Crystal Engineering of Reticular Materials for Gas‐ and Liquid‐Phase Hydrocarbon Separation

ABSTRACT

Crystal engineering focuses upon the design, properties, and applications of crystals, whereas reticular chemistry involves linking molecular building blocks to create network structures. The intersection of these areas is evident in the number of systematic studies of structure/function relationships concerning porous coordination networks (PCNs) and covalent organic frameworks (COFs). PCNs and COFs are inherently modular in nature and therefore amenable to systematic fine‐tuning of both pore size and chemistry in a manner that is infeasible for other classes of porous solid. This review highlights how this exquisite control over pore size and chemistry has enabled the development of a new generation of physisorbents that are effective in the context of industrially relevant hydrocarbon (HC) separations. The motivation behind such reticular sorbents is the need to replace today's energy‐intensive HC separation methods with more sustainable alternatives. Physisorbents are attractive in this context as they can offer the high selectivity needed for trace removal of impurities along with relatively low energy of recycling. This review details how crystal engineering strategies offer precise control of pore size and chemistry to enable HC selectivity to reach hitherto unprecedented levels. Nevertheless, despite these property advances, challenges remain to be addressed before commercial adoption becomes feasible.

Crystal engineering enables systematic study of structure/function relationships as exemplified by pore engineering of reticular sorbents, including porous coordination networks and covalent organic frameworks. This review assesses such studies applied across the full scope of industrially relevant hydrocarbon separations to provide insight into how reticular sorbents offer low energy alternatives to conventional separation processes like cryogenic distillation and solvent extraction.

Abbreviations

AUM

azolate ultramicroporous material

BET

Brunauer–Emmett–Teller

CNs

coordination networks

COFs

covalent organic frameworks

CSD

Cambridge Structural Database

DCB

dynamic column breakthrough

DFT

dispersion‐corrected density functional theory

ELM

elastic layer‐structured MOF

EU

European Union

FT‐IR

fourier transform‐infrared spectroscopy

GC

gas chromatography

HCs

hydrocarbons

HUM

hybrid ultramicroporous material

IAST

ideal adsorbed solution theory

LP

large pore

MAF

metal azolate framework

MBB

molecular building block

MCM

mobil composition of matter

MD

molecular dynamics

MIL

Matériaux de l′Institut Lavoisier

MIP

Materials of the Institute of Porous Materials of Paris

MOFs

metal–organic frameworks

MOMs

metal–organic materials

MOP

metal–organic polyhedra

MTO

methanol‐to‐olefins

MUF

Massey University Framework

NG

natural gas

NP

narrow pore

PCCs

porous coordination cages

PCNs

porous coordination networks

PCP

porous coordination polymer

PET

polyethylene terephthalate

PIET

photoinduced electron transfer

Q _st

isosteric enthalpy of adsorption

RBBs

rod building blocks

RE

rare‐earth

RH

relative humidity

RON

research octane number

SBB

supermolecular building block

SCXRD

single‐crystal X‐ray diffraction

SIFSIX

hexafluorosilicate

SMB

simulated moving bed

SSST

synergistic sorbent separation technology

TGA

thermogravimetric analysis

TIFSIX

hexafluorotitanate

UMCs

unsaturated metal centeres

VOC

volatile organic compounds

VT‐PXRD

variable temperature‐powder X‐ray diffraction

ZIFs

zeolitic imidazolate frameworks

1. Introduction

The chemical industry, with an annual turnover of approximately $6.2 trillion in 2024, excluding pharmaceutical chemicals (≈$1.6 trillion in 2024), accounts for around 5.4% of global GDP (≈ $113.8 trillion) [1, 2]. It is also one of the most energy‐intensive industries, with around 40% of its energy consumption dedicated to separation and purification processes [3]. Overall, the energy footprint of chemical separations and purifications has been estimated to consume nearly 15% of global energy consumption [4]. A key driver of this high energy footprint is that the current state‐of‐the‐art relies upon utilization of energy‐intensive purification methods like cryogenic separation, fractional distillation, catalytic cracking, solvent extraction, and combustion [5, 6]. With the demand for chemical commodities expected to triple by 2050 [7], this energy footprint must be addressed. Energy‐efficient approaches to commodity separation and purification are therefore urgently needed, including for hydrocarbons (HCs), the highest volume products of the chemical industry.

Efforts towards decarbonization and energy efficiency are gaining momentum, with European Union (EU) initiatives to reduce greenhouse gas emissions and improve energy efficiency being part of broader climate goals [8]. Further, innovations in chemical processing, such as increased reliance on emission reduction methods like carbon capture, process electrification, and improved recycling, are being developed to help the sector meet its ambitious sustainability targets that are aligned with the EU's zero pollution ambition [9, 10]. For example, let us consider HC production, in particular ethylene (C₂H₄) purification, a crucial step in producing polyethylene, the World's most widely used plastic. C₂H₄ serves as a key feedstock for not only plastics but also detergents and coatings. Global production is already nearing 250 million tons per year, with projections suggesting that capacity will double by 2050 [11, 12]. However, producing C₂H₄ is energy‐intensive, particularly during the downstream purification stages where impurities like acetylene (C₂H₂), CO₂, and ethane (C₂H₆) must be removed. C₂H₄’s environmental impact is significant, contributing nearly 13% of the petrochemical industry's total CO₂ emissions. The primary culprit is the steam cracking process, which releases around 1.2 tons of CO₂ for every ton of C₂H₄ produced [12, 13]. The industry is under pressure to reconcile the expected growth in HC production with global net‐zero carbon goals. Advanced technologies and machine learning are now being employed to identify more energy‐efficient separation processes, but large‐scale adoption is essential to achieve energy and emissions reduction targets. The widespread implementation of low‐cost, energy‐efficient separation methods will be crucial for the sector's sustainability goals. In the current “age of gas”—where gases and vapors have largely replaced liquids as fuels and feedstocks in the chemical industry—the strategic adoption of alternative separation technologies is vital [3]. Replacing energy‐intensive methods like fractionation, cryogenic separation, and solvent extraction with more energy‐efficient solutions, such as physisorption and its low‐energy recycling footprint, is key to this transition [3, 6].

Several classes of solid adsorbents, such as zeolites, activated carbon, and hierarchical inorganic solids, including mesoporous silica materials like Santa Barbara Amorphous‐15 (SBA‐15) and Mobil Composition of Matter No. 41 (MCM‐41), have been investigated for HC separation. However, the performance of such physisorbents has been suboptimal due to the difficulty in precisely fine‐tuning pore structure (size, shape) and environment (pore chemistry) at the molecular level. HC separations are particularly challenging for these conventional physisorbents, as they tend to suffer from weak binding affinities, especially when dealing with HCs that exhibit similar physicochemical properties.

In this context, the potential utility of advanced physisorbents in HC production from chemical feedstocks is evident (Table 1). Researchers in crystal engineering, materials science, and process engineering are actively addressing the development of innovative, energy‐efficient HC separation processes, and developing the materials that will enable these innovative processes. These efforts are increasingly driven by physisorbents that are designed from first principles and fine‐tuned by exploiting their modular compositions. This article delves into the key advances behind this new generation of crystalline reticular physisorbents, highlighting the significant progress that has been made in performance metrics relevant to HC separation and purification processes and the underlying reasons for these enhancements in performance. Amorphous porous materials, inorganic sorbents, and membranes are excluded from the scope of this review.

TABLE 1.

A comparative analysis of conventional industrial separation processes versus reticular sorbents‐based physisorption.

Separation task	Conventional industrial route	Typical energy consumption and key constraints	Reticular sorbents‐based physisorption	Potential impact (displace/complement) and advantages
C2H4/C2H6	Cryogenic distillation	Energy: Very High (∼200–400 MJ/ton C2H4 ) Constraints: High capital cost, operates at cryogenic temperatures (−25°C to −50°C) and high pressure (∼20–30 bar), complex and energy‐intensive due to similar boiling points.	Adsorptive Separation: Equilibrium‐based: Using PCNs (e.g., Mg‐MOF‐74, Fe‐MOF‐74) with higher affinity for C2H4 . Kinetics‐based: Using PCNs/COFs with precise pore sizes to selectively diffuse C2H4 .	Complement/Displace in niche cases. Advantages: Can operate at near‐ambient temperatures, potentially significantly lowering energy use. Suitable for small‐scale and/or purifier units.
C3H6/C3H8 (Propylene/Propane)	Cryogenic distillation	Energy: Extremely High (∼300–500 MJ/ton C3H6 ) Constraints: Even more demanding than C2 separation due to closer relative volatility. Requires very tall distillation columns (over 200 trays).	Adsorptive Separation: Using rigid PCNs/COFs with pore chemistry/size for one‐step, high‐purity C3H6 capture.	Potential for Displacement (Long‐term). Advantages: Promises a one‐step, low‐energy alternative. Recent materials exhibit record selectivity and capacity, potentially replacing this energy‐intensive process with low‐energy physisorbent alternatives.
CO2 Capture (Post‐Combustion)	Amine scrubbing	Energy: High (∼3.5–4.5 GJ/ton CO2 for regeneration) Constraints: High energy penalty for solvent regeneration (120‐140°C), solvent degradation (oxidation, corrosion), environmental concerns with waste streams.	Physisorption on Solid Sorbents: Using PCNs/COFs (e.g., Mg‐MOF‐74, SIFSIX‐3‐Cu, COF‐102) with high CO2 capacity at flue gas conditions (∼40–60°C, 1 bar).	Complement, aiming for Displacement. Advantages: Much lower regeneration energy (80‐100°C via Temperature Swing Adsorption – TSA), faster adsorption kinetics, no corrosive liquids.
CO2 Capture (Pre‐Combustion)	Physical solvents (e.g., Selexol)	Energy: Moderate‐High (for solvent regeneration and refrigeration) Constraints: Operates at high pressure (∼20–70 bar), co‐absorption of other gases (H2S).	Pressure Swing Adsorption (PSA): Using PCNs/COFs (e.g., Ni‐MOF‐74, COF‐105) with high CO2 capacity at elevated pressures.	Strong Potential to Displace. Advantages: PSA cycles are highly compatible with high‐pressure streams. PCNs/COFs can offer higher selectivity and capacity over physical solvents, simplifying the process.
CH4/N2 (Natural Gas Purification)	Cryogenic distillation	Energy: High Constraints: Technically challenging and expensive due to the very low boiling points of both components. Requires significant compression and cooling.	Adsorptive Separation: Using PCNs (e.g., SIFSIX‐3‐Cu, Ni‐MOF‐74) with tailored pores for strong N2 binding over CH4.	Displace (Primary target for new tech). Advantages: Cryogenic distillation is rarely used due to cost; current methods (e.g., PSA with zeolites) are inefficient. PCNs/COFs with high N2 uptake offer a potentially transformative, low‐energy solution.
Hydrocarbon Separations (Xylene Isomers)	Crystallization/simulated moving bed (SMB) chromatography	Energy: High (for refrigeration in crystallization; complex operation for SMB) Constraints: Crystallization is batch‐wise; SMB is complex with continuous counter‐current flow of solid and liquid.	Adsorptive Separation: Using rigid PCNs/COFs (e.g., MIL‐125, COF‐5) with shape‐selective pores to differentiate between isomers (e.g., p‐xylene vs. o‐xylene/m‐xylene).	Displace (Demonstrated success). Advantages: PCN‐based benchmark separations of xylene isomers is a landmark success. It offers a continuous, more efficient, and lower‐energy process than crystallization.

2. Crystal Engineering of Reticular Materials

Crystal engineering is the field of chemistry that studies the design, properties, and applications of crystalline solids and has rapidly evolved from its initial focus on crystal design to its current emphasis on controlling bulk properties [14]. It was not until 1989 that principles for design of coordination networks (CNs) were presented by Hoskins and Robson, in an article that laid the foundation for the design of new CNs by following the principles of self‐assembly and crystal engineering to generate a diamondoid (dia) topology network [15]. They followed up in 1990 with an article that introduced the important concept of modular self‐assembly through the “node‐and‐linker” approach [16] to generate an interpenetrated square lattice (sql) topology network that laid the groundwork for future development of reticular crystalline materials, including porous CNs (PCNs) [17, 18, 19]. A linker refers to a ligand (organic or inorganic) that connects two adjacent nodes, a node being any chemical entity that has three or more points of connection [20]. As illustrated in Scheme 1, the role of the linker is to propagate the inherent connectivity and geometry of the node. In such a manner, families of CNs based upon topologies such as dia, sql, primitive cubic (pcu), or honeycomb (hcb) are accessible through use of metal cations as 4‐connected (4‐c) tetrahedral, 4‐c square planar, 6‐c octahedral or 3‐c trigonal nodes, respectively. Metal clusters and organic moieties with three or more connections were also soon introduced as nodes, followed by high connectivity polyhedral nodes that can generate high symmetry CNs, e.g., face centered cubic (fcu) topology networks can be generated by 12‐c nodes. Since selected nodes typically pre‐existed as molecular compounds, the term molecular building block (MBB) was coined [21], or, in the case of polyhedral nodes, supermolecular building block (SBB) [22]. The network structures depicted in Scheme 1 are representative examples of network structures that can be designed by the node and linker crystal engineering approach and, as we now know, there are many potential nodes and many linkers, thus far resulting in >131,682 CNs archived in the metal‐organic framework (MOF) subset of the Cambridge Structural Database (CSD), specifically, the CSD release 2025.1 (version 6.00, build 0.617, from May 2025 [23]. Inspired by Robson's node and linker crystal engineering approach introduced in 1989 [16], the design of families of CNs advanced rapidly in the 1990s. In 1994, Fujita and colleagues reported a non‐interpenetrated variant of Robson's 1990 CN with sql topology [16], and MacGillivray et al. reported interpenetrated dia and hcb nets [24]. The dia and sql CNs are based upon 4,4′‐bipyridine (4,4′‐bipy) linkers and 4,4′‐bipy is the most commonly used linker in the MOF subset of the CSD with 8592 entries (CSD release 2025.1, version 6.00, build 0.617, from May 2025). Fujita's sql net was studied by single‐crystal X‐ray diffraction (SCXRD), which revealed clathration of o‐dibromobenzene [25]. Such sql nets remain of topical interest as they are amenable to fine‐tuning through metal, linker or counterion substitution. That multiple moieties can be substituted is profound from a crystal engineering perspective as second generation, Gen‐2, materials can be formed by systematic variation of composition that in turn modifies pore size, shape and chemistry [17]. The following year, Moore and Lee coined the term MBB [21], which can be defined as a known molecule or coordination complex that can be used to construct network solids. MBB‐based nodes have since been widely employed in the construction of CN families, as exemplified by CNs containing metal carboxylate clusters, including 4‐c nodes such as the “paddlewheel” cluster, M₂(RCOO)₄. sql topology CNs based upon paddlewheel clusters of formula [M₂(1,4‐bdc)₂]_n, 1,4‐bdc = 1,4‐benzenedicarboxylate, were reported to be porous by Mori et al. (M = Cu) and Yaghi et al. (M = Zn) in 1997 [26] and 1998 [27], respectively. Also in 1997, Kitagawa's group reported the methane sorption properties of [Co₂(4,4′‐bpy)₃(NO₃)₄]_n, the first report of permanent porosity in CNs [28]. Shortly thereafter in 1999, the first PCN with ultra‐high surface area, HKUST‐1 ([Cu₃(BTC)₂]_n, a.k.a., Cu‐BTC, BTC = 1,3,5‐benzene‐tricarboxylate) [29], was reported. HKUST‐1 is comprised paddlewheel MBB nodes and demonstrated that PCNs can exhibit properties previously unattainable in earlier classes of physisorbents. The gravimetric surface area record set by HKUST‐1 was soon eclipsed by MOF‐5, also reported in 1999 [30].

SCHEME 1.

A schematic illustration of the “node and linker” approach to the crystal engineering of reticular networks exemplified by self‐assembly of linkers ① and commonly used nodes ②‐⑥ (3‐, 4‐, 4‐, 6‐ and 12‐ connected nodes, respectively).

Distinct from MOFs in composition are “hybrid” CNs comprised of combinations inorganic and organic linker ligands. The prototypal CN of this type, [Zn(4,4′‐bipy)₂(SiF₆)]_n, SIFSIX‐1‐Zn, was reported as a non‐interpenetrated pcu topology CN in 1995 by Zaworotko's group [31] and the Cu variant was subsequently studied for its methane storage properties by Kitagawa et al. [32] Such CNs are also highly modular as they can be regarded as pillared sql nets, meaning that there are at least three variables (linker, pillar and metal) for fine‐tuning of pore size, shape and chemistry. An ultramicroporous variant, [Zn(pyrazine)₂(SiF₆)]_n, SIFSIX‐3‐Zn, was found to offer exceptional performance for trace capture of CO₂ and is the prototypal hybrid ultramicroporous material, HUM [33], although such structures had been known since the 1990s [34]. HUMs marked an important step forward in terms of HC separation as their narrow pores, when integrated with fluoro‐rich pore walls, afforded tight binding enhanced by strong electrostatics, thereby enabling new benchmarks for trace capture of several light HCs, including C₂H₂ [35, 36]. Gen‐2 HUMs have resulted in further improvements in adsorptive separation performance by at least two orders of magnitude [37, 38, 39, 40]. Pillaring of [M₂(1,4‐bdc)₂]_n sql nets was also used as the design principle to prepare a large family of MOFs derived from DMOF‐1, DMOF‐1 being comprised of [M₂(1,4‐bdc)₂]_n nets pillared by dizazabicylooctane, dabco [41].

Several design approaches to PCNs were therefore established by 2000, although several notable developments followed. In 2003, X. M. Chen and colleagues introduced the first metal azolate framework (MAF) with zeolitic (sodalite, SOD) topology (MAF‐3), [Zn(bim)₂] (bim = benzimidazolate) [42]. Chen's work spawned development of the family of adsorbents now commonly known as zeolitic imidazolate frameworks (ZIFs) [43, 44]. Also in 2003, Yaghi and O'Keeffe introduced and defined the concept of reticular synthesis, i.e. the design of reticular materials using pre‐determined building blocks to form families of extended solid‐state network architectures [45]. In 2004, Zaworotko's group introduced the idea supermolecular building blocks (SBBs), CNs built by self‐assembly of metal–organic polyhedra (MOPs) [22]. SBBs can offer high connectivity (≥8) and typically better control over network topology than low connectivity nodes. Prototypal examples of SBBs that form PCNs include 12‐c [M₆(bdc)₁₂]¹²⁻ (M = Ni, Co) [46] and the 24‐c “nanoball” SBB, Cu₂₄(1,3‐bdc)₂₄ [47]. In 2005, Yaghi's group introduced covalent organic frameworks (COFs), porous networks comprised solely of organic nodes and linkers, by following the same design principles used to design reticular CNs [48]. Also in 2005, O'Keeffe and Yaghi reported CNs featuring metal‐based rod building blocks (RBBs) [49]. The coordination environment of RBBs in 3D CNs (e.g., the M‐MOF‐74 family based upon Zn‐MOF‐74 ([Zn₂(dobdc)]_n; H₄dobdc = 2,5‐dioxido‐1,4‐benzenedicarboxylic acid) [49]) typically precludes interpenetration, thereby enabling extra‐large porosity by extending the linker length.

With design rules in place by the mid‐2000s, attention shifted to structure‐function relationships. While coordination complexes and organic molecules have long been known for their ability to include guest molecules (clathrate formation) [50, 51], often involving transformations between closed and open phases (switching), the first examples of 2D and 3D switching CNs were not reported until the early 21^st century [52, 53]. Werner complexes, the prototypical coordination compounds, had their guest‐clathration capabilities systematically explored by Schaeffer in 1957 [51]. In 1969, Barrer's group reported sorption isotherms of Werner complexes upon exposure to different gases and vapors [50]. Specifically, when [Co(etpy)₄(NCS)₂] (etpy = 4‐ethylpyridine) was exposed to benzene, toluene, and xylenes, switching (stepped) adsorption isotherms were observed for some gases, later attributed to phase transformations between non‐porous and large pore forms of the sorbent [50, 54]. In 1997, the first gas sorption studies of PCNs that verified permanent porosity were reported by the group of Kitagawa [28], the 3D CN, {[M₂(4,4′‐bpy)₃(NO₃)₄](H₂O)_x}_n (M = Co., x = 4; M = Ni, x = 4; M = Zn, x = 2). Mori [26] studied the 2D CN, [Cu₂(1,4‐bdc)₂]_n, and shortly thereafter, Yaghi's group reported the sorption properties of the Zn(II) analogue, [Zn₂(1,4‐bdc)₂]_n, MOF‐2 [27]. In 1998, Kitagawa and colleagues introduced the concept of third‐generation CNs or flexible CNs [55], which later became recognized as “soft” or stimuli‐responsive CNs [56, 57]. The first CNs with ultra‐high surface areas, HKUST‐1, [Cu₃(1,3,5‐benzenetricaboxylate)₂] _n [29], and MOF‐5 [30] ([Zn₄O(bdc)₃, bdc = benzene‐1,4‐dicarboxylate]), were reported in 1999. HKUST‐1 and MOF‐5, the latter with a specific surface area >3000 m²/g, demonstrated that PCNs can exhibit properties that were hitherto unattainable in previous classes of sorbents. The study of porous materials with extra‐large surface areas remains a topical subject. The groups of Hupp and Kaskel extended the upper limit of Brunauer–Emmett–Teller (BET) surface area to approximately 7000 m²/g in NU‐110 [Cu₃(L⁶–₍₁₁₀₎)(H₂O)₃]_n (L⁶–₍₁₁₀₎ = the hexa‐anion) [58], and 7800 m²/g in DUT‐60 (Zn₄O(bbc)_4/3(bcpbd), bbc = 1,3,5‐tris(4′‐carboxy [1,1′‐biphenyl]‐4‐yl)benzene, bcpbd = 1,4‐bis‐p‐carboxyphenylbuta‐1,3‐diene), respectively.

HC separation was one of the first potential applications addressed by reticular sorbents, and a large body of subsequent work has provided insight into selective binding sites and the supramolecular chemistry that drives sorbent‐sorbate binding (Scheme 3). For example, Kitagawa's group in 2004 reported on the first “shape responsive” sorbent which exhibited induced fit binding of benzene, {[Cu₂(pzdc)₂(bpy)]}_n (pzdc = pyrazine‐2,3‐dicarboxylate), CPL‐2 [59]. In 2005, the same group reported C₂H₂/CO₂ selectivity in CPL‐1, i.e., [Cu₂(pzdc)₂(pyz)] _n , (pyz = pyrazine) [60]. The crystallographically determined C₂H₂ binding sites in CPL‐1 channels revealed H‐bonding between non‐coordinated oxygen atoms of CPL‐1 and C₂H₂ [60]. In 2001, Kitagawa's prediction of “third‐generation” flexible porous coordination polymers, PCPs [55], was validated by Kaneko et al. when they studied the two‐dimensional (2D) CN Cu(4,4′‐bpy)₂(BF₄)₂ (ELM‐11(Cu), ELM = Elastic Layered Material, or sql‐1‐Cu‐BF₄ ) [52]. In 2007, Kajiro and Kanoh's work on ELM‐12, [Cu(bpy)₂(OTf)₂]_n (MOF; OTf⁻ = triflate, also known as sql‐1‐Cu‐OTf [61]) confirmed that the observed guest‐induced stepped isotherms (type F‐II isotherm) [54] were a manifestation of a “breathing effect” from a less open phase to a more open phase. Such structural flexibility can also play a role in HC separation. That same year, Kitagawa's group reported benzene (Bz)‐selective binding over cyclohexane (Cy) driven by CH···π interactions between the flexible host framework {[Zn(μ ₄‐TCNQ−TCNQ)bpy] (TCNQ = tetracyanoquinodimethane) and Bz molecules [62]. Although unsaturated metal centers (UMCs) had been previously reported in HKUST‐1 [29] and Zn‐MOF‐74 [49], they were not utilized as sorbate binding sites until Matzger's work on CO₂ capture at low pressure in 2008 [63], followed shortly thereafter by B. Chen's work on C₂H₂ capture by HKUST‐1 [64]. Subsequently, Yaghi's and B. Chen's groups independently investigated the influence of UMCs on CO₂‐selective and C₂H₂‐selective separation over N₂ and CO₂, respectively, in Mg‐MOF‐74, [Mg₂(dobdc)]_n and UTSA‐74, [Zn₂(H₂O)(dobdc)·0.5H₂O]_n [65]. Long's group also exploited UMCs for C₂H₆ and propane (C₃H₈) purification driven by interactions between UMCs and olefinic HCs in Fe‐MOF‐74, e.g., [Fe₂(dobdc)] [66]. However, in sorbents that exploit UMCs for separation, the coordination of atmospheric water by UMCs can hinder real‐world applications [67, 68, 69].

SCHEME 3.

Key discoveries concerning structure‐function relationships in physisorbents with particular emphasis upon binding sites that drive strong performance for hydrocarbon separation and purification. Reprinted with permissions from ref [28, 29, 30, 33, 50, 55, 59, 61, 62, 63, 64, 66, 70, 71, 72, 73, 74, 75, 79, 80]. Copyright 1969, Royal Society of Chemistry; copyright 1997, Wiley‐VCH Verlag GmbH & Co. KGaA, Weinheim; copyright 1998, The Chemical Society of Japan; Copyright 1999, The American Association for the Advancement of Science; Copyright 1999, Springer Nature; Copyright 2004, The American Association for the Advancement of Science; Copyright 2007, The American Association for the Advancement of Science; Copyright 2007, The American Association for the Advancement of Science; Copyright 2008, The American Association for the Advancement of Science; Copyright 2009, The American Association for the Advancement of Science; Copyright 2012, The American Association for the Advancement of Science; copyright 2013, Springer Nature; Copyright 2013, The American Association for the Advancement of Science; Copyright 2019, The American Association for the Advancement of Science; Copyright 2019, The American Association for the Advancement of Science; Copyright 2020, The American Association for the Advancement of Science; Copyright 2021, The American Association for the Advancement of Science; Copyright 2022, Wiley‐VCH Verlag GmbH & Co. KGaA, Weinheim; Copyright 2023, Springer Nature; Copyright 2024, The American Association for the Advancement of Science.

As mentioned earlier, the Cu analog of SIFSIX‐1‐Zn, SIFSIX‐1‐Cu, exhibits excellent methane (CH₄) storage performance as detailed by Kitagawa's group in 2000 [32]. The stronger electrostatics associated with inorganic linker ligands was fully exploited when gas separation selectivity an order of magnitude better than the previous best‐performing CO₂ selective physisorbents was observed in an ultramicroporous variant of SIFSIX‐1‐Zn [31]. This material, the prototypical HUM SIFSIX‐3‐Zn ([Zn(pyrazine)₂(SiF₆)] _n , was reported in 2013 to be the first physisorbent with ultrahigh (>1000 at ≤10,000 ppm) trace CO₂/N₂ selectivity and also set a new benchmark for CO₂/CH₄ selectivity (feed ratio = 1:1, temperature: 298 K, total pressure: 1 bar) [33]. The related pcu topology sorbent, SIFSIX‐2‐Cu ([Cu(dpa)₂(SiF₆)] _n , dpa = 4,4′‐dipyridylacetylene), and its interpenetrated variant, SIFSIX‐2‐Cu‐i, highlighted the profound impact of pore size and pore chemistry on sorption properties, with the latter being the first sorbent of any type with high enough selectivity to enable removal of trace (1%) C₂H₂ from C₂H₄ [35]. Subsequently, by using three ultramicroporous sorbents in series (each suited for a specific trace gas impurity), Zaworotko and B. Chen's groups demonstrated that synergistic sorbent separation technology produces polymer‐grade C₂H₄ (purity >99.9%) in a single‐step process [70].

In 2013, Long's group reported that the triangular‐shaped pores of Fe₂(BDP)₃ (BDP²⁻ = 1,4‐benzenedipyrazolate) enabled separation of hexane isomers [71]. An example of fine‐tuning pore size by crystal engineering was reported by Feng's group in 2020, whereby a pore‐space‐partitioning strategy was introduced to adjust the pore size of Fe‐MOF‐74 for C₂H₆/C₂H₄ separation [72]. Isostructural to MOF‐5, CUB‐5, [Zn₄O(1,4‐cdc)₃], features the substituted aliphatic linker cdc, cubane‐1,4‐dicarboxylate, and methine units orient towards the pores, providing adsorption sites for the π and CH moieties of Bz [73]. In 2021, Shimizu and co‐workers reported an easily scalable, low‐cost, moderately hydrophobic MOF, CALF‐20 ([Zn₂(1,2,4‐triazolate)₂(oxalate)]), with properties suitable for CO₂ capture from flue gas. Under humid conditions, CALF‐20 preferentially adsorbs CO₂ and can be readily regenerated [74].

In 2022, Rosseinsky's group introduced the use of machine learning to explore MOFs for evaluating guest accessibility, which is crucial for several HC separations [75]. This approach has the potential to be an effective tool for screening thousands of hypothetical and literature‐reported MOFs from the CSD MOF database [76, 77], including the Computation‐Ready, Experimental (CoRE) MOF database [78], to identify the most likely separations that can be targeted by particular MOFs. More recently, in 2023, a cage‐on MOF strategy was applied to incorporate porous coordination cages (PCCs) onto the external surface of a MOF for selective dye adsorption using PCN‐222@PCCs (PCN‐222, Zr₆(μ₃‐OH)₈(OH)₈‐(TCPP)₂, TCPP = meso‐tetra(4‐carboxyphenyl)porphyrin) and MIL‐101@PCCs (MIL‐101, Cr₃F(H₂O)₂O(1,4‐bdc)₃, 1, 4‐bdc = 1,4‐benzendicarboxylic acid) [79]. In 2024, Xing's group investigated a molecular sieving mechanism in a cage‐type structure ZU‐609 (Cu(eds)(dps)₂, eds = 1,2‐ethanedisulfonate, dps = 4,4′‐dipyridylsulfide) with 1D channel‐shaped narrow windows and 3D cage‐shaped large voids. ZU‐609 broke the trade‐off between separation selectivity and capacity [80], aligned with the Zaworotko group's earlier 2021 report on overcoming this Catch‐22 [81]. The limited diffusion path functions as a sieve to exclude large molecules, while the coexisting large channel allows for the rapid diffusion of adsorbed molecules.

As indicated by Scheme 2, the key design concepts that afforded the most common families of PCNs were developed over two decades ago, and the separation potential of such materials started to be addressed shortly thereafter. This review focuses upon advances made in terms of properties and insight with respect to HC separation by physisorbents (Scheme 3), with particular emphasis upon two distinct classes of crystalline adsorbents, PCNs (including MOFs and HUMs) and COFs. Therefore, in the following sections we highlight examples of MOFs, HUMs and COFs for HC separation in the gas phase (C1‐C4 for the following separations: CH₄/N₂, N₂/CH₄, CO₂/CH₄, C₂H₂/C₂H₄, CO₂/C₂H₂, C₂H₄/C₂H₆, C₂H₆/C₂H₄, C₃H₄/C₃H₆, C₃H₆/C₃H₈, multicomponent HC mixtures including those comprising CO₂ and C4 isomers) and the liquid phase (separation of C6 isomers and C8 isomers). To obtain insight into the observed separation performance, the top‐performing reticular sorbents for each binary/multicomponent mixture are identified and analyzed with regard to their structure‐property relationships. Nine features have been identified by us and others as playing a key role in observed HC separation performance: H‐bonding; electrostatics; kinetics; flexibility‐driven induced fit; Van der Waals interactions; π interactions (C‐H···π interactions and π···π interactions); polarizability; pore size; UMCs. The relevance of each parameter to the top‐performing reticular sorbents for each separation is evaluated through radar plots in each of the following sections.

SCHEME 2.

Chronology of key discoveries in the crystal engineering of reticular materials. Reprinted with permissions from ref [15, 16, 21, 22, 25, 26, 27, 28, 29, 30, 31, 41, 42, 45, 48, 49]. Copyright 1989, Royal Society of Chemistry; Copyright 1990, Royal Society of Chemistry; Copyright 1994, The American Association for the Advancement of Science; Copyright 1995, Springer Nature; Copyright 1995, Wiley‐VCH Verlag GmbH & Co. KGaA, Weinheim; copyright 1997, Wiley‐VCH Verlag GmbH & Co. KGaA, Weinheim; Copyright 1997, The Chemical Society of Japan; Copyright 1998, The American Association for the Advancement of Science; Copyright 1999, The American Association for the Advancement of Science; copyright 1999, Springer Nature; Copyright 2003, Springer Nature; Copyright 2003, Springer Nature; Copyright 2004, Wiley‐VCH Verlag GmbH & Co. KGaA, Weinheim; Copyright 2004, The American Association for the Advancement of Science; Copyright 2005, The American Association for the Advancement of Science; Copyright 2005, The American Association for the Advancement of Science.

As a general rule, when comparing separation performances in tabular form, this article ranks the top‐performing reticular sorbents in order of decreasing adsorption selectivity, from top to bottom (Tables 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15). For binary separations, feed compositions were typically reported using equimolar ratios (1:1, v/v), though other variations were also examined depending on the context, including 1:2, 1:9, 1:99, 0.5:99, and 2:98. Most hydrocarbon separation studies were conducted at 298 K, although the full temperature range studied extends from 233 K to 523 K.

TABLE 2.

Key physicochemical properties of adsorptives relevant to HC purification [82, 83, 84, 85, 86].

Molecule	Structure	Kinetic diameter (Å)	Molecular size (Å3)	Boiling point (K)	Polarizability (× 10−25/cm3)	Dipole moment (×1018/esu cm2)	Quadrupole moment (×1026/esu cm2)
CH4		3.8	3.7 × 3.7 × 3.7	111.6	26	0	0
N2		3.64	—	77.3	17.6	0	1.52
CO2		3.3	3.18 × 3.33 × 5.36	194.7	29.11	0	−4.30
C2H2		3.3	3.32 × 3.34 × 5.70	189.3	33.3‐39.3	0	7.50
C2H4		4.163	3.28 × 4.18 × 4.84	169.4	42.52	0	1.50
C2H6		4.443	4.08 × 3.81 × 4.82	184.5	44.3‐44.7	0	0.65
C3H4 (MA)		4.2	4.16 × 4.01 × 6.51	250	55.5	0.75	—
C3H4 (PD)		—	6.2 × 3.8 × 3.8	239	56.9	0	—
C3H6		4.678	4.16 × 4.65 × 6.44	225.4	62.6	0.366	—
C3H8		4.3‐5.118	4.02 × 4.52 × 6.61	231.1	62.9‐63.7	0.084	—
C4H6		5.2	3.4 × 5.36 × 7.84	268.6	—	0	—
i‐C4H8		4.23	4.52 × 5.76 × 6.71	276.87	—	0.253	—
n‐C4H8		4.83	4.17 × 5.05 × 7.87	266.77	81	0.3	—
2‐C4H8		4.49	4.2 × 5.5 × 7.6	276.87	—	0.253	—
i‐C4H10		5.278		261.34	81.4‐82.9	0.132	—
n‐C4H10		4.687	4.2 × 4.6 × 8.1	272.66	82	0.05	—
C6H6		5.349‐5.85	6.6 × 7.3 × 3.3	353.24	100‐107.4	0	—
c‐C6H12		6.0‐6.182	7.2 × 6.6 × 5.0	343.93	107.7‐110	0	—
n‐C6H14 (n‐HEX)		4.3	9.7 × 4.5 × 4.0	341.88	119	0	—
i‐C6H14 (2MP)		5.5	9.2 × 6.4 × 5.3	333.4	—	0.1	—
neo‐C6H14 (22DMB)		6.2	8.0 × 6.7 × 5.9	322.87	—	—	—
3‐methylpentane (3MP)		5.5	9.3 × 6.2 × 5.2	336.40	—	—	—
2,3‐Dimethylbutane(23DMB)		5.6	7.8 × 6.7 × 5.3	331	—	—	—
Ethylbenzene		6.7	9.3 × 7.0 × 4.0	409.36	142	0.59	—
p‐Xylene		6.7	9.0 × 6.5 × 4.0	411.53	137‐149	0.1	—
m‐Xylene		7.1	8.4 × 7.2 × 4.0	412.34	142	0.37	—
o‐Xylene		7.4	7.7 × 7.2 × 4.0	417.59	141‐149	0.640	—

Abbreviations: MA: methylacetylene; PD: propadiene; n‐HEX: n‐hexane; 3MP: 3‐methylpentane; 2MP: 2‐methylpentan; 22DMB: 2,2‐dimethylbutane; 23DMB: 2,3‐dimethylbutane.

