Separation and Purification of Hydrocarbons with Porous Materials

Abstract This Minireview focuses on the developments of the adsorptive separation of methane/nitrogen, ethene/ethane, propene/propane mixtures as well as on the separation of C8 aromatics (i.e. xylene isomers) with a wide variety of materials, including carbonaceous materials, zeolites, metal–organic frameworks, and porous organic frameworks. Some recent important developments for these adsorptive separations are also highlighted. The advantages and disadvantages of each material category are discussed and guidelines for the design of improved materials are proposed. Furthermore, challenges and future developments of each material type and separation processes are discussed.


Introduction
Hydrocarbons play asignificant role in the global energy structure as well as in the chemical and petrochemical industries.They are not only important fuels,such as gasoline, diesel and kerosene,b ut also vital feedstocks for the production of many of our chemical products and materials, including pharmaceuticals,c oatings,a nd plastics. [1] However, hydrocarbons from crude oil or natural gas always exist as mixtures,which need to be separated and purified to asingle component for the further production of,f or example, plastics. [2] Separation of hydrocarbons having comparable sizes and molecular structures,a nd consequently similar physicochemical properties,i ss till ac hallenging endeavor. Adsorbent-based separation technologies are potential alternatives to the current industrially used cryogenic distillation. [3] In this Minireview,wemainly focus on the separation and purification of hydrocarbons from methane and olefins (i.e.e thene and propene) to C 8 aromatics (i.e.x ylene isomers).
Natural gas is becoming more and more important in the global energy structure.I ti sr eported that in 2018, global natural gas consumption grew at ar ate of 5.3 %, one of the strongest rates of growth since 1984. [4] With the increasing standards of living and related consumption, the global demand for natural gas has also risen sharply.T he primary constituent of natural gas is methane,w hile it also contains am ixture of impurities,i ncluding C 2+ hydrocarbons such as ethane and propane,aswell as N 2 ,CO 2 ,water,and hydrogen sulfide or other sulfur compounds. [5] All raw natural gases containing these contaminants require some treatment to meet pipeline specifications,t ypically > 90 %m ethane. [6] Furthermore,t he exploitation of unconventional sources of natural gases,like shale gas,coalbed methane,methane from anaerobic wastewater treatment plants,a nd landfill gas,h as greatly increased the accessible reserves of natural gas and these sources have become important in filling the gap between demand and supply.Ingeneral, water, N 2 ,CO 2 ,and sulfur are also typical impurities found in those unconventional natural gases.T oi mprove the purity and upgrade the energy content of the natural gases,these contaminants need to be removed. Among them, the very similar physicochem-ical properties of CH 4 and N 2 makes CH 4 /N 2 separation one of the most challenging and key separations for natural gas utilization.
Olefin/paraffin separation and the separation of xylene isomers are listed among the "seven chemical separations to change the world." [1c] Ethene and propene are critical petrochemical feedstocks,a nd the starting chemicals for the most widely produced synthetic plastics worldwide,n amely polyethylene (PE) and polypropylene (PP). Usually,e thene and propene are produced by the steam cracking of hydrocarbons ranging from ethane to vacuum gas oils (VGOs), in which many other hydrocarbons,such as ethane and propane,also co-exist with ethene and propene. [7] Because the production of PE and PP requires polymer-grade (> 99.5 %) ethene and propene, removal of ethane and propane are essential. Currently, ethene/ethane and propene/propane separation rely on the energy-intensive cryogenic distillation, which is performed in large columns containing over 100 trays. [8] Theannual energy consumption in ethene and propene purification alone accounts for 0.3 %o ft he global energy use. [1c] Adsorptionbased separation is believed to be an energy-and costefficient alternative technology to accomplish this highly energy-consuming process. C 8 aromatics,consisting of the three xylene isomers, paraxylene (PX), ortho-xylene (OX), and meta-xylene (MX), and ethylbenzene (EB), are raw materials for the synthesis of many important chemical intermediates.P X, the starting material for the synthesis of terephthalic acid, is the most valuable commodity among C 8 aromatics,b ecause terephthalic acid is the key precursor for production of polymers such as polyethylene terephthalate (PET), and polyester.MX is oxidized to synthesize isophthalic acid, which can be used as aco-monomer in the production of PET-based resins blends. OX is basically used in the production of phthalic anhydride. EB is utilized in polystyrene (PS) production after undergoing acatalytic dehydrogenation to styrene.C 8 aromatics are mainly produced by the catalytic reforming of crude oil, gasoline pyrolysis,a nd toluene disproportionation, which always generate am ixture that must be further separated. [9] This Minireview focuses on the developments of the adsorptive separation of methane/nitrogen, ethene/ethane,propene/propane mixtures as well as on the separation of C 8 aromatics (i.e.xylene isomers) with awide variety of materials,i ncluding carbonaceous materials,z eolites,metal-organic frameworks,a nd porous organic frameworks.Some recent important developments for these adsorptive separations are also highlighted. The advantages and disadvantages of each material category are discussed and guidelines for the design of improved materials are proposed. Furthermore,challenges and future developments of each material type and separation processes are discussed.
Owing to the similar boiling points,m elting points,a nd structures of the C 8 aromatics,s eparation of these isomers is expensive and very energy intensive. [10] Currently,i ndustrial separation of xylene isomers is mainly accomplished by crystallization and selective adsorption on zeolites.However, the development of adsorbents with higher efficiency is important and great efforts have been made worldwide.
Owing to the tremendous prospects of adsorptive separation technology,avast variety of porous materials have been explored for the separation and purification of hydrocarbon mixtures.P orous materials,such as activated carbons,carbon molecular sieves,z eolites,a ctivated aluminas,s ilica gels, polymer resins,m etal-organic frameworks (MOFs), metal organic cages,a nd porous organic frameworks (POFs), have been extensively studied for adsorptive separations. [3c, 11] Porous carbons have been among the most investigated porous adsorbents with low cost, high specific surface areas, and high stability.Italso has been used, for example,for CH 4 / N 2 separation, ethene/ethane separation, and propene/propane separation. Earlier studies mainly focused on the methods like loading other chemicals into porous carbons to promote the separation performance. [11a, 12] In recent years, porous carbons derived from biomass have also attracted great attention, [13] while some heteroatom-doped porous carbon materials have also received considerable attention for hydrocarbon separations. [14] Thediscovery of natural zeolites and the development of synthetic zeolites greatly broadened the available range of adsorbents and represents one of the major breakthroughs in gas adsorption and separation. [15] So far,the number of zeolite framework types officially registered by the International Zeolite Association (IZA) is more than 250. With the great availability of zeolite framework structures,h igh stability, easy synthesis process,a nd low cost, zeolites have been broadly used for catalysis,i on exchange,a nd adsorptionseparation in the chemical industry. [16] Noteworthy developments include improved adsorbents for xylene separation (in the so-called Heavy Parex Process) and the development of Li + -exchanged low silica zeolite Xasthe adsorbent of choice for the pressure swing-adsorption (PSA) process to separate oxygen, nitrogen, and carbon dioxide from air. [17] Zeolite research is ongoing with the possibility of preparing them in awide range of chemical compositions and with awide variety of framework structures.F or hydrocarbon separations,b oth traditional zeolites,o ften after proper modifications,a nd newly synthesized zeolites have been developed as useful adsorbents.
In the past decades,t he emergence of MOFs and POFs has brought new life to the field of hydrocarbon separation and purification. MOFs are well-defined microporous crystalline materials,w hich consist of inorganic nodes (i.e.m etal ions and their clusters) and organic linkers. [18] Given the broad range of inorganic nodes and organic linkers,t here is currently alarge family of synthetic MOFs available.D ue to the inherent diversity,t he tunable pore geometry,a nd easy functionalization, MOFs exhibit great potential for the separation of hydrocarbon mixtures by various separation mechanisms. [2b,c,11f] Along with MOFs,later discovered POFs also exhibit exceptional porosity and higher chemical stability and have gained great attention for gas storage,s eparation, catalysis,a nd electronics applications. [19] POFs assembled from organic building blocks via strong covalent bonds can be divided into two subcategories:crystalline,including covalent organic frameworks (COFs), and amorphous,like conjugated microporous polymers (CMPs) and hypercrosslinked poly- mers (HCPs), porous organic cages (POCs), covalent triazine frameworks (CTFs), and porous aromatic frameworks (PAFs). With high stability,p orosity,a nd designable structures,POFs have exhibited great potential for gas separation processes. [3c] During the past three decades,alarge amount of research has been devoted to the adsorptive separation and purification of hydrocarbons.R esearch on porous materials for hydrocarbon separation and purification has experienced an explosive growth (Scheme 1). In this Minireview,wesummarize the latest trends and related developments of hydrocarbon purification and separation ranging from methane/ nitrogen, ethene/ethane,p ropene/propane separation to the separation of C 8 aromatics with porous adsorbents.T his article is organized according to the different hydrocarbon separation systems with af urther section focusing on the different porous materials and the different separation mechanisms.T he future trends and challenges for each type of porous material are also discussed.

