73

0.99

0.95

225.71

BEWCUD

10.59 – 8.54

1.61

1.01

222.52

LEPKOJ

10.45 – 5.89

0.93

0.93

221.69

ZEZFIV

10.69 – 9.42

1.36

0.39

218.87

DUPVER

9.93 – 6.47

0.92

0.98

216.16

CUBBEI

11.32 – 8.35

1.12

1.01

215.02

SAHYIK

14.95 - 7.84

1.33

1.02

214.27

ADUWON

16.38 – 6.00

1.48

0.98

213.76

PICZAE

8.79 – 4.97

0.97

0.95

212.95

FIQCEN

13.19 – 6.67

0.82

1.03

212.02

RAVVUH

17.18 – 16.38

1.23

0.92

211.82

ADUWIH

9.28 – 6.55

0.99

1.02

210.25

AGAXIP

8.05 – 6.78

0.84

0.99

209.65



Page 26 of 31

Table 4. Adsorption, diffusion and membrane selectivity data together with gas diffusivities and permeabilities for MOF membranes which surpass Robeson’s 2008 upper bound for O2/N2 separation MOF

D O 2 ,self (cm2/s) D N 2 ,self (cm2/s) PO

2

(Barrer) PN 2 (Barrer) Sads ( O 2 / N 2 ) Sdiff ( O 2 / N 2 ) Smem ( O 2 / N 2 )

AGAXIP

7.21 × 10-4

1.70 × 10-4

8.94 × 105 2.14 × 105

0.99

4.23

4.19

BIPSUQ

1.40 × 10-3

9.29 × 10-4

5.47 × 105 3.65 × 105

1.00

1.50

1.50

BORYOY

1.02 × 10-3

4.38 × 10-4

4.59 × 105 2.00 × 105

0.98

2.34

2.29

CATDIL

1.34 × 10-4

6.19 × 10-5

4.35 × 104 1.61 × 104

1.26

2.16

2.72

CAYRIE

2.68 × 10-4

1.08 × 10-4

9.63 × 104 3.10 × 104

1.25

2.48

3.10

CUBBEI

5.60 × 10-4

3.27 × 10-4

3.38 × 105 1.97 × 105

1.00

1.71

1.71

DEPJIR02

2.34 × 10-4

1.43 × 10-4

1.49 × 105 7.56 × 104

1.20

1.63

1.96

DOGBEI

1.20 × 10-3

4.75 × 10-4

4.64 × 105 1.85 × 105

0.99

2.53

2.50

FIFNUE

4.49 × 10-5

2.05 × 10-5

4.57 × 104 1.79 × 104

1.17

2.19

2.56

FIFPAM

2.71 × 10-5

1.08 × 10-5

2.71 × 104 9.01 × 103

1.19

2.52

3.00

HIKSIF

3.63 × 10-4

2.23 × 10-4

3.29 × 105 2.06 × 105

0.98

1.63

1.60

RAVWOC

4.88 × 10-3

4.44 × 10-4

1.82 × 106 1.64 × 105

1.01

10.99

11.10

RAVXAP

1.31 × 10-3

1.01 × 10-3

4.85 × 105 3.53 × 105

1.06

1.30

1.38

RAVXIX

5.60 × 10-3

2.48 × 10-3

1.62 × 106 7.15 × 105

1.00

2.25

2.25

SADZUV

5.32 × 10-5

2.81 × 10-5

3.13 × 104 1.34 × 104

1.23

1.90

2.34

SAPBIW

1.67 × 10-3

7.89 × 10-4

4.25 × 105 1.91 × 105

1.05

2.12

2.23

WIWHUG

1.09 × 10-4

6.72 × 10-5

2.77 × 105 1.42 × 105

1.20

1.62

1.94


8

(a)

6

2

N2 uptake (mol/kg)

4 Simulations/Cu-BTC Simulations/Cu-BTC(Karra et al., 2010) Simulations/IRMOF-3 Simulations/IRMOF-3(Karra et al., 2010) Simulations/IRMOF-1 Experiments/IRMOF-1(Siberio-Perez et al., 2007)

2

0 0

20

40

60

80

100

22

(b) Simulations (Moghadam et al., 2018)

20

R2=0.99 ANUGIA ANUGUM BICDAU DIDDOK HIGRIA HIHNUJ ICALOP KEFBEE MOCKAR WEBKOF

18 16 14 12 12

20

14

16

18

20

22

Our predictions for WCO (mol/kg)

Pressure (bar)

O2 uptake (mol/kg)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Literature data for WCO (mol/kg)

Page 27 of 31

2

(c)