TABLE 3.

CH₄/N₂ and N₂/CH₄ binary separations. The following parameters are listed for comparison: pore size; BET surface area (S_BET); single‐component gas uptakes; adsorption enthalpies (Q _st); adsorption selectivity; regeneration temperature; and attributed mechanisms. Reticular sorbents are listed in decreasing order of binary selectivity for CH₄‐selective sorbents (above) and N₂‐selective sorbents (underneath).

Binary	Adsorbent, network dimensionality (nD)	Pore size e (Å)	CH4 uptake at 1 bar (mmol/g)	N2 uptake at 1 bar (mmol/g)	Q st (kJ/mol)	Binary Selectivity a	Regeneration temperature (K)	SBET e (m2/g)	Mechanism	Refs.
CH4/N2	Ni(ina)2 , 3D	5.0 × 5.8	1.82	0.53	28.0/17.9	15.8	298	470	C‐H···π interactions	[87]
	Al‐CDC, 3D	5.4 × 5.4	1.43	0.23	27.5/18.6	13.1	—	380	Van der Waals interactions	[88]
	Co3(C4O4)2(OH)2 , 3D	4.1 × 4.3	0.4	0.19	25.1/18.1	12.5	373	—	Van der Waals, electrostatics	[89]
	Al‐fumarate, 3D	5.7 × 6	1.14	0.2	15.9/6.3	11.7	393	1130	C‐H···π interactions	[90]
	SBMOF‐1, 3D	7.3 × 7.3	0.92	0.18	23.2/16.1	11.5	—	149.7	Van der Waals interactions	[91]
	STAM‐1, 3D	5 × 5; 7 × 7	0.63	0.11	20.0/15.0	10.8	298	110	Van der Waals interactions	[92]
	ATC‐Cu, 3D	4.43 × 5.39	2.9	0.75	26.8/16.0	9.7	298	600	Electrostatics, van der Waals interactions	[93]
	ROD‐8a, 3D	6.5 × 11.8; 8.5 × 9.5	0.77	0.13	16.7	9.1	—	369.3	—	[94]
	NKMOF‐8‐Me, 3D	5.7 × 5.7	1.76	0.31	28.0/18.9	9.0	298	626.1	Van der Waals interactions	[95]
	Ni(3‐ain)2 , 3D	5 × 5	2.1	0.65	30.9/17.0	9.0	298	615.5	C‐H···π interactions	[87]
	MIL‐160, 3D	5 × 5	0.47	0.14	11/7	8.9	393	1104	C‐H···π interactions	[90]
	Al‐FUM‐Me, 3D	5.0 e	1.21	0.23	24.06/17.41	8.6	298	611.8	C–H···O interactions	[96]
	Cu(ina)2 , 3D	4.7 d	0.827	0.121	17.5/–	8.34 b	—	251.8 d	Polarizability, electrostatics	[97]
	Ni‐Qc‐5, 3D	3.6 f	1.26	0.28	31.5/23.5	7.4	298 vacuum	700 f	Van der Waals interactions, electrostatics	[98]
	Ni‐MA‐BPY, 3D	5.1 e	1.0	0.21	23.5/19.6	7.4	298	464 e	CH4 polarizability	[99]
	Co(AIP)(BPY)0.5 , 3D	8.8 × 7.9 [100]	1.03	0.26	21.2/–	7.3	—	—	Size sieving	[101]
	Co‐MA‐BPY, 3D	5.1 e	0.92	0.2	22.8/18.4	7.2	298	451 e	CH4 polarizability	[99]
	CAU‐10‐H, 3D	5 × 5	0.74	0.2	7.5/5	7.2	393	469	C‐H···π interactions	[90]
	MIL‐53(Al), 3D	7.3 × 7.7	0.57	0.15	15/5	7.1	393	915	C‐H···π interactions	[90]
	ZIF‐94, 3D	9.1 × 9.1	1.5	0.36	23.9/20.0	7	298	597	Size sieving	[102]
	Ni(OAc)2L, 3D	2.5 × 19; 2 × 2	1.76	0.47	27.8/19.8	7.0	—	1168	C‐H⋯arene interactions	[103]
	Cu(hfipbb)(H2hfipbb)0.5 , 3D	4.6 e	0.47	0.13	24.1/19.8	6.9 b	—	105	Pore size	[104]
	Ni‐BPZ, 3D	5.8 × 8.6	1.56	0.4	23/18	6.6	393	790	C‐H···π interactions	[105]
	TUT‐100, 3D	8.0 e	2.02	0.31	23.7/19.6	6.3	—	604	Pore size	[106]
	[Ni3(HCOO)6], 3D	5 × 6	2.06	1.56	24.82‐19.33/–	6.1 b	298	233	Surface polarizability	[107]
	MIL‐120Al, 3D	5.4 × 4.7	1.5	0.47	20.87/17.83	6.0	298	343	C‐H···π interactions	[108]
	Ni‐BTC, 3D	9 × 9	1.67	0.31	—	5.1	—	1088.15	Polarizability	[109]
	[Co3(HCOO)6], 3D	5 × 6	1.39	0.77	20.03/19.7	5.1 b	298	294	Surface polarizability	[107]
	Al‐FUM, 3D	6.3 e	0.91	0.23	20.60/16.16	5.1	298	1023.3	—	[96]
	UiO‐66‐Br2 , 3D	7‐8 e	0.72	0.19	—	5.06	298	622	Polarizable bromine at UMCs	[110]
	UTSA‐30a, 3D	3.2 × 3.2	0.63	0.165	—	5	—	592	Pore size	[111]
	[Mg3(OOCH)6]2 , 3D	4.9 d	2.64	1.52	24.9‐25.5/–	4.9‐5.3	433	—	UMCs	[112]
	CFAb‐FumMOF‐2, 3D	—	0.95	0.22	16.7/–	4.94	—	1116.92	C‐H···O interactions	[113]
	CFAb‐FumMOF‐1, 3D	5.9 e	0.98	0.22	18.64/–	4.87	—	1073.08	C‐H···O interactions	[113]
	Cu(bpy)2(OTf)2 , 2D	7.2	0.25	0.07	19.6/16.0	4.8	323	205.6	Flexibility	[61, 114, 115]
	CFAs‐FumMOF‐1, 3D	6.8 e	0.98	0.25	17.84/–	4.73	—	1124.86	C‐H···O interactions	[113]
	CFAs‐FumMOF‐2, 3D	—	0.92	0.22	17.34/–	4.73	—	1163.94	C‐H···O interactions	[113]
	[Cu(Me‐4py‐trz‐ia)] n , 3D	5.5 × 5.5	1.12	0.30	18.0/12.0	4.2	—	1473	—	[116]
	Ni(2‐ain)2 , 3D	4.8 × 4.8	0.3	0.1	22.9/11.1	4.2	298	—	—	[87]
	ZIF‐93, 3D	17.9 × 17.9	0.52	0.16	15.8/12.5	3.6	298	846	—	[102]
	ZIF‐68, 3D	10.3 e	0.40	0.121	15.69/11.88	3.5	—	1283 g	Electrostatics	[117]
	Al‐BDC, 3D	8.2 e	0.71	0.23	18.81/15.52	3.4	298	1335.6	—	[96]
	Cu(bpy)2(OTf)2 , 3D	3.9 × 3.9	0.35	0.11	11.0/10.0	3.2	323	225.4	Flexibility	[61, 114, 115]
	Co/DOBDC, 3D	11‐12	1.91	0.12	19.6/–	3.2	298	1089.3	UMCs	[118]
	Al‐NDC, 3D	9.2 e	0.48	0.15	14.89/11.06	3.1	298	1545.2	—	[96]
	ZIF‐69, 3D	7.8 e	0.50	0.138	16.24/12.76	3.0	—	1251 g	Electrostatics	[117]
	MIL‐100(Cr), 3D	5 × 5; 8 × 8	0.60	0.170	—	3.0	298	1528.7	UMCs	[118]
	ZIF‐8, 3D	11.6 × 11.6	0.35	0.14	12.4/9.8	2.8 c	298	1425	—	[102, 119, 120]
	MIL‐101(Cr) (293K), 3D	5‐12 and 18–35 e	0.65	0.129	15.7/12.0	2.7	—	2560	—	[121]
	ZIF‐90, 3D	11.2 × 11.2	0.47	0.20	15.9/12.5	2.3	298	1059	—	[102]
	Al‐BPDC, 3D	11.5 e	0.26	0.12	11.75/9.83	2.2	298	1729.9	—	[96]
	Ni‐MOF‐74, 3D	12 d	2.54	1.92	20.3/–	1.35	298	1405.1 d	UMCs	[97]
N2/CH4	V‐MOF‐74, 3D (Computational study)	12 × 12	—	—	—	—	—	—	UMCs	[122]
	Fe‐MOF‐74, 3D (Computational study)	12 × 12	—	—	5.3/5.5	—	—	—	UMCs	[122]
	V2Cl2.8(btdd), 3D	—	—	1.9	35/56	72 h	—	1930	UMCs	[123]
	MIL‐100(Cr) (283K), 3D	5 × 5; 9 × 9	0.8	1.64	27/39	8	283	2040	UMCs	[124]
	TYUT‐96Cr, 3D	6 e	1.25	1.03	27/37	4.6	—	—	UMCs	[125]
	MIL‐101(Cr)−NO2 , 3D	7.9, 20.2 and 20.7 e	0.57	0.72	–/30.01	2.8	298	1907	UMCs	[126]
	MIL‐101(Cr) (283K), 3D	16 × 16	0.91	0.96	20/40	1.8	—	4000	UMCs	[127]

^a For v/v = 1/1 equimolar mixture, predicted by IAST.

^b Calculated from the ratio of Henry constants.

^c Determined from theoretical calculations.

^d Pore size distribution and BET surface area calculated from the Ar adsorption isotherm at 87K.

^e Pore size distribution and BET surface area calculated from the N₂ adsorption isotherm at 77K.

^f Pore size distribution and BET surface area calculated from the CO₂ adsorption isotherm at 195K.

^g Accessible surface area.

^h For v/v = 2/98 mixture, predicted by IAST.

TABLE 4.

CO₂/CH₄ binary separations. The following parameters are listed for comparison: pore size; BET surface area (S_BET); single‐component gas uptakes; adsorption enthalpies (Q _st); adsorption selectivity; regeneration temperature; and attributed mechanisms. Reticular sorbents are listed in decreasing order of binary selectivity for CO₂‐selective sorbents (above) and CH₄‐selective sorbents (below).

Adsorbent, network dimensionality (nD)	Pore size (Å)	CH4 uptake at 1 bar (mmol/g)	CO2 uptake at 1 bar (mmol/g)	Q st (CO2/CH4) (kJ/mol)	Selectivity a	Regeneration temperature (K)	SBET (m2/g)	Mechanism	Refs.
Cu‐F‐pymo, 3D	3.3 × 3.3	—	1.61	29.1/–	> 107	He flow at 298 K	146	Molecular sieving	[128]
MUF‐16, 3D	3.6 × 7.6	0.04	2.13	32.3/–	6690	Vacuum or inert gas	214	H‐bonding	[129]
Qc‐5‐Cu‐sql‐β, 3D	3.3 × 3.3	0.06	2.16	36/–	3300	—	222	Molecular sieving	[130]
In(aip)2 , 3D	3.4‐3.6	—	1.27	34.3/–	1808	—	223.5	Molecular sieving, ‐NH2	[131]
MUF‐16(Ni), 3D	—	0.09	2.13	37.3/–	1220	Vacuum or inert gas	204	H‐bonding	[129]
PEI‐mediated amine‐MIL‐101(Cr)‐c, 3D	—	∼0.05	3.6	—	931 (0.5 bar)	—	96.4	UMCs	[132]
MUF‐16(Mn), 3D	—	0.09	2.25	36.6/–	470	Vacuum or inert gas	205	H‐bonding	[129]
[Zn(odip)0.5(bpe)0.5], 3D	3.6	0.43	5.3	42.3/19.3	376	—	550	H‐bonding, Flexibility	[133]
NbOFFIVE‐1‐Ni, 3D	—	0.18	2.31	54/–	366	—	195	F‐CO2	[134]
UiO‐66(N10%‐Zr), 3D	4 b	0.25	2.1	35.7/5.9	326	He flow at 393 K	1648	N‐CO2, Kinetic	[135]
SIFSIX‐3‐Zn, 3D	3.84 × 3.84	—	2.54	45/	231	—	250	Electrostatics	[33]
TIFSIX‐3‐Ni, 3D	—	0.22	2.21	50/–	158	333	200	F‐CO2	[134]
ZU‐66, 3D	3.55 × 3.55	0.21	4.56	35/–	136	He flow at 298 K	177	Molecular sieving	[136]
SIFSIX‐14‐Cu‐i, 3D	3.6 × 3.6	0.17	4.71	37.7/–	46.7	He flow at 298 K	612	Molecular sieving	[137]
NJU‐Bai8, 3D	4 × 3.3	—	2.57	37.7/–	40.8 c	—	1103	N‐CO2	[138]
3D‐IL‐COF‐1b, 3D	—	—	—	—	35.1	—	537	CO2, ionic liquid	[139]
SIFSIX‐2‐Cu‐i, 3D	5.15 × 5.15	—	5.4	31.9/	33	—	735	Electrostatics	[35]
IRH‐3, 3D	4.19 × 4.19	0.07	2.7	—	27	—	525.4	N‐CO2	[140]
3D‐IL‐COF‐1, 3D	8.3	0.37%	5.34%	—	23.1	—	517	CO2, ionic liquid	[139]
3D‐IL‐COF‐2, 3D	10.7	0.85%	7.61%	—	22.3	—	653	CO2, ionic liquid	[139]
3D‐IL‐COF‐3, 3D	12.4	0.33%	4.93%	—	21.5	—	870	CO2, ionic liquid	[139]
Zn‐dmtrz‐mip, 3D	5.7 × 5.8	0.85	2.92	26.5/–	17.8	Inert gas	463.9	H‐bonding	[141]
TIFSIX‐2‐Cu‐i, 3D	—	0.76	4.22	35.8/–	16	—	590	F‐CO2	[134]
[Cu3(μ3‐OH)(PCA)3], 3D	3.8 × 9.7	0.67	2.93	31.5/–	15.9	—	583	UMCs	[142]
Er‐BDC‐NH2 , 3D	4.5 × 4.5, 6.0 × 6.0	0.31	1.92	26/17	15	—	539	Dipole‐quadrupole interaction	[143]
NJU‐Bai7, 3D	3.4 × 3.4	—	2.91	40.5/–	14.1 c	—	1155	Pore size	[138]
MIL‐53‐Al powder, 3D	7 b	0.64	2.58	23.27/16.98	14.42	Vacuum	1555	Van der Waals interactions	[144]
NJU‐Bai35, 3D	—	—	3.13	33.4/–	11.6	—	863	Molecular sieving	[145]
MIL‐53‐Al beads, 3D	8.2 × 8.2	0.51	2.13	21.48/16.20	11.41	Vacuum	981	Van der Waals interactions	[144]
ZIF‐78, 3D	7.1 × 7.1	0.67	2.32	—	10.6	—	620	Dipole‐quadrupole interactions	[146]
Mg‐MOF‐74, 3D	10.2	1.05	8.61	73/18.5	8	—	1174	UMCs	[147]
In‐MOF‐3, 3D	5.0 × 5.0	0.23	0.97	29.4/–	7.8 e	—	261	H‐bonding, π···π interactions	[148]
NiNi‐Pyz, 3D	3.8 × 3.9	4.13	1.44	30/–	7.2	—	453	Electrostatics	[149]
MIL‐53(Al), 3D (303K)	5.1	0.2	1.3	—	7.0	393	—	CO2, hydroxyl interaction	[150]
Sm‐BDC‐NH2 , 3D	4.5 × 4.5, 6.0 × 6.0	0.44	1.96	23/18	6.56	—	528	Dipole‐quadrupole interaction	[143]
Y‐BDC‐NH2 , 3D	4.5 × 4.5, 6.0 × 6.0	0.89	3.53	28/19	6.3	—	824	Dipole‐quadrupole interaction	[143]
CoNi‐Pyz, 3D	3.9 × 4.1	3.56	1.36	29.2/–	6.0	—	440	Electrostatics	[149]
MOF‐508b, 3D (303 K)	4.0 × 4.0	—	1.78	14.9/–	3‐6	—	—	Quadrupole interactions	[151]
MAF‐66, 3D	—	1.31	4.41	26/19	5.8	—	1014	N‐CO2	[152]
CNT@Cu3(BTC)2 , 3D	—	7.49	13.52(18 bar)	—	5.67	—	1458	—	[153]
In‐MOF‐1, 3D	8.0 × 8.0	0.24	0.92	24/–	5.5 e	—	122	H‐bonding, π···π interactions	[148]
3D‐COF‐1a, 3D	—	—	—	—	5.3	—	596	CO2, ionic liquid	[139]
In‐MOF‐2, 3D	6.0 × 6.0	0.50	1.26	26.4/–	5.2 e	—	255	H‐bonding, π···π interactions	[148]
Cu‐BTC powder, 3D (308K)	10	0.77	3.94	15.0/22.8	5.1	423	—	—	[154]
SYSU, 3D	6.3 × 6.3	—	3.11	28.2/–	4.7	—	1110	Pore size	[138]
MIL‐101 powder, 3D	6‐20 b	0.42	1.54	26.23/14.82	4.09	Vacuum	2758	Van der Waals interactions	[144]
MIL‐101 beads, 3D	6‐20 b	0.34	1.12	22.62/14.54	3.70	—	1963	Van der Waals interactions	[144]
Cu3(BTC)2 , 3D	6	4.47	6.71(18 bar)	—	2.28	—	1587	—	[153]
UMCM‐1, 3D	14	5.01	15.82(18 bar)	—	1.82	—	2932	—	[153]
ZIF‐8, 3D	11.6	3.28	7.04(18 bar)	—	1.32	—	1311	—	[153]
MOF‐177, 3D	10.6	4.2	17.0 (18 bar)	—	0.89	—	2157	—	[153]
UTSA‐28, 3D	3.2 × 4.5, 3.8 × 3.8	0.36	3	42.9/–	—	He flow at 323 K	330	Molecular sieving	[155]

^a For CO₂/CH₄ v/v = 1/1 equimolar mixture at 1 bar, 298 K, predicted by IAST.

^b Pore size distribution, BET surface area calculated from the N₂ adsorption isotherm at 77K.

^c Separation ratios at 273 K.

^d Pore size distribution, BET surface area calculated from the CO₂ adsorption isotherm at 195K.

^e For CO₂/CH₄ v/v = 15/85 equimolar mixture at 1 bar, 298 K, predicted by IAST.

TABLE 5.

C₂H₂/CO₂ binary separations. The following parameters are listed for comparison: pore size; BET surface area (S_BET); single‐component gas uptakes; adsorption enthalpies (Q _st); adsorption selectivity; regeneration temperature; and attributed mechanisms. Reticular sorbents are listed in decreasing order of binary selectivity for C₂H₂‐selective sorbents (above) and CO₂‐selective sorbents (below).

Adsorbent, network dimensionality (nD)	Pore size (Å)	C2H2 uptake at 1 bar (mmol/g)	CO2 uptake at 1 bar (mmol/g)	Q st (kJ/mol)	Selectivity a	Regeneration temperature	SBET (m2/g)	Mechanism	Refs.
SOFOUR‐TEPE‐Zn, 3D	3.63 × 3.81, 3.31 × 3.32	3.98	0.63	45.6/26.3	16833	—	410.96	H‐bonding, Van der Waals interactions	[156]
ZUL‐330, 2D	3.3 × 3.7	5.91	1.1	—	10086	—	352	H‐bonding, C−H···π interactions	[157]
ZUL‐430, 2D	3.0 × 3.3	5.74	1.12	—	3499	—	364	H‐bonding, C−H···π interactions	[157]
NCU‐100a (SIFSIX‐dps‐Cu), 3D	1.4 × 3.0	4.57	0.49	56.3/–	1787	He flow at 298 K	358	H‐bonding	[158]
UTSA‐300a, 2D	2.4 × 3.3	3.3	0.2	57.6/–	103	—	311	H‐bonding, van der Waals interactions	[159]
JNU‐1, 3D	8	2.9	—	47.6/–	285.6	393	818	Flexibility, dipole interactions	[160]
CuI@UiO‐66‐(COOH)2 , 3D	4.4, 5.0 g	2.44	0.85	74.5/28.9	185	—	302.3	UMCs	[161]
CPL‐1‐NH2 , 3D	3.8 × 4.4	1.84	0.21	50/32.4	119	—	103	H‐bonding	[162]
Ni(4‐DPDS)2CrO4 , 3D	5.2 × 5.2	2.99	1.79	75.4/37	67.7	He flow at 323 K	317	H‐bonding, C‐H···π interactions	[163]
ZNU‐1, 3D	4.4	3.4	1.7	54/44	56.6	—	532	H‐bonding	[164]
ATC‐Cu, 3D	4.43 × 5.39	5.0	4.02	79.1/–	53.6	—	600	UMCs, van der Waals interactions	[165]
ATC‐Cu (MOF‐11), 3D	4.4 × 5.4	5.01	4.02	79.1/–	53.6	—	600	UMCs, van der Waals interactions	[165]
ZJU‐74a, 3D	3.6 × 3.8	3.83	3.08	44.5/30	36.5	He flow at 373 K	694	H‐bonding, UMCs	[166]
SNNU‐277, 3D	5.5 × 5.5	3.64	1.2	42.6/30.5	32.8	—	459	C‐H···π, electrostatics	[167]
NKMOF‐1‐Ni, 3D	5.8 × 5.8	2.7	2.3	60.3/40.9	30	—	382	H‐bonding, π···π interactions	[168]
[Co 3(OH)2(HCOO)2(CPT)2], 3D	6.64 × 7.44	5.07	1.49	24.8/20.1	28	—	778	H‐bonding, electrostatics	[169]
SIFSIX‐21‐Ni, 3D	3.2 × 3.6	4.05	1.27	40.6/16.1	27.7	N2 flow at 333 K	871	Electrostatics, H‐bonding	[81]
CPL‐1, 2D	4.0 × 6.0	1.9	0.07	42.5/31.9	26 c	—	571	H‐bonding	[60]
ZJU‐196, 3D	5.1 × 5.1	3.7	0.4	39.2/–	25 d	—	—	Flexibility, pore size	[170]
MOF‐OH, 3D	4	3.04	1.20	17.5/–	25	—	150	Van der Waals interactions	[171]
FeNi‐M′MOF, 3D	4.15 × 4.27; 3.94 × 4.58	4.29	2.72	27/24.5	24	—	383	UMCs, π···π stacking interactions	[172]
SNNU‐98‐Mn, 3D	5.2 b	19.88	12.09	60.2/29.2	22.7	—	720.9	Electrostatics	[173]
[Ni3(HCOO)6]n , 3D	4.3 × 4.3	2.4	1.6	40.9/24.5	22	Vacuum at 403 K	289	Pore size, H‐bonding	[174]
Ni(4‐DPDS)2WO4 , 3D	5.0 × 5.0	2.05	1.43	77.2/40.5	20.9	He flow at 323 K	188	H‐bonding, C‐H···π interactions	[163]
DICRO‐4‐Ni‐i, 3D	6.2 × 6.6	1.9	1.0	37.7/33.9	18.2 d	—	398	H‐bonding	[175]
SNNU‐98‐Zn, 3D	4.2 b	14.64	7.90	41.3/35.9	17.4	—	398.2	Electrostatics	[173]
TCuCl, 3D	3.69 × 3.69	3.0	2.0	41/30.1	16.9	N2 flow at 333 K	167	Halogen···HC interactions	[176]
Ni(4‐DPDS)2MoO4 , 3D	5.2 × 5.2	2.77	1.99	47.8/39.8	14.5	He flow at 323 K	225	H‐bonding, C‐H···π interactions	[163]
SNNU‐98‐Co, 3D	4.8 b	13.79	8.05	53.5/34.7	14.4	—	328.3	Electrostatics	[173]
JLU‐MOF103, 3D	3.4 × 2.6, 4.2 × 2.6	4.3	1.9	31.7/25.8	14.0	—	697	H‐bonding	[177]
Pacs‐CoMOF‐2a	5.8 b × 6.6 b	5.40	2.81	34.2/–	13	—	196	UMCs	[178]
BSF‐3‐Co, 3D	6.79, 6.22, 6.33	3.85	2.41	42.7/22.4	12.7	He flow at 298 K	437	Electrostatics, H‐bonding	[179]
MIL‐100(Fe), 3D	5.5 × 8.6	5.3	2.5	65/–	12.5 d	—	2300	UMCs	[180]
ZJU‐40a, 3D	10.2, 9.6 × 22.3	9.64	3.34	34.5/–	11.5	—	2858	H‐bonding, pore size	[181]
Co‐MOF, 3D	—	6.47	2.68	33/27	11	—	973	H‐bonding	[182]
CAU‐10‐NH2 , 3D	3.8, 5.3	3.57	2.08	31.3/24.5	10.8	Vacuum at 333 K	403	Van der Waals interactions	[183]
GEFSIX‐21‐Cu, 3D	3.14 × 3.14	5.0	1.8	32.3/22.2	10.8	333 K	1014	H‐bonding	[184]
ZNU‐9, 3D	10.3 b	7.94	4.32	33.1/26.6	10.3	Ar flow at 323 K	1635	H‐bonding, van der Waals interactions	[185]
TIFSIX‐2‐Cu‐i, 3D	5.1 × 5.1	4.1	4.3	46/35.8	10 d	—	685	H‐bonding, electrostatics	[36]
MIL‐160, 3D	4.6 × 9.8	8.5	4.02	31.8/–	10	—	1138	H‐bonding	[186]
JCM‐1, 3D	12.5 × 3.9	3.3	1.7	36.9/33.4	10	373 K	550	Pore size, H‐bonding, dipolar interactions	[187]
ZJUT‐2a, 3D	3.2 × 3.2	3.4	2.2	41.5/35.5	10	—	350	H‐bonding, van der Waals interactions	[188]
TCuBr, 3D	3.59 × 3.59	2.8	2.0	36.6/30.7	9.5	N2 flow at 333 K	173	Halogen···HC interactions	[176]
TCuI, 3D	3.66 × 3.66	2.2	1.6	38.4/26.8	5.3	N2 flow at 333 K	250	Halogen···HC interactions	[176]
UTSA‐74a, 3D	8.0 × 8.0	4.8	3.2	31/25	9	—	830	UMCs	[65]
CTGU‐39‐a, 3D	5.0 × 5.2	2.92	1.99	26.8/19.2	8.4	He flow at 298 K	586.1	H‐bonding, π···π interactions	[189]
JUN‐4, 3D	4.8, 9.3 b	9.82	7.1	26.8/19.7	8.2	He flow at 298 K	1533	UMCs	[190]
SNNU‐150‐Al, 3D	8.5 b	4.33	1.98	29/24.8	7.27	—	—	UMCs	[191]
FJU‐22a, 3D	7.1 × 7.1	5.1	5.0	23/18.8	7.1 e	—	828	H‐bonding, van der Waals interactions	[192]
ZJU‐60a, 3D	4.4 × 5.4	6.7	3.3	17.6/15.2	6.7 e	—	1627	UMCs, pore size	[193]
NTU‐55, 3D	10.4 b	6.05	3.13	25.3/22	6.6 e	He flow at 333 K	2300	UMCs	[194]
UTSA‐83a, 2D	3.5 × 6.6	0.53	0.17	24.4/16.6	6.2	—	70 f	H‐bonding	[195]
MUF‐17, 3D	4.7 × 4.8	2.7	2.2	49.5/33.8	6	Vacuum at 333 K	247	H‐bonding, repulsive electrostatics	[196]
CPM‐107op, 3D	—	4.35	1.55	37/24	5.7	Vacuum	319	Flexibility	[197]
ZJUN‐13, 3D	6.8 b , 11.8 b	5.28	3.92	33.5/22.5	5.64	He flow at 333 K	1352	H‐bonding	[198]
PCP‐33, 3D	11 × 20	5.4	2.6	27.5/2.62	5.6 d	—	1248	UMCs	[199]
NUM‐20, 3D	8	1.74	0.79	38.83/37.14	5.4	—	319.17	Electrostatics, C‐H···π interactions	[200]
FJU‐6‐TATB, 3D	9.2‐17.1	4.9	4.5	29/26	5.3‐3.1	Vacuum at 353 K	1306	H‐bonding	[201]
HKUST‐1‐UF, 3D	—	7.82	4.84	27.7/23.5	5.1	He flow at 423 K	1877	H‐bonding, UMCs	[202]
UPC‐110, 3D	6 b	3.27	1.08	24.6/14	5.1	—	250	UMCs	[203]
JXNU‐5, 3D	4.6 b , 6.7 b	2.5	1.55	32.9/25.2	5	—	406	H‐bonding, electrostatics	[204]
HKUST‐1‐UF‐GLY0.3 , 3D	—	8.86	5.45	30.4/19.1	4.8	He flow at 423 K	2033	H‐bonding, UMCs	[202]
Ag NP@Fe2O3@ Zn‐MOF‐74, 3D	7‐10 b	6.7	5.13	—	4.73	—	936	π‐complexation, plasmon‐driven photothermal effect	[205]
CAU‐10‐CH3 , 3D	5.8 × 5.8	3.5	1.41	25.18/23.4	4.6	298 K	312	H‐bonding, C‐H···π interactions	[206]
SNNU‐45, 3D	4.5	5.98	4.33	40/27	4.5	He flow at 298 K	1006	UMCs	[207]
JLU‐MOF102, 3D	5.5 × 4.4, 6.8 × 3.3	5.6	3.3	23.3/20.6	4.5	—	1236	H‐bonding	[177]
UTSA‐220, 3D	4.5‐5.5, 3.1‐4.8	3.40	3.38	29/27	4.4	He flow at 298 K	577	H‐bonding	[208]
FJU‐89a, 3D	12 × 8	4.53	2.73	31/27.8	4.3	—	774	—	[209]
FJU‐90a, 3D	5.4 × 5.1	8.0	4.6	25.1/20.7	4.3	Vacuum	1572	H‐bonding, pore size	[210]
Cu2(ade)2(PA)2 , 3D	2 × 6	2.19	1.5	26.8/23.6	4.2	—	401	UMCs	[211]
FJU‐114, 3D	9.2	3.44	1.74	33/29.1	4.2	—	891	π···π interactions	[212]
TFT‐COF, 2D	7.5 × 7.5	1.76	1.39	76.8/25.5	4.18	—	689	H‐bonding	[213]
UPC‐120(Th), 3D	8.6 × 8.6, 8 × 8	1.89	1.54	24.13/23.29	4.11	He flow at 298 K	450	H‐bonding	[214]
ZJU‐199a, 3D	5‐7.5 b	5.71	2.78	38.5/29.0	4	—	987	Pore size, UMCs	[215]
Hex‐Zn‐MOF 1a, 3D	8.6 b , 9.8 b	3.18	2.21	39/27	4	—	770.3	H‐bonding, electrostatics	[216]
CAU‐10‐H, 3D	4.7 × 4.7	4.01	2.68	32.8/21.4	4	He flow at 298 K	627	C−H···π, dipole–dipole interactions	[217]
CAU‐10‐H, 3D	6.8 × 7.5	4.01	2.62	28/25	4.0	298 K	627	H‐bonding, C‐H···π interactions	[206]
mot‐Cu(Br‐BDC), 3D	4.2 × 4.7, 12 × 24.1	1.53	1.08	26.1/25.6	3.9	He flow at 298 K	303	UMCs, pore size	[218]
NBU‐3‐Mn/Fe, 3D	—	3.03	1.61	29/41	3.9	—	551	UMCs	[219]
UPC‐120(Zr), 3D	8.6 × 8.6, 8 × 8	2.14	1.53	25.8/14.03	3.71	He flow at 298 K	520.7	H‐bonding	[214]
Cu‐CPAH, 3D	6‐9 b	5.88	3.93	35.4/31.5	3.6	—	880	H‐bonding, UMCs	[220]
FJI‐H36, 3D	12.9 × 12.9, 8.4 × 10.2	6.46	4.2	36.1/29.2	3.5	—	899	H‐bonding, UMCs	[221]
UTSA‐68a, 3D	6.5 × 6.5; 7.5 × 9.5	3.13	1.77	25.8/–	3.4	—	1954	UMCs	[222]
CTGU‐40‐a, 3D	5 × 5	3.31	2.21	27.8/23.1	3.3	He flow at 298 K	758.4	H‐bonding, π···π interactions	[189]
TpPa‐F, 2D	11.9	5.22	2.52	38.1/28.2	3.3	—	1048	H‐bonding	[223]
SNNU‐98‐Ni, 3D	4.8 b	13.17	9.08	39.3/38.7	3.3	—	473.5	Electrostatics	[173]
UPC‐200(Al)‐F‐BIM, 3D	7 × 11	6.2	2.5	20.5/13.5	3.15	—	2212.8	Pore size	[224]
JNU‐1, 3D	16.3 × 6.6	2.7	2.2	13/–	3	—	818	UMCs, dipole interactions	[160]
QDU‐MOF‐1, 3D	7.76	6.3	—	72.52/39.89	2.95	—	1465	H‐bonding	[225]
LIFM‐210, 3D	5‐7.25 b	2.96	2.04	28.3/24.6	2.81	—	607	Cation···π interactions	[226]
JXNU‐18, 3D	5.8‐8.5 b	2.46	1.41	37.4/22.4	2.81	—	622	H‐bonding, C‐H···π interactions	[227]
Cu‐tztp MOF 1a, 3D	5.4‐8.6 b	5.02	3.35	38.3/26.2	2.7	—	798.9	H‐bonding, C‐H···π interactions	[228]
TpPa‐H, 2D	12.4	4.87	3.3	39. 9/–	2.7	—	1270	H‐bonding	[223]
NKM‐123, 3D	3.8 × 3.8	1.24	0.69	29.3/25.2	2.6	—	217.39	C‐H···π interactions, H‐bonding	[229]
FJU‐129, 3D	6.6, 10 b	3.14	1.33	24.78/24.76	2.5	—	1726	Electrostatics, C‐H···π interactions	[230]
Zn‐MOF‐74, 3D	11 × 11	5.5	5.4	22.1/25	2	—	1360	UMCs	[65]
ZJU‐30a, 3D	4.0 × 4.0; 5.6 × 5.6	2.31	1.87	31.3/–	1.7	—	228	UMC, pore size	[222]
Cu2(pzdc)2(pyz), 3D	4 × 6	2.0	—	42.5/31.9	—	—	—	H‐bonding	[60]

^a IAST selectivity at 1 bar for 1:1 (v/v) C₂H₂/CO₂.