Mechanisms, Performance Evaluation, and Computational Methods
Thea dsorptive separation by the above-mentioned porous materials is achieved by one of following three mechanisms:s teric,k inetic, or equilibrium effects. [20] Equilibrium separation processes are the most common, with av ast majority of adsorptive separation processes operating through the equilibrium adsorption of mixture components. Kinetic separation is achieved by virtue of the differences in diffusion rates of different molecules,ofwhich air separation with carbon molecular sieves is at ypical case.F or steric effects,s ize or shape sieving can be achieved in porous materials with suitable pore size and geometries,i nw hich small and properly shaped molecules can diffuse into the adsorbent, whereas molecules that are too large to enter the pores are totally excluded. Theu niform channels and aperture size in the well-defined crystalline structure offers zeolites and MOFs the possibility to separate different molecules by as ize exclusive effect. Tw ot ypical cases of steric separation can be found in 3A zeolite for solvent drying and 5A zeolite for the separation of linear paraffins from branched-chain and cyclic hydrocarbons. [21] Benefiting from avariety of ordered pore structures and adjustable porosities, molecular sieving of ethylene over ethane, [22] propene over propane, [23] acetylene over ethylene, [24] and methane and nitrogen from carbon dioxide [25] have been successfully achieved by the finetuning of some MOF structures.Adsorption selectivity is one of the most significant parameters for the evaluation of an adsorbent. Ideal adsorbed solution theory (IAST), as developed by Myers and Prausnitz, is the most widely accepted theory to predict mixed-gas adsorption isotherms,w hich are entirely based on single-component adsorption isotherms. [26] Accurately measured single-component isotherms and an excellent fitting of adsorption isotherms for the measured data are required for the application of IAST to predict adsorption selectivity. [27] Another method to determine the adsorption selectivity proposed by Knaebel takes the ratio of the Henrysl aw constants of the two components from the single-component adsorption isotherms,w hich is also called Henrysl aw selectivity. [27b] For an adsorbent based on kinetics,the kinetic selectivity is often calculated by the ratio of the diffusional time constants (D/r 2 , calculated by the short-time solution of the diffusion equation [28] )o ft he two gases.B reakthrough experiments are particularly useful in evaluating the practical separation performance of an adsorbent. In breakthrough experiments, af lowing gas with aw ell-defined concentration of one or more adsorbates in acarrier gas passes through afixed bed of porous adsorbents.Abreakthrough curve is the time-resolved effluent concentration of the adsorbate at the outlet of the fixed bed. Theb reakthrough results show clearly separation performance of the adsorbent.
Computational simulation has been widely applied in the study of adsorption and separation of porous adsorbents.One of the commonly used methods is quantum mechanics,such as ab initio and density functional theory (DFT). DFT has been av ery popular method over the past years,i nw hich the energy of am olecule can be determined in terms of the electron density instead of the electron wave function. [29] Quantum mechanical methods can be used to determine the optimized position of an adsorbate molecule within acertain cluster extracted from the crystal structure of the porous material as well as the adsorption energy for an adsorbate molecule with the optimized binding site within the cluster with high accuracy. [30] Another broadly used method is classical molecular simulations,s uch as molecular dynamics (MD) simulations and grand canonical Monte Carlo (GCMC) calculations.M Di sasimulation of the time-dependent behavior of am olecular system, [31] which can be used to investigate the dynamic properties of an adsorbate in an adsorbent. GCMC is the most widely used molecular simulation method, which can be used to simulate gas adsorption in adsorbent over aw ide range of temperatures and pressures. [30] Adsorption isotherms,a dsorption capacity, enthalpies of adsorption and selectivity (for mixtures) can be obtained from GCMC simulations.
Since there are thousands of porous materials (particularly MOFs), rapid screening of ideal and promising adsorbents can be difficult. High-throughput computational screening (HTCS) has emerged as ap owerful tool to make the fast evaluation and rational design of adsorbents feasible. [32] One screening strategy is high-throughput screening with molecular simulation or DFT calculations,w hich have played an important role in quickly identifying promising structures and accurately assessing adsorption and separation performances of porous adsorbents. [32a] However, the accumulation of tremendous volumes of simulated data and the rapid growth of adsorbents (mainly MOFs) make this screening strategy inefficient and huge computational resources as well as valuable research time would be wasted. [33] Another high-throughput screening strategy based on machine learning that can overcome the above-mentioned problems by the training of data has gradually received more and more attention. [32b, 33, 34] Angewandte Chemie 3. Methane-Nitrogen Separation CH 4 /N 2 separation is intrinsically difficult because of the close kinetic diameters and comparable polarizability of CH 4 and N 2 (Table 1). Foradsorption separation, according to the separation mechanisms,a dsorbents can be divided into two categories:1 )CH 4 -selective adsorbents,w hich exhibit stronger adsorption interactions and higher adsorption capacity for CH 4 than N 2 ;t he separation is typically based on the equilibrium mechanism;2 )N 2 -selective adsorbents,w hich preferentially adsorb N 2 over CH 4 ;t he separation is dominantly based on the kinetic effect or steric effect. We will now discuss these two groups of materials separately.