15 Simulations/Cu-BTC Simulations/Cu-BTC(DeCoste et al.,2014) Experiments/Cu-BTC(DeCoste et al.,2014) Simulations/NU-125 Simulations/NU-125(DeCoste et al.,2014) Experiments/NU-125(DeCoste et al.,2014) Simulations/UiO-66 Experiments/UiO-66(DeCoste et al.,2014)

10

5

0 0

20

40

60

80

100

120

140

Pressure (bar)

Figure 1. Comparison of our predictions with the literature for (a) N2 uptake in Cu-BTC, IRMOF-1 and IRMOF-3 at 298 K, (b) O2 working capacities of 10 MOFs at 298 K, (c) O2 uptake in Cu-BTC, NU-125 and UiO-66 at 303 K.



10-2

(a)

10-2 (b)

O2 N2

10-3

KHenry (mol/kg/Pa)

KHenry (mol/kg/Pa)

10-4 10-5 10-6

10-3 10-4 10-5 10-6

10-7 0

10

20

30

40

50

10-7

60

0

20

LCD (Å)

40 Q0st (kJ/mol)

2.0 (c) OREZES XACZEH

1.5

S0O /N 2 2

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 31

1.0

0.5

0.0 0

10

20

30

40

50

60

LCD (Å)

Figure 2. Relations between (a) KHenry of gases and LCDs of MOFs, (b) KHenry of gases and Q0st and (c) S0 and LCDs of MOFs at 298 K.


60

Page 29 of 31

2.0 (a)

Sads (O /N ) 2 2

1.8

1.6 1 bar 5 bar 10 bar 100 bar 140 bar

1.4

1.2 1.2

1.4

1.6

1.8

2.0

0

S (O /N ) 2 2 2.0

CATDEH CATDIL CAYRIE CAYROK CAYSIE DEPJIR02 FIFNUE FIFNUE01 FIFPAM FIFPAM01 MIXJOU OREZES SADZUV WIWHUG XACZEH

(b)

1.8

Sads (O /N ) 2 2

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1.6

1.4

1.2 100

101

102

Pressure (bar)

Figure 3. (a) Comparison of infinite dilution adsorption selectivities with mixture adsorption selectivities at 1, 5, 10, 100 and 140 bar and (b) Mixture adsorption selectivities of top 15 MOFs ranked based on S0 ( O 2 / N 2 ) up to 140 bar at 298 K.



300

(a)

2

WC of O2 (cm (STP)/cm )

60

(d)

3

R =0.90

40

2

R =0.68

200

3

WC of O2 (mol/kg)

80

20 0 0

2

4

6

8

10

100

0 0.2

0.4

Pore volume (cm /g) 300

(b)

2

R =0.85

20

0

(e)

200

100

0

0

00 00 00 00 00 00 10 20 30 40 50 60

Surface area (m /g) 2.0

0

(c)

Qst (kJ/mol)

1.5

5

1.0

14

(f)

1.5

0.5

20 25

0.0 0

20

40

O2/N2

16

0

10

S

O2/N2

00 00 00 00 00 00 10 20 30 40 50 60 2

2

0

2

R =0.69

Surface area (m /g) 2.0

1.0

3

40

0

0.8

3


WC of O2 (mol/kg)

60

0.6



3

S

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 31

1.0 0.5 0.0 0

60

100

200 3

300 3


WC of O2 (mol/kg)

Figure 4. Relations between (a) pore volume and gravimetric WC O 2 , (b) surface area and gravimetric WC O 2 , (c) gravimetric WC O 2 and S0 ( O 2 / N 2 ) , (d) porosity and volumetric WC O 2 , (e) surface area and volumetric WC O 2 , (f) volumetric WC O 2 and S0 ( O 2 / N 2 ) of MOFs at 298 K. The data points in (c) and (f) graphs are color coded by infinite dilution isosteric heat of adsorption of O2.


Page 31 of 31

2

10

Ro Ro 1

10

be

be

son

son

's u

's u

pp

pp

er

er

bo bo

un

un

2

2

Smem(O /N )

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


d-

d-

20 19

RAVWOC

08 AGAXIP CAYRIE DOGBEI RAVXIX

91

0

10

0

Top 15 having the highest S

O2/N2

Top 15 having the highest gravimetric WCO

2

Top 15 having the highest volumetric WCO

-1

10

2

-4

10

-2

10

0

10

2

4

10 10 PO (Barrer)

6

10

8

10

2

Figure 5. Membrane selectivity and O2 permeability of promising MOF membranes for O2/N2 separations.


Bio-compatible MOFs for Storage and Separation of O2: A Molecular

Recommend Documents