^b Determined from Horvath–Kawazoe method applied on N₂ isotherm at 77 K.

^c Uptake ratio at 0.01 bar for 270 K measurements.

^d C₂H₂/CO₂ uptake ratio at 0.5 bar.

^e IAST selectivity at 0.15 bar for 1:1 (v/v) C₂H₂/CO₂.

^f Determined from CO₂ isotherm at 195 K. S_BET = Brunauer–Emmett–Teller (BET) theory‐based surface areas from N₂ isotherm recorded at 77 K, unless otherwise mentioned.

^g Pore size calculated by Non‐Linear Density Functional Theory (NLDFT) method based on 87 K Ar isotherms.

TABLE 6.

C₂H₂/C₂H₄ binary separations. The following parameters are listed for comparison: pore size; BET surface area (S_BET); single‐component gas uptakes; adsorption enthalpies (Q _st); adsorption selectivity; regeneration temperature; and attributed mechanisms. Reticular sorbents are listed in decreasing order of binary selectivity for C₂H₂ selective sorbents (above) and C₂H₄ selective sorbents (below).

Adsorbent, network dimensionality (nD)	Pore size (Å)	C2H2 uptake at 1 bar (mmol/g)	C2H4 uptake at 1 bar (mmol/g)	Q st (kJ/mol)	Selectivity a	Regeneration temperature	SBET (m2/g)	Mechanism	Refs.
UTSA‐300a e , 2D	2.4 × 3.3	3.1	0.04	57.6/–	∼104 f	298 K	311	π···π interactions, H‐bonding	[159]
NCU‐100a e (UTSA‐300‐Cu), 2D	4.3 × 3.6	4.57	0.32	60.5/–	7291.3 f	He flow at 298 K	358	Flexibility, H‐bonding, Pore size	[231]
bnn‐1‐Ca‐H2O e , 3D	3.4 × 3.4	2.2	0.16	–/–	6966.4 f	Vacuum at 298 K	210	Molecular sieving	[232]
SIFSIX‐14‐Cu‐i e , 3D	3.4 × 3.4	1.8	0.6	40/37	6320 f	—	612	Molecular sieving, H‐bonding	[39]
UTSA‐200a, 3D	3.4 × 3.4	3.65	0.63	—	6320	—	612	Molecular sieving, H‐bonding	[39]
ZJU‐300a, 3D	4.1 d	5.37	2.39	61.1/39.7	1672	—	796	H‐bonding, supramolecular interactions	[233]
NKMOF‐1‐Ni, 3D	5.8 × 5.8	2.72	2.14	60.3/44.9	1272.6 c , 30 b	—	382	H‐bonding, π···π interactions	[168]
GeFSIX‐14‐Cu‐i e (ZU‐33), 3D	3.0 × 3.0	4.1	0.76	43.6/–	1100 f	He flow at 298 K	424	Size sieving, van der Waals interactions, H‐bonding	[234]
FJI‐W88, 3D	4.3 × 2.9	4.36	2.31	31.5/29.3	698	Ar flow at 353 K	487	Van der Waals interactions	[235]
TIFSIX‐14‐Cu‐i, 3D	3.6 × 3.6	3.78	1.41	54/40	229	—	425	Pore size, C‐H···π interactions	[236]
TIFSIX‐2‐Cu‐i, 3D	5.1 × 5.1	3.9	2.1	46/–	55 b , 212.2 b	—	685	H‐bonding	[36]
GeFSIX‐2‐Cu‐i, 3D	4.5 × 4.5	3.9	2.2	42.6/32	67	He flow at 298 K	424	Pore size, H‐bonding, Van der Waals interactions	[234]
sql‐SIFSIX‐bpe‐Zn, 2D	—	1.79	1.25	67.5/38.4	53.1	Vacuum at 313 K	—	H‐bonding, σ···π and π···π interactions	[237]
SIFSIX‐2‐Cu‐i, 3D	5.2 × 5.2	4.02	2.19	41.9/30.7	44.54, 41.01 b	—	735	Pore size, H‐bonding	[35]
ZNU‐6, 3D (278 K)	8.5 × 8.5, 4 × 4	8.06	6.5	37.2/29.0	43.8 c	Ar flow at 393 K	1330	Electrostatics	[238]
Ni‐gallate, 3D	3.5 × 4.9	3.59	2.1	46/–	43.7	He flow at 393 K	424	H‐bonding, pore size	[239]
NbOFFIVE‐2‐Ni‐i (ZU‐62‐Ni), 3D	3.0 × 3.9	3.0	0.8	43/–	37.2	298 K	404	Pore size	[240]
M′MOF‐3a, 3D	3.4 × 4.8	1.9	0.4	27.1/27.3	24.03, 34.17 b	—	110	Pore size, π···π interactions	[241]
CPL‐1, 3D	4.0 × 6.0	2.07	0.31	40.2/36.3	26.75	He flow at 393 K	414	Pore size, H‐bonding	[242]
Mg‐gallate, 3D	3.6 × 4.8	4.39	3.03	33/–	20.9	He flow at 393 K	559	Pore size, H‐bonding, Van der Waals interactions	[239]
UTSA‐100a, 3D	4.3 × 4.3	4.27	1.66	22/–	10.72, 19.55 b	353 K	970	Pore size, Van der Waals interactions	[243]
GeFSIX‐dps‐Cu e , 2D	1.8 × 2.6; 2.5 × 4.4	4.28	0.16	–/–	19 g	—	382	Size sieving, H‐bonding	[244]
UTSA‐60a, 3D	4.8 × 4.0	3.12	2.05	36/–	16	—	484	UMC, pore size	[245]
Co‐gallate	3.7 × 5.0	3.88	3.37	47/–	15	He flow at 393 K	475	Pore size, H‐bonding, Van der Waals interactions	[239]
ELM‐12, 2D	4.3 × 3.9	2.56	1.0	25.4/–	14.8	He flow at 373 K	706	Flexibility, H‐bonding	[246]
APPT‐Cd‐ClO4 −, 3D	11 × 11	1.75	0.44	28.5/3.7	14.71	298 K	205	H‐bonding	[247]
SIFSIX‐1‐Cu, 3D	8.0 × 8.0	8.5	4.11	30,37/23.5	8.82, 13.72	—	250	Pore size, H‐bonding, Van der Waals interactions	[35]
SIFSIX‐3‐Zn, 3D	4.2 × 4.2	3.64	2.24	21, 31/28.8	8.82, 13.72 b	He flow at 298 K	250	Pore size, H‐bonding	[35]
MECS‐5, 3D	4.5 × 5.3, 3.4 × 4.5, 1.8 × 3.9	3.85	1.14	26.1/26	12.6	—	964	Molecular sieving	[248]
CPL‐2, 3D	9.0 × 6.0	3.13	1.86	30.8/20.3	12	He flow at 393 K	495	Pore size, H‐bonding	[242]
Pacs‐CoMOF‐2a, 3D	5.8 × 6.6	5.4	2.81	34.2/–	11.5	—	196	UMCs	[178]
NTU‐92, 3D	3.0 × 3.2	1.47	0.58	35.5/24.5	10	Vacuum at 353 K	347	H‐bonding, C‐H···π interactions	[249]
UTSA‐220, 3D	4.5 × 4.1; 2.1 × 5.0	3.4	2.53	29/24	10, 8.8 b	He flow at 298 K	577	Pore size, H‐bonding	[208]
Na@COF‐ECUT‐1, 2D	8	4.00	2.20	19.21	9.41	—	149	UMCs	[250]
MUF‐17, 3D	3.1 × 3.5; 4.7 × 4.8	3.02	2.16	49.5/31.1	8.73	He flow at 403 K	247	Pore size, electrostatics	[196]
JCM‐1, 3D	3.9 × 12.5	3.34	1.56	36.9/34.2	8.1	He flow at 373 K	550	H‐bonding, pore size	[187]
NTU‐91, 3D	3.7 × 10.9	1.32	0.63	32.2/22.2	8	Vacuum at 353 K	287	—	[249]
Sr‐TCPE, 3D	5.2 × 4.3; 5.9 × 5.2	1.52	0.9	29/–	8	—	—	Pore size	[251]
ZJU‐198a, 3D	3.6 × 4.1; 2.1 × 5.0	3.25	2.95	26.1/37.4	7.2	He flow at 298 K	343.1	Pore size	[252]
NTU‐90, 3D	4 × 11.4	1.25	0.71	32/22	6.6	Vacuum at 353 K	321	—	[249]
COF‐ECUT‐1, 2D	12	2.47	1.28	7.68	6.33 (1:99)	—	306	H‐bonding	[250]
SIFSIX‐2‐Cu, 3D	10.5 × 10.5	5.38	2.02	26.3/20.8	6, 4.95	—	3140	H‐bonding	[35]
UTSA‐67a, 3D	3.3 × 3.3	5.13	2.81	32/	6	—	1136.7	Pore size	[253]
CPL‐5, 3D	11.0 × 6.0	3.01	1.84	31.3/19.1	6	He flow at 393 K	523	Pore size, H‐bonding	[242]
SIFSIX‐3‐Ni, 3D	4.2 × 4.2	3.3	1.75	30.5/30.3	5.03, 5.98	—	223	Pore size	[35]
NBU‐1, 3D	3.8 d	3.64	2.07	38.3/37.9	5.9 b	Vacuum, 323 K	368	UMCs	[254]
Ni‐DCPTP, 3D	6.7 d , 10 d	6.54	4.48	38.9/–	5.5	He flow at 298 K	857	H‐bonding	[255]
Al‐PyDC, 3D	5.8 d	8.24	3.44	35.3/27.8	4.3	N2 flow at 373 K	1134	H‐bonding, C‐H···π interactions	[256]
PAF‐120, 2D	24	2.27	1.22	37.5/343.3	4.1	—	801	Electrostatics	[257]
PAF‐110, 2D	17, 25	2.23	1.29	38.4/22.6	3.9	—	910	H‐bonding	[258]
HUST‐6, 3D	—	3.49	2.38	31.1/30.2	3.8	—	645.3	UMCs	[259]
CTF‐PO71, 2D	7, 14	3.3	2.3	28‐26/22‐19	2.8‐1.8	—	1401	Electrostatics	[260]
Fe‐BDC‐TPT‐BF4 , 3D		9.4	5.92	—	2.7	—	1307.8	Electrostatics	[261]
Mg‐MOF‐74, 3D	11 × 11	8.37	7.45	41/–	2.18	—	927	UMCs	[262]
NOTT‐300, 3D	6.5 × 6.5	6.34	4.28	32/16	2.17, 2.3 b	—	1370	H‐bonding, π···π, dipole interactions	[263]
Fe‐MOF‐74, 3D	11 × 11	6.8	6.1	46/–	2.08, 2.1 b	—	1350	UMCs	[66]
Co‐MOF‐74, 3D	11 × 11	8.17	7.02	45/	1.7	—	1018	UMCs	[262]
RPM3‐Zn, 3D	—	2.14	0.89	—	—	—	328	Flexibility	[264]
BUT‐11, 3D	11 d , 12.2 d	7.14	3.44	20/–	—	298	1233	Pore size, H‐bonding	[265]

^a IAST selectivity at 1 bar for 1:99 (v/v) C₂H₂/C₂H₄.

^b IAST selectivity at 1 bar for 1:1 (v/v) C₂H₂/C₂H₄.

^c IAST selectivity at lowest C₂H₂ loading for 1:99 (v/v) C₂H₂/C₂H₄.

^d Determined from Horvath–Kawazoe method or non‐local density functional theory applied on N₂ isotherm at 77 K.

^e IAST selectivities are qualitative, because of molecular sieving.

^f Not applicable because of virial fits not conforming to stepped isotherms obtained at 298 K and 273 K.

^g Uptake ratio at C₂H₂/C₂H₄ (0.1/0.9).

TABLE 7.

CO₂/C₂H₂ binary separations. The following parameters are listed for comparison: pore size; BET surface area (S_BET); single‐component gas uptakes; adsorption enthalpies (Q _st); adsorption selectivity; regeneration temperature; and attributed mechanisms. Reticular sorbents are listed in decreasing order of binary selectivity for CO₂ selective sorbents (above) and C₂H₂ selective sorbents (below).

Adsorbent, network dimensionality (nD)	Pore size (Å)	CO2 uptake at 1 bar (mmol/g)	C2H2 uptake at 1 bar (mmol/g)	Q st (kJ/mol)	S a	Regeneration temperature	SBET (m2/g)	Mechanism	Refs.
ALF, 3D	4.1 × 5.3	3.85	0.15	51/39	6.5 × 105	—	—	H‐bonding	[266]
Cu‐F‐pymo, 3D	3.4 × 3.4, 3.2 × 3.2	1.19	0.10	28.8/–	>105	He flow at 298 K	93	Molecular sieving	[267]
MUF‐4, 3D	2.2 × 2.2	3.17	—	24.4/–	3360 e	He flow at 298 K	1094	Electrostatics, pore size	[268]
Zn‐ox‐mtz, 3D	5.3 × 3.5, 5.2 × 5.7	3.07	0.25	43.02/–	1064.9	—	466	H‐bonding	[269]
PMOF‐1, 3D	5.5 × 6.8	2.38	0.33	—	694	—	—	H‐bonding	[270]
MUF‐16, 3D	3.6 × 7.6	2.13	0.18	32.3/25.8	510	—	214	H‐bonding	[129]
ZnAtzCO3 , 3D	2.9 × 5.1, 3.5 × 5.1	0.66	2.80	32.6/22.4	151	—	455.6	Electrostatics	[271]
Cd‐NP, 3D	3.2 × 3.2, 6.1 × 4.5 × 4.5	2.59	0.43	27.7/–	85	—	305 c	Electrostatics	[272]
MUF‐16 (Ni), 3D	3.6 × 7.6	2.13	0.34	37.3/–	46	Vacuum	204	H‐bonding	[129]
Ce(IV)‐MIL‐140‐4F, 3D	3.4‐4.3	4.92	1.85	39.5/27.4	>40	He flow at 298 K	360	F···C = O	[273]
MUF‐16 (Mn), 3D	3.6 × 7.6	2.25	0.43	36.6/–	31	Vacuum	205	H‐bonding	[129]
JNU‐5‐Me, 3D	2.5 × 2.5	4.7	0.25	37.4/1.7	17.7	Vacuum at 298 K	646	H‐bonding, van der Waals interactions	[274]
Tm(OH‐bdc), 3D	6.3 × 9.3; 6.3 × 10.6	5.8	2.0	45.2/17.8	17.5 b	—	923	H‐bonding	[275]
CD‐MOF‐2, 3D	4.2 × 4.2; 7.8 × 7.8; 17 × 17	2.7	2.0	67.2/25.8	16.6 b	He flow at 298 K	922	H‐bonding	[276]
Mg‐CUK‐1 (233 K), 3D	8.1 × 10.6	6.43	3.97	—	12.1	He flow at 298 K	500‐600	H‐bonding, C‐H···π interactions	[277]
Co‐CUK‐1 (233 K), 3D	8.1 × 10.6	7.59	5.31	43.5/33.3	9.5	He flow at 298 K	500‐600	H‐bonding, C‐H···π interactions	[277]
Mn(bdc)(dpe), 3D	3.3 × 3.5	2.1	0.3	29/27.6	8.8	—	535 c	Flexibility, C‐H···π interactions	[278]
SIFSIX‐3‐Ni, 3D	4.2 × 4.2	2.7	3.3	50.9/36.7	7.7	—	368	H‐bonding, electrostatics	[36]
CD‐MOF‐1, 3D	4.2 × 4.2; 7.8 × 7.8; 17 × 17	2.9	2.2	41.0/17.6	6.6 b	He flow at 298 K	1094	H‐bonding	[276]
PCP‐NH2‐ipa, 3D	4.4 × 4.8	3.21	1.94	36.6/26.8	6.4	Vacuum	—	—	[279]
Bi‐BTC, 3D	4 b	1.76	2.4	32.19/35.37	5.14	—	92.51	—	[280]
PCP‐NH2‐bdc, 3D	4.4 × 5.3	3.03	1.91	34.57/25.6	4.4	Vacuum	—	Electrostatics	[279]
Y‐ bptc, 3D	9.46 × 9.46, 4.2 × 4.2	2.48	1.17	31.5/–	4.1	—	—	Pore size, H‐bonding, kinetic	[281]
FJUT‐1, 3D	5.8, 6.9 b	5.95	4.84	31.88/25.97	4.06	—	1240	H‐bonding, C‐H···π interactions	[282]
Zn(atz)(BDC‐Cl4)0.5 , 3D	4.7 × 6.8	1.54	0.8	32.7/25.4	2.4	—	—	Electrostatics	[283]
Co(HL), 3D	6.2 b	10.69	6.25	—	1.7	—	773	Flexibility	[284]
Eu‐MOF, 3D	6.2 × 7.4	1.04	0.93	31.4/27.3	1.1	—	143	—	[285]
NbOFFIVE‐1‐Ni, 3D	3.0 × 3.9	2.2	2.4	54.6/34	0.9	—	280	—	[286]
AlFFIVE‐1‐Ni, 3D	3.2 × 3.2	2.75	4.6	47/38	0.6	—	258	—	[286]

^a IAST selectivity at 1 bar for CO₂/C₂H₂ (1:1) mixture.

^b IAST selectivity at 1 bar for CO₂/C₂H₂ (1:2) mixture.

^c Surface area calculated from CO₂ at 195 K or 273 K isotherm.

^d Kinetic selectivity at ambient temperature.

^e Kinetic selectivity is calculated from the ratio of the diffusion time constants for two gases.

TABLE 8.

C₂H₄/C₂H₆ binary separations. The following parameters are listed for comparison: pore size; BET surface area (S_BET); single‐component gas uptakes; adsorption enthalpies (Q _st); adsorption selectivity; regeneration temperature; and attributed mechanisms. Reticular sorbents are listed in decreasing order of binary selectivity for C₂H₄ selective sorbents (above) and C₂H₆ selective sorbents (below).

Adsorbent, network dimensionality (nD)	Pore size (Å)	C2H4 uptake at 1 bar (mmol/g)	C2H6 uptake at 1 bar (mmol/g)	Q st (kJ/mol)	Selectivity a	Regeneration temperature	SBET (m2/g)	Mechanism	Refs.
UTSA‐280, 3D	3.2 × 4.5; 3.8 × 3.8	2.5	0.098	34.1/–	104 b	He flow at 353 K	331	Molecular sieving, H‐bonding, π···π stacking, van der Waals interactions	[287]
NUS‐6(Hf)‐Ag, 3D	10, 17	2.02	1.35	56.5/29.8	106.3, 6 c	He flow at 298 K	1027	Cation···π	[288]
ITQ‐55, 3D	2.33 × 5.71, 3.08 × 5.71	1.28	0.76	—	100	—	—	Flexibility, pore shape	[289]
CuI@UiO‐66‐(COOH)2 , 3D	4.1 d	1.86	0.85	48.5/26.7	80.8	He flow at 413 K	320	Molecular sieving, π‐complexation	[290]
Co‐gallate, 3D	3.69 × 4.95	3.37	0.31	44/–	52	—	475	Molecular sieving	[291]
Mg‐gallate, 3D	3.56 × 4.84	3.03	0.26	39/–	37.3	—	559	Molecular sieving	[291]
Ni‐gallate, 3D	3.47 × 4.85	1.97	0.28	32/–	16.8	—	424	Molecular sieving	[291]
NOTT‐300, 3D	6.5 × 6.5	4.28	0.85	16/11	47.8	He flow at 373 K	1370	H‐bonding, π···π stacking, intermolecular dipole interactions	[263]
ZnAtzPO4@CMC, 3D (273 K)	5 × 5	0.64	1.01	—	34.7	—	—	Electrostatics	[292]
PAF‐1‐SO3Ag, 3D	8.0	4.06	2.23	106/–	27	—	783	π‐complexations	[293]
10 wt% Ag/CPL‐2, 3D	7‐11 d	0.9	0.15	—	26.1	He flow at 298 K	12	Electrostatics	[294]
Fe2(m‐dobdc), 3D	12	7.0	6.0	55/–	25	He flow at 298 K	1295	UMCs	[295]
Fe2(dobdc), 3D (318K)	11	6.02	—	45/25	13‐18	—	1350	UMCs	[66]
(Cr)‐MIL‐101‐SO3Ag, 3D e	–, 15–18 d	3.26, 4.32	1.47, 1.22	63/16, 120/–	9.7, 16	—	1374, 1253	UMCs, π‐complexation	[296, 297]
NaETS‐10, 3D	8.0	1.7	1.3	—	14	—	289	UMCs	[298]
Fe‐MOF‐74, 3D	11	6.28	5.10	47.5/25	13.6	—	1350	UMCs	[66]
Zeolite 13X, 3D	2.3	2.8	2.2	—	13.4	873 K	406	π‐cation interactions	[299]
ZnAtzPO4 , 3D	3.82 × 4.94	1.92	1.04	29.98/	12.4	He flow at 298 K	470	Kinetic	[300]
1.6AgM‐DS, 3D	—	3.37	0.94	59.2/–	9.5	378 K	846	π‐complexation	[301]
MIL‐101‐SO3Ag@CMC, 3D	—	2.2	1.0	49.5/18.0	7.0	N2 flow at 373 K	1320	UMCs	[302]
Cu(OPTz), 3D	3 × 3	2.3	0.4	27.7/–	6	—	—	Flexibility	[303]
GT‐18, 3D	3 × 7	0.6	0.2	—	6.8	—	—	Pore shape, diffusivities	[304]
Co‐MOF‐74, 3D	11	6.21	5.25	43.6/–	5.82	—	1341	UMCs	[305]
Mg‐MOF‐74, 3D	11	7.4	6.4	42/	5.6	—	927	UMCs	[262]
Ca(squarate), 3D	6.9 × 6.9	2.3	1.3	33.9/28.4	4.9	He flow at 353 K	224	π···π interactions, H‐bonding	[306]
Zeolite 5A, 3D	5.0	2.45	1.72	37/33	4.5	He flow at 553 K	457‐600	Cation···π interactions	[307]
NUS‐36, 3D	—	1.5	1.0	44/44	4.1	He flow at 298 K	79.1	Pore size, van der Waals interactions	[308]
HKUST‐1, 3D	10, 14	7.20	6.03	39/31	3.6	—	1500‐2100	UMCs	[262]
Co(VTTF), 2D (283K)	14.6 × 9.8	4.9	0.98	—	2.03	—	—	Flexibility	[309]
CuI‐MFU‐4 L, 3D	—	3.79	2.78	79.9/–	—	He flow at 413 K	2961	UMCs	[310]
ZNU‐10, 3D	8.59	2.13	1.57	26.4/23.2	1.68	Vacuum	1171	H‐bonding, van der Waals interactions	[311]
UiO‐66‐ADC, 3D	4.4	1.7	1.6	36/36	0.55	—	556	—	[308]
ZU‐901, 3D	3.4 × 4.2	1.55	0.26	39.28/24.2	—	Vacuum	229.86 f	H‐bonding	[312]

^a IAST selectivity at 1 bar for 1:1 (v/v) C₂H₄/C₂H₆.

^b IAST selectivities are qualitative, because of molecular sieving.

^c IAST selectivity at 0.01 bar for 1:1 (v/v) C₂H₄/C₂H₆.

^d Determined from Horvath–Kawazoe method applied on the N₂ isotherm at 77 K.

^e Two consecutive reports on this sorbent document distinct values that are included using a comma between them. S _BET = Brunauer–Emmett–Teller (BET) theory‐based surface areas from N₂ isotherm recorded at 77 K, unless otherwise mentioned.

^f Langmuir surface areas.

TABLE 9.

C₂H₆/C₂H₄ binary separations. The following parameters are listed for comparison: pore size; BET surface area (S_BET); single‐component gas uptakes; adsorption enthalpies (Q _st); adsorption selectivity; regeneration temperature; and attributed mechanisms. Reticular sorbents are listed in decreasing order of binary selectivity for C₂H₆ selective sorbents (above) and C₂H₄ selective sorbents (below).