Methane-Selective Adsorbents
Both CH 4 and N 2 are nonpolar molecules.F or most porous materials,CH 4 is always preferentially adsorbed over N 2 due to its higher polarizability. [36] Therefore,t he mechanism of CH 4 -selective adsorbents for CH 4 /N 2 separation is predominantly equilibrium-based. Porous carbons and MOFs are extensively studied porous materials for CH 4 /N 2 separation. [13b,e,14c,d, 37] Some zeolites have been tried and tested for CH 4 /N 2 separation as well. [38] 3.1.1. Carbonaceous Materials Carbonaceous materials have been used for CH 4 /N 2 separation for al ong time.I ne arlier studies,p orous carbons were treated by loading other chemicals,s uch as Br 2 ,I Cl, or MoO 2 ,o nto the adsorbents to improve their CH 4 /N 2 separation performance. [11a,12a] In recent years,h eteroatom-doped porous carbons and activated carbons derived from biomass have been explored. [13b,e,14c,39] N-rich microporous carbons derived from N-containing polymers were obtained by as olvent-free method;t hey possessed narrow pore size distributions (ca. 0.5-3 nm) and achieved an IAST CH 4 /N 2 selectivity of up to 5.1 at 298 Kand 1bar. [14a] High N-content porous carbons were also successfully synthesized from shrimp shells for CH 4 /N 2 separation of coal-bed gas. [14b] The obtained activated carbons exhibited an IAST CH 4 /N 2 selectivity of % 5at298 Kand 1bar. Surface functionalization also has as trong effect on the separation performance of porous carbons.I tw as found that the activated carbons derived from bamboo sawdust had ah igher adsorption capacity for CH 4 ,while oxygen-containing groups on activated carbons can improve the surface polarity and enhance the adsorption ability for N 2 ,a nd thus the material has al ower IAST CH 4 /N 2 selectivity. [13c] Zhong et al. chose rice as ac arbon source for making carbon-based adsorbents and found that carboxyl groups are the dominant surface groups and responsible for the enhanced IAST CH 4 /N 2 selectivity. [13d] More recently,Luetal. reported self-pillared 2D polymer and ultramicroporous carbon plates prepared by aone-pot multicomponent sequential assembly method. [40] With the narrow ultramicropore size distribution (4.8 ), the pillared polymer nanoplates exhibit ahighly competitive CH 4 /N 2 selectivity at lower CH 4 partial pressure.

Zeolites
Theseparation performance of zeolites can be affected by the window size,pore geometry,and cation distribution of the zeolite material. Theoretically,b ys urface modification and pore size adjustment of zeolites,g as mixtures with similar compositions can be separated. However,precise adjustment is rather difficult to achieve.T he earliest application of zeolites used for CH 4 /N 2 separation can be traced back to 1958. [41] Some traditional zeolites,i ncluding zeolite 4A, [42] zeolite 5A, [43] ZSM-5, [38c, 44] 13X, [45] HMOR, and chabazite, [42a] were studied and tested for the separation of CH 4 /N 2 .T he separation results indicated that their separation performances towards CH 4 /N 2 were rather desirable.C onsequently, traditional zeolites are not applicable in CH 4 /N 2 separation. On the other hand, modification of traditional zeolites can be an effective method to improve their CH 4 /N 2 separation performances.L iu et al. report as trategy of introducing subunits of ZIFs into zeolites Yand ZSM-5 to obtain effective adsorbents with advantages of both zeolites and MOFs ( Figure 1). [38e] Simulation results suggested that the incorporation of ZIF subunits (Zn-mIM, Zn-eIM, and Zn-pIM) may result in higher CH 4 /N 2 selectivities.E xperimental results validated that the incorporation of ZIF subunits into the zeolite structure lead to an increase in the IAST CH 4 /N 2 selectivity,which reached av alue of 8.4.

Metal-Organic Frameworks
Since the 20 th century,M OFs have been tested in great detail for their potential in CH 4 /N 2 separation [46] and progress has been made in recent years.V arious methods,i ncluding synthesizing MOFs with new framework structures,functionalizing MOFs with different groups,a nd reacting and combining MOFs with other chemicals or materials,h ave been investigated for the improvement of CH 4 /N 2 separation performance;t hese methods primarily aim at increasing the interaction between CH 4 and the adsorbent. AM OF-based methane nanotrap (ATC-Cu) was reported featuring oppositely adjacent open metal sites and dense alkyl groups that

Angewandte
Chemie can induce strong interactions with methane ( Figure 2). [47] Single-crystal X-ray diffraction experiments and molecular simulation studies indicated that ATC-Cu provides very strong binding sites for methane,a ttributing to aI AST CH 4 / N 2 selectivity up to 9.7 at 298 Ka nd 1bar.W oo et al. constructed anew MOF with a3Dframework with alternating large and small channels along the aa nd bd irections, which has an IAST CH 4 /N 2 selectivity of 7.0 at 298 Ka nd 1bar.
[37a] Several MOFs with layered and pillared structures proved to be efficient CH 4 /N 2 adsorbents due to their narrow and uniform pore networks.T wo isostructural MOF materials,Co-MA-BPY and Ni-MA-BPY,with intriguing pillar-layer structures,have prominent IAST CH 4 /N 2 selectivities of 7.2 and 7.4 (CH 4 /N 2 = 50/50,v/v), respectively,a t2 98 Ka nd 1bar. [48] Lately,the zinc-based pillar-layer MOF Zn 2 (5-aip) 2 (bpy) also has been effectively used for CH 4 /N 2 separation ( Figure 3). [37c] Molecular simulation indicated that within its narrow pore environment, the spheroidal molecular structure of CH 4 could be more adequately packed than the linear molecular structure of N 2 .A sar esult, as ample achieved IAST CH 4 / N 2 selectivity of up to 7.1 at 298 Ka nd 1bar. Based on the saturated C-H bonds as well as the corresponding trans corner in the ligand, Liu et al. constructed MOFs with specific cages to preferentially adsorb CH 4 molecules.InanAl-CDC MOF, the aliphatic ligand with low polarity that contains saturated C-H bonds may have arelatively strong interaction with CH 4 , leading to ahigh IAST CH 4 /N 2 selectivity of 13.1 at 298 Kand 1bar. [49] Modification of MOFs via compositing MOFs with other chemicals or materials is likewise an effective way to elevate their CH 4 /N 2 separation performance.D oping Mg 2+ into MIL-101 has been investigated to enhance the selective adsorption of CH 4 /N 2 of MIL-101.
[37d] Doping Mg 2+ increases the adsorption capacity of CH 4 and N 2 at different levels because doping the proper amount of Mg 2+ restrains the generation of Hbonds,which has apositive effect on methane gas adsorption. TheI AST CH 4 /N 2 selectivities upon doping MIL-101 with Mg 2+ also greatly increased, from % 2.1 for pristine MIL-101 to % 4.5 for MIL-101@12.8 %M g 2+ (CH 4 /     6 ], in CuBTC to examine the effect of methylation of ILs on the gas separation performance. [50] Compared to the corresponding selectivities of pristine CuBTC,CH 4

Porous Organic Frameworks
Porous organic frameworks with higher chemical stability and specific surface areas have also attracted attention in CH 4 /N 2 separation. [51] Rational modification with variable functionalities has been used to modify the physicochemical properties of POFs,aiming at improving their gas adsorption capacity and selectivity.Upon introduction of light metal ions to the porous aromatic framework (PAF), the adsorption affinity of PA F-26-COOM to CH 4 gases is enhanced compared to pristine PA F-26-COOH. [51a] Thei mproved performance of PA F-26-COOM and significantly higher CH 4 /N 2 selectivity is attributed to the strong interaction between CH 4 molecules and PA F-26-COOM. As eries of adamantane porous covalent triazine-based frameworks (PCTFs) with varying symmetry and functional group density were prepared by using aL ewis acid catalyst and as trong Brønsted acid catalyst, respectively. [51b] Them ost selective PCTF-7 is found to have ah igh CH 4 /N 2 selectivity of 7a t2 73 Ka nd 1bar.T able 2p rovides an overview of selected state-of-theart porous materials for CH 4 /N 2 separation.

