Dense Packing of Xenon in an Ultra-Microporous Metal−Organic Framework for Benchmark Xenon Capture and Separation

19 For adsorptive separation of xenon/krypton mixture, it still remains a daunting challenge to 20 target both high Xe adsorption and selectivity in a single adsorbent because of their close size 21 and inert nature. Herein, we report an ultra-microporous alkyl Cu-based MOF (MOF-11) with 22 suitable pore architecture for benchmark Xe capture and separation. MOF-11 features small 23 pore channels (4.4 Å) close to the kinetic diameter of Xe atom, in which a large number of 24 oppositely adjacent open Cu-metal sites and alkyl cavities are densely arranged along the pore 25 channels in staggered mode. Such unique pore system can not only introduce strong binding 26 affinity with Xe, but also enable the dense packing of Xe inside the pores. This material thus 27 shows the highest Xe storage density of 2826 g L -1 at ambient conditions, record-high Xe 28 uptake of 4.0 mmol g -1 (at 0.2 bar and 298 K) and one of the highest Xe/Kr selectivity (19.1), 29 significantly higher than most top-performing materials reported so far. The Xe molecules 30 trapped within MOF-11 were visually identified by single-crystal X-ray diffraction experiment, 31 which unveils the dense packing mechanism of Xe within the pores accountable for the record 32 Xe capture capacity. The breakthrough experiments confirm that MOF-11 exhibits the 33 exceptional separation performance for 20/80 Xe/Kr mixture with both record-high Xe uptake 34 capacity (3.46 mmol g -1 ) and Kr productivity (350 cm 3 g -1 ). 35


37
Highly efficient and energy-saving separation of noble gases such as xenon (Xe) and krypton 38 (Kr) is a very important but challenging process in the production of high-purity noble gases 39 reprocessing, which must be separated and sequestered safely to prevent the radioactive 48 contamination since they are both hazardous and 85 Kr has a long half-life of 10.8 years 3-5 . 49 Traditional cryogenic distillation is by far the most mature technology to separate Xe and Kr 50 from air or the off-gas streams during the UNF reprocessing, which is excessively costly and 51 energy-intensive 4 . To reduce energy consumption, the development of more efficient 52 separation technologies is highly desired for Xe/Kr separations. 53 Adsorptive separation using solid porous adsorbents has been considered as a more energy-54 efficient alternative due to the potential of greatly cutting energy consumption and operating 55 costs 6 . In this regard, porous metal−organic frameworks (MOFs) 7-13 , also known as porous 56 coordination polymers (PCPs), have emerged as a new class of promising adsorbents for 57 various gas separations owing to their highly tunable pore structures and functionalities [14][15][16][17][18][19][20][21][22][23][24] . 58 While a number of MOFs have been developed for efficient Xe/Kr separation 25-51 , it still 59 remains a grand challenge to target ultrahigh separation performance, since Xe and Kr are inert 60 atomic gases with no dipole or quadrupole moments. Typically, there exists an obvious trade-61 off between adsorption uptake and gas selectivity for Xe/Kr separation 27 . For example, some 62 large-pore MOFs (e.g., PCN-14 and HKUST) 25,35 exhibit high Xe adsorption capacity at 1 bar 63 but weak Xe binding affinity that cannot efficiently discriminate the two molecules, thus 64 resulting in low Xe/Kr selectivity. On the other hand, some ultra-microporous MOFs with 65 tailor-made pore size that is slightly larger than the kinetic diameter of Xe (4.047 Å), usually 66 exhibit the strong Xe binding affinity due to the pore confinement effect induced by the overlap 67 of van der Waals potentials, leading to the record high Xe/Kr selectivity reported so far. 68 However, their small pore volumes or low surface areas prone to result in relatively low Xe 69 uptake capacities, as exemplified by SBMOF-1 31 and Co-squarate 32 . To overcome this trade-70 off limitation, a better design is to construct suitable pore confinement densely decorated with 71 strong binding sites in ultra-microporous MOFs, enabling highly selective recognition and 72 efficient packing of targeted Xe molecules inside the pores. Thus, the strong binding sites can 73 strengthen the recognition of Xe over Kr and thus improve Xe/Kr selectivity; meanwhile, the 74 dense packing of the preferred Xe molecules is able to fully make use of pore spaces for high 75 Xe uptake density. This strategy has been well exemplified in the MOFs of SIFSIX-1-Cu 14 and 76 UTSA-16 23 for C2H2 or CO2 capture and separation, wherein the suitable pore confinement 77 enables the formation of gas clusters inside the pores to result in the dense packing of gas 78 molecules for targeting simultaneously high gas uptake and selectivity. However, compared 79 with those gases with dipole or quadrupole moments (e.g., CO2, SO2, and C2H2), inert atomic 80 Xe gas usually shows weak host-guest or guest-guest interactions during the adsorption process, 81 making the achievement of dense packing of Xe inside the pores more difficult and challenging.