Adsorbent, network dimensionality (nD)	Pore size (Å)	C2H6 uptake at 1 bar (mmol/g)	C2H4 uptake at 1 bar (mmol/g)	Q st (kJ/mol)	S (binary) b	Regeneration temperature	SBET (m2/g)	Mechanism	Refs.
NIIC‐20‐Bu, 3D	3.5, 11.6	2.5	1.4	24.1/22.9	15.4	—	1440	H‐bonding, C‐H···π interactions	[313]
MAF‐49, 3D	3.3 × 3.0	1.7	1.7	60/48	9	—	—	H‐bonding, electrostatics	[314]
NIIC‐20‐GI, 3D	3.5, 12	2.1	1.7	31.5/30.3	8.7	—	1600	—	[313]
NIIC‐20‐Pe, 3D	3.6, 11.5	2.2	1.6	25.4/24.3	8.4	—	1286	—	[313]
Fe2(O2)dobdc, 3D	14 × 14	3.32	2.54	66.8/36.5	4.4	298 K	1073	H‐bonding, Van der Waals interactions	[315]
NIIC‐20‐Pr, 3D	3.6, 11.8	2.4	1.9	34.2/32.4	4.0	—	2260	—	[313]
X‐dia‐1‐Ni0.89Co0.11 , 3D (273 K)	—	4.96	0.54	—	5.47	—	—	Flexibility	[316]
UTSA‐30, 3D	3.2 × 3.2	2.1	2.1	30/30	3.8	—	592	Pore size, shape	[111]
X‐dia‐1‐Ni, 3D (273 K)	—	5.54	1.58	23.4/20.5	3.51	—	—	Flexibility	[316]
NIIC‐20‐Et, 3D	6 × 3.5, 3.5 × 3.5	2.4	1.8	29.8/29.7	3.5	—	2700	—	[313]
Cu(Qc)2 , 2D	3.3 × 3.3	1.85	0.78	30/25.4	3.4	—	240	Flexibility, C‐H···π interactions	[317]
CuIn(3‐ain)4 , 3D	6.7	2.71	2.68	32.11/32.06	3.31	—	472	H‐bonding, C‐H···π interactions	[318]
ZIF‐62, 3D	5.1	1.65	—	—	3.3	—	994	—	[319]
F‐MOF‐2, 3D	6.5	1.2 c	1.0 c	37/–	3.3 c	—	390	H‐bonding	[320]
NPU‐3, 3D	12.2 a	3.36	2.22	18.71/17.79	3.21	He flow at 313 K	1834	H‐bonding, C‐H···π interactions	[321]
SBMOF‐2, 3D	3.6 × 3.6	2.8	2.7	32.3/29.2	3	—	195	C‐H···π interactions, H‐bonding	[322]
Co(AIN)2 , 3D	4.72 × 4.72	3.17	3.17	34.7/34.1	2.96	He flow at 298 K	450	C‐H···π interactions	[323]
Zn‐FBA, 3D	6.0 × 6.0	1.25	1.14	42.8/39.8	2.9	—	297	H‐bonding, van der Waals interactions	[324]
PCP‐IPA, 3D	4.7 × 5.6	2.5	2.13	37.73/20	2.80	N2 flow at 333 K, 337 K	486.7	H‐bonding, van der Waals interactions	[325]
ZJU‐121, 3D	3.7 × 3.7	3.1	3.19	47.1/43	2.74	He flow at 298 K	923	—	[326]
NKMOF‐8‐Br, 3D	7.11 × 6.15	4.22	3.67	40.8/33.6	2.65	373 K	352	C‐H···π interactions	[327]
Al‐MOFM15 , 3D	3 × 3, 7.7 × 7.7	2.23	1.29	45/32.8	2.51	298 K	793	H‐bonding, π···π, C‐H···π interactions	[328]
UiO‐66‐2CF3 , 3D	7.34 × 7.34	0.88	0.5	14.5/–	2.5	—	467	Pore size	[329]
Ni(IN)2 , 3D	5.06 × 4.37	3.05	0.89	34.5/33.3	2.44	433 K	520	C‐H···π interactions	[330]
Zn(ad)(int), 3D	8 a	2.32	2.22	33.2/28.9	2.4	—	617	H‐bonding, C‐H···π interactions	[331]
Ni(sdba)(dabco)0.5 , 3D	6.8 × 7.1, 4.9 × 3.0, 3.6 × 5.0	3.15	2.99	29.77/28.2	2.28	—	668	H‐bonding, C‐H···π interactions	[332]
Cu‐APC, 2D	5.1 × 7.2	1.56	1.55	29.2/28.8	2.27	298 K	655.5	Flexibility, H‐bonding, C‐H···π interactions	[333]
ZIF‐3, 3D	5, 8	6.0	5.5	28.5/24.2	2.22	—	2997	—	[334]
ZUL‐C4, 3D	6.00	1.7	2.9	31.6/24.3	2.21	N2 flow at 298 K	1190	Van der Waals, C–H···O dipolar interactions	[335]
ZIF‐1, 3D	6.5	1.48	—	—	2.2	—	1502	—	[319]
ZIF‐20, 3D	6.5, 15	2.16	—	—	2.2	—	1767	—	[319]
Zn(BDC)(H2BPZ), 3D	8.0‐10.2 a	3.63	3.29	31.8/23.2	2.2	—	906.5	H‐bonding, C‐H···π interactions	[336]
PCP‐IPA‐NH2 , 3D	3.3 × 5.2	1.7	1.9	–/32.4	2.19	—	387	H‐bonding, C‐H···π interactions	[337]
ZIF‐22, 3D	7, 14.8	2.3	—	—	2.15	—	1818	—	[319]
CuTiF6‐TPPY, 3D	5.0 × 8.0	5.32	4.32	34.2/29.6	2.12	He flow at 298 K	685	H‐bonding, Van der Waals interactions	[338]
Fe2(BDP)3 , 3D	4.9 × 4.9	2.25	2.11	23.9/19.8	2.1	—	1097	UMCs	[339]
ZIF‐96, 3D	17, 8.5	0.82	—	—	2.1	—	1890	—	[319]
CTF‐DCTC‐500, 3D	5.7, 7.9, 11.8	3.10	2.34	25.4/23.7	2.08	Vacuum at 298 K	916	Pore size	[340]
Tb‐MOF‐76(NH2), 3D	7.2 × 7.2	3.27	2.97	32.8/30.9	2.05	—	—	H‐bonding, C‐H···π interactions	[341]
dia‐4‐Co, 3D	7.8 × 9.9	4.6	4.07	24.56/23.37	2.04	Vacuum at 343 K	—	—	[342]
ZUL‐C3, 3D	6.16	1.5	2.3	29.5/19.9	2.04	N2 flow at 298 K	1200	Van der Waals interactions, C–H···O dipolar interactions	[335]
ZJU‐30, 3D	4.0 × 4.0; 5.6 × 5.6	2.1	2.0	29.7/28.1	2	—	228	Pore size, UMCs	[343]
ZJU‐120a, 3D	4.4 × 4.4	4.91	3.93	27.6/17	2	He flow at 298 K	1597	C‐H···π interactions	[326]
Ni(TMBDC)(DABCO)0.5 , 3D	5.9	5.45	5.02	39/31	2.0	Vacuum	894	—	[344]
ZIF‐2, 3D	5.5, 6.5	2.61	—	—	2.0	—	2746	—	[319]
ZIF‐68, 3D	10.5, 8.5	1.6	—	—	2.0	—	1831	—	[319]
HIAM‐210, 3D	4.02 a	2.34	2.12	31.24/22.72	2.0	He flow at 373 K	566	H‐bonding, Van der Waals interactions	[345]
In‐BH(im)3‐BDC, 3D	6.81 × 6.15	2.67	2.38	27.8/23.7	2.0	—	914	H‐bonding	[346]
MUF‐15, 3D	8.5 × 3.5, 7 × 3.8, 3.2 × 1.2	4.69	4.15	28.2/29.2	1.96	He flow at 343 K	1130	C‐H···π, Van der Waals interactions	[347]
Y‐BTC, 3D	7.0 × 7.0	3.5	3.1	22/19	1.92	—	933	UMCs	[348]
COF‐1, 2D	9	2.4	1.9	22.5/22.2	1.92	He flow at 298 K	819	π···π, C‐H···π interactions	[349]
PCN‐245, 3D	10 a	3.27	2.39	20.5/23.0	1.9	423 K	1743	Van der Waals interactions	[350]
PCN‐250, 3D	5.5 × 5.5; 9.6 × 9.6	5.2	4.2	23/21	1.9	—	1470	Pore size, van der Waals interactions	[351]
C‐PDA‐3, 3D	—	6.57	5.10	22/21.6	1.9	—	3160	H‐bonding, electrostatics	[352]
(Hf)DUT‐52, 3D	7.9 a , 10 a	4.02	3.22	25.6/24	1.9	—	1505	C‐H···π interactions	[353]
MIL‐53(Al)‐FA, 3D	7 a	3.7	3.5	34/30.5	1.9	—	1132	Van der Waals interactions	[354]
Ni‐MOF‐2, 3D	8.0	4.7	5.9	23.6/21.4	1.9	He flow at 298 K	1501	C‐H···π interactions, van der Waals interactions	[355]
HIAM‐102, 3D	4.6 × 4.6	1.97	2.14	32.1/27.9	1.9	He flow at 373 K	191	Van der Waals interactions	[356]
ZIF‐71, 3D	17	0.49	—	—	1.9	—	1224	—	[319]
Al‐PyDC, 3D	5.8 a	4.20	3.44	30.1/27.8	1.9	He flow at 373 K	1134	H‐bonding, C‐H···π interactions	[256]
MOF‐808‐Bzz, 3D	12, 14.8	2.20	1.43	29.87/26.43	1.9	—	1003	H‐bonding, C‐H···π interactions	[357]
Zn2(TM‐bdc)(dabco), 3D	—	5.23	4.86	31.5/30.2	1.89	—	962.5	Van der Waals interactions	[358]
NKMOF‐14‐PZ, 3D	8.5 × 8.5	3.4	5.6	21.3/19.2	1.89	—	1561	H‐bonding, C‐H···π interactions	[359]
NKMOF‐8‐Me, 3D	6.96 × 6.21	4.82	4.67	38.4/37.6	1.88	373 K	655	C‐H···π interactions	[327]
Zn(BTFM)(DABCO)0.5 , 3D	4.1‐6.6	2.74	2.01	27.2/24.7	1.88	N2 flow at 393 K	720	H‐bonding, C‐H···π interactions	[360]
Eu‐BTC, 3D	6.0 × 6.0	3.1	2.9	26/21	1.87	—	720	UMCs	[348]
Y‐TATB, 3D	7.5, 12	4.33	2.98	22.9/20.6	1.87	Vacuum at 393 K	1562	Van der Waals, C‐H···π interactions	[361]
FHI‐H11‐Me, 3D	—	2.59	2.08	38.9/25.9	1.85	—	—	H‐bonding, C‐H···π interactions	[362]
Zn‐ATA, 3D	3.5 f	1.08	1.20	32.5/27.4	1.84 e	298 K	344 g	H‐bonding, C‐H···π interactions	[363]
Cu(BTFM)(DABCO)0.5 , 3D	4.2‐5.8	3.02	2.09	25.3/22.7	1.83	N2 flow at 393 K	646	—	[360]
IRMOF‐8, 3D	11.0 × 11.0	4.1	2.9	54/50	1.8	—	1360	Electrostatics	[364]
Sm‐BTC, 3D	—	1.56	1.64	26/23	1.8	—	700	Van der Waals interactions	[348]
CPM‐80‐Fe, 3D	—	4.56	4.05	24.1/22.4	1.8	—	863	Van der Waals interactions, C‐H···π interactions	[365]
CPM‐80‐Co, 3D	—	4.33	3.86	22.5/22.6	1.8	—	895	Van der Waals interactions, C‐H···π interactions	[365]
CPM‐80‐Zn, 3D	—	4.77	4.24	24.6/23.1	1.8	—	995	Van der Waals interactions, C‐H···π interactions	[365]
CPM‐81‐Co, 3D	—	5.51	5.07	24.2/24.1	1.8	—	1015	Van der Waals interactions, C‐H···π interactions	[365]
CPM‐81‐Zn, 3D	—	4.46	4.13	25.5/22.6	1.8	—	907	Van der Waals interactions, C‐H···π interactions	[365]
ZIF‐97, 3D	16	0.83	—	—	1.8	—	1541	—	[319]
ZIF‐8, 3D	3.4 a	2.54	1.5	17.2/16.1	1.8	He flow at 373 K	1844	Kinetic	[366]
NUM‐7a, 3D	4.7 × 7.8	2.85	2.62	35.8/30.0	1.76	—	345	Van der Waals, C‐H···π interactions, H‐bonding	[367]
dia‐4‐Ni, 3D	7.8 × 9.9	4.46	4.01	24.85/23.59	1.76	Vacuum at 343 K	260	H‐bonding, C‐H···π interactions	[342]
CPM‐733, 3D	7.3 × 7.3	7.1	6.4	23.4/22.5	1.75	—	1328.5	Pore size	[72]
Ca(H2tcpb), 3D	5.5 × 5.5	2.78	2.67	35.1/26.8	1.75	—	200	—	[368]
Co2V‐bdc‐tpt, 3D	7.3 × 7.3	6.43	7.19	23.4/22.5	1.75	—	1162	UMCs	[72]
ZIF‐10, 3D	8.5	1.15	—	—	1.75	—	3385	—	[319]
MCOF‐1, 3D	6.4	3.3	2.8	30.0/27.5	1.73	He flow at 298 K	1045	—	[349]
Zn(sdba)(dabco)0.5 , 3D	6.8 × 7.1, 3.2 × 5.0, 3.8 × 5.1	2.83	2.68	28.96/27.94	1.72	—	224	H‐bonding, C‐H···π interactions	[332]
SBMOF‐1, 3D	4.2 × 4.2	1.3	1.3	36.3/35.0	1.7	—	145	C‐H···π interactions, H‐bonding	[322]
MFM‐300(In), 3D	6.8 a	5.1	4.9	30/28	1.7	—	1030	Van der Waals interactions	[369]
ZIF‐4, 3D	—	2.3	2.2	–/–	1.7	He flow at 373 K	300	Kinetic	[370]
JXNU‐9, 3D	6.4 a , 9.2 a , 11.7 a	3.61	2.45	23.6/21.4	1.7	—	1660	C‐H···π interactions	[371]
ScBPDC, 3D	8.43 a	3.42	2.41	16.4/15.4	1.7	N2 flow at 393 K	1777.7	C‐H···π interactions, H‐bonding	[372]
BUT‐10, 3D	12 a	4.76	3.56	24/22.5	1.7	—	1726	C‐H···π interactions	[373]
Ni‐4PyC, 3D	5‐9 a	3.84	3.5	29.07/37.32	1.7	He flow at 424 K	943	H‐bonding, C‐H···π interactions	[374]
Tb‐MOF‐76, 3D	7.9 × 7.9	2.8	3.0	25.2/22.4	1.7	—	—	H‐bonding, C‐H···π interaction	[341]
ZIF‐80, 3D	5.5, 9.5, 12	0.5	—	—	1.7	—	1571	—	[319]
ZIF‐93, 3D	17	0.67	—	—	1.7	—	1592	—	[319]
Zn2(DM‐bdc)(dabco), 3D	—	3.91	2.96	25.8/24.2	1.7	—	1133.8	Van der Waals interactions	[358]
Zn‐atz‐ipa, 3D	2.8 × 2.8, 5.5 × 5.5	1.81	1.80	45.8/40.0	1.7	He flow at 333 K	650	H‐bonding	[70]
PCN‐250(Fe2Zn), 3D	7	5.95	5.38	22.5/–	1.70	Vacuum at 423 K	1544	H‐bonding, C‐H···π interactions	[375]
UiO‐67‐(NH2)2 , 3D	5.2 × 5.2, 7.1 × 7.1, 11.5 × 11.5	5.32	4.32	26.5/24.5	1.7	—	2022	Van der Waals interactions	[376]
MIL‐53‐BDC, 3D	6.3	2.93	2.78	27.2/22.3	1.7	He flow at 298 K	630	O‐H···π, C‐H···π interactions	[377]
IRMOF‐6, 3D	9.4, 14.7	2.67	1.60	—	1.68	—	3066	Pore size	[339]
Ni(bdc)(ted)0.5 , 3D	4.0 × 4.0, 7.0 × 7.0	4.8	3.29	29/25	1.66	—	1905	Van der Waals interactions	[378]
ZIF‐69, 3D	40 a	2.2	1.74	26/23	1.65	—	882	—	[379]
DUT‐8 (Ni), 3D	18.37 × 18.37	5.3	2.97	25.8/24.1	1.65	He flow at 298 K	2440	C‐H···π interactions	[380]
PCP‐FDCA, 3D	3.8 × 5.2	3.0	3.2	35.9/33.3	1.65	—	542	H‐bonding, C‐H···π interactions	[337]
IRMOF‐7, 3D	10.5	1.27	0.93	—	1.65	—	3331	Pore size	[339]
UPC‐66, 3D	8.08 × 11.87, 10.37	2.72	2.36	15.61/17.83	1.65	—	483	C‐H···π interactions	[381]
NKU‐0821a, 3D	13.5 × 10.3	4.29	3.67	27.3/25.0	1.65	He flow at 373 K	962	H‐bonding, C‐H···π interactions	[382]
CPM‐233, 3D	6.8 × 6.8	7.4	6.5	25.4/25.0	1.64	—	1598	Pore size	[72]
Co(bdc)(ted)0.5 , 3D	5.1 × 3.7, 7.6 × 7.6	4.13	2.77	29/24	1.64	—	1708	Van der Waals interactions	[378]
Mg2V‐bdc‐tpt, 3D	6.8 × 6.8	6.58	7.45	27.3/26.7	1.64	—	1597	UMCs	[72]
LIFM‐XYY‐6, 3D	—	2.99	2.24	29.5/28.6	1.63	He flow at 358 K	1240	H‐bonding, C‐H···π, van der Waals interactions	[383]
ZSTU‐2, 3D	—	2.73	2.35	33/32	1.62	298 K	862	—	[384]
NUM‐9, 3D	4.5 × 4.5	2.48	2.32	35.75/32.32	1.61	—	330	C‐H···π interactions, H‐bonding	[385]
JUN‐2, 3D	3.7 × 3.7	4.1	3.6	29.4/26.7	1.6	Vacuum at 298 K	1219	Pore size, H‐bonding	[386]
PCN‐250(Fe), 3D	6.5	6.00	5.48	22.5/–	1.6	Vacuum at 423 K	1541	H‐bonding, C‐H···π interactions	[375]
ZIF‐7, 3D	3.0 × 3.0; 5.0 × 5.0	1.9	1.8	—	1.6	—	230	Flexibility	[387]
USTA‐38, 3D	4.6 × 6.6	4.6	3.3	24.4/29.4	1.6	—	1090	Pore size	[388]
[Ni(bdc)(ted)0.5], 3D	7.6 × 7.6; 5.1 × 3.7	5.0	3.4	21.5/18.4	1.6	—	1701	Van der Waals interactions	[389]
LIFM‐63, 3D	5.6, 8.6	3.0	2.1	25.8/–	1.6	298 K	1486	H‐bonding, C‐H···π interactions	[390]
ZIF‐6, 3D	6, 9	0.77	—	—	1.6	—	3401	—	[319]
NbU‐9, 3D	3.8 × 3.8, 6.4 × 6.4	3.39	3.04	35.8/–	1.6	—	671.3	UMCs	[391]
CPM‐82‐Zn, 3D	—	4.02	3.55	25.8/25.7	1.6	—	533	Van der Waals interactions, C‐H···π interactions	[365]
MAF‐X10(Cl), 3D	4.9 × 6.2	5.87	4.05	25.0/23.5	1.6	—	1751	C‐H···Cl, C‐H···π interactions	[392]
NKCOF‐21, 3D	14 a	4.4	3.3	26.2/23.6	1.6	373 K	1397	H‐bonding	[393]
NKU‐0210, 3D	8 a	4.73	—	21.8/–	1.6	He flow at 373 K	1290	He purging at 373 K	[394]
PCN‐250(Fe2Ni), 3D	6.5	6.19	5.80	22.2/–	1.57	Vacuum at 423 K	1666	H‐bonding, C‐H···π interactions	[375]
SNNU‐40, 3D	6.5 a	7.54	4.9	18/18.1	1.57	298 K	2233.8	H‐bonding, C‐H···π interactions	[395]
COF‐300, 3D	7.8	4.1	3.1	26.9/25.0	1.57	He flow at 298 K	1455	—	[349]
ZJNU‐115, 3D	8.58	4.21	3.75	28.2/27.7	1.56	—	1291	Van der Waals interactions, C‐H···π interactions	[396]
ZJNU‐7, 3D	5.0, 5.9, 7.3	4.13	3.80	29.7/29.3	1.56	—	1180	H‐bonding, C‐H···π interactions	[397]
Cu(bdc)(ted)0.5 , 3D	4.8 × 3.2, 7.5 × 7.5	3.68	2.53	29/24	1.55	—	1631	Van der Waals interactions	[378]
Zn(bdc)(ted)0.5 , 3D	4.8 × 3.2, 7.5 × 7.5	4.45	3.16	31/25	1.54	—	1781	Van der Waals interactions	[378]
NbU‐12, 3D	5 × 5	3.67	2.5	21.8/20.4	1.53	—	1395.6	C‐H···π interactions	[398]
MUV‐11, 3D	—	1.83	1.72	25/23	1.53	298 K	180	—	[384]
MIL‐53‐NDCA, 3D	8.4	4.24	3.12	24.2/17.0	1.53	He flow at 298 K	1590	O‐H···π, C‐H···π interactions	[377]
NPU‐2, 3D	10.5 a	4.44	3.45	19.64/18.18	1.52	He flow at 313 K	2133	H‐bonding, C‐H···π interactions	[321]
PCN‐250(Fe2Co), 3D	6.5	6.21	5.82	22.2/–	1.52	Vacuum at 423 K	1675	H‐bonding, C‐H···π interactions	[375]
CPM‐35, 3D	6.3 a	3.57	2.59	23.3/21.4	1.51	—	783	C‐H···π interactions	[399]
CPM‐223‐tpbz, 3D	6.8	6.88	6.25	21.9/23.3	1.51	—	1662	Pore size	[72]
UiO‐67, 3D	11.7 a , 16.1 a	4.43	2.95	—	1.5	—	1775	—	[373]
Cr‐BTC(O2), 3D	25, 29	3.3	2.89	37.2/24	1.5	Vacuum at 298 K	1135	H‐bonding, UMCs	[400]
UPC‐613, 3D	7 a , 12 a	2.55	2.31	31.83/28.51	1.5	—	853	UMCs	[401]
Ni2(HBTC)2(bpy)0.6(dabco)1.4 , 3D	5.3 a	4.8	4.6	33.5/31.7	1.5	—	1070	H‐bonding, Van der Waals interactions	[402]
CPM‐723, 3D	6.8	6.91	6.67	21.7/20.0	1.5	—	1369.8	Pore size	[72]
1a‐tz, 3D	7.3 × 11.8	3.4	3.3	35/33	1.5	—	845	H‐bonding	[403]
MIL‐142a, 3D	7.0 × 7.0	3.8	2.9	27.3/25.1	1.5	—	1580	Van der Waals interactions	[404]
MAF‐X10, 3D	4.9 × 6.8	5.07	3.57	23.4/22.4	1.5	—	1644	H‐bonding, C‐H···π interactions	[392]
NKCOF‐22, 3D	16 a	2.9	1.8	25.9/24.1	1.5	373 K	1580	—	[393]
COF‐320, 3D	13.5 × 6.2	2.4	1.8	26.9/25.1	1.49	He flow at 298 K	923	—	[349]
CPM‐736, 3D	5.9	4.03	3.88	30.4/30.1	1.48	—	472.5	Pore size	[72]
COF‐102, 3D	11.5	1.9	1.6	28.7/25.1	1.48	He flow at 298 K	33122	—	[349]
A‐66, 3D	5.6 × 6.8	6.56	4.45	23.7/22.6	1.48 h	—	—	H‐bonding, C‐H···π interactions	[405]
MIL‐53‐BPDC, 3D	10.9	2.97	2.07	22.1/22.1	1.47	He flow at 298 K	1780	O‐H···π, C‐H···π interactions	[377]
Azole‐Th‐1, 3D	10	4.5	3.6	28.6/26.1	1.46	He flow at 298 K	983	Van der Waals interactions	[406]
PCN‐250(Fe2Mn), 3D	6.5	5.53	5.24	21.4/–	1.45	Vacuum at 423 K	1391	H‐bonding, C‐H···π interactions	[375]
MIL‐125, 3D	—	4.83	3.98	23.6/17	1.43	298 K	1435	H‐bonding, C‐H···π interactions	[384]
DUT‐8 (Cu), 3D	18.47 × 18.47	4.74	2.44	25.5/18.3	1.43	He flow at 298 K	2370	C‐H···π interactions	[380]
CPM‐238, 3D	5.9	5.56	5.25	24.7/24.4	1.43	—	1444	Pore size	[72]
CPM‐738, 3D	5.9	1.67	4.56	27.9/26.5	1.42	—	1161.5	Pore size	[72]
Zn‐PNMI, 3D	6.4 × 6.4	1.6	1.7	24.5/23.8	1.42	Vacuum at 393 K	305	H‐bonding	[407]
CPM‐63 m, 3D	9.58	2.84	2.53	31.8/29.8	1.41	—	1023	Open Lewis basic N sites	[408]
In‐soc‐MOF‐1, 3D	7.65 × 5.65; 10 × 10	4.0	3.7	28.4/25.2	1.4	—	1223	Pore size, van der Waals interactions	[409]
UTSA‐33, 3D	5.4 × 6.5; 4.8 × 5.8	2.8	2.7	32/32	1.4	—	660	Pore size	[410]
UTSA‐35, 3D	7.7 × 5.8	2.4	2.1	30/28	1.4	—	742	Pore size, van der Waals interactions	[411]
UPC‐612, 3D	14 a , 20 a	3.58	2.8	22.39/16.94	1.4	—	2016	UMCs	[401]
Zr‐bptc, 3D	7 a , 13 a	3.26	3.08	27/24	1.4	Vacuum at 393 K	1085.3	Electrostatics	[412]
Ni(HBTC)(bpy), 3D	5.5 a	6.6	6.0	33.6/32.2	1.4	—	1474	H‐bonding, Van der Waals interactions	[402]
LIFM‐31, 3D	6.8, 11.8	2.4	1.8	26.9/–	1.4	298 K	1711	—	[390]
LIFM‐62, 3D	5.9, 11.7	2.6	2.0	24.7/–	1.4	298 K	1977	—	[390]
Mn‐PNMI, 3D	8.0 × 8.0	2.8	2.0	23.5/16.9	1.38	Vacuum at 393 K	818	H‐bonding	[407]
Dy‐BTC, 3D	—	1.90	1.90	32/27	1.37	—	947	Van der Waals interactions	[348]
NKCOF‐62, 3D	7 a	2.21	2.0	29.6/26.3	1.37	He flow at 373 K	714	Van der Waals interactions	[413]
MOF‐841, (273 K) 3D	5.1 × 5.1	3.4	4.7	24.8/20.8	1.35	He flow at 298 K	1426	C‐H···π interactions, van der Waals interactions	[414]
NPU‐1, 3D	7.4 a	4.5	3.2	29.1/23.95	1.32	He flow at 313 K	1580	H‐bonding, C‐H···π interactions	[321]
MOF‐545, 3D	13 a , 33 a	3.12	2.57	22.7/21.5	1.31	Vacuum at 393 K	2265.4	C‐H···π interactions, H‐bonding	[415]
Cu(ina)2 , 2D	4.1 × 4.1	2.0	1.9	—	1.3	—	228	—	[317]
MIL‐53 (Al), 3D (323K)	8.5 × 8.5, 2.6 × 13.6	2.05	1.69	22.5/20.5	1.3	—	—	—	[416]
MAF‐X10(Me), 3D	4.9 × 5.8	5.01	4.01	26.7/21.2	1.3	—	1782	H‐bonding, C‐H···π interactions	[392]
LIFM‐61, 3D	7.3, 11.8	1.7	1.4	24.1/–	1.3	298 K	1194	—	[390]
CPOC‐301, 3D	12 × 6	3.37	3.92	32.4/24.2	1.3	He flow at 373 K	1962	C‐H···π interactions	[417]
Co3(μ 3‐OH)(tipa)(bpy)1.5 , 3D	4.8 × 4.8	1.34	1.36	37.2/33.6	1.3	—	408.4	C‐H···π, π···π interaction	[418]
NKCOF‐23, 3D	17 a	2.7	2.2	24.3/23.0	1.3	373 K	1900	—	[393]
Cd‐PNMI, 3D	7.6 × 7.6	1.9	1.4	19.4/13.8	1.27	Vacuum at 393 K	264	H‐bonding	[407]
Zn‐atz‐oba, 3D	3.2 × 4.4	1.7	1.7	30/27	1.27	He flow at 333 K	710	H‐bonding	[419]
BUT‐151, 3D	11.2 × 11.2	3.99	3.72	31.06/23.95	1.26		1330	Pore size, C‐H···π interactions	[420]
DBA‐3D‐COF‐1, 3D	28	1.7	1.7	16.8/15.9	1.24	—	5083	—	[421]
MOF‐808‐Ind, 3D	12, 14.8	1.68	1.39	28.43/26	1.23	—	948	H‐bonding, C‐H···π interactions	[357]
MOF‐808‐Izo, 3D	12, 14.8	1.74	1.56	28.61/26.73	1.22	—	1130	H‐bonding, C‐H···π interactions	[357]
COF‐8, 2D	18.7	1.6	1.5	27.5/25.1	1.21	He flow at 298 K	1180	—	[349]
TJT‐100, 3D	8.7 × 11.6	3.7	3.4	29/25	1.2	—	890	H‐bonding, van der Waals interactions	[422]
LIFM‐28, 3D	6.8, 11.8	1.0	0.9	26.3/–	1.2	298 K	927	—	[390]
COF‐6, 2D	6.4	2.1	2.1	29.2/27.5	1.20	He flow at 298 K	706	—	[349]
NH2‐MIL‐125, 3D	—	4.69	4.41	25/23	1.18	298 K	1180	—	[384]
BUT‐150, 3D	13.0 × 13.0	4.30	4.26	23.87/23.68	1.15	—	1641	Pore size, C‐H···π interactions	[420]
Ni‐DBA‐3D‐COF, 3D	26	1.83	1.83	11.6/9.7	1.15	—	4763	UMCs	[421]
COF‐10, 2D	31.7	1.0	0.9	26.6/25.1	1.13	He flow at 298 K	2056	—	[349]
MOF‐808, 3D	7.3, 18	1.14	1.05	15.37/14.59	1.05	—	1574	—	[357]
CTF‐DCTC‐400, 3D	7.3, 11.8	1.82	1.68	22.7/22.0	1.04	Vacuum at 298 K	771	Pore size	[340]
Au‐PCM‐102, 3D	6.7 × 8.2	2.7	2.5	—	—	—	1449	Van der Waals interactions	[423]
Ag‐PCM‐102, 3D	6.7 × 8.2	3.5	2.9	—	—	—	1558	Van der Waals interactions	[423]
RPM3‐Zn, 3D	—	1.56	0.89	—	—	—	328	Flexibility, H‐bonding	[264]
IRMOF‐4, 3D	9.2	3.09	1.45	—	—	—	1438	Pore size	[339]
IRMOF‐5, 3D	7	1.73	0.91	—	—	—	861	Pore size	[339]
Cu(1,3‐bdc)(ted)0.5 , 3D	22	2.43	2.15	23.8/26.5	—	—	1673	H‐bonding, C‐H···π interactions	[424]
Co(1,3‐bdc)(ted)0.5 , 3D	—	2.58	2.13	34/29.3	—	—	1802	H‐bonding, C‐H···π interactions	[424]
UTSA‐34a, 3D	12.8	2.79	2.63	—	—	—	991	UMCs	[425]
FJI‐C4, 3D	5.5	2.96	2.74	32.7/33.1	—	—	690	Pore size	[426]
NTU‐24, 3D	0.86 × 0.86, 4.6 × 4.6	5.27	4.55	29/24	—	He flow at 298 K	1620	Pore size, shape	[427]
NTU‐25, 3D	1.22 × 1.22, 4.88 × 4.88	4.38	4.24	31/27	—	He flow at 298 K	1540	Pore size, shape	[427]
CPM‐63a, 3D	9.58	2.86	2.51	33.9/30	—	—	1127	Open Lewis basic N sites	[408]

^a Determined from Horvath–Kawazoe/DFT method applied on N₂ isotherm at 77 K.

^b IAST selectivity at 1 bar for 1:1 (v/v) C₂H₆/C₂H₄.

^c Adsorption pressure at 40kPa.

^d Adsorption pressure at 88.5 kPa.

^e IAST selectivity at 1 bar for 1:9 (v/v) C₂H₆/C₂H₄.

^f Pore size determined from CO₂ adsorption isotherms recorded at 195 K.

^g Langmuir surface areas determined from the corresponding CO₂ adsorption isotherms recorded at 195 K.

^h Uptake ratio.

TABLE 10.

C₃H₄/C₃H₆ binary separations. The following parameters are listed for comparison: pore size; BET surface area (S_BET); single‐component gas uptakes; adsorption enthalpies (Q _st); adsorption selectivity; regeneration temperature; and attributed mechanisms. Reticular sorbents are listed in decreasing order of binary selectivity for C₃H₄ selective sorbents (above) and C₃H₆ selective sorbents (below).

Adsorbent, network dimensionality (nD)	Pore size (Å)	C3H4 uptake at 1 bar (mmol/g)	C3H6 uptake at 1 bar (mmol/g)	Q st (kJ/mol)	Selectivity a	Regeneration temperature	SBET (m2/g)	Mechanism	Refs.
NKMOF‐1‐Ni, 3D	5.7 × 5.7	3.5	2.1	54.0/38.0	1217.8 e	—	374	H‐bonding, π···π interactions	[428]
NKMOF‐11, 3D	5.7 × 5.7	3.1	1.5	84.8/30.5	1074 d	He flow at 363 K	376	H‐bonding, π···π interactions	[429]
BUT‐306, 3D	3	1.32	0.11	—	636	He flow at 393 K	305	Dipole‐dipole, C‐H···π interactions	[430]
FJI‐W1, 3D	8 × 8	7.09	6.27	61.7/36.1	593 d	—	1376	H‐bonding	[431]
TIFSIX‐14‐Cu‐i, 3D	3.4 c	3.88	1.40	57/49	355	He flow at 338 K	481 c	Flexibility, H‐bonding, Van der Waals interactions	[432]
ZU‐16‐Co, 3D	3.62 × 3.62	2.45	2.1	87.5/–	248 d	Ar flow at 373 K	—	H‐bonding	[433]
sql‐NbOFFIVE‐bpe‐Cu‐AB, 2D	3.96 × 5.56	3.04	2.10	69/53	220	Vacuum at 333 K	295 d	H‐bonding	[434]
GeFSIX‐14‐Cu‐i (ZU‐33), 3D	3.4 × 3.4	3.34	1.5	54/–	217 d	He flow at 338 K	463 c	H‐bonding	[432]
Co‐gallate, 3D	4.0 c	3.21	1.49	82/47	152	He flow at 393 K	475	H‐bonding, Van der Waals interactions	[435]
UTSA‐200, 3D	3.4 × 4.2	3.58	1.20	62.3/45.4	149.5	—	617	Flexibility, pore size, H‐bonding	[436]
Ni‐gallate, 3D	3.9 c	2.65	0.90	84/53	113	He flow at 393 K	424	H‐bonding, Van der Waals interactions	[435]
ELM‐12, 2D	6.1 × 4.3 × 4.3, 6.8 × 4.0 × 4.2	2.74	1.38	60.6/15.8	84	He flow at 323 K	740 [437]	H‐bonding	[438]
ZJUT‐1a, 3D	7.5 × 3.7 × 3.7	3.97	1.25	38/–	70	—	222	H‐bonding, Van der Waals interactions	[439]
SIFSIX‐3‐Ni, 3D	3.8 × 5.0	2.98	2.67	68/47	76	He flow at 393 K	351	Pore size, shape, H‐bonding	[440]
Mg‐gallate, 3D	4.1 c	3.75	1.50	66/44	65	He flow at 393 K	559	H‐bonding, Van der Waals interactions	[435]
NTU‐88, 2D	4.4 × 7.9	3.84	0.07	29.4/27.5	52.7 f	—	420	Molecular sieving	[441]
ZU‐62 (NbOFFIVE‐2‐Cu‐i), 3D	6.75, 6.94, 7.20	3.60	2.60	71/56	48	—	476	H‐bonding, Van der Waals interactions	[37]
Cu(dps)2(GeF6), (GeFSIX‐dps‐Cu), 3D	1.8 × 2.6	3.7	0.08	50.2/–	46.5 f	—	382	Molecular sieving	[244]
CPOS‐1, 3D	5.3 × 5.3	0.93	0.36	44.2/31.6	43.9 d	—	336.2	Van der Waals interactions	[442]
Zeolite 5A (yNa+), 3D	4.4‐5.2 c	3.62	2.58	52/30	43	He flow at 298 K	503	Pore size	[443]
SIFSIX‐2‐Cu‐i, 3D	5.2 × 5.2	3.80	2.60	45/37	25	He flow at 298–308 K	586	Pore size, shape, H‐bonding	[440]
NTU‐100‐NH2 , 3D	5.9 b	6.7	5.58	43.5/40	23.2	—	947	H‐bonding	[444]
SIFSIX‐Cu‐TPA (ZNU‐2‐Si), 3D	6.27‐9.84 b	8.46	—	39.35/34.26	19.3	Ar flow at 298 K or vacuum	1339	H‐bonding, van der Waals interactions	[445]
TIFSIX‐Cu‐TPA (ZNU‐2‐Ti), 3D	6.56‐9.4 b	7.66	—	43.5/35	16.2	Ar flow at 298 K or vacuum	1380	H‐bonding, van der Waals interactions	[445]
NbOFFIVE‐Cu‐TPA (ZNU‐2‐Nb), 3D	7.85‐9.4 b	7.28	—	42/32.5	13.8	Ar flow at 298 K or vacuum	1281	H‐bonding, van der Waals interactions	[445]
ZNU‐2, 3D	8.5 × 8.5, 8.8 × 4	7.70	5.30	43/34.5	12.5	—	1380	Pore size, shape, H‐bonding, Van der Waals interactions	[446]
SIFSIX‐1‐Cu, 3D	8.0 × 8.0	8.76	5.90	49/35	9	He flow at 298–308 K	1128	Pore size, shape, H‐bonding	[440]
Cu‐FINA‐2, 3D	5.48 × 4.87	1.36	0.67	45.3/40.3	6.1	Vacuum	175.9	H‐bonding	[447]
BUT‐309, 3D	4.1 × 4.1	1.07	0.8	84.3/50.8	5.2	He flow at 298 K	—	H‐bonding	[448]
Cu‐FINA‐1, 3D	5.68 × 4.76	1.8	1.41	45.5/44.5	4.5	Vacuum	389.4	H‐bonding	[447]
MFM‐300(In), 3D	6.8 b	6.25	5.20	35/31	3.7	—	1030	Van der Waals interactions	[369]
SIFSIX‐3‐Ni, 3D	5.03 × 3.75	3.28	2.9	45/30	3.6	He flow at 298 K	368	Molecular sieving (Kinetic)	[449]
Cu‐INA, 3D	7.86 × 6.95	2.8	2.1	43/39	3.1	Vacuum	428.7	H‐bonding	[447]
JXNU‐6, 3D	10.4 × 10.4	5.07	3.57	39.9/27.1	3.1 d	—	856.3	H‐bonding	[450]
BUT‐310, 3D	9 × 9	10.8	7.8	32.3/30.6	2.73	298 K	1811	UMCs	[451]
Cu‐APC, 3D	5.1 × 7.2	3.75	1.73	29.9/25.7	2.6	298 K	655.5	H‐bonding, C‐H···π interactions	[333]
NbOFFIVE‐1‐Ni, 3D	4.66 × 3.21	2.99	2.39	38/30	2.4	He flow at 298 K	248	Molecular sieving (Kinetic)	[449]
NKCOF‐36, 3D	6, 9 b	8.49	6.49	39.7/37.4	2.22 (1/99)	373 K	1093	Van de Waals interactions, pore size	[115]
NKCOF‐37, 3D	6, 10 b	7.80	6.04	37.7/36.6	2.06 (1/99)	373 K	1838	Van de Waals interactions, pore size	[115]
COF‐1, 2D	15 × 15	4.61	3.61	–/–	1.59	—	640	—	[115]
JXNU‐15(NH2), 3D	10.9	12.1	9.8	24.3/20.9	1.48 d	—	2458	H‐bonding, C‐H···π interactions	[452]
NTU‐100‐NO2 , 3D	6.2 b	4.94	3.48	32.5/31	1.43	—	688	H‐bonding	[444]
JXNU‐15, 3D	11.8	13.9	11.2	17.8/15.7	1.28 d	—	2806	H‐bonding, C‐H···π interactions	[452]
BUT‐308, 3D	2.1 × 2.1	4.36	2.19	—	—	He flow at 298 K	—	H‐bonding	[448]
BUT‐305, 3D	—	1.98	1.96	—	—	He flow at 393 K	351	—	[430]
SIFSIX‐3‐Ni, 3D	3.6	3.32	2.73	45/30	—	He flow at 340 K	368	—	[453]
FCOF‐5, 3D	—	3.51	1.97	30.91/27.92	—	He flow 363 K	—	Flexibility	[454]

^a For v/v = 1/1 equimolar mixture, predicted by IAST.

^b Pore size, BET surface area calculated from the N₂ sorption isotherm at 77K.

^c Pore size, BET surface area calculated from the CO₂ sorption isotherm at 195K, 273K.

^d For v/v = 1/99 mixture, predicted by IAST.

^e For v/v = 0.5/99 mixture, predicted by IAST.

^f Uptake ratio.

^g Adsorption selectivity.

TABLE 11.

C₃H₆/C₃H₈ binary separations. The following parameters are listed for comparison: pore size; BET surface area (S_BET); single‐component gas uptakes; adsorption enthalpies (Q _st); adsorption selectivity; regeneration temperature; and attributed mechanisms. Reticular sorbents are listed in decreasing order of binary selectivity for C₃H₆ selective sorbents (above) and C₃H₈ selective sorbents (below).