Nitrogen-Selective Adsorbents
While the majority of adsorbents for CH 4 /N 2 separation are equilibrium-based, ap roportion of materials separate CH 4 /N 2 mixtures based on kinetic variance.D ue to the difference in the kinetic diameters of CH 4 and N 2 ,itshould be possible to separate them based on ak inetic mechanism. Therefore,m any studies on CH 4 /N 2 separation also focused on sorbents for kinetic separation (e.g.,c arbon molecular sieves,clinoptilolites and some MOFs).
Caron molecular sieves (CMS) is at ype of carbonaceous material, whose molecular sieving properties depends on their narrow and uniform pore size distribution. Based on their special pore textures,CMS have been successfully used for separating N 2 from CH 4 /N 2 mixtures. [11b,54] As early as 1991, the diffusion time constants of N 2 and CH 4 in CMS 3A were reported by Ma et al.;t he adsorbent was found to exhibit selectivity for N 2 /CH 4 separation. [55] In 2016, Li et al. used CMS to concentrate methane from raw gas of 10 %CH 4 to 79 %p urity during ah igh-pressure adsorption step with 93 %r ecovery. [54b] Recent research also found that CMS materials designed with proper microporosity would benefit practical coal mine methane upgrading. [54c] Clinoptilolites and titanosilicates are two promising adsorbents for the kinetic separation of N 2 /CH 4 mixtures.C linoptilolites are naturally occurring zeolites with a2 Dc hannel structure formed by eight-membered rings and ten-membered rings.The location, number, and type of cations in these channels have ah eavy impact on the selectivity and adsorption rate of gases. [56] The separation performances of clinoptilolites modified by cation exchange are different depending on the metal cations. Among the many metal cations,N a + ,M g 2+ ,L i + ,a nd Ni 2+ clinoptilolites showed favorable kinetic selectivity for possible N 2 /CH 4 kinetic separation. Modified titanosilicate molecular sieves,i ncluding ETS-4, ETS-10, and UPRM-5, are another representative class of adsorbents with desirable N 2 / CH 4 kinetic selectivities. [57] MOFs with flexible framework structures have attracted enormous attention and some flexible MOFs preferentially adsorb N 2 over CH 4 .Z hou et al. synthesized am esh-adjustable MAMS-1 from H 2 (bbdc) and Ni(NO 3 ) 2 for gas separations.Asshown in Figure 4, MAMS-1 has aflexible structure and its gates open linearly as temperature increases. [58] The adjustable mesh of MAMS-1 makes it possible to separate gases with kinetic diameters in the range of 2.9 to 5.0 .A t 113 K, the gate of MAMS-1 opens to about 3.7 ,w hich is wide enough to allow N 2 (3.64 )toenter the chambers,but CH 4 (3.8 )s tays in the hydrophilic channels.
Featuring low-energy, p-symmetric orbitals capable of accepting electron density,N 2 is aw eakly p-acidic species. Based on quantum mechanical computations,t he team of Jeffrey R. Long predicted that V-MOF-74 can be used to separate dinitrogen from methane due to the selective p backbonding interactions between the vanadium(II) cation centers in this MOF and the unoccupied p*o rbitals of N 2 . [59] This insight provides new MOF targets to synthesize.I n2 013, inspired by biomimetic nitrogen fixation to produce ammonia, they found that mesoporous MOF containing accessible Cr III sites is able to thermodynamically capture N 2 selectively  [60] Recently,Long and his group reported the synthesis of am etal-organic framework with exposed vanadium(II) sites,w hich engages p-acidic gases via backbonding interactions. [61] TheN 2 /CH 4 separation performance of V II -MOF has been verified. Specifically,t he btdd 2À ligand was used to synthesis V 2 Cl 2.8 (btdd) instead of more common carboxylatecontaining ligands to achieve better square-pyramidal vanadium(II) centers.T he IAST N 2 /CH 4 selectivity values of V 2 Cl 2.8 (btdd) are exceptional for low N 2 concentrations at 1bar total pressure (Figure 5c). TheN 2 /CH 4 selectivity is 38 for a2 0:80 N 2 /CH 4 mixture at 25 8 8Ca nd at 2:98 N 2 /CH 4 ,t he selectivity reaches 72. Incorporating such p-basic metal centers into porous materials offers ah andle for capturing and activating key molecular species within next-generation adsorbents.

Olefin-Paraffin Separation
Ethene/ethane and propene/propane separations are the most important separation processes among olefin/paraffin mixtures as ethene and propene are the most important raw materials in the petrochemical industry.T he very similar boiling point and the small variations in the condensabilities of these molecules (ethene/ethane and propene/propane) lead to great challenges in those separations,and also makes the currently used cryogenic distillation based on different vapor pressures and boiling points very energy-intensive. Consequently,a dsorption-based separation processes are explored and developed.

Ethene-Selective Adsorbents
With the presence of p electrons,h igher quadrupole moment, and smaller molecular size,e thene is more easily adsorbed and also adsorbed in higher amounts than ethane by most of the developed adsorbents. [62] In this regard, al arge number of porous adsorbents exhibit excellent performance for ethene/ethane separation based on three possible mechanisms:e quilibrium-based, kinetic-based, and size exclusion mechanisms.
Equilibrium-based mechanism.Ethene/ethane separation achieved by thermodynamically driven separation is one of the most common and popular cases.O wing to the pcomplexation effect between unsaturated hydrocarbons and metal ions (mostly Cu I and Ag I ), ethene undergoes selective p-complexation with metal ions in adsorbents (porous carbons,z eolites,M OFs,a nd POFs) or the open metal sites (OMSs) in MOFs and is thus separated from ethanecontaining mixtures.One effective method, the incorporation of metal ions into the pores of adsorbents,h as been implemented in and well proven for different types of porous materials,i ncluding zeolites,p orous carbons,a nd MOFs. Zeolites,i np articular cation-exchanged zeolites,h ave been successfully used for ethene/ethane separation. [63] Porous carbons and mesoporous silica were also used as supports for Cu I and Ag I loading to separate ethene/ethane. [12b,64]   . a) Structure of asingle vanadiumsite in V 2 Cl 2.8 (btdd) after dosing with 700 mbar of N 2 ,a sdetermined from analysis of powder X-ray diffraction data. Cyan, green, blue and gray spheres represent V, Cl, Nand Catoms, respectively;a40 %-occupied terminal chloride ligand has been omitted. b) Adsorption isotherms for N 2 (blue) and CH 4 (black) collected at 25 8 8Ci nV 2 Cl 2.8 (btdd). c) IAST selectivity values calculated at 25, 35, and 45 8 8Cfor various N 2 :CH 4 ratios at atotal pressure of 1bar.Adapted from ref. [61].