82
Therefore, such dense packing strategy has not been realized yet for Xe/Kr separation.

83
To achieve the desired dense packing of Xe within porous materials, an ideal material would 84 have an appropriate pore system that densely and orderly arranges a large number of strong 85 binding sites (e.g., polar groups or open metal centers) combined with an optimal pore size 86 close to the kinetic diameter of Xe. We target this matter herein in an ultra-microporous alkyl 87 Cu-based MOF-11 (also called ATC-Cu) 13,24 that achieves the dense packing of Xe molecules 88 for benchmark Xe capture and separation. We discovered that MOF-11 features small pore 89 channels with suitable pore dimensions (4.4 Å) close to the kinetic diameter of Xe atom (4.047 90 Å), in which a large number of oppositely adjacent open Cu-metal sites and alkyl cavities are 91 densely arranged along the channels in staggered mode to provide dense and strong adsorption 92 sites for Xe capture (Fig. 1). Single-crystal X-ray diffraction studies on Xe loaded MOF-11 93 crystals (Xe@MOF-11) identified that this unique pore system can not only introduce the 94 strong Xe-framework interactions, but also enable the dense packing of Xe atoms within the 95 pores due to the efficient Xe···Xe interactions. MOF-11 thereby exhibits the highest Xe storage 96 density of 2826 g L -1 at 1 bar and 298 K, the record-high Xe uptake of 4.0 mmol g -1 at 0.2 bar 97 and 298 K, and very high Xe/Kr selectivity of 19.1 among all the reported materials at the same 98 conditions. Dynamic breakthrough experiments confirm the benchmark separation 99 performance of MOF-11 on Xe/Kr (20/80) gas mixtures, affording both the unprecedentedly 100 high Xe adsorption capacity (3.46 mmol g -1 ) and Kr productivity (350 cm 3 g -1 ).

105
The phase purity of bulk MOF-11 sample was confirmed by matching the experimental and 106 simulated powder X-ray diffraction patterns (PXRD; Supplementary Fig. 2). The dehydrated 107 crystal structure of MOF-11 is depicted in Fig. 1. The framework of MOF-11 consists of 108 paddle-wheel dinuclear Cu2(COO)4 secondary building units (SBUs) linked by the ATC 4-109 organic linkers to form three-dimensional (3D) network, in which each ATC 4ligand binds with 110 four SBUs and one SBU is bridged by four ATC linkers (Fig. 1a). As a result, MOF-11 exhibits 111 a (4,4)-connected PtS structure topology ( Supplementary Fig. 1). As depicted in Fig. 1b and 1c, subtracting van der Waals radius, thereby providing a considerable dual Coulombic interaction 119 within the pores and potentially offering a strong binding site for Xe (Fig. 1b). In addition, 120 there exist two kinds of aliphatic hydrocarbon cavities (cavity II and cavity III) constructed by 121 four enclosed ATC linkers, as shown in Fig.1b and 1c. Eight or twelve hydrogen atoms 122 originated from four ATC linkers point toward the center of the aliphatic hydrocarbon cavities.