Adsorbent, network dimensionality (nD)	Pore size (Å)	C3H6 uptake at 1 bar (mmol/g)	C3H8 uptake at 1 bar (mmol/g)	Q st (kJ/mol)	Selectivity a	Regeneration temperature	SBET (m2/g)	Mechanism	Refs.
HAF‐1, 3D	4.14	1.22	0.04	65.88/–	1.67 ×107	He flow at 423 K	410.6	Molecular sieving, H‐bonding, C‐H⋅⋅⋅π interactions	[455]
UTSA‐400, 3D	3.2 × 3.2	1.84	0.05	60.5/–	107	He flow at 298 K	226 b	H‐bonding, van der Waals interactions	[456]
NTU‐85‐WNT, 3D	4.57 × 4.57	0.45	0.003	49.9/–	1570	Vacuum at 323 K	220	H‐bonding	[457]
Zn(ox)0.5(trz), 3D (323K)	2.9	2.29	—	—	1565	—	546	Kinetics (Pore size)	[458]
Cd‐HFDPA, 3D	7.3 × 7.3	1.16	0.96	—	827	—	171.98	π complexations, electrostatics	[459]
MIP‐203, 3D (303 K)	6.3	0.33	0.45	—	551.4	—	5.8	Flexibility, H‐bonding, electrostatics	[460]
JNU‐3a, 3D	5.6 e	2.62	2.14	29.3/16.1	513	He flow at 298 K	588	Flexibility, molecular sieving	[461]
Co‐gallate, 3D	4.2 × 5.1	1.79	0.14	41/–	330	—	486.8	H‐bonding, O‐H⋅⋅⋅π, van der Waals interactions	[462]
FDC‐4, 3D	3.76 × 3.76	5.43	0.88	35/–	318	—	—	Molecular sieving	[463]
CAU‐10‐OMe, 3D	—	1.4	0.3	—	276	—	179	Flexibility	[464]
Zn(ox)0.5(atrz), 3D (323K)	3.3	1.64	—	—	220	—	521	Kinetics (Pore size)	[458]
ELM‐12, 2D	4 × 4	1.48	1.36	30/28	204	—	—	Kinetics	[465]
HIAM‐301, 3D	4.6 d	3.16	0.27	27/–	150	—	579	H‐bonding, π⋅⋅⋅π interactions	[466]
1to, 3D	4.7–5.0	3.07	1.70	38.1/9	80	He flow at 423 K	1360	π‐complexation, van der Waals interactions	[467]
[Zn(2‐cim)2], 3D	12.5 × 12.5	2.26	2.39	34/30	60	—	—	Kinetics (Pore size)	[468]
Milli‐Zn‐ATA, 3D	3.7	1.0	—	44.9/34.8	60 e	N2 flow at 373 K	349	π⋅⋅⋅π interactions	[469]
ATC‐Cu, 3D	6 × 6	3.62	3.16	69.86/39.92	58.7	Vacuum	—	UMCs	[470]
Fe2(m‐dobdc), 3D	—	7.5	6.0	73/42	55	—	—	UMCs	[295]
ZJU‐75a, 3D	4.1 × 4.4	3.31	2.33	65.9/33.1	54.2	He flow at 323 K	391	UMCs	[471]
Co‐MOF‐74, 3D	11‐12	7.29	—	49/–	45	C3H8 Flow	—	UMCs	[472]
1tm, 3D	—	2.30	1.92	18.2/28.5	40.8	He flow at 423 K	1171	H‐bonding, C‐H⋅⋅⋅π interactions	[467]
Mn2(m‐dobdc), 3D	—	7.5	6.0	70/38	40	—	—	UMCs	[295]
Co2(m‐dobdc), 3D	—	7.5	6.0	53/52.5	39	298 K	—	UMCs	[295]
NTU‐65‐CoTi, 3D	3.5 × 4.2, 4.7 × 6.0	2.56 g	0.07 g	—	37.05 h	—	1078.5	Flexibility	[473]
Ni2(m‐dobdc), 3D	—	7.5	6.0	55/45	35	298 K	—	UMCs	[295]
ZnAtzPO4 , 3D	3.8	1.1	0.3	27.5/–	31	N2 flow at 298 K	420	H‐bonding	[474]
Ni(AIP)(BPY)0.5 , 3D	6.0	1.94	0.45	48.7/–	31	—	355	π‐complexations	[475]
Co(AIP)(BPY)0.5 , 3D	6.6	1.99	0.49	41.79/–	21	—	260	π‐complexations	[475]
Co2(5‐aip)2(bpy), 3D	—	2.0	0.5	42.37/33.68	21	He flow at 423 K	—	C‐H⋅⋅⋅π interactions	[476]
Zn(AIP)(BPY)0.5 , 3D	7.1	1.87	0.59	40.67/–	20	—	344	π‐complexations	[475]
Zn2(5‐aip)2(bpy), 3D	—	1.91	0.76	46.2/25.5	19.8	423 K	—	UMCs	[477]
MFM‐520, 3D	—	2.33	2.03	53.4/43.2	17	He flow at 318 K	313	H‐bonding, π⋅⋅⋅π interactions	[478]
MOF‐74‐Mn, 3D	11	7.2	6.0	54/35	16.6	—	1797	UMCs	[305]
1tp, 3D	—	2.00	1.50	15.2/7	15.3	He flow at 423 K	893	H‐bonding, C‐H⋅⋅⋅π interactions	[467]
GeFSIX‐3‐Ni, 3D	4.42	2.0	0.55	42/22.8	15	He flow at 373 K	246	H‐bonding, π⋅⋅⋅π interactions	[479]
MOF‐74‐Fe, 3D	11	6.9	6.2	44/33	14.7	—	1536	UMCs	[66]
1, 3D	5.3, 6.4	1.92	2.13	11.8/20	14.3 h	He flow at 423 K	859	π‐complexations, C‐H⋅⋅⋅π, van der Waals interactions	[467]
Fe‐MOF‐74, 3D (318K)	12 × 12	6.24	5.19	44/33	13.6	—	—	UMCs	[295]
Cu(0.6)@MIL‐100(Fe), 3D	5.5	—	—	—	13.2	298 K	1490	UMCs	[480]
BTO‐MOF, 3D	3.05 × 3.05	0.59	0.45	—	12	—	283	Kinetics (Pore size)	[481]
DBTO‐MOF, 3D	3.34 × 3.34	1.0	0.84	—	11	—	457	Kinetics (Pore size)	[481]
Ni‐NP, 3D	6.3	3.57	2.13	57/32.3	10.5	He flow at 373 K	543	UMCs	[482]
MOF‐74‐Ni, 3D	11	7.0	5.7	52/36	10.4	—	1532	UMCs	[305]
GeFSIX‐3‐Co, 3D	4.72	1.9	1.09	—	10	He flow at 373 K	—	H‐bonding, π⋅⋅⋅π interactions	[479]
NTU‐65‐CoZr, 3D	3.2 × 4.1, 4.8 × 5.5	0.38 g	0.04 g	—	8.91 h	—	537.3	Flexibility	[473]
MIP‐202, 3D	6.3	0.30	0.13	—	8.8	—	13.1	H‐bonding, electrostatics	[460]
MAF‐23‐O, 3D	—	1.34	1	54/34	8.8	—	—	H‐bonding	[483]
MOF‐74‐Co, 3D	11	6.8	5.9	51/35	8.6	—	1438	UMCs	[305]
NiNi‐Pyz, 3D	3.9 × 3.8	3.26	2.77	52/–	7.8	—	448.7	UMCs, C‐H⋅⋅⋅π interactions	[484]
NTU‐85‐F, 3D	7.031	0.886	1.02	61.2/55.3	7.0	—	350	H‐bonding	[457]
AGTU‐3a, 3D	—	0.5	1.2	68/34	7	—	227	UMCs	[485]
MIL‐101(Cr)‐SO3H‐Ag, 3D	15	4.3	3.0	—	6.0	—	1253	UMCs	[296]
MOF‐74‐Mg, 3D	11	7.5	6.0	47.5/35	5.5	—	1835	UMCs	[305]
Py1/3@Cu‐BTC, 3D	8.5	7.0	6.7	—	5.5	—	1510	Electrostatics	[486]
[Zn2(BDC‐ Cl)2(Py2TTz)] n , 3D	5.4, 4.9	3.32	1.62	—	5.2	He flow at 298 K	441	Flexibility	[487]
SIFSIX‐2‐Cu‐i, 3D	4.7	2.65	—	35.82/20	5.0	338 K	—	H‐bonding, π⋅⋅⋅π interactions	[488]
FePt‐M′MOF, 3D	4.2	2.1	2.05	46.3/–	4.7	—	432	UMCs	[489]
NJU‐Bai8, 3D	4.0 × 3.3	2.89	2.89	—	4.6	Vacuum at 338 K	1048	Flexibility	[490]
FeNi‐Pyz, 3D	4.0 × 4.0	3.08	3	51.4/–	4.3	—	—	UMCs, C‐H⋅⋅⋅π interactions	[484]
CoNi‐Pyz, 3D	3.9 × 4.1	3.38	3.19	51/–	4.2	—	—	UMCs, C‐H⋅⋅⋅π interactions	[484]
GeFSIX‐2‐Cu‐i, 3D	4.5	2.69	1.8	36.25/20.37	4.0	338 K	—	H‐bonding, π⋅⋅⋅π interactions	[488]
MOF‐74‐Zn, 3D	11	6.3	5.5	46/34	3.9	—	1277	UMCs	[305]
MUF‐17, 3D	5.7	2.44	2.02	49/17.5	3.8	—	310	Electrostatics	[491]
NKU‐FlexMOF‐1, 3D	—	3.20	2.33	61.1/52.5	2.31	—	952	Flexibility, electrostatics	[492]
MIL‐100(Cr)‐DAA, 3D	—	6.5	7.5	49.5/36	2	298 K	3501.6	Electrostatics	[493]
[Co(mpba)2], 3D	5.4 × 3.9	1.13	1.10	—	1.81	—	220 f	Flexibility	[494]
NTU‐65‐FeTi, 3D	3.8 × 4.1, 5.0 × 6.0	0.06 g	0.03 g	—	1.74 h	—	931.8	Flexibility	[473]
CuxOy @HP‐Cu‐BTC(CTAB), 3D	7, 12.9	3.42	3.06	18.38/19.47	1.66	—	817.5	H‐bonding, van der Waals, π⋅⋅⋅π interactions	[495]
[Mn(mpba)2], 3D	3.1 × 3.0	1.42	1.33	—	1.64	—	231 f	Flexibility	[494]
CuxOy@HP‐Cu‐BTC(SDBS), 3D (313K)	8, 9	6.09	5.03	1.95/25.44	1.53	—	620	H‐bonding, van der Waals, π⋅⋅⋅π interactions	[495]
NTU‐65‐FeZr, 3D	3.5 × 4.3, 5.0 × 5.6	0.03 g	0.02 g	—	1.45 h	—	1066.1	Flexibility	[473]
ZIF‐7, 3D	—	2.4	2.3	—	1.25	—	—	Flexibility	[387]
MAF‐23, 3D	3.6	1.23	0.85	60.2/57.9	1.2	—	—	H‐bonding	[483]
MIL‐101(Cr), 3D	—	7.9	6.2	36.9/36.3	1.2	298 K	3179.7	Electrostatics	[493]
MIL‐101‐Cr‐SO3H, 3D	15	4.5	3.8	—	1.1	—	1856	UMCs	[296]
CuxOy @HP‐Cu‐BTC(N,N), 3D	—	4.74	4.59	3.61/18.33	0.94	—	1194.1	H‐bonding, van der Waals, π⋅⋅⋅π interactions	[495]
CPL‐1, 3D	10–13	1.82	0.29	—	—	298 K	330	Flexibility	[496]
Zn3(OH)2(pzdc)(atz), 3D	4.0 × 2.1, 3.7 × 1.3	2.1	—	—	—	—	—	Flexibility, molecular sieving	[497]
Y‐dbai, 3D	4.4	2.57	0.1	55/–	—	Ar flow at 423 K	405.4	Molecular sieving	[498]
Y‐abtc, 3D	4.7 c	2	0.07	50/–	—	—	427	Pore size	[499]
Cu‐BTC, 3D	—	8.0	6.8	41.9/28.7	—	—	1500–2100	Cation⋅⋅⋅π	[500]
MIL‐100(Fe), 3D	—	1.3	0.6	—	—	—	2266	UMCs	[501]
Ftw‐MOF‐ABTC, 3D	4.44 × 2.35	2.30	2.46	48.8/38	—	—	—	Kinetics	[502]
KAUST‐7 (NbOFFIVE‐1‐Ni), 3D	4.75 × 3.05	0.6	0.1	57.4/–	—	He flow at 298 K	280	H‐bonding	[38]
Zn‐ATZ‐IP(OH)‐a, 3D	2.7 × 2.7	0.65	0.05	55.2/–	—	—	270	Pore size, N‐H⋅⋅⋅π	[503]

^a IAST selectivity at 1 bar for 1:1 (v/v) C₃H₆/ C₃H₈.

^b The Brunauer–Emmett–Teller surface area calculated from CO₂ isotherm at 298K.

^c Pore size calculated from CO₂ isotherm at 298K.

^d Pore size calculated from N₂ isotherm at 77K.

^e Pore size calculated from Ar isotherm at 87K.

^f Langmuir surface area.

^g Uptake at 0.5 bar.

^h Uptake ratio.

ⁱ Separation factor.

TABLE 12.

C₃H₈/C₃H₆ binary separations. The following parameters are listed for comparison: pore size; BET surface area (S_BET); single‐component gas uptakes; adsorption enthalpies (Q _st); adsorption selectivity; regeneration temperature; and attributed mechanisms. Reticular sorbents are listed in decreasing order of binary selectivity for C₃H₈ selective sorbents (above) and C₃H₆ selective sorbents (below).

Adsorbent, network dimensionality (nD)	Pore size (Å)	C3H8 uptake at 1 bar (mmol/g)	C3H6 uptake at 1 bar (mmol/g)	Q st (kJ/mol)	Selectivity (binary) a	Regeneration temperature	SBET (m2/g)	Mechanism	Refs.
Ni(ADC)(TED)0.5 , 3D	4.8 b	2.32	2.11	65.3/56.5	6.4	He flow at 298 K	679	C‐H⋅⋅⋅π interactions	[504]
Ni(bpe)2(WO4), 3D	5.6 × 5.6	0.67	0.57	42/36	2.75	Ar flow at 328 K	—	Pore shape, electrostatics	[505]
JUN‐90, 3D	5.1 × 5.15, 5.6 × 5.6	2.41	2.19	38.2/33.7	2.7	—	294	H‐bonding, C‐H⋅⋅⋅π interactions	[506]
PCP‐IPA, 3D	4.7 × 5.6	2.23	2.25	50.94/43.36	2.48	N2 flow at 373 K	486.7	H‐bonding, van der Waals interactions	[325]
FDMOF‐2, 3D	6.0 b	5.04	4.15	34.6/30.9	2.18	298 K	1011	C‐H⋅⋅⋅π interactions, H‐bonding	[507]
NUM‐7, 3D	4.7 × 7.8	2.98	3.1	40.03/38.19	1.7	—	345 [367]	C‐H⋅⋅⋅π interactions	[508]
WOFOUR‐1‐Ni, 3D	5.6 × 5.6	1.16	0.89	42/36	1.6	Ar flow at 328 K	315 [509]	Van der Waals interactions	[505]
Zr‐BPYDC, 3D	11.8 b	6.8	7.2	32.5/29.5	1.6	—	2080	Van der Waals interactions	[510]
MoFOUR‐1‐Ni, 3D	—	1.07	1.12	—	1.58	Ar flow at 328 K	456 [511]	—	[505]
JNU‐9‐CH3 , 3D	9.5	2.96	2.92	31.1/–	1.5	Vacuum at 298 K	716	Van der Waals interactions	[512]
CPM‐734c, 3D	9.3 b	8.73	9.05	31.5/30.8	1.44	—	1944	Pore size, van der Waals interactions	[513]
HIAM‐402, 3D	8–12	∼5.8–6.25	∼5.8–6.25	34.5/31.2	1.43	—	1442	H‐bonding	[514]
ZIF‐8, 3D	3.4 × 3.4, 12 × 12	4.4	4.5	—	1.4	—	1844	—	[366]
BUT‐10, 3D	10 b	6.25	6.43	32.7/30	1.4	—	1726	H‐bonding, C‐H⋅⋅⋅π interactions	[515]
g‐C3N4@Zr‐BPDC, 3D	10.9 b	8.9	8.9	29/27.5	1.4	—	2409	Van der Waals interactions	[510]
Zr‐bpy, 3D	11.2 b	8.21	8.46	32/–	1.25	—	1605	—	[515]
CPM‐736t, 3D	10.9 b	10.9	11.5	25.2/24.7	1.25	—	2087	Pore size, van der Waals interactions	[513]
Zr‐BPDC, 3D	10.9 b	8.8	8.4	47/46	1.2	—	2094	Van der Waals interactions	[510]
FDMOF‐1, 3D	6.9 b	5.45	5.31	28.5/27.1	1.19	298 K	1176	—	[507]
UiO‐67, 3D	11.2 b	9.38	9.8	36.3/–	1.09	—	1775	—	[515]
Zn‐DMOF, 3D	7.5 b	6.79	7.23	24.7/23.1	1.06	298 K	1880	—	[507]
CPM‐734t, 3D	8.6 b	8.4	9.33	30.6/28.7	0.99	—	1727	Pore size, van der Waals interactions	[513]
ZIF‐67, 3D	3.3 × 3.3	3.5	3.5	—	—	He flow at 298 K	1500	Flexibility	[505, 516]
ZU‐609, 3D	4.2 × 5.1	2.34	0.13	43/–	—	N2 flow at 298 K	380	Molecular sieving	[80]

^a For v/v = 1/99 equimolar mixture, predicted by IAST.

^b Pore size, BET surface area calculated from the N₂ sorption isotherm at 77K.

TABLE 13.

C1‐C3 multicomponent gas mixtures separations in reticular sorbents. The following parameters are listed for comparison: pore size; BET surface area (S_BET); single‐component gas uptakes; adsorption enthalpies (Q _st); adsorption selectivity; regeneration temperature; and attributed mechanisms.

Adsorbent, network dimensionality (nD)	Pore size (Å)	Gas uptake at 1 bar (mmolg−1)	Q st (kJmol−1)	Components	Selectivity a	Regeneration temperature (K)	SBET (m2g−1)	Mechanism	Refs.
TIFSIX‐2‐Cu‐i, 3D	5.1 × 5.1	CO2: 4.27 C2H2: 4.38 C2H4: 2.75 C2H6: 2.2	CO2: 35.8 C2H2: 46.3 C2H4: 35.9 C2H6: 34.5	C2H2/C2H4 (1/1)	48.8	He flow at 333 K	685	H‐bonding	[70]
				C2H2/C2H6 (1/1)	97.8
				C2H2/CO2 (1/1)	6.1
SIFSIX‐3‐Ni, 3D	4.2 × 4.2	CO2: 2.7 C2H2: 3.59 C2H4: 1.98 C2H6: 1.53	CO2: 50.9 C2H2: 36.7 C2H4: 31.7 C2H6: 23.7	CO2/C2H2 (1/1)	6.9	He flow at 333 K	230	Electrostatics	[70]
				CO2/C2H4 (1/1)	103
				CO2/C2H6 (1/1)	308
Zn‐atz‐ipa, 3D	5.6 × 5.6	CO2: 1.90 C2H2: 1.99 C2H4: 1.80 C2H6: 1.81	CO2: 31.5 C2H2: 37.5 C2H4: 40.0 C2H6: 45.8	C2H6/C2H2 (1/1)	2	He flow at 333 K	650	H‐bonding	[70]
				C2H6/C2H4 (1/1)	1.7
				C2H6/CO2 (1/1)	5
MCOF‐1, 3D	6.4	CH4: 0.40 C2H4: 1.61 C2H6: 1.96 C3H8: 2.46	15 19 41 –	C2H4/CH4 (1/1)	26	—	874	Pore size	[517]
				C2H6/CH4 (1/1)	88
				C3H8/CH4 (1/1)	1800
UTSA‐34b, 3D	12.8	C2H2: 5.36 CH4: 0.89 C2H4: 4.24 C2H6: 4.24	C2H2: 49 CH4: 20 C2H4: 31 C2H6: 31	C2H2/CH4 (1/1)	24	—	991.4	UMCs	[425]
				C2H4/CH4 (1/1)	17
				C2H6/CH4 (1/1)	17
UTSA‐35a, 3D	7.7 × 5.8	CH4: 0.31 C2H2: 3.85 C2H4: 2.71 C2H6: 3.26 C3H6: 6.18 C3H8: 5.84	17.2 29.4 28.0 30.0 33 41.8	C3H6/CH4 (1/1)	80	—	742.7	Van der Waals interactions and pore size	[411]
				C3H8/CH4 (1/1)	75
				C2H2/CH4 (1/1)	21
				C2H4/CH4 (1/1)	8
				C2H6/CH4 (1/1)	15
[La(BTB)H2O], 3D	10	(At 10kPa) CH4: 0.31 CO2: 3.85 C2H4: 2.71 C2H6: 3.26	(At 10kPa) 0.38 8.9 6.6 6.2	CO2/CH4 (1/1)	—	—	1024	UMCs	[518]
				C2H4/CH4 (1/1)	12 (273K)
				C2H6/CH4 (1/1)	22 (273K)
TJT‐100, 3D	8.7 × 11.6	C2H4: 3.4 C2H6: 3.66	C2H2: 31 C2H4: 29 C2H6: 25	C2H6/C2H4 (1/99)	1.2	He flow at 423 K	890	H‐bonding and van der Waals interactions	[422]
				C2H2/C2H4 (1/99)	1.8
FNU‐2, 3D	6, 8, 12 b	CH4: 0.59 C2H6: 1.58 C3H8: 1.51	CH4: 6 C2H6: 10.5 C3H8: 12.4	C3H8/CH4 (5/85)	638.9	—	193.3	Van der Waals interactions	[519]
				C2H6/CH4 (10/85)	43.9
ZNU‐6, 3D	8.22, 10.76	CO2: 4.76 C2H2: 8.06 C2H4: 4.76	CO2: 37.2 C2H2: 37.1 C2H4: 29	C2H2/C2H4	8.19	Ar flow at 393 K	1330.3	π···π interactions	[520]
				CO2/C2H4	7.84
NKCOF‐62, 2D	8 × 8	C2H2: 2.54 C2H4: 1.88 C2H6: 2.1	C2H2: 30.5 C2H4: 26.3 C2H6: 29.6	C2H2/C2H4 (1/1)	1.3	—	714	Pore size, noncovalent interactions	[413]
				C2H6/C2H4 (1/1)	1.37
MOF‐303, 3D	5.8 c	C2H2: 7.91 C2H6: 5.01	C2H2: 31.7 C2H4: 24.3 C2H6: 25.1	C2H2/C2H4 (1/99)	2.4	—	1244	H‐bonding, C‐H···π interactions	[521]
				C2H6/C2H4 (1/1)	1.7
FJI‐W‐66a, 3D	5.2 × 4.4	C2H2: 1.88 C2H4: 1.29 C2H6: 1.32	C2H2: 42.36 C2H4: 38.49 C2H6: 40.48	C2H2/C2H4 (1/99)	2.31	—	329	Flexibility	[522]
				C2H6/C2H4 (1/99)	1.4
Zn‐fa‐atz, 3D	5.5 × 4.9	CO2: 2.8 C2H2: 2.65 C2H4: 2.1 C2H6: 2.05	CO2: 30.2 C2H2: 30.6 C2H4: 29.3 C2H6: 35.9	CO2/C2H4 (1/1)	1.4	—	—	H‐bonding	[523]
				C2H2/C2H4 (1/1)	1.5
				C2H6/C2H4 (1/1)	1.4
NTUniv‐58, 3D (308 K)	4‐6 b	C2H2: 1.85 C2H4: 1.52 C2H6: 1.43	C2H2: 35.1 C2H4: 33.9	C2H2/C2H4 (1/99)	6.2	—	442	H‐bonding, π···π interactions	[524]
NTUniv‐59, 3D (308 K)	4‐6 b	C2H2: 1.85 C2H4: 1.43 C2H6: 1.46	C2H2: 39 C2H4: 32	C2H2/C2H4 (1/99)	17.2	—	258	H‐bonding, π···π interactions	[524]
LIFM‐XYY‐1, 3D	—	C2H2: 1.85 C2H4: 1.43 C2H6: 1.46	C2H2: 29.6 C2H4: 28.8 C2H6: 29.5	C2H6/C2H4 (1/1)	1.51	He flow at 358 K	1343	H‐bonding, C‐H···π interactions	[383]
				C2H2/C2H4 (1/99)	1.07
LIFM‐XYY‐2, 3D	—	C2H2: 3.26 C2H4: 2.92 C2H6: 3.65	C2H2: 33.1 C2H4: 28.8 C2H6: 29.5	C2H6/C2H4 (1/1)	1.48	He flow at 358 K	1724	H‐bonding, C‐H···π interactions	[383]
				C2H2/C2H4 (1/99)	1.51
LIFM‐XYY‐3, 3D	—	C2H2: 2.73 C2H4: 2.10 C2H6: 2.53	C2H2: 31.0 C2H4: 28.5 C2H6: 29.6	C2H6/C2H4 (1/1)	1.33	He flow at 358 K	1274	H‐bonding, C‐H···π interactions	[383]
				C2H2/C2H4 (1/99)	1.42
LIFM‐XYY‐4, 3D	—	C2H2: 3.54 C2H4: 2.74 C2H6: 3.89	C2H2: 27.2 C2H4: 25.0 C2H6: 29.3	C2H6/C2H4 (1/1)	1.68	He flow at 358 K	1743	H‐bonding, C‐H···π interactions	[383]
				C2H2/C2H4 (1/99)	1.41
LIFM‐XYY‐5, 3D	—	C2H2: 3.07 C2H4: 2.56 C2H6: 3.39	C2H2: 25.9 C2H4: 18.3 C2H6: 26.7	C2H6/C2H4 (1/1)	1.53	He flow at 358 K	1694	H‐bonding, C‐H···π interactions	[383]
				C2H2/C2H4 (1/99)	1.31
LIFM‐XYY‐6, 3D	—	C2H2: 2.97 C2H4: 2.24 C2H6: 2.99	C2H2: 37.1 C2H4: 31.5 C2H6: 32.2	C2H6/C2H4 (1/1)	1.63	He flow at 358 K	1240	H‐bonding, C‐H···π interactions	[383]
				C2H2/C2H4 (1/99)	1.53
LIFM‐XYY‐7, 3D	—	C2H2: 4.86 C2H4: 4.07 C2H6: 5.23	C2H2: 22.3 C2H4: 23.0 C2H6: 24.9	C2H6/C2H4 (1/1)	1.50	He flow at 358 K	2079	H‐bonding, C‐H···π interactions	[383]
				C2H2/C2H4 (1/99)	1.24
LIFM‐XYY‐8, 3D	—	C2H2: 3.81 C2H4: 3.30 C2H6: 4.21	C2H2: 25.9 C2H4: 26.7 C2H6: 27.4	C2H6/C2H4 (1/1)	1.47	He flow at 358 K	2104	H‐bonding, C‐H···π interactions	[383]
				C2H2/C2H4 (1/99)	1.14
LIFM‐28, 3D	11.1 × 11.1	C2H2: 0.87 C2H4: 0.78 C2H6: 0.97	C2H2: 26.6 C2H4: 26.0 C2H6: 28.8	C2H6/C2H4 (1/1)	1.24	He flow at 358 K	862	—	[383]
				C2H2/C2H4 (1/99)	1.04
PCN‐700, 3D	5.6 × 5.6	C2H2: 0.94 C2H4: 0.87 C2H6: 1.28	C2H2: 25.6 C2H4: 23.6 C2H6: 26.3	C2H6/C2H4 (1/1)	1.53	He flow at 358 K	849	—	[383]
				C2H2/C2H4 (1/99)	1.07
NPU‐1, 3D	7.4 c	C2H2: 5.09 C2H4: 4.2 C2H6: 4.5	C2H2: 27.8 C2H4: 23.9 C2H6: 29.1	C2H6/C2H4 (1/1)	1.32	He flow at 313 K	1580	H‐bonding	[321]
				C2H2/C2H4 (1/99)	1.40
Azole‐Th‐1, 3D	9.2 c	C2H2: 3.63 C2H4: 3.62 C2H6: 4.47	C2H4: 26.1 C2H6: 28.6	C2H6/C2H4 (1/1)	1.46	He flow at 298 K	983	C‐H···π interactions	[406]
				C2H2/C2H4 (1/99)	1.00
UiO‐67(NH2)2 , 3D	5.2 × 5.2, 7.1 × 7.1, 11.5 × 11.5	C2H2: 5.9 C2H4: 4.32 C2H6: 5.32	C2H2: 27.4 C2H4: 24.5 C2H6: 26.5	C2H6/C2H4 (1/1)	1.70	—	2815	H‐bonding, C‐H···π interactions	[376]
				C2H2/C2H4 (1/99)	2.10
UPC‐612, 3D	14, 20 c	C2H2: 3.0 C2H4: 2.92 C2H6: 3.58	C2H2: 23.94 C2H4: 16.94 C2H6: 22.39	C2H6/C2H4 (1/1)	1.40	—	2016	—	[401]
				C2H2/C2H4 (1/99)	1.08
UPC‐613, 3D	7, 12 c	C2H2: 2.83 C2H4: 2.31 C2H6: 2.55	C2H2: 30.4 C2H4: 28.5 C2H6: 31.8	C2H6/C2H4 (1/1)	1.47	—	853	—	[401]
				C2H2/C2H4 (1/99)	1.38
NUM‐9a, 3D	4.5 × 4.5	C2H2: 2.33 C2H4: 2.23 C2H6: 2.48	C2H2: 35.8 C2H4: 32.3 C2H6: 35.8	C2H6/C2H4 (10/90)	1.62	—	330	H‐bonding, C‐H···π interactions	[385]
				C2H2/C2H4 (1/99)	1.48
Zn(ad)(int), 3D	8 c	C2H2: 3.04 C2H4: 2.21 C2H6: 2.32	C2H2: 34.7 C2H4: 29 C2H6: 33.3	C2H6/C2H4 (1/1)	2.4	—	617	H‐bonding, π···π, C‐H···π interactions	[331]
				C2H2/C2H4 (1/99)	1.61
CuTiF6‐TPPY, 3D	5 × 5, 8 × 8	C2H2: 3.62 C2H4: 2.42 C2H6: 2.82	C2H2: 36.5 C2H4: 29.6 C2H6: 34.2	C2H6/C2H4 (1/1)	2.12	He flow a 298 K	685	H‐bonding, C‐H···π, van der Waals interactions	[338]
				C2H2/C2H4 (1/99)	5.03
UPC‐66‐a, 3D (293 K)	6.01 × 13.84, 7.04	C2H2: 3.18 C2H4: 2.41 C2H6: 2.76	C2H2: 17.6 C2H4: 17.8 C2H6: 15.6	C2H6/C2H4 (1/1)	1.65	—	483	C‐H···π interactions	[381]
				C2H2/C2H4 (1/1)	1.05
[Zn(BDC)(H2BPZ)] n , 3D	8‐10.2	C2H2: 4.46 C2H4: 2.59 C2H6: 3.63	C2H2: 28.7 C2H4: 23.2 C2H6: 31.8	C2H6/C2H4 (1/1)	2.20	—	906.5	H‐bonding, C‐H···π interactions	[336]
				C2H2/C2H4 (1/99)	1.60
Zn‐atz‐oba, 3D	3.2, 4.4	C2H2: 2.77 C2H4: 2.02 C2H6: 2.04	C2H2: 27.5 C2H4: 27 C2H6: 30	C2H6/C2H4 (1/1)	1.27	He flow at 333 K	710.7	H‐bonding	[419]
				C2H2/C2H4 (1/99)	1.43
Zn‐ATA, 3D	4.1 × 3.8	C2H2: 2 C2H4: 1.20 C2H6: 1.08	C2H2: 24.5 C2H4: 27.4 C2H6: 32.5	C2H6/C2H4 (1/1)	1.84	N2 flow at 373 K	344	H‐bonding, C‐H···π interactions	[363]
				C2H2/C2H4 (1/99)	1.81
ZJNU‐7, 3D	5.0, 5.9, 7.3	C2H2: 5.04 C2H4: 3.8 C2H6: 4.13	C2H2: 34.1 C2H4: 29.3 C2H6: 29.7	C2H6/C2H4 (1/1)	1.56	—	1180	H‐bonding, C‐H···π interactions, π‐complexion	[397]
				C2H2/C2H4 (1/99)	2.31
ZJNU‐115, 3D	8.58	C2H2: 4.73 C2H4: 3.75 C2H6: 4.21	C2H2: 29.2 C2H4: 27.7 C2H6: 28.2	C2H6/C2H4 (1/1)	1.56	—	1291	C‐H···π interactions, π‐complexion	[396]
				C2H2/C2H4 (1/99)	2.05
BFFOUR‐Cu‐dpds, 2D	3.9 × 4.0, 3.3 × 3.8	C2H4: 1.31 C2H6: 0.1 C3H6: 1.23 C3H8: 0.09 n‐C4H8: 0.96 n‐C4H10: 0.09	C2H4: 34.44 C3H6: 31.13 n‐C4H8: 33.37	C2H4/C2H6 (1/1)	68.8	He flow at 333 K	140	H‐bonding, C‐H···π interactions	[525]
				C3H6/C3H8 (1/1)	108.4
				n‐C4H8/n‐C4H10 (1/1)	22.9
NTUniv‐63, 3D	3.4 × 3.4, 4.4 × 4.4	C2H2: 2.73 C2H4: 2.07 C2H6: 2.05	C2H2: 34.2 C2H4: 32.7 C2H6: 33.9	C2H2/C2H4 (1/99)	7.3	—	518	H‐bonding, N‐H···π interactions	[526]
				C2H6/C2H4 (1/99)	1.7
FJI‐Y9, 3D	9‐12 c	C2H2: 4.51 C2H4: 3.93 C2H6: 4.69 C3H6: 5.85	C2H2: 27.2 C2H4: 23.7 C2H6: 24.5 C3H6: 28.7	C2H2/C2H4 (1/99)	3.8	—	1420	H‐bonding, C‐H···π interactions	[527]
				C2H6/C2H4 (10/90)	1.79
Zr‐TCA, 3D	6 × 7	C2H2: 2.78 C2H4: 2.02 C2H6: 2.28	C2H2: 43.8 C2H4: 23.9 C2H6: 35.3	C2H2/C2H4 (1/99)	5.64	—	350.9	H‐bonding, C‐H···π, polarizability	[528]
				C2H6/C2H4 (10/90)	2.72
CAU‐23, 3D	6.3 × 7.6	C2H2: 4.7 C2H4: 3.8 C2H6: 4.0	C2H2: 32.8 C2H4: 25.4 C2H6: 28.4	C2H2/C2H4 (1/15)	1.50	He flow at 298 K	1240	Van der Waals interactions	[529]
				C2H6/C2H4 (1/1)	1.54
3D‐IL‐COF‐1, 3D	8.3	—	—	CO2/N2 (1/1)	24.6	—	517	—	[139]
				CO2/CH4 (1/1)	23.1
3D‐IL‐COF‐2, 3D	10.7	—	—	CO2/N2 (1/1)	24.0	—	653	—	[139]
				CO2/CH4 (1/1)	22.3
3D‐IL‐COF‐3, 3D	12.4	—	—	CO2/N2 (1/1)	24.4	—	870	—	[139]
				CO2/CH4 (1/1)	21.5
3D‐COF‐1a, 3D	—	—	—	CO2/N2 (1/1)	7.1	—	596	—	[139]
				CO2/CH4 (1/1)	5.3
3D‐IL‐COF‐1b, 3D	—	—	—	CO2/N2 (1/1)	43.6	—	537	—	[139]
				CO2/CH4 (1/1)	35.1
TpTa‐NO2 , 2D	15 × 15	CO2: 2.02 C2H2: 2.85 C2H4: 1.4	CO2: 36.26 C2H2: 42.57 C2H4: 33.07	CO2/N2 (15/85)	125.23	—	398	H‐bonding, van der Waals interactions	[530]
				C2H2/C2H4 (1/99)	37.96
				C2H2/CO2 (10/90)	10.58
				CO2/C2H4 (10/90)	2.45
CMOM‐7, 3D	5.3 b , 8.9 c	CH4: 0.68 C2H6: 3.10 C3H8: 3.40	CH4: 20.9 C2H6: 27.9 C3H8: 31.2	C3H8/C2H6 (1/1)	13.3	He flow at 298 K	861.38	H‐bond, C‐H···π interactions	[531]
				C2H6/CH4 (1/1)	12.5
				C3H8/CH4 (1/1)	40.1
GNU‐1, 3D	2.3 × 11.2	CH4: 1.12 C2H6: 4.6 C3H8: 6.64	CH4: 14.7 C2H6: 29.2 C3H8: 35.5	C3H8/CH4 (5/85)	330.1	—	1297	H‐bonding, electrostatics, C‐H···π interactions	[532]
				C2H6/CH4 (10/85)	17.5
MOF‐303, 3D	5‐7	CH4: 0.86 C2H6: 4.96 C3H8: 4.74	CH4: 19 C2H6: 24 C3H8: 34	C3H8/CH4 (5/85)	5114	N2 flow at 333 K	1220	H‐bonding, pore size	[533]
				C2H6/CH4 (10/85)	26
MIL‐160, 3D	5‐7	CH4: 0.94 C2H6: 4.65 C3H8: 5.08	CH4: 19 C2H6: 28 C3H8: 35	C3H8/CH4 (5/85)	174	N2 flow at 333 K	1188	H‐bonding, pore size	[533]
				C2H6/CH4 (10/85)	20
UiO‐66, 3D	8.6, 11.1	CH4: 0.60 C2H6: 1.67 C3H8: 1.79	C2H6: 26.76 C3H8: 32.69	C3H8/CH4 (1/1)	65	He flow at 373 K	1305	C‐H···π interactions	[534]
				C2H6/CH4 (1/1)	8
UiO‐66‐Naph, 3D	5.5, 10.5	CH4: 0.38 C2H6: 1.24 C3H8: 1.39	C2H6: 28.29 C3H8: 37.92	C3H8/CH4 (1/1)	741	He flow at 373 K	881	C‐H···π interactions	[534]
				C2H6/CH4 (1/1)	32
DUT‐52, 3D	10.3, 17.3	CH4: 0.54 C2H6: 1.89 C3H8: 2.21	C2H6: 24.57 C3H8: 28.89	C3H8/CH4 (1/1)	48	He flow at 373 K	1641	C‐H···π interactions	[534]
				C2H6/CH4 (1/1)	5
UiO‐66‐Anth, 3D	5.7, 10.1	CH4: 0.21 C2H6: 0.70 C3H8: 0.90	C2H6: 28.95 C3H8: 35.16	C3H8/CH4 (1/1)	535	He flow at 373 K	676	C‐H···π interactions	[534]
				C2H6/CH4 (1/1)	32
CFA‐1, 3D	10‐12 c	CH4: 0.21 C2H6: 0.70 C3H8: 0.90	CH4: 18.3 C2H6: 29.3 C3H8: 33.1	C3H8/CH4 (5/85)	11.7	—	1903.3	Polarizability, H‐bonding	[535]
				C2H6/CH4 (10/85)	225.1
CFA‐1‐NiCl2‐2.3, 3D	10‐12 c	CH4: 0.21 C2H6: 0.70 C3H8: 0.90	CH4: 16.7 C2H6: 28.0 C3H8: 32.3	C3H8/CH4 (5/85)	15.2	—	2090.7	Polarizability, H‐bonding	[535]
				C2H6/CH4 (10/85)	382.7
Ni‐FDMOF, 3D	7.2 c	CH4: 0.67 C3H6: 8.08 C3H8: 8.08	CH4: 16.6 C3H6: 32.8 C3H8: 35.2	C3H8/CH4	123.8	423 K	1110	H‐bonding	[536]
				C3H6/CH4	79.5
Zn‐BPZ‐SA, 3D	6.4‐8.4	C2H4: 2.85 C3H6: 3.05 C2H6: 2.97 C3H8: 2.73	C2H4: 23.13 C3H6: 33.65 C2H6: 26.4 C3H8: 32.7	C3H6/C2H4 (1/1)	4.8	—	925	H‐bonding, C‐H···π, π···π interactions	[537]
				C3H8/C2H6 (1/1)	4.3
Fe‐pyz, 3D	3.8 × 3.2	CH4: 0.75 C2H6: 1.67 C3H8: 2.75	CH4: 16 C2H6: 52 C3H8: 32	C3H8/CH4 (5/85)	89	—	361	H‐bonding, van der Waals interactions	[538]
				C2H6/CH4 (10/85)	23
Co‐pyz, 3D	3.9 × 3.3	CH4: 1.00 C2H6: 1.67 C3H8: 1.63	CH4: 20 C2H6: 41 C3H8: 50	C3H8/CH4 (5/85)	78	—	458	H‐bonding, van der Waals interactions	[538]
				C2H6/CH4 (10/85)	22
Ni‐pyz, 3D	3.8 × 3.9	CH4: 2.94 C2H6: 3.13 C3H8: 3.31	CH4: 23 C2H6: 38 C3H8: 70	C3H8/CH4 (5/85)	53	—	572	H‐bonding, van der Waals interactions	[538]
				C2H6/CH4 (10/85)	18
Zn(bdc)(ted)0.5 , 3D	7.5 × 7.5, 4.8 × 3.2	CH4: 0.70 C2H6: 4.9 C3H8: 6.6 C4H10: 6.9	CH4: 19.8 C2H6: 21.3 C3H8: 24.2 C4H10: 29.6	C4H10/CH4 (1/1)	180	—	1904	Van der Waals interactions	[539]
				C3H8/CH4 (1/1)	67
				C2H6/CH4 (1/1)	13
TIFSIX‐Cu‐TPA, 3D	2.4 × 2.4	CH4: 0.68 C2H6: 4.4 C3H8: 4.91	CH4: 18.5 C2H6: 27.3 C3H8: 29.4	C3H8/CH4 (1/1)	68.6	Ar flow at 373 K	—	Van der Waals interactions	[540]
				C2H6/CH4 (1/1)	16.2
Co‐MOF, 3D	4.3 × 4.3	CH4: 0.74 C2H6: 2.62 C3H8: 2.65	CH4: 29.67 C2H6: 37.17 C3H8: 38.21	C3H8/CH4 (1/1)	290	He flow at 373 K	345.8	H‐bonding, C‐H···π, van der Waals interactions	[541]
				C2H6/CH4 (1/1)	26
SU‐100, 3D	7.1 × 7.6	CH4: 0.66 CO2: 2.24 C2H2: 2.63 C2H6: 1.69	CH4: 23.3 CO2: 32.3 C2H2: 41.7 C2H6: 33.2	C2H2/CH4 (1/1)	48.2	He flow at 343 K	388	H‐bonding, C‐H···π, π···π interactions	[542]
				C2H6/CH4 (1/1)	39.3
				CO2/CH4 (1/1)	38.7
Cu‐IPA, 3D	9.3 c	CH4: 0.81 C2H6: 2.57 C3H8: 3.10	CH4: 11.4 C2H6: 39.3 C3H8: 43.9	C3H8/CH4 (5/85)	296	373 K	640.2	Van der Waals interactions	[543]
				C2H6/CH4 (10/85)	35
Cd(II)‐MOF, 3D	4.2 c	CH4: 0.73 C2H6: 3.19 C3H8: 3.02	CH4: 27.7 C2H6: 33.3 C3H8: 50.5	C3H8/CH4 (5/85)	233.8	—	327.9	H‐bonding, van der Waals interactions	[544]
				C2H6/CH4 (10/85)	34.3
(Me2NH2) [Ni3(μ 3‐OH) (CF3‐BPDC‐CF3)3(tpt)] n , 3D	11.5 × 11.5, 9.3 × 9.3	CH4: 0.46 C2H6: 2.78 C3H8: 5.09	CH4: 7.74 C2H6: 20.2 C3H8: 29.2	CH4/C3H8 (85/5)	60.1	—	1495	H‐bonding, C‐H···π, interactions	[545]
				CH4/C2H6 (85/10)	7.1
C‐PVDC‐800, 3D	5.2‐5.3	CH4: 1.54 C2H6: 5.29 C3H8: 5.17	CH4: 19.5 C2H6: 35.6 C3H8: 78.1	C3H8/CH4 (1/1)	3387	He flow at 423 K	1087	Kinetics	[546]
				C2H6/CH4 (1/1)	75
ANPC‐1‐800, 3D	—	CH4: 1.45 C2H6: 6.84 C3H8: 9.74	CH4: 16.7 C2H6: 23.7 C3H8: 23.3	C3H8/CH4 (1/1)	110.4	Vacuum	2836	Van der Waals interactions	[547]
				C2H6/CH4 (1/1)	14.5
ANPC‐2‐700, 3D	—	CH4: 1.12 C2H6: 4.88 C3H8: 8.80	CH4: 15.4 C2H6: 23.6 C3H8: 28.9	C3H8/CH4 (1/1)	162.5	Vacuum	2729	Van der Waals interactions	[547]
				C2H6/CH4 (1/1)	13.5
ANPC‐2‐800, 3D	—	CH4: 1.15 C2H6: 4.94 C3H8: 11.5	CH4: 15.6 C2H6: 22.6 C3H8: 26.8	C3H8/CH4 (1/1)	120.2	Vacuum	3177	Van der Waals interactions	[547]
				C2H6/CH4 (1/1)	11.9
BSF‐1, 3D	—	CH4: 0.66 C2H6: 1.17 C3H8: 0.99	CH4: 23.7 C2H6: 28.6 C3H8: 33.7	C3H8/CH4 (1/1)	353	—	535	Electrostatics, H‐bonding	[548]
				C2H6/CH4 (1/1)	23
FJI‐C4, 3D	5.9 × 5.9	CH4: 1.15 C2H6: 2.21 C3H8: 1.63	CH4: 23.1 C2H6: 40.9 C3H8: 42.9	C3H8/CH4 (1/1)	293.4	—	690	Pore size, electrostatics, π···π interactions	[426]
				C2H6/CH4 (1/1)	39.7
FJI‐H22, 3D	12.8 × 15.08, 6.05 × 11.83	CH4: 0.88 C2H6: 1.49 C3H8: 1.10	—	C3H8/CH4 (1/1)	145.23	—	483	—	[549]
				C2H6/CH4 (1/1)	11.95
JLU‐Liu5, 3D	—	CH4: 1.00 C2H6: 2.37 C3H8: 1.59	CH4: 19 C2H6: 29.8 C3H8: 21.6	C3H8/CH4 (1/1)	107.8	—	707	UMCs	[550]
				C2H6/CH4 (1/1)	17.6
JLU‐Liu6, 3D	—	CH4: 0.81 C2H6: 1.63 C3H8: 1.30	CH4: 24.9 C2H6: 46.5 C3H8: 12.3	C3H8/CH4 (1/1)	274.6	—	544	UMCs	[550]
				C2H6/CH4 (1/1)	20.4
JLU‐Liu7, 3D	5.2 × 10.8	CH4: 1.06 C2H6: 3.57 C3H8: 2.57	CH4: 20.9 C2H6: 34.8 C3H8: 28.5	C3H8/CH4 (1/1)	128.5	—	879	—	[551]
				C2H6/CH4 (1/1)	50.4
JLU‐Liu38, 3D	8.6‐11	CH4: 0.48 C2H6: 4.96 C3H8: 8.39	CH4: 17.5 C3H8: 29.2	C3H8/CH4 (1/1)	98	—	1784	Van der Waals interactions	[552]
				C2H6/CH4 (1/1)	12.5
JUC‐100, 3D	14	CH4: 0.64 C2H6: 3.07 C3H8: 3.09	CH4: 27.1 C2H6: 26.1	C3H8/CH4 (1/1)	80	—	2040	—	[553]
				C2H6/CH4 (1/1)	11
JUC‐103, 3D	10	CH4: 0.73 C2H6: 2.85 C3H8: 2.77	CH4: 23.5 C2H6: 22.6	C3H8/CH4 (1/1)	55	—	1484	—	[553]
				C2H6/CH4 (1/1)	8
JUC‐106, 3D	8	CH4: 0.51 C2H6: 2.61 C3H8: 2.59	CH4: 26.1 C2H6: 24.2	C3H8/CH4 (1/1)	75	—	1122	—	[553]
				C2H6/CH4 (1/1)	13
UTSA‐35a, 3D	3.4 × 3.8, 2.0 × 3.8, 3.1 × 4.4	CH4: 0.43 C2H6: 2.43 C3H8: 2.97	CH4: 24.4 C2H6: 36.1	C3H8/CH4 (1/1)	80	—	806	Molecular sieving	[554]
				C2H6/CH4 (1/1)	20
MFM‐202a, 3D	9 × 9	CH4: 0.45 C2H6: 4.21 C3H8: 6.76	CH4: 19 C2H6: 21 C3H8: 39	C3H8/CH4 (1/1)	87	—	2220	—	[555]
				C2H6/CH4 (1/1)	10
InOF‐1, 3D	7.4 × 7.4	CH4: 0.64 C2H6: 4.14 C3H8: 4.25	CH4: 18.8 C2H6: 25.1 C3H8: 31.4	C3H8/CH4 (1/1)	90	—	1109	Electrostatics	[556]
				C2H6/CH4 (1/1)	17
RT‐MIL‐100(Fe), 3D	—	CH4: 0.36 C2H6: 2.22 C3H8: 6.78	CH4: 19 C2H6: 23 C3H8: 26	C3H8/CH4 (5/85)	33.3	—	2482	H‐bonding	[557]
				C2H6/CH4 (10/85)	6
MIL‐101‐Cr, 3D	10.5, 16.9, 32.2	CH4: 0.49 C2H6: 1.59 C3H8: 3.35	C2H6: 22.2 C3H8: 29.2	C3H8/CH4 (1/1)	84.3	He flow at 373 K	2961	H‐bonding, C‐H···π interactions	[558]
				C2H6/CH4 (1/1)	22.5
MIL‐101‐Fe, 3D	11, 15.5, 32.2	CH4: 0.45 C2H6: 1.25 C3H8: 3.29	C2H6: 25.6 C3H8: 34.5	C3H8/CH4 (1/1)	24.9	He flow at 373 K	2617	H‐bonding, C‐H···π interactions	[558]
				C2H6/CH4 (1/1)	15.4
MIL‐101‐Fe‐NH2 , 3D	11, 16.8, 24.4, 31	CH4: 0.46 C2H6: 1.35 C3H8: 3.32	C2H6: 24.6 C3H8: 28.3	C3H8/CH4 (5/85)	42.5	He flow at 373 K	2648	H‐bonding, C‐H···π interactions	[558]
				C2H6/CH4 (10/85)	11.6
A‐AC‐3, 3D	20‐40 c	CH4: 1.38 C2H6: 7.09 C3H8: 11.34	—	C3H8/CH4 (1/4)	76.6	—	2928	—	[559]
				C2H6/CH4 (1/4)	16.9
A‐AC‐4, 3D	20‐40 c	CH4: 1.18 C2H6: 6.59 C3H8: 11.76	—	C3H8/CH4 (1/4)	88.8	—	3131	—	[559]
				C2H6/CH4 (1/4)	15.1
sPI‐A‐H, 3D	3.5, 4.8 b	CH4: 0.39 C2H6: 1.59 C3H8: 2.00	—	C3H8/CH4 (1/1)	66.7	—	665	Polarizability, π···π interactions	[560]
				C2H6/CH4 (1/1)	13.3
UiO‐67, 3D	10.9, 13.58	CH4: 0.56 C2H6: 4.26 C3H8: 9.50	CH4: 22 C2H6: 32.3 C3H8: 47.5	C3H8/CH4 (5/85)	73.7	—	2591	—	[561]
				C2H6/CH4 (10/85)	8.1
GTGU‐15, 3D	13.8, 20 c	CH4: 0.40 C2H6: 2.13 C3H8: 12.13	CH4: 19 C2H6: 21.3 C3H8: 21.7	C3H8/CH4 (5/85)	32.61	Vacuum at 393 K	3164	H‐bonding	[562]
				C2H6/CH4 (10/85)	5.12
JLU‐Liu22, 3D	6.8 × 6.8	CH4: 0.71 C2H6: 3.3 C3H8: 4.15	CH4: 11 C2H6: 30.5 C3H8: 30.3	C3H8/CH4 (5/85)	144.89	—	1487	UMCs, Polarizability	[563]
				C2H6/CH4 (10/85)	5.5
MIL‐142A, 3D	7, 10	CH4: 0.54 C2H6: 3.82 C3H8: 5.32	CH4: 13.7 C2H6: 25.5 C3H8: 26.6	C3H8/CH4 (5/85)	85.51	Vacuum at 423 K	1425	Van der Waals interactions	[564]
				C2H6/CH4 (10/85)	8.6
Ni(TMBDC)(DABCO)0.5 , 3D	5 c	C2H6: 5.81 C3H8: 5.54	CH4: 14 C2H6: 36 C3H8: 59	C3H8/CH4 (5/85)	274	N2 flow at 373 K	940	Van der Waals interactions	[565]
				C2H6/CH4 (10/85)	29
NKM‐101, 3D	5.1 × 5.1	—	CH4: 17.2 C2H6: 25 C3H8: 23.8	C3H8/CH4 (5/85)	189.66	—	552	—	[566]
				C2H6/CH4 (10/85)	18.19
ZUL‐C2, 3D	4.76 × 4.76	C2H6: 2.82 C3H8: 2.52	CH4: 71 C2H6: 45 C3H8: 5.32	C3H8/CH4 (5/85)	741	—	462	H‐bonding, van der Waals interactions	[567]
				C2H6/CH4 (10/85)	82
ZUL‐C1, 3D	4.99 × 4.99	C2H6: 2.95 C3H8: 2.72	CH4: 54 C2H6: 33 C3H8: 23	C3H8/CH4 (5/85)	158	—	504	H‐bonding, van der Waals interactions	[567]
				C2H6/CH4 (10/85)	28
BSF‐2, 3D	—	CH4: 0.48 C2H6: 1.22 C3H8: 1.77	CH4: 23.5 C2H6: 32.8 C3H8: 39.7	C3H8/CH4 (5/85)	681	He flow at 313 K	403	H‐bonding	[568]
				C2H6/CH4 (10/85)	25