Angewandte Chemie
MOFs and POFs,i nw hich Cu I is chelated by organic linkers ( Figure 6) and that are functionalized with Ag I by grafting or sulfonate functionalization, were shown to have superior selectivity towards ethene. [19e,65] Another approach is to achieve p-complexation between open metal sites in MOFs and ethene.M OFs can contain coordinatively unsaturated sites or open metal sites when vacant Lewis acid sites on the metal ions or cluster nodes have been generated. [66] Due to the p-interactions between the electron-rich p-orbital in olefins and the vacant s-orbital of the open metal site,MOFs with open metal sites preferentially adsorb olefins over paraffin, achieving outstanding olefin/paraffin separation performance.T he open metal sites in coordinatively unsaturated MOFs have selective interactions with olefins via pcomplexation. [67] CuBTC is the first MOF with open metal sites for the efficient separation of ethene/ethane. [68] M 2 -(dobdc) frameworks (also MOF-74, M: Mn, Fe, Co,N i, Zn; dobdc 4À :2 ,5-dioxido-1,4-benzenedicarboxylate) with ah igh density of open metal sites were also explored and exhibited great potential in ethene/ethane separation. [69] Thes trong pinteraction leads to high gas uptake and separation selectivity, but the regeneration of these adsorbents can also be difficult and highly energy consuming.
Kinetic-based mechanism.Based on the difference in the shape and size of ethene and ethane molecules,an umber of adsorbents have been developed for ethene/ethane separation. Fora dsorbents based on kinetics,p ore dimensions and pore shape play dominant role in the overall separation performance.Corma et al. synthesize apure silica zeolite with large heart-shaped cages and framework flexibility,which can kinetically separate ethene from ethane with an exceptional selectivity of % 100 (Figure 7). [70] When ethene molecules enter the center of the 8-ring window of ITQ-55, the window size will expand from 2.38 of the empty structure to 3.08 (Figure 7b). However,the adsorption capacity is also limited by its contracted aperture.Arobust MOF GT-18 with optimum pore size and shape was synthesized through am ixed-linker strategy and displayed promising diffusion selectivity toward ethylene. [71] With ab enzotriazole (BTA)/benzimidazole (BIM) linker synthesis ratio of 4:1, GT-18 was obtained, featuring a1 0-ring window and flexible pore apertures ( % 3 ). Thephosphate-anion pillared MOF ZnAtz-PO 4 decorated with electronegative groups was also reported;i ts periodically expanded and contracted aperture enables effective trapping of C 2 H 4 and impedes the diffusion of C 2 H 6 . An equilibrium-kinetics synergetic effect was observed in this MOF,w hich displayed acombined selectivity of 32.4 at 273 Kand 1bar. [72] Size exclusion mechanism.S ize exclusion or molecular sieving is an ideal approach for separation, by which only small and properly shaped molecules can diffuse into the adsorbent, allowing ah ighly selective separation based on molecular size or shape cut-off.S tudies aimed at molecular sieving mainly focus on the difference between the kinetic diameters of adsorbate molecules,t he pore size of the adsorbent, and differentiation of van der Waals molecular dimensions and molecular cross-section. Focusing on the difference in molecular size and shape of ethene and ethane, Chen et al. reported an ultramicroporous MOF [Ca(C 4 O 4 )-(H 2 O)] with aperture sizes (with slightly different shapes of 3.2 4.5 2 and 3.8 3.8 2 )s imilar to the size of ethene molecules but smaller than the ethane molecules. [22] Owing to the good size/shape match and its highly rigid pore structure, [Ca(C 4 O 4 )(H 2 O)] can act as amolecular sieve to prevent the

Angewandte
Chemie transport of ethane inside its channels.The low cost of the raw materials (calcium nitrate and squaric acid) and the high stability of [Ca(C 4 O 4 )(H 2 O)] make it promising for industrial applications.H owever,w hether the coordinated water molecules will be lost after long-term use and the effect of this on the structural stability remains to be considered. Considering the differentiation of molecular cross-section size,afamily of gallate-based MOFs,M -gallate (M = Ni, Mg, Co) featuring 3D interconnected zigzag channels and aperture sizes of 3.47-3.69 ,a lso exhibited ideal exclusion of ethene and ethane ( Figure 8). [73] Thes pecial channels and pores can ideally separate ethene (3.28 4.18 4.84 )a nd ethane (3.81 4.08 4.82 )t hrough molecular cross-section size differentiation (Figure 8c). Consequently,f or Co-gallate,a nu nprecedented IAST ethene/ethane selectivity of 52 was achieved at 298 Ka nd 1bar for equimolar ethene/ethane mixtures.T able 3p rovides an overview of selected state-ofthe-art porous materials for C 2 H 4 /C 2 H 6 separation.

