Highly Productive C3H4/C3H6 Trace Separation by a Packing Polymorph of a Layered Hybrid Ultramicroporous Material

Ultramicroporous materials can be highly effective at trace gas separations when they offer a high density of selective binding sites. Herein, we report that sql-NbOFFIVE-bpe-Cu, a new variant of a previously reported ultramicroporous square lattice, sql, topology material, sql-SIFSIX-bpe-Zn, can exist in two polymorphs. These polymorphs, sql-NbOFFIVE-bpe-Cu-AA (AA) and sql-NbOFFIVE-bpe-Cu-AB (AB), exhibit AAAA and ABAB packing of the sql layers, respectively. Whereas NbOFFIVE-bpe-Cu-AA (AA) is isostructural with sql-SIFSIX-bpe-Zn, each exhibiting intrinsic 1D channels, sql-NbOFFIVE-bpe-Cu-AB (AB) has two types of channels, the intrinsic channels and extrinsic channels between the sql networks. Gas and temperature induced transformations of the two polymorphs of sql-NbOFFIVE-bpe-Cu were investigated by pure gas sorption, single-crystal X-ray diffraction (SCXRD), variable temperature powder X-ray diffraction (VT-PXRD), and synchrotron PXRD. We observed that the extrinsic pore structure of AB resulted in properties with potential for selective C3H4/C3H6 separation. Subsequent dynamic gas breakthrough measurements revealed exceptional experimental C3H4/C3H6 selectivity (270) and a new benchmark for productivity (118 mmol g–1) of polymer grade C3H6 (purity >99.99%) from a 1:99 C3H4/C3H6 mixture. Structural analysis, gas sorption studies, and gas adsorption kinetics enabled us to determine that a binding “sweet spot” for C3H4 in the extrinsic pores is behind the benchmark separation performance. Density-functional theory (DFT) calculations and Canonical Monte Carlo (CMC) simulations provided further insight into the binding sites of C3H4 and C3H6 molecules within these two hybrid ultramicroporous materials, HUMs. These results highlight, to our knowledge for the first time, how pore engineering through the study of packing polymorphism in layered materials can dramatically change the separation performance of a physisorbent.

S4 step-size of 0.0167113° and a scan time of 50 seconds per step. Crude data were analyzed using the X'Pert HighScore Plus™ software V 4.1 (PANalytical, The Netherlands). The sample was heated up to 453 K under N2 atmosphere.
Aluminium pans and a flow rate of 60 cm 3 min -1 for the nitrogen gas were used for the experiments. The data was collected in the High Resolution Dynamic mode with a sensitivity of 1.0, a resolution of 4.0, and a temperature ramp of 10 °C min -1 up to 550 °C. The data was evaluated using the T.A. Universal Analysis suite for Windows XP/Vista Version 4.5A.

Dynamic Vapor Sorption (DVS)
Dynamic water vapour sorption studies were performed on the sample using a dynamic vapour sorption system (Surface Measurement Systems, DVS Adventure) which gravimetrically measures the uptake and loss of vapour using air as a carrier gas. Pure water was used as the adsorbate for these measurements and temperature was maintained at 298 K by enclosing the system in a temperature-controlled incubator. Prior to measurement, the sample was in-situ activated at 100 o C, 0 RH. The mass of the sample was determined by comparison to an empty reference pan and recorded by a high resolution microbalance with a precision of 0.1 µg.
Sorption isotherm was measured from 0 to 95% RH step-wise with a convergence equilibrium criterion dm/dt =0.05 %/min. The minimum and maximum equilibration times for each step were 10 and 360 min, respectively. The kinetics were measured between two points 0 and 30% RH with a convergence equilibrium criterion dm/dt =0.05 %/min. The recyclability test was done by performing 10 cycles, each cycle consisting of 15 min adsorption step (90% RH) and 30 min desorption step (0% RH).

Gas sorption measurements.
For gas sorption experiments, gases were used as received from BOC Gases Ireland: He (99.999%), N2 (99.9995%), CO2 (99.995%), C3H4 (97.0%), C3H6 (99.5%). Before sorption measurements, activation of sql-NbOFFIVE-bpe-Cu-AB and sql-NbOFFIVE-bpe-Cu-AA were achieved by degassing the air-dried samples on a SmartVacPrep™ using dynamic vacuum and heating for 12 h (the sample heated from RT to 333 K with a ramp rate of 1°C min -1 ). About 100 mg of activated samples were used for the measurements on Micromeritics Tristar II 3030 or Micromeritics 3Flex surface area and pore size analyzer 3500. A Julabo temperature controller was used to maintain a constant temperature in the bath throughout the experiment. The bath temperatures of 273 and 298 K were precisely controlled with a Julabo ME (v.2) recirculating control system containing a mixture of ethylene glycol and water. The low temperatures at 77 K and 195 K were controlled by a 4 L Dewar filled with liquid N2 and dry ice/acetone, respectively. At every interval of two independent isotherms recorded for any sorbent, samples were regenerated by degassing over 30 min under high vacuum at 333 K, before commencing the next sorption experiment.