^a IAST selectivity at 1 bar 298 K.

^b Pore size distribution determined from Horvath–Kawazoe method applied on CO₂ isotherm at 195 K.

^c Pore size distribution calculated based on NLDFT model from N₂ sorption at 77K.

^d Henry's selectivity

TABLE 14.

C4 isomers (C₄H₆/i‐C₄H₈/n‐C₄H₈/cis‐C₄H₈/i‐C₄H₁₀/n‐C₄H₁₀) binary or multicomponent separations. The following parameters are listed for comparison: pore size; BET surface area (S_BET); single‐component gas uptakes; adsorption enthalpies (Q _st); uptake ratio; analysis method, regeneration temperature; and attributed mechanisms. Reticular sorbents are listed in decreasing order of binary or multicomponent uptake ratio for C4 isomers.

Adsorbent, network dimensionality (nD)	Pore size (Å)	Components	Uptake ratio (v/v = 1/1)	Capacity (mmol/g)	Q st (kJ/mol)	Analyzation method	Regeneration temperature	SBET (m2/g)	Mechanism	Refs.
Y‐fcu‐MOF, 3D	4.7 × 4.7	n‐C4H10/ i‐C4H10	+ꝏ	n‐C4H10: 1.97 i‐C4H10: 0.05	—	Breakthrough	—	691 a	Molecular sieving	[569]
CALF‐20, 3D	4.68	n‐C4H8/i‐C4H8	1.67 × 107	n‐C4H8: 2.87 i‐C4H8: 0.11	n‐C4H8: 86.49 i‐C4H8: 0.21	Breakthrough	—	—	Molecular sieving, H‐bonding	[570]
IPM‐101, 3D	3.9 × 3.9	n‐C4H10/i‐C4H10	2241 e	n‐C4H8: 2.3	n‐C4H8: 39.8	Breakthrough, sorption	373 K	342.8	H‐bonding	[571]
ZU‐36‐Co (GeFSIX‐3‐Co), 3D	3.82 × 3.82, 5.25 × 5.25	n‐C4H8/i‐C4H8	2050 f	n‐C4H8: 2.35 i‐C4H8: 0.17 n‐C4H10: 2.20 i‐C4H10: 0.12	—	Breakthrough, sorption	He flow at 373 K	257.2	H‐bonding; Van der Waals interactions	[572]
		n‐C4H10/i‐C4H10	140.4 f		—
ZIF‐8/DMPU‐water slurry, 3D	11.6, 3.4 [573]	i‐C4H10/n‐C4H10	890 g	n‐C4H10: 4.0	—	Breakthrough, sorption	323.15 K	—	Kinetics	[574]
NUIG2, 2D	5.5, 3.7	C4H10/C3H8	2.21	—	C4H10: 38.91	Sorption	—	—	Hydrophobic interactions	[575]
		C4H10/C2H2	20.9
		C4H10/CH4	648
TMOF‐1, 3D	4.5 × 4.5	C4H6/i‐C4H8	519.2 e	n‐C4H8: 1.05 C4H6: 1.65 i‐C4H8: 0.2	—	Breakthrough, sorption	He flow at 423 K	318	H‐bonding	[576]
		n‐C4H8/i‐C4H8	5.42 (93.2 e )
		C4H6/n‐C4H8	1.5 (5.95 e )
MnINA, 3D	4.62	n‐C4H8/i‐C4H8	327.7	n‐C4H8: 1.79 i‐C4H8: 1.76	—	Breakthrough, sorption	He flow at 373 K	—	C−H···π interactions	[577]
ZU‐601, 3D	6.7 × 2.2	C4H6/n‐C4H8	207	C4H6: 2.9 n‐C4H8: 0.014 trans‐C4H8: – iso‐C4H8: 0.022	—	Breakthrough, sorption	Vacuum at 393 K	223	H‐bonding	[578]
		C4H6/trans‐C4H8	—
FJI‐H38, 3D	8.74 × 10.54	n‐C4H10/i‐C4H10	246.75	n‐C4H10:2.39 i‐C4H10: ∼0	n‐C4H10: 20.61	Breakthrough	—	296.2	H‐bonding	[579]
Znpyc‐CF3 , 3D	7.8 d	n‐C4H10/i‐C4H10	187	n‐C4H10: 2.48 i‐C4H10: 1.17	n‐C4H10: 34.8 i‐C4H10: 25.2	Breakthrough	—	948	H‐bonding	[580]
ZU‐619, 3D	9.4 × 9.4	C4H6/i‐C4H8	72.58 e	n‐C4H8: 0.68 C4H6: 1.58 i‐C4H8: 0.32	—	Breakthrough, sorption	He flow at 423 K	154	H‐bonding	[576]
		n‐C4H8/i‐C4H8	1.89 (6.36 e )
		C4H6/n‐C4H8	2.55 (11.42 e )
Y‐abtc‐160, 3D	—	C4H6/n‐C4H8	40.93	C4H6: 2.21 n‐C4H8: 0.054 i‐C4H8: 0.035	—	Breakthrough, sorption	—	332.2	Pore size, H‐bonding, π⋅⋅⋅π interactions	[581]
		C4H6/i‐C4H8	63.14
Mn‐bpdc MOF, 3D	4.2 × 6.1	C4H6/n‐C4H8	40.0	C4H6: 0.04 n‐C4H8:0.98 i‐C4H8:0.03 n‐C4H10:0.01 i‐C4H10:∼0	—	Breakthrough, sorption	298 K	159	H‐bonding	[582]
		C4H6/i‐C4H8	45.0
COF‐300 (dia‐c5), 3D	5.9 × 5.9	i‐C4H8/trans‐C4H8	38.4	n‐C4H8: 7.8 trans‐C4H8: 1.1 i‐C4H8: 10.7 n‐C4H10: 6.0 i‐C4H10: 8.0 C4H6: 9.3	n‐C4H8: 13.1 trans‐C4H8: 2.6 cis‐C4H8: 2.6 i‐C4H8: 20.0 n‐C4H10: 11.8 i‐C4H10: 16.5 C4H6: 15.5	Sorption	—	5546	Pore shape, π···π interactions	[583]
		i‐C4H8/cis‐C4H8	35.8
		i‐C4H8/n‐C4H10	13.4
		i‐C4H8/n‐C4H8	6.3
		i‐C4H8/i‐C4H10	2.6
		i‐C4H8/C4H6	2.5
[Ag2(o‐Hmpba)2(o‐H2mpba)2], 2D	1.8 × 4.4	n‐C4H10/i‐C4H10	23	n‐C4H10: 0.9 i‐C4H10: 0.012	—	Breakthrough, sorption	—	0.99	Molecular‐Sieving, flexibility	[584]
BFFOUR‐Cu‐dpds, 2D	3.9 × 4.0	n‐C4H8/n‐C4H10	22.9 f	n‐C4H8: 0.70	n‐C4H8: 33.37	Breakthrough, sorption	333 K	140	C‐H∙∙∙F, C‐H∙∙∙π interactions	[525]
TaOFFIVE‐3‐Ni (ZU‐96), 3D	3.2‐4.7	C4H6/i‐C4H8	20.83	C4H6: 1.90 n‐C4H8:0.98 i‐C4H8:0.09	—	Breakthrough, sorption	—	—	H‐bonding	[585]
Y‐dbai, 3D	4.4 × 4.4	C4H6/i‐C4H8	20.6	C4H6: 2.88 n‐C4H8: 1.07 i‐C4H8: 0.14	C4H6: 58 n‐C4H8: 46	Breakthrough, sorption	Ar flow at 423 K	—	C─H···π, C─H···O, C─H···N interactions	[586]
		1‐C4H8/i‐C4H8	7.6
ZU‐602, 3D	6.8 × 2.5	C4H6/n‐C4H8	66	C4H6: 1.45 n‐C4H8: 0.022 trans‐C4H8: 0.144 i‐C4H8: 0.022	—	Breakthrough, sorption	Vacuum at 393 K	123	H‐bonding	[578]
		C4H6/trans‐C4H8	10.1
CuINA, 3D	4.84	n‐C4H8/i‐C4H8	7.1	n‐C4H8: 1.56 i‐C4H8: 1.50	—	Breakthrough, sorption	He flow at 373 K	—	C−H···π interactions	[577]
SOFOUR‐DPDS‐Ni, 2D	5.3 × 4.5 × 5.8	C4H6/i‐C4H8	9.9 (2321.8 e )	C4H6: 1.36 n‐C4H8: 1.25 i‐C4H8: 0.17	C4H6: 77 n‐C4H8: 38	Breakthrough, sorption	He flow at 343 K	270	Van der Waals interactions, H‐bonding	[587]
		n‐C4H8/i‐C4H8	8.7 (233.5 e )
KAUST‐7 (NbOFFIVE‐1‐Ni), 3D	3.14 × 3.14	C4H6/n‐C4H8	3.18	C4H6: 1.55 n‐C4H8: 0.49 i‐C4H8: 0	—	Breakthrough, sorption	Ar flow at 378 K	205.96 b	Pore size	[588]
		C4H6/i‐C4H8	10.4
ZU‐36‐Ni (GeFSIX‐3‐Ni), 3D	4.75 × 4.75	trans‐C4H8/cis‐C4H8	7	trans‐C4H8: 2.45	trans‐C4H8: 42.0	Breakthrough, sorption	298 K	313 c	π–π interactions	[589]
GeFSIX‐14‐Cu‐i (ZU‐33), 3D	3.0 × 4.2	C4H6/n‐C4H8	4.68	C4H6:2.67 n‐C4H8: 0.57 i‐C4H8:0.42	—	Breakthrough	—	424.1 c	Molecular recognition, molecular sieving	[590]
		C4H6/i‐C4H8	6.40
		n‐C4H8/i‐C4H8	1.37
KAUST‐8 (AlFFIVE‐1‐Ni), 3D	5.01 × 5.01	C4H6/i‐C4H8	5.19	C4H6: 2.49 n‐C4H8: 1.94 i‐C4H8: 0.48	—	Breakthrough, sorption	Ar flow at 378 K	240	Pore size, H‐bonding	[588]
		n‐C4H8/i‐C4H8	3.81
Ni2(m‐dobdc), 3D	9.8 [591]	n‐C4H8/cis‐C4H8 f (2.3±0.1 g )	2.4	—	—	Breakthrough, sorption	—	1321 [592]	UMCs	[593]
		n‐C4H8/trans‐C4H8 f (2.1±0.1 g )	6.0
		n‐C4H8/(cis‐C4H8, trans‐C4H8)	2.8
NbFSIX‐2‐Cu‐i (ZU‐52), 3D	4.1 × 4.3	C4H6/n‐C4H8	1.15	C4H6:2.64 n‐C4H8: 2.26 i‐C4H8:0.48	—	Breakthrough	—	440.0 c	Molecular recognition, molecular sieving	[590]
		C4H6/i‐C4H8	5.74
		n‐C4H8/i‐C4H8	5.00
Co2(m‐dobdc), 3D	9.8 [591]	n‐C4H8/cis‐C4H8 f	2.9	—	—	Breakthrough, sorption	—	1264 [592]	UMCs	[593]
		n‐C4H8/trans‐C4H8 f	5.7
SIFSIX‐3‐Ni, 3D	3.8 × 3.8 [594]	C4H6/n‐C4H8	1.00	C4H6:2.46 n‐C4H8: 2.45 i‐C4H8:0.52	—	Breakthrough	—	368 [137]	—	[590]
		C4H6/i‐C4H8	4.73
		n‐C4H8/i‐C4H8	4.73
ZU‐36‐Fe (GeFSIX‐3‐Fe), 3D	4.85 × 4.85	trans‐C4H8/cis‐C4H8	3.6	trans‐C4H8: 1.81	trans‐C4H8: 61.8	Breakthrough, sorption	298 K	295 c	Pore size	[589]
TMA‐VT‐1, 3D	4.3 × 4.3	C4H6/n‐C4H8	1.14	C4H6: 0.80 n‐C4H8: 0.70 n‐C4H10: 0.25	—	Breakthrough, sorption	He flow	38	Van der Waals interactions	[595]
		C4H6/n‐C4H10	3.20
Mg‐gallate, 3D	3.6 × 4.6	trans‐C4H8/cis‐C4H8	3.2	trans‐C4H8: 1.85 cis‐C4H8: 0.58	—	Breakthrough	He flow at 393 K	559 [291]	Shape selective, C···H–O interactions	[596]
Co‐gallate, 3D	3.7 × 5.0 [291]	trans‐C4H8/cis‐C4H8	2.9	trans‐C4H8: 1.67 cis‐C4H8: 0.58	—	Breakthrough	He flow at 393 K	475 [291]	Shape selective, C···H–O interactions	[596]
GeFSIX‐2‐Cu‐i (ZU‐32), 3D	4.5 × 4.5	C4H6/n‐C4H8	1.12	C4H6:3.67 n‐C4H8: 3.3 i‐C4H8:1.25	—	Breakthrough	—	467.6 c	Molecular recognition, molecular sieving	[590]
		C4H6/i‐C4H8	2.94
		n‐C4H8/i‐C4H8	2.26
TMA‐VT‐5, 3D	7.4 × 7.4, 3.9 × 3.9	C4H6/n‐C4H8	1.29	C4H6: 0.97 n‐C4H8: 0.75 n‐C4H10: 0.37	—	Breakthrough, sorption	He flow	5	Van der Waals interactions	[595]
		C4H6/n‐C4H10	2.62
Ni‐gallate, 3D	3.5 × 4.9 [291]	trans‐C4H8/cis‐C4H8	2.5	trans‐C4H8: 1.09 cis‐C4H8: 0.44	—	Breakthrough	He flow at 393 K	424 [291]	Pore shape, C···H–O interactions	[596]
K‐VT‐1, 3D	4.3 × 4.3	C4H6/n‐C4H8	1.16	C4H6: 1.0 n‐C4H8: 0.86 n‐C4H10: 0.54	—	Breakthrough, sorption	He flow	235	Van der Waals interactions	[595]
		C4H6/n‐C4H10	1.85
TIFSIX‐Cu‐TPB, 3D	7.02 × 7.10, 8.28 × 8.31	C4H6/i‐C4H8	1.31 (2.6 f )	C4H6: 3.35 n‐C4H8: 2.96 i‐C4H8: 2.56	C4H6: 61.4 n‐C4H8: 55.3 i‐C4H8: 59.4	Breakthrough, sorption	He flow at 333 K	426.8	C−H···π, π···π interactions, H bonding	[597]
		C4H6/n‐C4H8	1.13 (2.2 f )
SIFSIX‐Cu‐TPB, 3D	6.96 × 7.12, 8.37 × 8.43	C4H6/i‐C4H8	1.32 (1.7 f )	C4H6: 3.23 n‐C4H8: 2.71 i‐C4H8: 2.44	C4H6: 62.4 n‐C4H8: 55.8 i‐C4H8: 59.2	Breakthrough, sorption	He flow at 333 K	257.3	C−H···π, π···π interactions, H bonding	[597]
		C4H6/n‐C4H8	1.19 (1.5 f )
TIFSIX‐2‐Cu‐i, 3D	4.8 × 4.8	C4H6/n‐C4H8	1.10	C4H6:4.05 n‐C4H8: 3.67 i‐C4H8:2.90	—	Breakthrough	—	480.5 [598]	H‐bonding	[590]
		C4H6/i‐C4H8	1.40
		n‐C4H8/i‐C4H8	1.27
SIFSIX‐2‐Cu‐i, 3D	4.7 × 4.7	C4H6/n‐C4H8	1.3	C4H6:3.99 n‐C4H8: 3.06 i‐C4H8:1.74	—	Breakthrough	—	503 [137]	H‐bonding	[590]
		C4H6/i‐C4H8	2.29
		n‐C4H8/i‐C4H8	1.76
Mn2(m‐dobdc), 3D	9.8 [591]	n‐C4H8/cis‐C4H8 f	2.0	—	—	Breakthrough, sorption	—	1349 [592]	UMCs	[593]
Fe2(m‐dobdc), 3D	9.8 [591]	n‐C4H8/cis‐C4H8 f	2.0	—	—	Breakthrough, sorption	—	1295 [592]	UMCs	[593]
Zn‐bzc‐2CH3 , 3D	4.13 × 4.13	n‐C4H10/i‐C4H10	n‐C4H10 selective	n‐C4H10:2.38 i‐C4H10: 0.03	—	Breakthrough, sorption	—	387.2	Van der Waals interactions	[599]
Zr‐fum‐fcu‐100, 3D	4.7 × 4.7, 5.2 × 5.2, 7.6 × 7.6	n‐C4H10/i‐C4H10	n‐C4H10 selective	n‐C4H10:1.23	—	Breakthrough, sorption	—	718	Pore size	[600]
SIFSIX‐1‐Cu, 3D	9.5 × 9.5 [601]	C4H6/n‐C4H8	1.10	C4H6:6.75 n‐C4H8: 6.43 i‐C4H8:6.28	—	Breakthrough	—	1178 [602]	—	[590]
		C4H6/i‐C4H8	1.13
		n‐C4H8/i‐C4H8	1.00
Cu3(BTC)2 (HKUST‐1), 3D	9 × 9	i‐C4H8/ i‐C4H10	i‐C4H8 selective	i‐C4H8: 6.5 i‐C4H10:5.1	i‐C4H10: ‐42	Breakthrough	—	1055 [603]	Van der Waals interactions	[604]
ZJNU‐30a, 3D	13.6, 21.6 d	trans‐C4H8/cis‐C4H8/n‐C4H8/i‐C4H8	cis‐C4H8 selective	—	—	Breakthrough	—	1570	—	[605]
ZJNU‐80a, 3D	9.2 d	C4H6/n‐C4H8	1.00	C4H6:8.35 n‐C4H8: 8.35 i‐C4H8:6.80	—	Adsorption	—	2279	Pore size	[606]
		C4H6/i‐C4H8	1.23
		n‐C4H8/i‐C4H8	1.23
NOTT‐101a, 3D	10.2 d	C4H6/n‐C4H8	1.00	C4H6:9.28 n‐C4H8: 9.28 i‐C4H8:9.15	—	Adsorption	—	2755	Pore size	[606]
		C4H6/i‐C4H8	1.01
		n‐C4H8/i‐C4H8	1.01
Zn(Hmpba)2 , 3D	5.4 × 6.4	C4H6/n‐C4H8	1.11	C4H6:3.10 n‐C4H8: 2.80 i‐C4H8:2.80	—	Breakthrough	—	—	Flexibility	[607]
		C4H6/i‐C4H8	1.11
		n‐C4H8/i‐C4H8	1
Cu4(m4‐O)(m2‐ OH)2(Me2trz‐pba)4 , 3D	4.5 × 5.5	n‐C4H8/i‐C4H8	0.94	n‐C4H8: 4.10 i‐C4H8:4.35	n‐C4H8: 49.5 i‐C4H10:46.4 n‐C4H8: 45.6 i‐C4H8: 40.4	Adsorption	—	—	Flexibility	[608]
[Zn(NO2ip)(dpe)] n (SD‐65), 3D	Discrete voids	C4H6/i‐C4H8/n‐C4H8/2‐C4H8/i‐C4H10/n‐C4H10	C4H6 selective	C4H6: 1.16	—	Breakthrough	298 K	—	C−H···π interactions	[609]
Y‐fum‐fcu‐MOF, 3D	4.7 × 4.7	C4H6/i‐C4H8/n‐C4H8/cis‐C4H8/trans‐C4H8	trans‐C4H8 selective	—	—	GC, breakthrough	N2 flow at 423 K	835	Pore shape, size	[610]
[Zn2(btm)2], 3D	3.6 × 3.6	C4H6/i‐C4H8/n‐C4H8/n‐C4H10	i‐C4H8 selective	C4H6:2.0 i‐C4H8: 2.0 n‐C4H8:2.0 n‐C4H10: 2.0	—	Breakthrough	—	622 b [611]	UMCs, pore size, shape	[82]
CMS‐PMOF‐1, 3D	5 × 5	n‐C4H10/i‐C4H10	n‐C4H10 selective	n‐C4H10: 1.9 i‐C4H10: 1.43	—	Breakthrough, sorption	—	799	Molecular sieving	[612]
UiO‐66 (Zr), 3D	8 × 8, 11 × 11	n‐C4H10/i‐C4H10	n‐C4H10 selective	—	—	NMR	—	1145	Kinetics	[613]
RUB‐41, 3D	2.7 × 5, 4 × 6.5	trans‐C4H8/cis‐C4H8/n‐C4H8	—	—	—	Breakthrough, sorption	—	—	Pore size, shape	[614]

^a BET surface area was calculated by Ar sorption isotherm.