Ethane-Selective Adsorbents
Ethene/ethane separation by ethene-selective adsorbents to obtain high-purity ethene requires atwo-step "adsorption-desorption" process and subsequent multiple adsorptiondesorption purification cycles,w hich are highly energyconsuming.T herefore,e thane-selective adsorbents would be more desirable and efficient for ethene/ethane separation, because ethene would be obtained directly which would thus be more energy-saving.However,the development of ethaneselective adsorbents is ac hallenging task. Ethene shows al arger quadrupole moment than ethane (ethene:1 . 50 10 À26 esu cm 2 ,e thane:0 .65 10 À26 esu cm 2 ), while ethane has ah igher polarizability (ethene:4 2.52 10 À25 cm 3 ,e thane: 44.7 10 À25 cm 3 ) (Table 1). Therefore,unlike ethene-selective adsorbents with strong adsorption sites for ethene,e thaneselective adsorbents always suffer from poor selectivity due to the lack of strong adsorption sites.I ns pite the challenges, research on ethane-selective adsorbents,predominantly MOF and POF materials,h as been made significant progress over the several past years. [74] Strengthening binding affinity towards ethane,enhancing host-guest (ethane) interactions,a nd decreasing or preventing strong adsorption sites for ethene are the most used and effective strategies to improve the separation performance of ethane-selective adsorbents.T he proper positioning of multiple electronegative and electropositive functional groups on the pore surface of the material MAF-49 resulted in multiple and stronger CÀH···N hydrogen bonds between ethane molecules and the MAF-49 framework, leading to the preferential adsorption of ethane over ethene and ah igh IAST ethane/ethene selectivity of 9f or equimolar ethane/ ethene mixtures with ar elatively low ethane uptake (1.7 mmol g À1 )a t3 16 Ka nd 1bar. [74a] By controlling pore structures,t he material Cu(Qc) 2 (Qc À = quinolone-5-carboxylate) with aw eakly polar pore surface exhibited selfadaptive sorption behavior for ethane and thus higher binding affinity towards ethane over ethene. [74b] At 298 Kand 1bar,it presented aI AST selectivity of 3.4 for equimolar ethane/ ethene mixtures.Chen et al. reported the microporous MOF Fe 2 (O 2 )(dobdc) (dobdc 4À :2 ,5-dioxido-1,4-benzenedicarboxylate), which displayed highly selective separation of ethane/ ethene. [74c] TheF e-peroxo sites on the pore surface of Fe 2 (O 2 )(dobdc) ap lay key role in the recognition of ethane, which resulted in the adsorption of alarger amount of ethane than of ethene (Figure 9). At 298 Ka nd 1bar,t he ethane uptake and the selectivity for an equimolar ethane/ethene mixture were 3.32 mmol g À1 and 4.4, respectively.
Then ovel hydrogen-bonded organic framework (HOF) HOF-76a has been exploited for ethane/ethene separation ( Figure 10). [62] Thenonpolar/inert pore surface together with the triangular channel-like pores within HOF-76a enables it to preferentially adsorb ethane over ethene.T he IAST selectivity for an equimolar ethane/ethene mixture is 2, at 296 Kand 1bar. Theethane-selective MOF JNU-2 with cagelike cavities interconnected by asmall aperture ( % 3.7 )has been reported by Li et al. [74d] Themultiple CÀH···O hydrogen bonds between ethane and the precise arrangement of oxygen atoms on the small aperture resulted in an enhanced ethaneselectivity over ethene.V ia ap ore-space-partition (PSP) strategy,af amily of heterometallic vanadium and titanium MOFs were synthesized. [74e] Thet otal annihilation of open metal sites in these MOFs could be beneficial for the ethane-  It is worth mentioning that high-throughput computational screening (HTCS) methods have become au seful approach to screen adsorbents for gas adsorption and separations.I n2 016, molecular simulation based computational screening of 278 different MOFs was performed to simulate their separation of ethane/ethene. [75] Several MOFs were predicted to exhibit higher adsorption capacities and selectivities than zeolites under similar conditions.W ith the continuous increase in the number of adsorbents,l arger databases have been screened. Al arge set of MOFs was computationally screened by first excluding MOFs with disordered atoms,o pen metal sites,a nd ap ore limiting diameter < 3.8 to identify ideal adsorbents for ethane/ ethene separation. [76] 16 ideal MOFs with ethane/ethene selectivity ! 2.16 and ethane uptakes ! 0.54 mmol g À1 were identified. In addition to molecular simulation based HTCS, am ore effective machine learning based HTCS is beginning to emerge for hydrocarbon separations. [77] Xi et al. created am odeling library of 425 UiO-66 materials with al arge variety of missing-linker defects to explore the effect of defects in MOFs. [77a] They proved that machine learning could be an efficient way to elucidate how the defects control the performance of UiO-66 in adsorption, separation, and mechanical stability.T his study also concluded that the concentration of defects is more important than their distribution in the overall separation performance of UiO-66 materials.Finally,they provided astraightforward guide to access ap rivileged defect-containing UiO-66 material with optimal properties (Figure 11). By combining machine learning algorithms with molecular simulation, they screened the hypothetical metal-organic framework (h-MOF) database for ethane/ethene separation materials. [77b] Finally,f our h-MOF materials with the best ethane/ethene separation performance were identified by using the random forest (RF) algorithm. With the increasing number of possible porous materials,H TCS has been regarded as av aluable tool to identify the top-performing materials from the large database of candidates and will accelerate the rational design and development of highly efficient adsorbents.T able 4p rovides an overview of selected state-of-the-art porous materials for C 2 H 6 /C 2 H 4 separation.

Propene/Propane Separation
Equilibrium-based mechanism.L ike ethene/ethane separation, aproportion of propene/propane separation in porous materials are based on p-complexation between propene and   (b), and the close van der Waals contacts within the corner surface of triangular channel-likep ores, as obtained by DFT calculations (C, dark gray;O,r ed;H ,white);C ÀH···p interactions highlighted by red dashed bonds,c )Adsorption isotherms of C 2 H 6 (red) and C 2 H 4 (black) for HOF-76a at 296 K. d) IAST selectivity of HOF-76a from C 2 H 6 /C 2 H 4 (50/50 and 10/90) gas mixtures at 296 K. Adapted from ref. [62]. In recent years,anumber of materials have been explored for propene/propane separation based on p-complexation and they have displayed efficient propene/propane separation performances:t he Ag I -decorated porous polyimide material MPI-Ag, [65d] MFU-4L with incorporated Cu I sites, [78] MIL-101(Cr) loaded with cuprous oxide nanoparticles,Cu I -loaded MIL-101(Fe), HCPs doped with Ag I ,a luminosilicates with calcium cations, [79] titanate nanotubes containing Cu I , [80] Ag Idoped microporous carbon, [64b] M-MOF-74 (M = Co,Mn, and Mg) with high densities of open metal sites, [81] AGTU-3a with open Ag I sites, [82] and many more.I na ddition to p-complexation, some other adsorbents based on the equilibrium mechanism have been developed as well. In 2019, Zhang et al. synthesized the MOF MAF-23-O by selective aerobic oxidation of the soft methylene bridges of the organic ligands of MAF-23 ( Figure 12). [83] Theo xidation leads to more rigid carbonyl bridges and gives additional guest recognition sites, improving both thermodynamic and kinetic adsorption selectivity.A nI AST selectivity of 8a nd breakthrough selectivity of 15 for equimolar propene/propane mixtures was achieved for MAF-23-O,a t2 98 Ka nd 1bar.R ecently,b yg rafting pyrrole onto Cu-BTC,L ie tals ynthesized Pyr@Cu-BTC. [84] Since propene can be preferentially adsorbed to pyrrole by electrostatic interactions,P yr@Cu-BTC displayed ah igh IAST propene/propane selectivity of 8.3 for equimolar propene/propane mixtures,a t2 98 Ka nd 1bar.Z hou et al. designed propane-selective adsorbents by surface tuning and replacing the ligand. [85] They reported ahigh propane capacity of 8.79 mmol g À1 for g-C 3 N 4 @Zr-BPDC and as uperior propane/propylene selectivity of 1.5 for Zr-BPYDC at 298 Kand 1bar.
There have also been af ew computational screening studies on propene/propane separation. ADFT-derived force field was applied to describe the adsorption of C 2 -C 3 olefins and paraffins in CuBTC.T his method was then extended to evaluate 94 related Cu-OMS MOFs for propene/propane separation and 18 MOFs with attractive separation performances were identified. [86] Later, approximately 1m illion crystal structures of MOFs,ZIFs,a nd zeolites were screened by Han et al. for propene/propane separation. [87] GCMC simulations were performed to simulate the selectivity,working capacity,a nd physical properties of those porous materials.W ebelieve that with the rapid development of computational techniques,H TCS will play am ore important role in materials screening for propene/propane separation.
Kinetic-based mechanism.K inetic-based separation of propene/propane mixtures makes up ac onsiderable portion of propene/propane adsorption separations.U ntil now,o nly ahandful of pure silica or high silica zeolites,like Si-CHA, [88] 4A, [89] ITQ-12, [90] ITQ-32, [91] and ZSM-58, [92] have been reported for the kinetic separation of propene/propane mixtures,w hile the synthesis of these zeolites is relatively difficult. Decoration/modification has proved to be an effective strategy to improve the separation performance of the traditional zeolites.T he directional decoration of the 12membered rings of traditional mordenite with ZIF fragments was found to greatly enhance its kinetic propene/propane selectivity. [93] One of the decorated mordenites achieved ah igh kinetic selectivity of 139 at 298 K, indicating the effectiveness of the decoration strategy.
Over the past decades,anumber of MOF materials have been investigated for the kinetic separation of propane and propene. [94] Thef irst and most-studied example is ZIF-8, which showed ak inetic selectivity of 125 at 303 K ( Figure 13 a). [94a] It is also reported that for ZIF-67 (ZIF-8 (Co)), cobalt promotes amore rigid framework and slightly smaller