123
The average distance between these hydrogen atoms and center of the cavity is only 3.7 Å and 124 3.5 Å, respectively, which provide potential adsorption sites for Xe capture. Most importantly, 125 we found that cavity I and cavity II are staggered with each other along the a and b axis, and Xe molecule. The total Xe uptake at 298 K and 1 bar can reach up to 4.95 mmol g -1 , which is 148 notably higher than most of porous materials such as Ag-MOF-303 (3.25 mmol g -1 ) 37 , CC3

149
(2.44 mmol g -1 ) 53 , SBMOF-1 (1.4 mmol g -1 ) 31 , and Co-squarate (1.34 mmol g -1 ) 32 . At 0.2 bar, 150 which is an indicator of the Xe capture ability of adsorbents from a 20/80 Xe/Kr mixture, MOF-151 11 exhibits a record high Xe uptake of 4.0 mmol g -1 at 298 K, significantly higher than all the 152 reported benchmark materials including MOF-74-Co (2.57 mmol g -1 ) 42 , Ag-MOF-303 (2.06 153 mmol g -1 ) 37 , SBMOF-1 (1.27 mmol g -1 ) 31 and Co-squarate (1.18 mmol g -1 ) 32 , setting a new 154 benchmark for Xe capture uptake at 0.2 bar. Considering the relatively low BET surface areas 155 (602.8 m 2 g -1 ) and pore volume (0.23 cm 3 g -1 ), the density of adsorbed Xe inside MOF-11 was 156 calculated to be ultrahigh of 2826 g L -1 at 298K and 1 bar, which is 480 times larger than the 157 density of gaseous Xe (5.89 g L -1 at 1 bar and 273K) and just slightly lower than the liquid Xe 158 density (3058 g L -1 at 1 bar and 165 K). To the best of our knowledge, this value is the highest 159 among all the reported materials so far, notably higher than that of SBMOF-1 (2042 g L -1 ) 31 ,  with Xe. After the as-synthesized MOF-11 crystals in the capillary glass tube were activated at 213 150 ℃ under vacuum, the capillary glass tube was filled by pure Xe gas to 1 bar at 298K and 214 sealed. Single-crystal X-ray diffraction data collected at room temperature revealed the crystal 215 structure of Xe@MOF-11 with a formula of Cu2(C14H12O8)Xe3. As shown in Fig. 3, MOF-11 216 was found to exhibit three binding sites (site-1, site-2 and site-3) for Xe molecules. Fig. 3a   217 shows that Xe atom is located in the middle of the two oppositely Cu-paddle wheels (site-1), 218 which dually interacts with the two metal sites with a short Xe···Cu distance of 2.950 Å. This 219 value is much shorter than the metal−Xe interaction distance found in MOF-74-Co (Xe···Co,  (Fig. 3d). In addition, Xe atom in site-3 also strongly interacts with Xe atom in site-I with  (Fig. 3e), thus accounting for the record-high Xe adsorption capacity and storage 246 density observed in MOF-11. The SCXRD data indicates that the adsorbed Xe amount in site-247 1, site-2 and site-3 corresponds to 2.2, 2.2 and 2.2 mmol g −1 gas uptake, wherein the total value 248 of 6.6 mmol g −1 is very close to the experimentally Xe uptake (5.28 mmol g −1 ) at 273 K and 1 249 bar. and 101 cm 3 g -1 ) 29 and ZJU-62 (2.88 mmol g -1 and 206 cm 3 g -1 at 273 K) 28 . MOF-11 can be 274 easily regenerated by vacuum or purging with helium for around 60 min at 298 K due to its 275 relatively low Qst value of Xe and Kr (29.4 kJ mol -1 and 23.5 kJ mol -1 ). During the desorption 276 process, high-purity Xe (99%) could be produced with a high productivity of 66.8 cm 3 cm -3 (a 277 recovery rate of ~ 61.5% for all the desorbed Xe). Multiple breakthrough tests for Xe/Kr 278 mixtures confirmed that MOF-11 maintains the excellent Xe adsorption capacity and 279 separation performance within three continuous cycles, indicating its good recyclability for 280 Xe/Kr separation (Supplementary Fig. 19).