^b Langmuir surface area calculated from CO₂ sorption at 195 K or 298 K.

^c BET surface area calculated from CO₂ sorption isotherms at 196 K or 273 K.

^d Pore size calculated from N₂ isotherm at 77 K.

^e Henry selectivity.

^f Ideal Adsorbed Solution Theory (IAST) selectivity.

^g Selectivity calculated based on mass ratio.

TABLE 15.

C6 alkane (n‐HEX, 2MP, 3MP, 22DMB, 23DMB) and cyclic (Bz, Cy) isomers binary or multicomponent separations. The following parameters are listed for comparison: pore size; BET surface area (S_BET); single‐component gas uptakes; adsorption enthalpies (Q _st); selectivity; analysis method, regeneration temperature; and attributed mechanisms. Reticular sorbents are listed in decreasing order of binary or multicomponent uptake ratio for C6 isomers.

	Adsorbent, network dimensionality (nD)	Pore size (Å)	Components	Binary selectivity a	Capacity (mmol/g)	Q st (kJ/mol)	Analyzation method	Regeneration temperature (K)	SBET (m2/g)	Mechanism	Refs.
C6 alkane isomers	UU‐200, 3D (303K)	5 × 5	n‐HEX/22DMB	1×105	n‐HEX: 1.69 3MP: 1.14 22DMB: 0.09	n‐HEX: 53.3 3MP: 44.9	Breakthrough	—	482 h	Molecular sieving, C‐H···O H‐bonding, C‐H···π interactions	[615]
			3MP/22DMB	3×104
			n‐HEX/3MP	160
	ZUL‐C5, 3D	4.61	n‐HEX/22DMB	863	n‐HEX: 1.70 (303 K) 3MP: 1.71 (303 K) 22DMB: 0.37 (303 K)	—	Breakthrough	N2, 423 K	611	Van der Waals interactions	[616]
			3MP/22DMB	124
	CopzNi, 3D	7.11 × 7.22	2MP/22DMB	420 (303K)	n‐HEX (363K): 1.88	n‐HEX: 78.8 2MP: 45.2	GC	N2, 423 K	607	Molecular sieving	[617]
			n‐HEX/2MP	171.7 (363K)
	ZUL−C6, 3D	4.94 × 5.76	n‐HEX/22DMB	172	n‐HEX: 1.08 3MP: 0.86	—	Breakthrough	—	607	Flexibility	[618]
	NipzNi, 3D	7.10 × 7.27	2MP/22DMB	27.3(303K)	n‐HEX (363K): 1.49	n‐HEX: 65.2 2MP: 32.1	GC	N2, 393K	664	Molecular sieving	[617]
			n‐HEX/2MP	73.7 (363K)
	JNU‐80‐LP, 3D	6.5	2PMP/22DMB	45.3 g	n‐HEX: 2.1 2MP: 1.7	—	GC	—	670	Molecular sieving	[619]
	Zn‐tcpt, 3D	4.9 f	3MP/22DMB	28.5 g	n‐HEX: 3.25 3MP: 2.14	—	Breakthrough	—	1135	Pore size	[620]
			n‐HEX/22DMB	43.3 g
	Mn‐dhbq, 3D (303K, 393K)	5.58 f	3MP/23DMB	34.57 g (303K) 22.18 g (333K)	n‐HEX: 1.53	—	Breakthrough	N2 flow at 423K	429	Polarizability, H‐bonding	[621]
	Zeolite 5A, 3D	4 × 4 [622]	n‐HEX/3MP	29 b	—	—	GC	—	650	Molecular sieving	[623]
			n‐HEX/22DMB	18 b
			3MP/22DMB	1.0 b
	{[Fe3(μ3‐O)](6FDCA)3·2DMF}, 3D (303K)	8.45 × 11.02	n‐HEX/22DMB	27.83 f	n‐HEX: 0.79 g 3MP: 0.58 g 22DMB: 0.43 g	—	Breakthrough	298	269.5	Molecular sieving	[624]
			n‐HEX/3MP	8.66 f
			3MP/22DMB	3.82 f
	BNF‐100a, 3D	14.3	n‐HEX/22DMB	13.6 g	n‐HEX: 1.33 (303K) 3MP: 1.23 (303K) 22DMB: 0.1 (303K)	—	Breakthrough	—	483.1	C‐H⋯π interactions	[625]
			3MP/22DMB	12.5 g
	Zr6(μ3‐O)4(μ3‐OH)4(bptc)3 , 3D	4.5	n‐HEX/3MP	13 g	3MP: 1.23 (423K) 3MP: 0 22DMB: 0	n‐HEX: 48	Breakthrough	—	1030	Van de Waals interactions	[626]
	ZIF‐8, 3D (398 K)	11.6, 3.4	n‐HEX/3MP	3.7 b	—	—	GC	—	1813	Kinetics	[623]
			n‐HEX/22DMB	11.9 b
			3MP/22DMB	9.2 b
	Zn(BPZ(NH2)2), 3D	5.4 × 5.4	n‐HEX/2MP	3.42 j	—	n‐HEX: 61.7 2MP: 59.8 3MP: 67.9 23DMB: 54.2 22DMB: 56.9	GC	—	155	Molecular sieving	[627]
			n‐HEX/3MP	4.77 j
			n‐HEX/23DMB	8.12 j
			n‐HEX/22DMB	10.76 j
	Zn‐Me2BPZ, 3D	4.9 × 4.9	n‐HEX/2MP	2.42 j	—	n‐HEX: 85.0 2MP: 82.9 3MP: 80.7 23DMB: 79.3 22DMB: 71.9	GC	—	123	Molecular sieving	[627]
			n‐HEX/3MP	3.28 j
			n‐HEX/23DMB	5.27 j
			n‐HEX/22DMB	9.48 j
	Cu‐MoO4‐TPA, 3D	4.96 f	n‐HEX/23DMB	10.5 g	n‐HEX: 1.2 2MP: 0.42 3MP: 0.42 23DMB: 0.12 22DMB: 0.03	—	Breakthrough	—	217	C‐H···O interactions	[628]
			n‐HEX/2MP	2.61
			n‐HEX/3MP	2.67
	IM‐22, 3D	—	n‐HEX/3MP	1.7 b	—	—	GC	—	575	Kinetics	[623]
			n‐HEX/22DMB	8.4 b
			3MP/22DMB	5.3 b
	ZIF‐8, 3D (423K)	11.6, 3.4	n‐HEX/2MP	5.8 b	n‐HEX: 5.9, 2MP: 1.05	—	GC	—	1285.2	Kinetics	[573]
	MIL‐140C, 3D	5.7 × 5.7	n‐HEX/3MP	7 g	—	n‐HEX: 71.5	Breakthrough	—	—	Kinetics	[629]
	UiO‐66(Zr), 3D	6 × 6	22DMB/n‐HEX	7.6 d	—	—	GC	—	885 [630]	Pore size, shape	[631]
			23DMB/n‐HEX	5.2 d
			n‐HEX/2MP	0.9 d
	HKUST‐1, 3D	5.6 × 5.6	22DMB/n‐HEX	0.5 d	—	—	GC	—	1055 [603]	UMCs	[631]
			23DMB/n‐HEX	0.61 d
			n‐HEX/2MP	6.8 d
	MIL‐125(Ti), 3D	6.13 × 6.13, 12.55 × 12.55 [632]	22DMB/n‐HEX	0.82 d	—	—	GC	—	1446 [632]	Pore shape	[631]
			23DMB/n‐HEX	0.79 d
			n‐HEX/2MP	3.9 d
	MIL‐125(Ti)‐NH2 , 3D (373K)	5‐7	23DMB/22DMB	1.4 g	—	—	Breakthrough	—	1550	Molecular sieving	[633]
			3MP/22DMB	1.8 g
			n‐HEX/22DMB	2.6 g
	MIL‐140B, 3D	4.0. × 4.0	n‐HEX/3MP	5 g	—	n‐HEX: 84.5	Breakthrough	—	—	Kinetics	[629]
	UiO‐66‐Br, 3D (423K)	—	22DMB/n‐HEX	2.9 b	—	—	Breakthrough	—	600	—	[634]
			23DMB/n‐HEX	2.9 b
			23MP/n‐HEX	2.9 b
	UiO‐66‐NO2 , 3D (423K)	—	22DMB/n‐HEX	1.3 b	—	—	Breakthrough	—	650	—	[634]
			23DMB/n‐HEX	1.3 b
			23MP/n‐HEX	1.1 b
	UiO‐66‐NH2 , 3D (423K)	—	22DMB/n‐HEX	1.2 b	—	—	Breakthrough	—	670	—	[634]
			23DMB/n‐HEX	1.2 b
			23MP/n‐HEX	1.1 b
	HIAM‐410LI, 3D	4 × 4	3MP/22DMB	1.31 b	3MP: 1.50 3MP: 1.24 22DMB: 1.03	3MP: 90.1 3MP: 55.3 22DMB: 27.9	Breakthrough	—	832	Pore size, shape	[635]
	MIL‐100(Cr)‐MEDA, 3D (343 K)	6.6 × 2.5 × 1.8	23DMB/22DMB	1.4 g	n‐HEX: 0.57 3MP: 0.35 23DMB: 0.30 22DMB: 0.21	—	Breakthrough	—	2400	Molecular sieving	[633]
			3MP/22DMB	1.8
			n‐HEX/23DMB	2.7
	ZIF‐76, 3D	5.4, 12.2 [636]	3MP/n‐HEX	0.8 b	—	—	GC	—	1560	Pore shape	[623]
			22DMB/n‐HEX	0.7 b
			22DMB/3MP	0.8 b
	[Zn2(HBDC)2(dmtrz)2], 3D	6.7	n‐HEX/3MP/22DMB	n‐HEX selective	n‐HEX: 13.3wt% c 3MP: 12.1wt% c 22DMB: 3.5wt% c	—	GC	—	552	Molecular sieving	[637]
	MIL‐53(Fe), 3D (CF3)2 , 3D	5.6 × 5.6 (LP), 4.2 × 4.2 (NP)	n‐HEX/3MP/22DMB	n‐HEX selective	—	—	GC	—	12	Flexibility, molecular sieving	[638]
	UiO‐66, 3D	7 × 7	n‐HEX/3MP/22DMB/23DMB	n‐HEX selective	—	—	GC	—	885 [630]	Van der Waals interactions	[639]
	MOF‐CJ3 , 3D	11.6 × 11.6	n‐HEX/3MP/22DMB	n‐HEX selective	—	—	GC	—	525	Van der Waals interactions	[640]
	MIL‐101(Cr), 3D	16 × 16, 12 × 12	n‐HEX/2MP/22DMB/23DMB	n‐HEX selective	n‐HEX: 10.2 2MP: 9.2 22DMB: 8.6 23DMB: 8.4	—	GC	—	3990	Van der Waals interactions, pore size	[641]
	Fe2(BDP)3 , 3D (433K)	13.2 × 13.2	n‐HEX/2MP/3MP/22DMB/23DMB	n‐HEX selective	n‐HEX: 1.32, 2MP: 1.18, 3MP: 1.27, 23DMB: 1.37, 22DMB: 1.22	—	GC	—	1230	Van der Waals interactions	[71]
	Ca(H2tcpb), 3D (333K)	5.5 × 5.5	n‐HEX/2MP/3MP/22DMB	n‐HEX selective	n‐HEX: 1.14, 3MP: 0.12, 22DMB: 0	—	GC	N2, 423K	220	Molecular sieving	[642]
	Zr6(μ3‐O)4(μ3‐OH)4(bptc)3 , 3D	14.6 × 17.2	n‐HEX/3MP/23DMB	n‐HEX selective (423K)	n‐HEX: 1.51, 3MP: 0.29, 23DMB: 0	n‐HEX: 48	GC	—	1030	Molecular sieving	[626]
	Al‐fumarate, 3D (453K)	5.7 × 6.0	n‐HEX/3MP/23DMB	n‐HEX selective	—	—	GC	—	879.5 e	Van der Waals interactions	[643]
	MIL‐127(Fe), 3D	6 × 6, 10 × 10	n‐HEX/3MP/23DMB/22DMB	n‐HEX selective	—	—	Breakthrough	—	1400	Molecular sieving	[633]
	HIAM‐203, 3D (423K)	4.82 × 4.82	n‐HEX/3MP/22DMB	—	n‐HEX selective	—	GC	—	499	H‐bonding, pore size	[644]
	ZIF‐69, 3D (423K)	7.8 × 7.8	—	—	n‐HEX: 3.95, 3MP: 1.16	—	GC	—	845.1	Kinetics	[573]
	Zr‐abtc, 3D	7 × 7	n‐HEX/3MP/22DMB	n‐HEX selective	n‐HEX: 3.31 3MP: 2.14 22DMB: 1.51	—	Breakthrough	—	1300	Pore size	[645]
	Al‐bttotb, 3D	5.6 × 5.6	n‐HEX/3MP/22DMB	n‐HEX selective	n‐HEX: 1.75 3MP: 1.09 22DMB: 0.14	—	Breakthrough	—	600	Pore size	[645]
	Co‐FA, 3D	5.5 × 5.5	n‐HEX/3MP/22DMB	n‐HEX selective	n‐HEX: 1.22 3MP: 1.37 22DMB: 0.31	—	Breakthrough	—	350	Pore size	[645]
	ZSM‐5, 3D	5.1 × 5.5, 5.4 × 5.6	n‐HEX/3MP/22DMB	n‐HEX selective	n‐HEX: 1.31 3MP: 0.32 22DMB: 0.17	—	Breakthrough	—	360	Pore size	[645]
	Zeolite 5A, 3D	5, 11.4	n‐HEX/3MP/22DMB	n‐HEX selective	n‐HEX: 1.69 3MP: 0.32 22DMB: 0.32	—	Breakthrough	—	650	Pore size	[645]
	NU‐2002, 3D	4.7 × 4.7	n‐HEX/2MP/3MP/23DMB/22DMB	n‐HEX+2MP+3MP selective	n‐HEX: 1.41 2MP: 1.54 3MP: 1.56 23DMB: 1.67 22DMB: 1.31	—	Vapor sorption isotherms	—	405	Pore size	[646]
	NU‐2200, 3D	5 × 5	n‐HEX/2MP/3MP/23DMB/22DMB	n‐HEX selective	n‐HEX: 1.32 2MP: 1.03 3MP: 0.89 23DMB: 0.21 22DMB: 0.17	—	GC	—	355	Pore size, shape	[647]
	ZIF‐8‐90‐S1, 3D	3.1	—	n‐HEX selective	2.82	—	Vapor sorption isotherms	—	1374	Van der Waals interactions	[648]
	ZIF‐8‐90‐S5, 3D	8.8	—	n‐HEX selective	3.10	—	Vapor sorption isotherms	—	1208	Van der Waals interactions	[648]
	ZIF‐8, 3D (373K)	3.4, 11.4	n‐HEX/2MP/23DMB/22DMB	n‐HEX selective	n‐HEX: 2.6 2MP: 2.3 23DMB: 1.8 22DMB: 0.2	—	Vapor sorption isotherms	—	1800	Molecular sieving	[649]
	Zn2(bdc‐R)2(bpy), 3D	—	n‐HEX/3MP/22DMB	n‐HEX selective	n‐HEX: 1.68 3MP: 0.48 22DMB: ∼0	—	Vapor sorption isotherms	—	—	Flexibility, π···π interactions	[650]
	MIP‐214, 3D	6.1 × 6.1, 2.6 × 2.6	n‐HEX/2MP/3MP/23DMB/22DMB	—	n‐HEX: 1.41 2MP: 0.32 3MP: 0.21 23DMB: 0.15 22DMB: 0.1	—	Breakthrough	—	1140	Pore shape, size	[651]
	NU‐2004, 3D	5 × 5	n‐HEX/2MP/3MP/23DMB/22DMB	—	n‐HEX: 1.47 2MP: 0.74 3MP: 0.8 23DMB: 0.27 22DMB: 0.27	—	Vapor sorption isotherms	—	435	Molecular sieving	[652]
	NU‐2005, 3D	5 × 5	n‐HEX/2MP/3MP/23DMB/22DMB	—	—	—	Vapor sorption isotherms	—	380	Molecular sieving	[652]
	HIAM‐318, 3D	5.4 × 5.4	n‐HEX/3MP/22DMB	n‐HEX, 3MP selective	n‐HEX: 1.49 3MP: 1.34 22DMB: 0.22	—	Breakthrough	—	599	Pore size	[653]
	CAU‐21‐ODB, 3D	4.6 × 4.6	n‐HEX/3MP/22DMB	n‐HEX selective	n‐HEX: 1.64	—	Breakthrough	—	275	Pore size, shape	[654]
	BTAPa‐CF3 , 3D	6.3 × 6.3, 14.5 × 14.5	n‐HEX/3MP/22DMB/23DMB	n‐HEX selective	—	—	GC	—	953	Pore size	[655]
	COK‐18, 3D (373K)	11 × 11	n‐HEX/2MP/3MP/23DMB/22DMB	n‐HEX selective	—	—	GC	—	800	Polarizability	[656]
	Zn(bdc)(4‐4′‐bpy)0.5 (MOF‐508a), 3D	4 × 4	n‐HEX/2MP/22DMB	n‐HEX selective	—	—	GC	—	553.7 [657]	Pore size, shape	[658]
	Zn(bdc)(dabco)0.5 , 3D	7.5 × 7.5, 3.8 × 4.7	n‐HEX/3MP/22DMB	n‐HEX selective	—	—	Breakthrough	—	1450 [41]	Pore size, Van der Waals interactions	[659]
	Al‐bttotb, 3D (303K)	5.6 f	n‐HEX/2MP/3MP/22DMB/23DMB	n‐HEX selective	n‐HEX: 1.75 3MP: 1.09 23DMB: 0.49	—	Breakthrough	—	572	Pore size	[660]
	Zr6O4(OH)4(bptc)3 , 3D (423K)	4.5 × 4.5, 12 × 12	n‐HEX/3MP/23DMB	ns‐HEX selective	n‐HEX: 1.51 3MP: 0.17 23DMB: 0.03	48 (n‐HEX)	Breakthrough	—	1030	Van der Waals interactions	[626]
	Zr6O4(OH)8(H2O)4(abtc)2 , 3D (423K)	7 × 7	n‐HEX/3MP/23DMB	n‐HEX selective	n‐HEX: 1.21 3MP: 1.04 23DMB: 0.52	—	Breakthrough	—	1318	Van der Waals interactions	[626]
	HIAM‐302, 3D	7.6 × 7.6	n‐HEX/3MP/22DMB	n‐HEX selective	n‐HEX: 1.91 3MP: 1.09 22DMB: 0	—	Breakthrough	—	388	Van der Waals interactions	[661]
	ZIF‐8, 3D	3.4, 11.4	n‐HEX/2MP/22DMB/23DMB	n‐HEX selective	—	—	GC	—	—	Van der Waals interactions	[662]
	MIL‐140A, 3D	3.2 × 3.2	n‐HEX/2MP/3MP/22DMB/23DMB	n‐HEX selective	—		Breakthrough	—	423	Van der Waals interactions	[663]
	MIL‐140B, 3D	4.0 × 4.0	n‐HEX/2MP/3MP/22DMB/23DMB	n‐HEX selective	—	n‐HEX: 54.6 2MP: 51.2 3MP: 47.5 23DMB: 44.4 23DMB: 41.3	Breakthrough	—	472	Van der Waals interactions	[663]
	MIL‐140C, 3D	5.7 × 5.7	n‐HEX/2MP/3MP/22DMB/23DMB	n‐HEX selective	—		Breakthrough	—	782	Van der Waals interactions	[663]
	[Zn2(PdTCPP)(Bpa)]n , 3D (333K)	5.7 × 9.8	n‐HEX/3MP	n‐HEX selective	n‐HEX: 1.39 3MP: 0	—	Breakthrough	—	—	Flexibility	[664]
	Ca(H2tcpb), 3D (393K)	5.0‐6.2	n‐HEX/3MP/22DMB	n‐HEX selective	n‐HEX: 1.02 3MP: 0	—	Breakthrough	—	220	Molecular size	[642]
	ZIF‐8, 3D	11.6, 3.4 [573]	n‐HEX/2MP/3MP/22DMB/23DMB	n‐HEX selective	—	—	Breakthrough	—	—	Kinetic	[665]
	ZIF‐8, 3D	11.6, 3.4	n‐HEX/3MP/22DMB	n‐HEX selective	—	—	Breakthrough	—	1950	Kinetic, diffusion	[666]
	Zr‐MIL‐140B, 3D (343K)	3.8 × 3.8	n‐HEX/2MP/3MP/22DMB/23DMB	n‐HEX selective	n‐HEX: 0.25 2MP: 0.11 3MP: 0.08 22DMB: 0.02 23DMB: 0.03	n‐HEX: 55.1 2MP: 52.8 3MP: 44.3 22DMB: 40.1 23DMB: 43.5	Vapor sorption isotherms	—	—	Pore shape	[667]
	UiO‐66, 3D (343K‐473K)	7, 11, 8	22DMB/23DMB/3MP/n‐HEX	22DMB selective	n‐HEX: 0.87 3MP: 2.64 22DMB: 2.81 23DMB: 3.04 e	—	Breakthrough	—	1050	Pore size	[668]
	MIL‐160(Al), 3D	5.8 × 5.8	n‐HEX/2MP/3MP/22DMB/23DMB	n‐HEX selective	—	—	Breakthrough	—	—	Van der Waals interactions	[669]
	CAU‐10‐H, 3D (303K)	6.2, 11.6	n‐HEX/3MP/22DMB	n‐HEX selective	n‐HEX: 1.56 3MP: 1.10 22DMB: 0.63	—	Breakthrough	N2 flow at 423K	483	Van der Waals interactions	[670]
	CAU‐10‐Br, 3D (303K)	4.1, 11.6	n‐HEX/3MP/22DMB	n‐HEX selective	n‐HEX: 0.55 3MP: 0.13 22DMB: 0.05	—	Breakthrough	N2 flow at 423K	463	Van der Waals interactions	[670]
	CAU‐10‐H/Br, 3D (303K)	5.4, 11.6	n‐HEX/3MP/22DMB	n‐HEX selective	n‐HEX: 1.46 3MP: 0.70 22DMB: 0.09	—	Breakthrough	N2 flow at 423 K	362	Van der Waals interactions	[670]
	[Zn9(tba)9(dabco)3], 3D	11.92 × 11.92	n‐HEX/3MP/22DMB/23DMB	n‐HEX selective	—	—	Breakthrough	—	1175	Shape sieving, shape, diffusion	[671]
	JNU‐2, 3D	7.4 ×7.4	n‐HEX/2MP/3MP/23DMB/22DMB	n‐HEX selective	—	—	Breakthrough	Vacuum at 473 K	—	Kinetics	[672]
	[Zn2(L‐AlaPyr)2], 3D	—	n‐HEX/2MP/3MP/23DMB/22DMB	n‐HEX selective	—	—	NMR, GC	—	—	Pore shape	[673]
C6 cyclic isomers	CoV‐FA‐TPA, 3D	—	Bz/Cy	235.4	Bz: 2.47 Cy: 0.14	—	NMR	He flow at 333 K	887	π⋯π interactions	[674]
	MFM‐68‐Cl2 , 3D	9.3	Bz/Cy	277	Bz: 4.62	Bz: 60	Breakthrough	N2 flow at 353 K	—	C‐H⋯π interactions	[675]
	Ni3(OH)(Ina)3(BDC)1.5 , 3D	7 e	Bz/Cy	145 g	Bz: 2.9 Cy: 0.02	—	Vapor sorption isotherms	—	1255	π⋯π interactions	[676]
	MAF‐7Br, 3D	3.4 × 3.4	Bz/Cy	113	Bz: 2.55 Cy: 0.17	—	GC	—	—	Van der Waals interactions	[677]
	MAF‐40, 3D	8.2 × 10.2	Bz/Cy	40	Bz: 0.72 Cy: 0.46	—	GC	—	338	Kinetic	[678]
	Mn‐MOF‐74, 3D	11 × 11	Bz/Cy	37.5 g	Bz: 9.38 Cy: 0.25	—	Vapor sorption isotherms	—	1500‐1600	UMCs	[679]
	CUB‐30, 3D	8, 18, 21	Cy/Bz	28.5	—	—	Breakthrough	—	2930	Pore size	[680]
	Zn‐BDC, 2D	5.3 × 5.3	Bz/Cy	22.6	Bz: 2.15 Cy: 0.1	—	Breakthrough	—	308	Pore size	[681]
	[Cu(etz)] (MAF‐2), 3D	9 × 9	Bz/Cy	21 g	Bz: 10.5 Cy: 0.5	—	Vapor sorption isotherms	—	—	Flexibility, C‐H⋯π interactions	[682]
	Cu‐1‐NO3 , 1D to 2D	5.7 j	Bz/Cy	18.9 g	Bz: 1.7 Cy: 0.09	—	Vapor sorption isotherms	—	—	C‐H⋯π, π⋯π interactions	[683]
	[Zn(ip)(bpa)]n (CID‐23), 3D	8.6 × 8.6	Bz/Cy	13.3 g	Bz: 1.2 Cy: 0.09	—	Vapor sorption isotherms	—	553 h	Flexibility, pore size	[684]
	LiZn‐bdc‐bpy, 3D	7 × 7	Bz/Cy	12.6 g	Bz: 2.89 Cy: 0.23	—	Vapor sorption isotherms	—	742	C‐H⋯π, π⋯π, C‐H⋯O interactions	[685]
	MAF‐24 β, 3D	3.3 × 3.3	Bz/Cy	10.8 g	Bz: 10.24 Cy: 0.95	—	Vapor sorption isotherms	—	444 h	Pore size, C‐H⋯π interactions	[686]
	DAT‐MOF‐1, 3D	6.71 × 7.08	Bz/Cy	8.4 g	Bz: 1.51 Cy: 0.18	—	Vapor sorption isotherms	—	—	π‐electron deficient	[687]
	Cd‐ATAIA, 3D	11.3 × 5.92, 11.99 × 20.08, 5.21 × 4.91	Bz/Cy	6.7 g	Bz: 2.36 Cy: 0.35	—	Vapor sorption isotherms	—	62	π‐electron deficient	[688]
	Cd‐dtztp, 3D	10.0‐15.9 f	Bz/Cy	6.0 g	Bz: 6.52 Cy: 1.09	—	Vapor sorption isotherms	—	575.8	Pore size, C‐H⋯π, π⋯π interactions	[689]
	ZnL, j 3D	1.75 × 3.65, 0.32 × 2.22	Bz/Cy	5.1 g	Bz: 5.6 Cy: 1.1	—	Vapor sorption isotherms	—	—	C‐H⋯π interactions	[690]
	[Zn(μ4‐TCNQ−TCNQ)bpy], 3D	10.1 × 10.1	Bz/Cy	4 g	Bz: 3.6 Cy: 0.9	—	Vapour sorption isotherms	—	—	Size, C‐H⋯π interactions	[62]
	ZnL′, k 3D	9.11	Bz/Cy	3.3 g	Bz: 1.4 Cy: 0.45	—	Vapor sorption isotherms	—	—	π⋯π interactions	[691]
	CUB‐5, 3D	12 b	Bz/Cy	2.7	—	—	Vapor sorption isotherms	—	3107	C‐H···π interactions	[692]
	MUF‐77, 3D	8.2, 17.5, 20.5	Cy/Bz	2.45	—	—	Breakthrough	—	3604	Pore size	[680]
	3DL‐MOF‐1 ([Zn4O(pdc)3]), 3D	11 b	Bz/Cy	2.3	—	—	Vapor sorption isotherms	—	2378	C‐H···π interactions	[692]
	Mn(TCNQ−TCNQ)bpy, 3D	10.1 × 10.1	Bz/Cy	2.2 g	Bz: 3.7 Cy: 1.7	—	Vapor sorption isotherms	—	—	C‐H⋯π interactions	[693]
	Cu2I2(BTTP4)], 3D	10 × 10	Bz/Cy	2.2 g	Bz: 2.6 Cy: 1.2	—	Vapor sorption isotherms	—	496 i	π⋯π interaction	[694]
	CUB‐5, 3D	7, 12 f	Bz/n‐HEX/2MP/23DMB	Bz selective	Benzene: 7.3 n‐HEX: 4.9 2MP: 5.1 23DMB: 4.7	—	Breakthrough	—	2614	C‐H···π interactions	[73]
	BUT‐53, 3D	7.8 × 7.8	Bz/air	Trace Bz capture	—	—	Breakthrough	N2 flow at 393 K	811	C‐H···π, C‐H···N interactions	[695]
	BUT‐54, 3D	10.0 × 10.0	Bz/air	Trace Bz capture	—	—	Breakthrough	N2 flow at 393 K	1128	C‐H···π, C‐H···N interactions	[695]
	BUT‐55, 3D	8.0 × 8.0	Bz/air	Trace Bz capture	—	—	Breakthrough	N2 flow at 393 K	873	C‐H···π, C‐H···N interactions	[695]
	BUT‐56, 3D	8.0 × 8.0	Bz/air	Trace Bz capture	—	—	Breakthrough	N2 flow at 393 K	897	C‐H···π, C‐H···N interactions	[695]
	BUT‐57, 3D	11.5 × 11.5	Bz/air	Trace Bz capture	—	—	Breakthrough	N2 flow at 393 K	970	C‐H···π, C‐H···N interactions	[695]
	BUT‐58, 3D	8.0 × 8.0	Bz/air	Trace Bz capture	—	—	Breakthrough	N2 flow at 393 K	849	C‐H···π, C‐H···N interactions	[695]
	MIL‐125‐Zn, 3D		Bz/air	Trace Bz capture	—	—	Breakthrough	N2 flow at 353 K	—	UMCs, electrostatics	[696]
	2‐bpe‐Zn, 3D	12.76 × 8.39	Bz/Cy	—	Bz: 5.5 Cy: ∼0	—	Vapor sorption isotherms	—	—	π⋯π interactions	[697]
	1‐bpy‐Zn, 3D	10.49 × 8.34	Bz/Cy	—	Bz: 2.32 Cy: ∼0	—	Vapor sorption isotherms	—	—	π⋯π interactions	[697]
	MFOF‐1, 3D	7.2, 10.3 b	Bz/Cy	1.5 g	Bz: 20.38 Cy: 13.18	—	Vapor sorption isotherms	—	2287	π‐electron deficient	[698]
	IPM‐MOF‐110, 3D	6.17 × 7.68	Bz/Cy	—	Bz: 4.57 Cy: ∼0	—	Vapor sorption isotherms	—	937	π‐electron deficient	[699]
	Cu‐terpy‐I, 3D	10.56 × 7.32	Bz/Cy	—	Bz: 2.77 Cy: ∼0	—	Vapor sorption isotherms	—	—	C‐H⋯π, π⋯π interactions	[700]
	Cu‐bpp‐BF4 , 3D	—	Bz/Cy	—	Bz: 11 Cy: ∼0	Bz: 50	Vapor sorption isotherms	—	14	Flexibility, C‐H⋯π interactions	[701]

^a IAST Selectivity calculated from mixtures of binary (v/v = 1/1) C6 compounds.

^b Selectivity calculated based on binary breakthrough curves.

^c Capacity of weight precent based on breakthrough data (wt%).

^d Calculated selectivity by binary of 50:50 n‐HEX and its branched isomer.

^e BET surface area was calculated by Ar sorption isotherm.

^f Pore size and BET surface area were calculated by N₂ isotherm at 77K.

^g Uptake ratio.

^h Langmuir surface area.

ⁱ Pore size and BET surface area calculated from CO₂ sorption at 195K.

^j Separation coefficient. ^j L = (R,R)‐(−)‐N,N′‐bis(3‐tert‐butyl‐5‐(4‐ethynylpyridyl)salicylidene)‐1,2‐diaminocyclohexane.

^k L′ = 1,2,4,5‐Benzene tetracarboxylic anhydride.

3. HC Separation in the Gas Phase

Under ambient conditions (298 K, 1 bar), most of the light HCs (C1‐C4 HCs, i.e., CH₄, C₂H₂, C₂H₄, C₂H₆, C₃H₄, C₃H₆, C₃H₈, C₄H₆, C₄H₈, and C₄H₁₀, and their isomers) exist in the gas phase, the others being volatile liquids. Their utility in the commodity chemicals sector requires a high level of purity, which in turn requires energy‐ and capital‐intensive separation technologies. Herein, we highlight the potential of physisorbents to enable adsorptive separation of C1–C4 HCs by focusing upon the nature of the binding sites in the best‐performing physisorbents in the context of purifying a specific HC from a binary mixture or a multicomponent feed gas stream.

3.1. Methane Separation

CH₄ offers advantages over other fossil fuels, thanks to its ample natural reserves, low cost, and high fuel value of 57.1 kJ/mol [105]. However, CH₄ is primarily obtained from natural gas (NG), which is typically contaminated with impurities such as carbon dioxide (CO₂) and nitrogen (N₂), among others. These impurities can lower its heating value and energy content. Selective separation of CH₄ from mixtures containing CO₂ or N₂ can therefore enhance the calorific value of low‐quality NG. NG sweetening (purification) is relevant herein as sorbents tend to strongly adsorb (and thereby remove) acidic gases like H₂S and CO₂ from NG [702]. Moreover, biogas prepared by anaerobic decomposition of organic waste is primarily composed of CH₄ and CO₂ [703].

For adsorptive purification of CH₄, adsorbents can be categorized into three groups: (1) CH₄–selective adsorbents, which exhibit stronger interactions for CH₄ over N₂, separation being typically driven by thermodynamic equilibrium. As detailed above, in 1997, Kondo et al. reported the first example of a 3D MOF with permanent porosity; CH₄/N₂ separation selectivity was observed [28]. (2) N₂‐selective adsorbents, which selectively adsorb N₂ over CH₄, performance typically enabled by UMCs or sieving. In 2007, a prototypical MOF for N₂/CH₄ separation was reported by S. Ma et al. by utilizing a molecular sieving mechanism [704]. (3) CO₂‐selective adsorbents offer a higher affinity for CO₂ than CH₄, typically driven by sorbate‐sorbent interactions enhanced by control of pore chemistry and pore size. Zaworotko and co‐workers reported that HUMs such as SIFSIX‐3‐Zn offer optimal pore size and favorable electrostatic interactions from inorganic anions, which are highly effective for trace CO₂/CH₄ separation by offering an order of magnitude improvement in selectivity versus previous benchmark reticular sorbents for this separation [33].