Angewandte
Chemie windows than zinc in the isostructural ZIF-8, and as ar esult UIF-67 displays the opposite kinetic propane/propene separation. [95] Aq uite innovative concept was also proposed in that the pore architecture of the ZIF-8 material can be mechanically tuned by the application of external pressure up to 1Gpa, resulting in as ignificant enhancement in the propylene/propane diffusion selectivity (Figure 13 b). [96] Gate-opening effect.The gate-opening effect is known for metal-organic frameworks,p roviding them with advantageous properties gas separation. Afew MOFs displaying gateopening behavior have been explored for propene/propane separation. Thef lexible pillared-layer framework CPL-1 exhibited at hermo-responsive gate-opening behavior towards propene rather than propane,a nd thus the adsorptive separation of propene over propane was achieved. [97] It is also confirmed that hydrogen bonding plays an important role in the adsorptive separation of propene over propane for CPL-1. Li et al. also reported that the flexible MOF NJU-Bai8 can separate propene and propane based on ag ateopening effect over aw ide temperature range from 298 to 348 K( Figure 14). [98] It exhibited ah igher propene/propane selectivity at lower pressure (7.2 at 0.05 bar) and exceeded 4 in the range of total pressure up to 100 kPa, at 298 K.
Size exclusion mechanism.Suitable porosity and precisely controlled channels are required to achieve propene/propane separation via the size exclusion mechanism, which is particularly challenging.U pt on ow,q uite af ew MOFs have been reported to achieve the ideal propene/propane separation by the molecular sieving effect. By controlling the surface chemistry and pore size (usually substitutions of the organic ligand and inorganic nodes), the MOF structures can be finetuned, thereby achieving ideal molecular sieving properties for gas separations.T he ultra-microporous fluorinated MOF NbOFFIVE-1-Ni (also referred to as KAUST-7), which can fully sieve propane from propene/propane mixtures under ambient conditions,was reported. [23] Thesieving effect is attributed to the selected bulkier (NbOF 5 ) 2À hindering the rotation of the pyrazine moieties,and thus dictating the pore aperture size and its maximum opening. However, due to the flexible aperture,t he effect can only be achieved at low pressures.A nother MOF material, Y-abtc (with abtc = 3,3',5,5'-azobenzene-tetracarboxylates) with cage-like pores, was prepared by at opology-guided design strategy;i tc ould adsorb propene,b ut completely excluded propane (Figure 15). [99] Precise tuning of the pore aperture and optimal pore dimensions for propene/propane separation was achieved by replacing secondary building units and through the judicious selection of structure topology,inorganic nodes,and organic linkers.T able 5p rovides an overview of selected state-of-the-art porous materials for C 3 H 6 /C 3 H 8 separation.

Separation of C 8 Aromatics
Them ixture at the top of the xylene splitter typically contains about 19 %e thylbenzene,4 4% m-xylene,2 0% oxylene,a nd 17 % p-xylene,a nd must be separated into the individual isomers for specific end-use,a sm entioned before. [100] Thes eparation of xylene isomers is highly challenging as ar esult of the similar physicochemical properties of these isomers (Table 1). In particular, the extremely close boiling points makes it impracticable to separate them efficiently by distillation, because of the large number of theoretical plates required. [101] Thea dsorptive strategy is the main technology for industrial separation of xylene isomers, of which about 60 %PXseparation is performed by selective adsorption with zeolites.I ndustrially,s eparation of PX is mainly performed on large-scale simulated moving bed (SMB) units,w hich has been implemented in three industrial-scale processes (i.e., UOP'sP arex, ToraysA romax, and IFP'sE luxyl). [102] TheP arex process,f irst commercialized by UOP for the production of PX in 1971, was pioneering in applying the principle of adsorption separation on an industrial scale.Later, in the early 1970s,the Aromax process was developed by Toray Industries,a nd in 1994, IFP commercialized the Eluxyl adsorption process.I nr efineries, the amount of PX in xylene mixtures varies from 17 to 24 wt %. By these technologies,aliquid mixture can be separated by SMBs around 453 Ka nd 9bar, achieving the isolation of PX with ap urity grade superior to 99 %. [103] The adsorbents employed in SMBs are FAU-type zeolites Xa nd Y, ion-exchanged with, for example,N a + ,K + ,a nd Ba 2+ . [104] Besides the FAUz eolite framework structure,o ther zeolite framework structures,l ike MFI [105] and MOR, [106] also have been tested for PX separation. Thedominant role adsorbents play in these separation process makes it necessary to explore and develop highly efficient adsorbents.A dvanced porous materials,l ike MOFs and COFs,h ave been extensively explored for separation of xylene isomers. [107] However,f or zeolite-based separation materials not much progress has been made in recent years,a lthough it should be clear that their durability and thermal stability offer many prospects (mainly in industrial applications) over the more recently explored MOFs and POFs.
Polyukhov et al. synthesized ZIF-8 with the stable nitroxide TEMPO ((2,2,6,6-tetramethylpiperidin-1-yl)oxyl) permanently entrapped in the pores ( Figure 16). [108] Thediffusion of xylene isomer molecules in the ZIF-8 cavities could be significantly modulated by changing the temperature within the range 298-333 K, as the window size of the TEMPO@-ZIF-8 changes with temperature.InTEMPO@ZIF-8, PX can be easily separated from OX and MX with MX/OXs eparation efficiencies up to 93-95 %a t2 98 K. MOFs also can be used as astationary phase for the chromatographic separation of xylene isomers and ethylbenzene. [109] Functionalized Zr-BTB nanosheets with an untwisted stacking mode were examined for chromatographic separation. [110] Thes tacking was untwisted and ordered sub-nanometer pores were created by preheating;acolumn coated with the untwisted nanosheets exhibited practical PX selectivity from xylene isomers. In 2018, Long et al. reported two MOFs,C o 2 (dobdc) and Co 2 (m-dobdc), with unsaturated cobalt(II) sites for the separation of C 8 aromatic isomers;their separation performances depended on differences in interactions between each isomer and the two open cobalt(II) centers. [35c] All four isomers can interact with both the unsaturated cobalt(II) sites and the linker aromatic rings in Co 2 (dobdc) through arene pp interactions with the dobdc 4À linker.Owing to the different strengths of its binding affinities to those isomers (OX > EB > MX > PX), Co 2 (dobdc) can separate all four isomers effectively.I nc ontrast, Co 2 (m-dobdc) has similar binding to MX and EB and can thus distinguish only three of the four isomers.T his indicates that the subtle structural differences between Co 2 (dobdc) and Co 2 (m-dobdc) have ah uge impact on their adsorption properties.Recently,byrefining the pore size at sub-Angstrom resolution (7.4-6.3 in steps of 0.2 ), Schrçder et al. reported aseries of MFM-300(M) (M = In, V, Fe,A l) for the separation of xylene isomers at room temperature,a chieving selectivity (PX < OX < MX) and [a] Zr-BPYDC is propane-selective. The selectivity is C 3 H 8 /C 3 H 6 selectivity and uptake is C 3 H 8 uptake. Figure 16. a) Schematic of the self-assembly of TEMPO@ZIF-8. b) Kinetic curves a(t)for TEMPO@ZIF-8 with different xylenes (indicated) at room temperature. c) Schematic of the separation of xylene isomers in TEMPO@ZIF-8. Adapted from ref. [108].