281
To evaluate the performance of MOF-11 on capture and separation of Xe and Kr from the 282 off-gases in UNF reprocessing, we further carried out the breakthrough experiments using a 283 low Xe-Kr concentration gas mixture (400 ppm Xe, 40 ppm Kr balanced with dry air). After 284 initially purging with He for 5 min, the target gas mixture was injected into the column. As added into a Teflon-lined autoclave and heated at 573 K for 18 hours. The green crystals of  were obtained at the bottom of the vial. The as-synthesized samples were filtrated and washed with 316 methanol, and then exchanged with methanol for 2 days prior to activation. 317 Characterization. A X′Pert PRO PANalytical diffractometer (Cu Kɑ, 45 kV and 40 mA) was applied 318 to obtain the powder X-ray diffraction (PXRD) patterns (2ɵ = 5~50 o ). Thermogarvimeteric analysis 319 (TGA) was performed on a TGA-550 (TA instrument) analyzer under N2 gas flow from 30 ℃ to 800 ℃ 320 in an aluminum crucible with a heating rate of 10 ℃ min -1 using 3~6 mg of products.. 321 Gas adsorption measurements. The gas sorption isotherms were measured using a surface 322 characterization analyzer (Micrometritics 3Flex). The as-synthesized MOF-11 samples were degassed 323 at 150 ℃ for 18 hours to yield guest-free phase before sorption measurements. The N2 sorption 324 isotherms at 77 K were measured using a liquid nitrogen bath. The gas sorption isotherms at different 325 temperatures were collected using a Micromeritics's ISO Controller (Thermoelectric Cooled Dewar) to 326 maintain the temperatures. 327 Single-crystal diffraction and structure determination. A single crystal of MOF-11 was selected and 328 putted into a capillary glass tube with a 0.1mm inner diameter. This capillary glass tube was activated 329 at 150 ℃ under vacuum for 24 hours. And then the capillary glass tube was filled by pure Xe gas up to 330 1 bar at 298K and then sealed. The single-crystal data on the Xe-loaded activated MOF-11 crystal were 331 collected using an Agilent Supernova CCD diffractometer equipped with graphite-monochromatic 332 enhanced Mo-Kα radiation (l = 0.71073 Å) at room temperature. The datasets were corrected by 333 empirical absorption correction using spherical harmonics, implemented in the SCALE3 ABSPACK 334 scaling algorithm. The structure was solved by direct methods and refined by full matrix least-squares 335 methods with the SHELX-97 program package. 55 The crystal data are summarized in Supplementary 336 Table 4. 337 Breakthrough measurements. The breakthrough experiments were performed in a home-built 338 dynamic gas breakthrough equipment (Supplementary Fig. 20). The activated MOF-11 sample (~210 339 mg) was packed into a steel column (the steel column is about 12 cm long with a 4-mm inner and 6.4-340 mm outer diameter) with silica wool filling the void space. The column was activated by heating at 423 341 K under vacuum for 12 h and then purged by helium gas with a flow rate of (2.5 mL min -1 ) for 2 h 342 before the column temperature was decreased to 298 K. A circulator bath was used to maintain the 343 columns at 298 K. The flow of helium gas was stopped while target gas mixtures (a 20/80 Xe/Kr binary 344 gas mixture or 400 ppmv Xe 40 ppmv, Kr balanced with dry air) were fed to the column with a flow 345 rate of 2.5 mL min -1 . The flow of helium and targeted gas mixture was controlled by two mass flow 346 controllers. The downstream was monitored by a Hiden mass spectrometer (HPR 20). The desorption 347 test of MOF-11 was carried out using the same dynamic gas breakthrough equipment, while the 348 downstream was monitored by a gas chromatography (SCION GC-450) with thermal conductivity 349 detector (TCD, detection limit 0.1 ppm). The standard gases were used to calibrate the concentration of 350 the outlet gas. After the breakthrough tests, MOF-11 samples in the steel column can be readily 351 generated by sweeping He gas (10 mL min -1 ) at room temperature for 40 min or in high vacuum at room 352 temperature for 60 min. 353 The adsorption capacity was estimated from the breakthrough curves using the following equation: 354 Where nads is the adsorption capacity of the gas i, F is the total molar flow, Ci is the concentration of the 356 gas i entering the column and the ti is the time corresponding to the gas i, which is estimated from the 357 breakthrough profile.