Breakthrough experiments.
In typical breakthrough experiments, MeOH exchanged sql-NbOFFIVE-bpe-Cu-AB (~0.5 g) or MeOH exchanged sql-NbOFFIVE-bpe-Cu-AA (~0.35 g) was placed in quartz tubing (8 mm diameter; 8 mm x 6 mm x 400 mm) to form fixed beds. First, the adsorbent bed was purged under a 20 cm 3 min -1 flow of He gas at 353 K for 6 hours prior to breakthrough experiment.
Upon cooling to room temperature, the gas flow was switched to the desired C3H4/C3H6 gas mixture compositions (1:99), maintained at a total flow rate of 1.0 cm 3 min -1 . Then, 1:99 C3H4/C3H6 binary breakthrough experiments were conducted at 298 K. The outlet composition S6 was continuously monitored by a Shimadzu Nexis GC-2030 gas chromatograph until complete breakthrough was achieved.
The C3H6 productivity (q) is defined by the breakthrough amount of C3H6, which is calculated by integration of the breakthrough curves f(t) during a period from τ1 toτ2 where the C3H6 purity is higher than or equal to a threshold value p:

Single Crystal X-ray diffraction
Single crystal X-ray diffraction data were collected on a Bruker Quest diffractometer equipped with a IμS microfocus X-ray source Cu Kα, (λ = 1.54178 Å); Mo Kα, (λ = 0.71073 Å); Synchrotron (λ = 0.4859 Å), and CMOS detector. APEX3 or APEX4 was used for collecting, indexing, integrating and scaling the data. 1 An open-flow nitrogen attachment (Oxford Cryosystems) was used for low temperature measurements. Absorption corrections were performed by multi-scan method. 2 Space groups were determined using XPREP 3 as implemented in APEX3. All the scaled data were solved using intrinsic phasing method (XT) 4 and refined on F 2 using SHELXL 5 inbuilt in OLEX2 v1.5 (2020) program. 6 All non-hydrogen atoms present in the frameworks were refined anisotropically. Hydrogen atoms were located at idealized positions from the molecular geometry and refined isotropically with thermal parameters based on the equivalent displacement parameters of their carriers. Crystallographic data reported in this paper are summarized in Tables S4 and Table S5. These crystal structures have been deposited to the Cambridge Crystallographic Data Centre (CCDC 2154469-2154473, and 1973753).

Fitting of experimental data on pure component isotherms
The isotherm data for C3H4 in the flexible MOF at 298 K were fitted with the 2-site Langmuir-Freundlich model, 8 where we distinguish two distinct adsorption sites A and B: The isotherm data for C3H6 in the flexible MOF at 298 K were fitted with the 3-site Langmuir-Freundlich model, 9,10 where we distinguish three distinct adsorption sites A, B, and C: The unary isotherm fit parameters are provided in Tables S6 and S7.

Adsorption Selectivity Calculations.
The detailed methodology for calculating the amount of A and B adsorption from a mixture by ideal adsorbed solution theory (IAST) is described elsewhere. 11 The adsorption selectivity is finally defined as: where qi (i = A or B) is the uptake quantity in the mixture and pi is the feeding partial pressure of component i.

Breakthrough simulations
The performance of industrial fixed bed adsorbers is dictated by a combination of adsorption selectivity and uptake capacity. Transient breakthrough simulations were carried out for 1/99 C3H4/C3H6 mixtures in the flexible MOF operating at a total pressure of 100 kPa, and temperatures of 298 K, using the methodology described in earlier publications. [12][13][14][15][16] In these simulations the intra-crystalline diffusional influences are considered to be of negligible importance. The breakthrough simulations are plotted as follows. The y-axis is the , , . The x-axis is 11. Separation factor / Separation selectivity calculations.
The amount of adsorbed gas i (qi) is calculated from the breakthrough curve as follows: Here, ! is the influent flow rate of gas (cm 3 min -1 ), $ is the effluent flow rate of gas (cm 3 min -1 ), Vdead is the dead volume of the system (cm 3 ), " is the adsorption time (min) and m is the mass of the sorbent (g). 17 On approximation, this simplifies to: is the total flow rate of gas (cm 3 min -1 ), ! is the partial pressure of gas i (bar) and ∆ is the time for initial breakthrough of gas i to occur (min). 18 The separation factor, also known as separation selectivity (aAC) for the breakthrough experiment i.e. breakthrough derived selectivity is determined as follows: yi is the partial pressure of gas i in the gas mixture. In the case where one gas component has negligible adsorption, the amount of gas adsorbed is treated as ≤ 1 cm 3 for calculations.  (Table S11 and S12, Figures S31 and S32

S15
Optical Image Figure S1. Optical image of as-synthesized sql-NbOFFIVE-bpe-Cu-AA-α.               Figure S29. Schematic of dynamic gas breakthrough separation experimental setup, including gas mixing unit, gravimetric gas uptake analyser and gas separation analyser. 32     Greek letters n Freundlich exponent, dimensionless t time, dimensionless