Angewandte
Chemie separation factors of 4.6-18 for PX and MX. [111] Xing al. also reported aZU-61 MOF with accessible and rotational Lewis basic sites for the adaptive molecular discrimination of xylene isomers ( Figure 17). [112] Through the rotation of fluorine atoms,t he anionic sites can adapt to the shape specific isomers,t herefore enabling ZU-61 to efficiently separate xylene isomers.Z U-61 exhibited ap referential adsorption sequence of OX > MX > PX and both high MX uptake capacity (3.4 mmol g À1 ,7 .1 mbar) and MX/PX separation selectivity (2.9, obtained from breakthrough curves) were achieved at 333 K. TheH TCS method has also been used to investigate the separation of xylene isomers.S holl et al. performed GCMC simulations for % 4700 MOF structures from the Computation-Ready,E xperimental MOF database to identify PX-selective MOFs. [113] Thet wo best-performing MOFs (MIL-140B and MOF-48) were synthesized and evaluated by breakthrough experiments and modeling.T he PX selectivities of MIL-140B and MOF-48 are lower than the simulated results but exceeded that of zeolite BaX. The diversity,high stability,and adsorption capacity of POFs also endow them with anumber of benefits for the separation of xylene isomers. Host-guest interactions between POFs and xylene isomers can by tuned by choosing suitable external crosslinkers and building blocks. Thet riptycene-like microporous organic polymer POP-1 was reported for the separation of xylene isomers.
[109c] Thew eak CH/p interactions were successfully used to tune the host-guest interactions and to achieve separation of xylene isomers.I n2 019, COFs were first reported by Huang and co-workers for the separation of xylene isomers and ethylbenzene. [114] Tw opairs of microporous 3D salen-and Zn(salen)-based COFs served as the stationary phase and were examined for the chromatographic separation of C 8 aromatics.A ll four COFs present as evenfold interpenetrated diamondoid open framework with wide tubular channels of about 7.8 decorated with salen or Zn (salen) units (Figure 18 a-c). C 8 aromatic molecules stack in the COFs by an edge-to-face configuration.
Only the two methyl groups of OX can interact with the polar salen groups of the four COFs,and the hydrogen bond lengths between OX and COFs are shorter than with other isomer molecules.O ne salen-COF exhibited excellent column efficiency and precision. Ther etention time remained the same   [a] Uptake measured at 318 K.

Angewandte
Chemie for EB and xylene isomers when the injected masses increased from 3t o3 0mg, and only as light decrease in selectivity was observed as temperature increased from 20 to 32 8 8C (Figure 18 d-f). Table 6provides an overview of selected state-of-the-art porous materials for separation of xylene isomers.

Concluding Remarks, Challenges, and Outlook
Thepast three decades have witnessed the fast growth and development of adsorptive purification and separation of hydrocarbons with awide range of porous materials.W ehave summarized and discussed the recent developments of porous materials research for the separation of methane/nitrogen, ethene/ethane,p ropene/propane,a nd C 8 aromatics mixtures. Great progress has been made in adsorptive separation of hydrocarbons and the research on aw ide variety of porous materials with different compositions and structures is going at full blast. Figure 19 provides as ummary of the reported selectivities and volumetric uptakes of typical porous materials for the separation of methane/nitrogen, ethene/ethane, and propene/propane.T raditional materials,l ike zeolites, have played and will continue to play an important role in gas separations.S ince several zeolites have been successfully applied in industrial processes,z eolites have obvious advantages for practical application of hydrocarbon separations. Modifications of zeolites and the synthesis of new zeolite materials with novel framework structures and chemical compositions are the two strategies for the development of new,better performing zeolite-based adsorbents.Incontrast, for another traditional material, porous carbons and activated carbons,t he irregular structures as well as the difficulty in tuning pore shapes and sizes greatly limit their practical application and separation performance.A dvanced porous materials,i ncluding MOFs and POFs,e xhibit great potential and appear quite promising for the purification and separation of hydrocarbons ( Figure 19). They are currently being researched in great detail. Thep recise control of both their surface chemistry and aperture size has proven to be asuccessful strategy,o ffering some MOF and POF materials extraordinary separation selectivities.
Although progress has been made,t he separation of hydrocarbons using porous materials still faces many chal- Figure 19. Comparison of a) CH 4 /N 2 adsorption selectivity and volumetricC H 4 uptake, b) C 2 H 4 /C 2 H 6 adsorption selectivity and volumetricC 2 H 4 uptake, c) C 2 H 6 /C 2 H 4 adsorptions electivity and volumetric C 2 H 6 uptake and d) C 3 H 6 /C 3 H 8 adsorption selectivity and volumetricC 3 H 6 uptake in typical porous materials.

Angewandte
Chemie lenges for future successful applications.These challenges can be summarized as follows: 1) One of the greatest challenges involves developing adsorbents that perform well in terms of both selectivity and adsorption capacity.A lthough MOFs and POFs are promising,a ttention should still be paid to traditional materials.Z eolites,M OFs,a nd POFs are all promising candidates;h owever, for zeolites,e specially for modified zeolites,t heir adsorption capacities need to be improved. Many MOFs and POFs possess either high adsorption capacities or high selectivities.Studying the structural and chemical properties of porous materials and understanding the influence of slight differences in their physicochemical properties should be further evaluated. Furthermore,e mphasis should also be put on exploiting new MOFs and POFs and industrializing the best ones in practical applications,w hich also require the proper shaping of these porous materials.T he combination of traditional and advanced porous materials also would be an effective strategy for the preparation of novel adsorbents. 2) Developing ethane-selective and propane-selective adsorbents is still difficult. More emphasis should be put on the development of highly efficient ethane-selective and propane-selective adsorbents for light olefin separation systems.F or xylene isomers,b esides improvement in the selectivity of PX-selective adsorbents,the development of MX-adsorption materials for the removal of MX in PXcontaining gas streams is also challenging but appealing. 3) Compared with the more classic use of single-component gas adsorption isotherms,t he dynamic breakthrough experiment has now become abasic and common evaluation method for porous adsorbents.I na ddition to developing novel and efficient porous materials,i nt he future, more properties of adsorbents,especially those related to industrial applications,i ncluding stability,m echanical properties,s haping, and regeneration, should be used as the criteria for the evaluation of adsorbents. 4) Advances in computational methods have already assisted in elucidating adsorption and separation mechanisms, including predicting the structure of materials and guiding the design of new adsorbents.H igh-throughput computational screening can play arole in accelerating the design and development of highly efficient adsorbents.However, there is ag ap between experimental synthesis methods and computational methods.F or example,t he proposed hypothetical structures are often difficult or impossible to synthesize experimentally.C omputational works should take at horough consideration of both experimentally synthesized structures and hypothetical framework structures.W ith the rapid development of computer science, computational screening methods,i ncluding machine learning, will continue to play asignificant role in boosting the design, development, and application of porous materials in hydrocarbon separation and purification.
Despite many potential challenges,there is no doubt that important improvements will be achieved in the field of adsorptive separation and purification of hydrocarbons.W e can expect that in the near future,m ore efforts will be devoted to the implementation of porous materials,MOFs in particular,i nl arge-scale industrial applications,t hereby focusing on improving the stability in the presence of gasphase impurities and reducing the overall cost of adsorbents as well as process design.