3.1.1. CH₄‐Selective Separation (CH₄/N₂)

Separating CH₄ from N₂ is a challenge due to their similar kinetic diameters (CH₄: 3.80 Å and N₂: 3.64 Å) and comparable polarizabilities (CH₄: 26.0 × 10⁻²⁵ cm³ and N₂: 17.6 × 10⁻²⁵ cm³) (Table 2). In adsorptive separation, as both N₂ and CH₄ are nonpolar molecules, most CNs tend to selectively adsorb CH₄ over N₂ because of the slightly higher polarizability of CH₄, but typically with relatively low selectivity (Table 3). Table 3 presents the leading reticular sorbents reported to date for CH₄/N₂ and N₂/CH₄ separation, arranged in decreasing order of adsorption selectivity. Table 3 reveals little correlation between selectivity and pore size, surface area, or single‐component gas uptake. The mechanisms for the six top‐performing CH₄‐selective reticular sorbents are primarily attributed to Van der Waals interactions, while all N₂‐selective sorbents are driven by UMC binding sites. The current top‐performing sorbents are discussed below.

In 2018, Bao's group synthesized an ultramicroporous CN featuring a pore size of 4.1 × 4.3 Ų, [Co₃(C₄O₄)₂(OH)₂] (C₄O₄ ²⁻  =  squarate) (Figure 1a) [89]. Thanks to the binding between CH₄ molecules and hydroxo groups that line the pores of [Co₃(C₄O₄)₂(OH)₂], an IAST CH₄/N₂ selectivity of 12.5 and CH₄ uptake of 0.4 mmol/g at 298 K were achieved.

FIGURE 1.

Examples of top‐performing reticular sorbents for CH₄/N₂ separation based upon selectivity: (a) [Co₃(C₄O₄)₂(OH)₂] [89] (b) Al‐CDC [88] (c) Ni(ina)₂ [87] with feed ratio, temperature, and total pressure set at 1:1, 298 K, and 1 bar, respectively. (Reprinted with permission from ref. [89]: copyright 2018, American Institute of Chemical Engineers; ref. [88]: Copyright 2020, The Royal Society of Chemistry; ref. [87]: copyright 2022, Wiley‐VCH GmbH.).

In 2020, a family of CNs known as M‐CDC (CDC = trans‐1,4‐cyclohexanedicarboxylic acid) featuring various metal centers demonstrated impressive separation performance for CH₄/N₂ mixtures (Figure 1b) [88]. Al‐CDC, in particular, exhibited IAST selectivity of 13.1 at 298 K and CH₄ uptake of 1.43 mmol/g, a result supported by dynamic column breakthrough experiments. Computational studies indicated that strong van der Waals interactions between pore walls and CH₄ molecules were enabled by confinement.

In 2022, Q. Yang and colleagues reported a new benchmark for CH₄/N₂ IAST selectivity (15.8) in Ni(ina)₂ (Figure 1c, feed ratio = 1:1, temperature: 298 K, total pressure: 1 bar) [87]. Ni(ina)₂ , featuring a pore size of 5.0 × 5.8 Ų, also exhibited exceptional CH₄ uptake of 1.82 mmol/g at 298 K. Theoretical calculations and analysis of the CH₄‐loaded single‐crystal structures revealed a CH₄‐selective binding site rich with C─H···π interactions. With favorable thermal and moisture stability, low cost, and scalability, Ni(ina)₂ is therefore a promising candidate for further development.

3.1.2. N₂‐Selective Separation (N₂/CH₄)

Whereas CH₄ selectivity over N₂ is the most common situation, utilizing CH₄ would necessitate an additional desorption process. There are, however, CNs that preferentially adsorb N₂ over CH₄ by relying upon reticular sorbent‐sorbate interactions (Table 3).

In 2014, Long and colleagues predicted the potential of V‐MOF‐74 for separating N₂ from CH₄ through quantum‐mechanical computations. This prediction was based on the presence of selective back‐bonding interactions between the unsaturated vanadium ions in V‐MOF‐74 and the empty π* orbitals of N₂ (Figure 2a) [122]. In 2020, they successfully synthesized another MOF with exposed V(II) sites, which facilitated the formation of back‐bonding interactions with weakly π‐acidic N₂ [123]. The separation performance for N₂/CH₄ mixtures by V(II)‐MOF revealed an exceptionally high N₂/CH₄ IAST selectivity at lower N₂ concentrations (72 at an N₂:CH₄ volumetric ratio of 2:98). Introducing π‐basic metal centers into porous adsorbents is likely to be an effective general design strategy for selective capture of π‐acidic molecules. (Figure 2b).

FIGURE 2.

Examples of top‐performing N₂/CH₄ separating reticular sorbents that exhibit high selectivity for N₂ over CH₄: (a) V‐MOF‐74 [122] (b) V(II)‐MOF [123] (c) MIL‐100(Cr) [124] (d) MIL‐101(Cr) [127] with feed ratio, temperature, and total pressure set at 1:1, 298 K, and 1 bar, respectively. (Reprinted with permission from ref. [122]: Copyright 2013, American Chemical Society; ref. [123]: Copyright 2020, Springer Nature; ref. [124]: Copyright 2017 Springer Nature; ref. [127]: Copyright 2017, American Chemical Society.).

In 2017, drawing inspiration from biomimetic and metal‐dinitrogen chemistry, Chang and co‐workers reported a mesoporous N₂‐selective CN, MIL‐100(Cr), featuring accessible Cr(III) sites. MIL‐101(Cr) demonstrated a capacity to selectively capture N₂ over CH₄ and O₂ based on thermodynamic equilibrium (Figure 2c) [124]. The calculated IAST selectivity was 8, and N₂ uptake was determined to be 1.64 mmolg⁻¹ at 298 K. The N₂‐selective binding mechanism was elucidated through a combination of experimental and computational studies, which suggested that strong binding affinity between N₂ and the unsaturated Cr(III) sites drives separation performance. Similarly, in 2017 and 2021, two isostructural CNs, MIL‐101(Cr) [127] and MIL‐101(Cr)‐NO₂ [126], respectively, were reported to offer excellent N₂/CH₄ separation efficiencies (Figure 2d). In both cases, the separation performance was attributed to N₂ selective interactions enabled by the unsaturated Cr(III) sites.

3.1.3. CO₂‐Selective Separation (CO₂/CH₄)

Thanks to the difference in kinetic diameters between CO₂ (3.3 Å) and CH₄ (3.8 Å), molecular sieving could be feasible for the separation of CO₂ from CH₄. Further, the difference in polarizability between CO₂ (29.11 × 10⁻²⁵ cm³) and CH₄ (26 × 10⁻²⁵ cm³), although small, favors CO₂ (Table 4). Table 4 presents the leading reticular sorbents reported to date for CO₂/CH₄ separation, arranged in decreasing order of adsorption selectivity. The sorbents with relatively high selectivity in Table 4 were found to have pore sizes ranging from 3 to 4 Å. Moreover, molecular sieving, as noted in Table 4, facilitated the highest selectivity values for CO₂/CH₄ separation. The Q _st values for CO₂ of the top‐performing sorbents with a molecular sieving mechanism are generally below 40 kJmol⁻¹, resulting in mild regeneration conditions (inert gas flow at 298 K or under vacuum). However, the CO₂ uptake in the ten top‐performing reticular sorbents (with pore sizes of around 3.6 Å) is 5.3 mmol/g, whereas for sorbents with larger pores of around 10.2 Å, the CO₂ uptake increases to 8.61 mmol/g. Nonetheless, due to the trade‐off [81], the separation performance of the large poor is poor, with a selectivity of only 8. Four representative examples with high selectivity and optimal pore sizes (3–4 Å) are discussed below to provide insight into design principles for CO₂‐selective sorbents.

In 2013, Zaworotko's group reported three hybrid CNs, SIFSIX‐2‐Cu (pore size 13.05 Å), SIFSIX‐2‐Cu‐i (5.15 Å), SIFSIX‐3‐Zn (pore size 3.84 Å), with SIFSIX‐3‐Zn being the parent of the first family of physisorbents with ultrahigh (>1000 at ≤10 000 ppm) trace CO₂ selectivity (Figure 3a) [33]. The use of crystal engineering to fine‐tune pore chemistry and size in this family of HUMs with coordinatively saturated metal centers and periodically arrayed hexafluorosilicate (SIFSIX) anions enables a “sweet spot” of kinetics and thermodynamics that offers high volumetric uptake even at low CO₂ partial pressure (0.0004‐0.15 bar).

FIGURE 3.

Examples of adsorbents that offer high CO₂/CH₄ selectivity: (a) SIFSIX‐3‐Zn [33] (b) Qc‐5‐Cu‐sql‐α [130] (c) MUF‐16 [129] (d) Cu‐F‐pymo [128] with feed ratio, temperature, and total pressure set at 1:1, 298 K, and 1 bar, respectively. (Reprinted with permission from ref. [130]: copyright 2016, WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim; ref. [129]: copyright 2021, Springer Nature; ref. [128]: copyright 2022, Elsevier B.V.).

In 2016, Zaworotko and colleagues reported on the sorption properties of 2‐fold interpenetrated dia, Qc‐5‐M‐dia (M = Co, Ni, Zn, and Cu; CN), and sql topology, Qc‐5‐Cu‐sql‐α, CNs (Figure 3b) [130]. Qc‐5‐Cu‐sql‐α, with 3.3 × 3.3 Ų pores, enabled molecular sieving and high CO₂/CH₄ IAST selectivity of 3300, coupled with CO₂ uptake of 2.16 mmol/g at 298 K. Qc‐5‐Cu‐sql‐α also exhibited stability in the presence of moisture and maintained its separation efficiency even in the presence of water vapor.

In 2021, Telfer's group studied a series of robust and scalable CNs, [M(Haip)₂] (M = Co, Ni, Mn; H₂aip = 5‐aminoisophthalic acid), MUF‐16(Co) (MUF‐16), MUF‐16(Mn), and MUF‐16(Ni), featuring 3.6 × 7.6 Ų pores (Figure 3c) [129]. MUF‐16 and MUF‐16(Ni) exhibited high IAST selectivity values of 6690 and 1220, respectively, towards equimolar CO₂/CH₄ mixtures, and CO₂ uptake as high as 2.13 mmol/g at 298 K. Computational studies revealed strong interactions between CO₂ and pore walls facilitated by hydrogen bonds formed between the oxygen atoms of CO₂ molecules and amino/phenyl groups.

In 2022, a CO₂/CH₄ IAST selectivity of >10⁷ was reported for Cu‐F‐pymo (Figure 3d) [128]. The observed molecular sieving effect was attributed to ultramicropores of 3.3 Å diameter. Computational analyses revealed that the oxygen moieties distributed along the pore walls enhanced binding affinity toward CO₂ through electrostatic and hydrogen‐bonding interactions. Cycling dynamic column breakthrough experiments validated the CO₂ separation performance and revealed a dynamic CO₂ capture capacity of 1078 mmol/kg for an equimolar CO₂/CH₄ (v/v = 1/1) mixture.

To enable C1 separations, i.e., CH₄/N₂, N₂/CH₄, CO₂/CH₄, nine mechanisms have been identified, either as a single mechanism or a combination of two or more mechanisms: C─H···π interactions; van der Waals interactions; electrostatics; H‐bonding; polarizability; UMCs; flexibility; kinetics; pore size (molecular sieving). We note that supramolecular chemistry or “chemistry beyond the molecule” relies on non‐covalent interactions to enable the formation of larger, organized structures and that it encompasses a wide range of non‐covalent forces [705]. Thinking “supramolecularly” can provide insight and design rules to advance the functionality of MOFs for gas adsorption and separation. In particular, by exploiting host–guest interactions, the pore structure of MOFs can be engineered to yield materials with superior adsorption capacity and selectivity compared to conventional adsorbents.

The contributions of these mechanisms in the top‐performing adsorbents for CO₂/N₂, N₂/CO₂, and CO₂/CH₄ binary separation are collated as a radar plot in Figure 4.

FIGURE 4.

Radar plot illustrating the relative contributions of nine separation mechanisms among the 15 top‐performing CH₄/N₂ and CO₂/CH₄ selective adsorbents as well as five adsorbents that exhibit N₂/CH₄ adsorption selectivity. Every concentric decagon denotes an individual sorbent, with the central decagon symbolizing one sorbent.

For CH₄/N₂ separation, the higher polarizability of CH₄ over N₂ and, especially, van der Waals interactions (tight binding sites in confined space), were observed with the highest frequency. Meanwhile, the presence of negatively charged atoms when coupled with the right pore size can further enhance CH₄ selectivity. Conversely, for N₂/CH₄ separation, the difference in polarizability seems to be of little relevance. Rather, UMCs were observed to play a prominent role in N₂‐selective separation. Specifically, the unoccupied π* orbitals of N₂ can interact with unsaturated metals that line pore walls. For CO₂/CH₄ separation, the relatively large difference in molecular sizes of CO₂ (3.18 × 3.33 × 5.36 ų) and CH₄ (3.7 × 3.7 × 3.7 ų) can enable molecular sieving mechanisms to afford ultra‐high CO₂ selectivity, as exemplified by the pair of ideal molecular splitters, Qc‐5‐Cu‐sql‐β [130] and SIFSIX‐14‐Cu‐i [137]. Also, H‐bonding interactions between the pore walls and the oxygen atoms of CO₂ were found to enhance CO₂/CH₄ separation in MUF‐16 [129].

3.2. C2 Separations

C2 HCs are commodities used in the production of value‐added chemicals in the petrochemical industry [706]. Their production results in impurities, which must be removed. Adsorptive separation has emerged as an alternative to current purification technologies, thanks to its potential for reduced energy footprints. In a seminal discovery by the Kitagawa group in 2005, the 2D CN CPL‐1 was reported to exhibit C₂H₂/CO₂ selectivity of 26 and, therefore, offer potential utility in adsorptive separations [60]. In essence, this work introduced a new paradigm to separate binary gases with similar physicochemical properties, i.e., precise control of pore size and chemistry in PCNs. Several early studies involving structural flexibility and optimal pore size/chemistry were demonstrated for C₂H₂/C₂H₄ separation. In 2011, B. Chen's group reported adsorption selectivity of up to 5.2 for the narrow pore variants of a flexible PCN, [Zn₃(CDC)₃{Cu(SalPycy)}] _n (M′MOF‐3, CDC = 1,4‐cyclohexanedicarboxylate) [241]. In 2016, the groups of H. Xing, B. Chen and Zaworotko reported that the rigid physisorbent SIFSIX‐2‐Cu‐i exhibits C₂H₂ selectivity of 44.5 over C₂H₄ [35], an order of magnitude beyond the previous benchmark (feed ratio = 1:1, temperature: 298 K, total pressure: 1 bar). In 2018, the first example of a C₂H₄ sieve over C₂H₆, UTSA‐280, was reported by B. Chen's group. It exhibited ultrahigh adsorption selectivity of ≈ 4100 [287]. With respect to C₂H₆/C₂H₄ “inverse” separation, an azolate ultramicroporous material (AUM), MAF‐49 (MAF: metal azolate framework) was reported to exhibit reverse selectivity for C₂H₆ by J. P. Zhang and X. M. Chen's group in 2015 [314]. Given the differences in separation mechanisms (molecular sieving, flexible sorbents that open only for specific sorbates, and rigid sorbents), comparing calculated selectivity should be regarded only as an indicator of separation performance under dynamic breakthrough conditions.

3.2.1. C₂H₂‐Selective Separation

3.2.1.1. C₂H₂/CO₂ and C₂H₂/C₂H₄ Separation

C₂H₂ is an important feedstock in the chemical industry, serving as a precursor to a wide spectrum of commodity chemicals, including polyurethane and polyesters. However, industrial C₂H₂ production tends to entail elevated temperatures when it is produced by the breakdown of petroleum gas or the combustion of natural gas. This results in contaminants such as CO₂ or C₂H₄ [707, 708, 709]. Since these gases exhibit similar physicochemical properties and comparable molecular sizes (Table 2), the production of high‐purity C₂H₂ from C₂H₂/CO₂ and/or C₂H₂/C₂H₄ mixtures remains a challenge and has been subjected to in‐depth study (Tables 5 and 6) [710]. Tables 5 and 6 present the leading reticular sorbents thus far reported for C₂H₂/CO₂ and C₂H₂/C₂H₄ separation, respectively, arranged in decreasing order of adsorption selectivity. The sorbents with the highest C₂H₂/CO₂ and C₂H₂/C₂H₄ selectivity in Tables 5 and 6 were found to have pore sizes ranging from 3 to 5 Å, although small pore sizes do not always offer a guarantee of high selectivity. A typical trade‐off is evident in Tables 5 and 6: sorbents with larger pore sizes (11‐18 Å) tend to exhibit relatively high uptakes of C₂H₂, CO₂, or C₂H₄ (up to 6.46 mmol/g for C₂H₂, 4.2 mmol/g for CO₂, and 7.45 mmol/g for C₂H₄), but their selectivities are low (< 3.5). For C₂H₂‐selective sorbents in both tables, H‐bonding as a separation mechanism is frequently observed and can result in relatively high Q _st(C₂H₂) and selectivity. Due to their structural characteristics, HUMs with optimal pore sizes and abundant H‐acceptor sites from SiF₆ ²⁻ anions on the pore walls offer advantages over other reticular sorbents for C₂H₂‐selective sorption. Four representative examples (particularly HUMs) with high selectivity from both C₂H₂/CO₂ and C₂H₂/C₂H₄ binary systems are discussed to illustrate the C₂H₂‐binding separation mechanisms (Tables 5, 6).

Reticular sorbents that can adapt their pore structures to better accommodate a specific guest molecule through an induced fit mechanism also offer potential for separation and purification applications. In 2019, D. Li and colleagues reported that the flexible MOF JNU‐1 exhibited induced fit for C₂H₂ with a high adsorption enthalpy of 47.6 kJ mol⁻¹ [160]. The diamond‐shaped channel presents UMCs that cooperatively bind to C₂H₂ molecules. Consequently, a high C₂H₂/CO₂ selectivity of 285.6, along with a high C₂H₂ uptake of 2.9 mmol g⁻¹ at 298 K, was determined.

In 2022, B. Chen and colleagues reported that the HUM [Cu(dps)₂(SiF₆)] _n (SIFSIX‐dps‐Cu, SIFSIX = SiF₆ ²⁻, dps = 4.4′‐dipyridylsulfide) exhibits a pore confinement effect (pore size = 1.4 × 3.0 Ų) that distinguishes between C₂H₂ and CO₂ (Figure 5a) [158]. At 298 K, SIFSIX‐dps‐Cu adsorbed a high amount of C₂H₂ (4.57 mmol/g), and equimolar C₂H₂/CO₂ IAST selectivity was determined to be 1787. Computational studies indicated multiple H‐bonding interactions between C₂H₂ and SiF₆ ²⁻ within both inter‐ and intralayer spaces.

FIGURE 5.

Examples of top‐performing reticular sorbents for C₂H₂/CO₂ separation: (a) SIFSIX‐dps‐Cu [158], (b) SOFOUR‐TEPE‐Zn [156], (c) ZUL‐330 [157], and (d) UTSA‐300a [159] with feed ratio, temperature, and total pressure set at 1:1, 298 K, and 1 bar, respectively. (Reprinted with permission from ref. [158]: copyright 2022, Springer Nature; ref. [156]: copyright 2023, Wiley‐VCH GmbH; ref. [157]: copyright 2023, American Chemical Society; ref. [159]: copyright 2023, American Chemical Society.).

In 2023, taking inspiration from the prototypal sulfate‐pillared HUM, SOFOUR‐1‐Zn, SOFOUR‐TEPE‐Zn exploited 1,1,2,2‐tetra(pyridin‐4‐yl) ethene (TEPE) to create a negatively charged pore surface (Figure 5b) [156]. SOFOUR‐TEPE‐Zn exhibited high IAST selectivity of 16,833 and C₂H₂ uptake of 3.98 mmol/g. Computational studies indicated multiple H‐bonding interactions between C₂H₂ molecules and pore walls. Conversely, only weak van der Waals interactions were observed for CO₂.

In 2023, two‐layered HUMs, ZUL‐330 (SiFSIX‐dpm‐Cu, [Cu(4,4′‐dipyridylmethane)₂(SiF₆)] _n ) and ZUL‐430 (GeFSIX‐dpm‐Cu, [Cu(4,4′‐dipyridylmethane)₂(GeF₆)] _n ) were reported by Q. Yang and co‐workers (Figure 5c). Compared to ZUL‐430, ZUL‐330 exhibited high C₂H₂ uptake (5.91 mmol/g for ZUL‐330; 5.74 mmol/g for ZUL‐430) and outstanding C₂H₂/CO₂ IAST selectivity (10086 for ZUL‐330; 3499 for ZUL‐430) [157]. Subtle structural differences in these sorbents resulted in significant differences in separation performance. Dispersion‐corrected density functional theory (DFT‐D) calculations were consistent with the superior separation performance of ZUL‐330. Specifically, H‐bonding and van der Waals interactions in ZUL‐330 and ZUL‐430 resulted in higher C₂H₂ binding energy for ZUL‐330 (86.1 kJ/mol) than ZUL‐430 (82.4 kJ/mol), and lower CO₂ binding energy for ZUL‐330 (44.5 kJ/mol) than ZUL‐430 (45.8 kJ/mol).

In 2017, B. Chen et al. reported a flexible layered HUM with 2.4 × 3.3 Ų pores (open phase), [Zn(dps)₂(SiF₆)] _n (UTSA‐300, dps = 4,4′‐dipyridylsulfide). UTSA‐300 exhibited high IAST selectivity of 1000 for C₂H₂/CO₂ and 10000 for C₂H₂/C₂H₄ (Figure 5d) [159]. Closed‐pore and open‐pore phases were structurally characterized for UTSA‐300. Strong C─H···F and π···π stacking interactions were observed in closed‐pore UTSA‐300a, resulting in the shrinkage of the structure. As pressure increased, C₂H₂ molecules were observed to bind to two hexafluorosilicate F atoms, thereby breaking intra‐network hydrogen bonds and enabling transformation to the open‐pore structure. As a result, UTSA‐300a was found to exhibit high C₂H₂ uptake (4.57 mmol/g) accompanied by exclusion of C₂H₄ and CO₂ under ambient conditions. In 2020, B. Chen et al. reported an isostructural HUM, UTSA‐300‐Cu, [Cu(dps)₂(SiF₆)] _n (also known as NCU‐100) [231]. Substitution from zinc(II) to copper(II) resulted in longer Cu‐F distances that expanded the closed pore cavities from 3.9 × 3.5 Ų to 4.3 × 3.6 Ų (Figure 6a). The Cu analog was found to offer more optimal binding with IAST C₂H₂/ C₂H₄ selectivity of 7291.3 (molecular sieving) and C₂H₂ uptake of 4.57 mmol g⁻¹.

FIGURE 6.

Examples of top‐performing reticular sorbents for C₂H₂/C₂H₄ separation with high selectivity: (a) UTSA‐300‐Cu [231], (b) SIFSIX‐14‐Cu‐i [39], (c) bnn‐1‐Ca‐H₂O [232], (d) sql‐SIFSIX‐bpe‐Zn [237] with feed ratio, temperature, and total pressure set at 1:1, 298 K, and 1 bar, respectively. (Reprinted with permission from ref. [231]: copyright 2020, American Chemical Society; ref. [39]: copyright 2017, Wiley‐VCH GmbH; ref. [232]: copyright 2020, Wiley‐VCH GmbH; ref. [237]: copyright 2021, Wiley‐VCH GmbH.).

In 2017, the second generation (Gen‐2) HUM, SIFSIX‐14‐Cu‐i ([Cu(azpy)₂(SiF₆)] _n , azpy = 4,4′‐azopyridine), was reported by Zaworotko, B. Chen, W. Zhou and Q. Ren et al. [39] The pore size of 3.4 Å was found to effectively block C₂H₄ molecules while adsorbing C₂H₂ (1.8 mmol g⁻¹ at 298 K, Figure 6b)). An IAST C₂H₂/C₂H₄ selectivity of 6320 was observed, and high‐purity (99.9999%) C₂H₄ was produced during dynamic column breakthrough experiments.

In 2020, Zaworotko et al. reported Ca(HBTC)·H₂O (H₃BTC = trimesic acid), bnn‐1‐Ca‐H₂O, an ultramicroporous MOF (3.4 Å) that exhibits a sieving effect for C₂H₂ and IAST selectivity of 6966.4 (Figure 6c) [232]. Computational studies indicated that the coordinated water molecules in bnn‐1‐Ca‐H₂O enabled the binding of C₂H₂.

In 2021, Zaworotko et al. reported a 2D layered HUM with benchmark C₂H₂ affinity (isosteric heat of adsorption (Q _st) of 67.5 kJ/mol). In this work, the title sorbent, sql‐SIFSIX‐bpe‐Zn, was found to exist in four phases: as‐synthesized (α); activated (β); two C₂H₂‐induced phases (β′ and γ) [237]. The high selectivity over C₂H₂ resulted from the induced fit of C₂H₂, resulting in a C₂H₂/C₂H₄ selectivity of 53.1 (feed ratio v/v = 1: 99, IAST, 298 K, total pressure: 1 bar). Structural insight revealed H‐bonding interactions between C₂H₂ and SiF₆ ²⁻ on site I alongside C‐H···F and C‐C···π & C─H···π interactions at site II (Figure 6d).

3.2.2. CO₂‐Selective Separation (Inverse Separation)

3.2.2.1. CO₂/C₂H₂ Separation

To obtain pure C₂H₂, the design of C₂H₂‐selective sorbents requires a desorption step, which in turn increases energy consumption and operational expenses. Adsorbents that selectively capture CO₂ instead of C₂H₂ would therefore be desirable as they could enable the direct production of high‐purity C₂H₂ during the adsorption process. In recent years, a number of CO₂‐selective CNs with high CO₂/C₂H₂ selectivity have been reported (Table 7). Table 7 presents the leading reticular sorbents reported to date for CO₂/C₂H₂ separation, listed in decreasing order of adsorption selectivity. The sorbents with relatively high selectivity were found to feature pore sizes ranging from 3 to 8 Å. Upon analyzing the performance metrics in Table 7, it is evident that the sorbents with the highest CO₂ selectivity over C₂H₂ primarily rely on H‐bonding between the sorbate and pore walls. The Q _st values for the top ten sorbents range from 25 to 51 kJ/mol, requiring relatively low regeneration energy. Five studies reporting reticular sorbents with high CO₂/C₂H₂ selectivity are discussed below.

In 2021, Guo et al. reported a photoinduced electron transfer (PIET) strategy with PMOF‐1, which features a photochromically active bipyridinium zwitterion site. This site exhibits high CO₂ selectivity, making it well‐suited for the separation of CO₂/C₂H₂ mixtures [270]. Induced by light, a photocontrollable gate effect was observed in PMOF‐1. Zwitterions have intrinsic electric field gradients due to the presence of electropositive and electronegative groups, and thus are good candidates to construct CO₂‐selective CNs [711]. Structural data revealed that CO₂ adsorption is enabled by the PIET process that increases CO₂‐framework interactions while maintaining the intrinsic electric field gradient (Figure 7a). Weaker C₂H₂ adsorption results from hydrogen bonds involving C₂H₂.

FIGURE 7.

Examples of top‐performing reticular sorbents for CO₂/C₂H₂ separation with high selectivity: (a) PMOF‐1 [270], (b) MUF‐16 [129], (c) ALF [266], and (d) MFU‐4 [268] with feed ratio, temperature, and total pressure set at 1:1, 298 K, and 1 bar, respectively. (Reprinted with permission from ref. [270]: Copyright 2021, Wiley‐VCH GmbH; ref. [129]: Copyright 2021, Springer Nature; ref. [266]: Copyright 2023, American Chemical Society; ref. [268]: Copyright 2023, Wiley‐VCH GmbH.).

Considering the lower polarizability of CO₂ than HCs, in 2021, B. Chen et al. reported the ultramicroporous PCN Cu‐F‐pymo ([Cu(5‐F‐pymo)₂]·1.25 H₂O, F‐pymo = 5‐fluoropyrimidin‐2‐olate) with pore size 3.4 × 3.4 Ų for inverse CO₂/C₂H₂ separation [267]. IAST selectivity around10⁵ suggests molecular sieving. Structural analysis of partially dehydrated Cu‐F‐pymo revealed residual water molecules occupying the C₂H₂‐preferential site, further diminishing C₂H₂ sorption. Supported by sorption experiments and modelling studies, the sieving effect precludes coadsorption of C₂H₂ and ensures high purity and productivity in cycling column breakthrough tests.

In 2021, Telfer et al. reported MUF‐16 (MUF = Massey University Framework) MOFs with selectivity for CO₂ over C₂H₂ [129]. These MOFs have optimal pore size and electrostatic moieties for CO₂ binding as well as other favorable noncovalent interactions (Figure 7b). Specifically, the amino and phenyl groups on the pore wall of MUF‐16 interact with the electropositive oxygen atoms of CO₂ molecules via H‐bonding. The equimolar CO₂/C₂H₂ IAST selectivity of MUF‐16 was found to be 510 with a CO₂ uptake of 2.13 mmolg⁻¹.

In 2023, Zhao et al. also used a molecular recognition strategy for CO₂‐selective sorption with the ultramicroporous adsorbent Al(HCOO)₃ (ALF), achieving strong CO₂ capture and C₂H₂ rejection [266]. The CO₂/C₂H₂ selectivity is facilitated by optimal pore size (4.1 × 5.3 Ų) and H‐bonding, with all HCs rejected. In situ FTIR spectroscopy of CO₂‐loaded ALF revealed that the CO₂ molecules within ALF interact with the electropositive CH moieties of formate groups (Figure 7c). ALF exhibited high selectivity of 6.5 × 10⁵ with CO₂ uptake of 3.85 mmol/g at 298 K.

In 2023, Wade et al. reported on MFU‐4 (Zn₅Cl₄(bbta)₃, bbta = benzo‐1,2,4,5‐bistriazolate), which exhibited a size‐sieving effect that effectively excluded C₂H₂ through narrow pore windows (2.2 × 2.2 Ų) formed by Zn–Cl groups (Figure 7d) [268]. Column breakthrough experiments demonstrated a high kinetic selectivity of 3360, determined from the ratio of the diffusion time constants for CO₂ and C₂H₂. Additionally, the MFU‐4‐packed column was easily regenerated by purging with helium at room temperature.

As presented in Figure 8, eight mechanisms have been asserted in the 15 top‐performing reticular sorbents that demonstrate C₂H₂/C₂H₄, C₂H₂/CO₂, and CO₂/C₂H₂ adsorption selectivity: pore size (e.g., ultramicropores inducing partial and/or complete size sieving of C₂H₂ over C₂H₄ or CO₂ over C₂H₂); π–π interactions (including π···π and C–H···π interactions); framework flexibility (gate opening or switching selective to C₂H₂); H‐bonding; electrostatics; dipole‐dipole interactions; van der Waals interactions; UMCs. A radar plot for these three binary mixtures (Figure 8) indicates that hydrogen bonding between sorbates and sorbents tends to play a crucial role in both C₂H₂‐ and CO₂‐selective separations. In essence, the acidic alkyne hydrogen atoms in C₂H₂ and the electropositive oxygen atoms in CO₂ make it feasible for adsorbents embedded with hydrogen donors and acceptors to be highly selective. Additionally, optimal pore size is important for C₂H₂/C₂H₄ and CO₂/C₂H₂ mixtures, with molecular sieving achieving ultra‐high selectivity in a number of examples as discussed above. When combined with moderate regeneration conditions, the advances summarized in this section indicate that adsorbents that exhibit a molecular sieving effect offer strong potential for further development in commodity purification applications.

FIGURE 8.

Radar plot showing the contributions of eight separation mechanisms (and their combinations) among the 15 top‐performing reticular sorbents that exhibit binary C₂H₂/C₂H₄, C₂H₂/CO₂, and CO₂/C₂H₂ adsorption selectivities. Each concentric octagon represents two sorbents, while the central octagon represents one.

3.2.3. C₂H₄‐Selective Separation

3.2.3.1. C₂H₄/C₂H₆ Separation

C₂H₄ is one of the most widely produced chemicals in the petrochemical industry, serving as the precursor to a wide range of solvents, polymers, and chemical commodities, including plastics and coatings [712, 713]. Industrially, C₂H₄ is manufactured mainly by steam cracking of naphtha or dehydrogenation of C₂H₆, invariably resulting in the formation of C₂H₆ as a byproduct [714, 715]. The production of C₂H₄ requires cryogenic separation from C₂H₆, an energy‐consuming step. This drives the development of adsorptive separation of C₂H₄ from C₂H₆ using solid sorbents. In recent years, several reticular sorbents have been reported for C₂H₄/C₂H₆ separation to produce polymer‐grade (>99.9% purity) C₂H₄ (Table 8). Table 8 presents the leading sorbents reported to date for C₂H₄/C₂H₆ separation, listed in decreasing order of adsorption selectivity. The sorbents with the highest selectivity in Table 8 were found to have a wide range of pore sizes ranging from 3.8 to 11 Å. Among the top ten reticular sorbents, some of those with smaller pores (3–5 Å) exhibited high selectivity due to molecular sieving. Other sorbents also demonstrated high selectivity, with high Q _st values for C₂H₄. Five studies reporting reticular sorbents with high selectivity are discussed below.

In 2017, Bereciartua et al. reported kinetic separation of C₂H₄/C₂H₆ in a flexible zeolite (ITQ‐55) [289]. ITQ‐55 possesses two parallel zig‐zag pore channels with a window aperture of 2.33 × 5.71 Ų, constraining C₂H₄ in this channel induced expansion to 3.08 × 5.71 Ų as indicated by ab initio molecular dynamics simulations (Figure 9a). Kinetic studies revealed that ITQ‐55 exhibits faster adsorption for C₂H₄ than C₂H₆ and selectivity of ∼100 with C₂H₄ uptake of 1.28 mmol g⁻¹ at 298 K. This kinetic separation of a C₂H₄/C₂H₆ mixture was also demonstrated by dynamic column breakthrough (DCB) experiments.

FIGURE 9.

Examples of top‐performing reticular sorbents for C₂H₄/C₂H₆ separation with high selectivity: (a) Zeolite ITQ‐55 [289], (b) NUS‐6(Hf)‐Ag [288], (c) UTSA‐280 [287], and (d) M‐gallate (M = Ni, Mg, Co) [291] with feed ratio, temperature, and total pressure set at 1:1, 298 K, and 1 bar, respectively. (Reprinted with permission from ref. [289]: Copyright 2018, American Association for the Advancement of Science; ref. [288]: Copyright 2017, American Chemical Society; ref. [287]: Copyright 2018, Springer Nature; ref. [291]: Copyright 2020, American Chemical Society.).

In the same year, a silver‐modified Hf‐MOF, NUS‐6(Hf)‐Ag, demonstrated olefin selectivity and effectiveness for C₂H₄/C₂H₆ separation (Figure 9b) [288]. High IAST selectivity of 106 for equimolar C₂H₄/C₂H₆ was obtained with C₂H₄ uptake of 2.02 mmol/g at 298 K. Despite the relatively large pores dimensions (10×17 Å), C₂H₄ and C₂H₆ molecules were distinguished by interactions between C₂H₄ and Ag⁺ cations driven by electron donation from C₂H₄ to Ag⁺ and concomitant back bonding from the d orbital of Ag⁺ to the vacant π* orbital of C₂H₄. NUS‐6(Hf)‐Ag demonstrated good recyclability without any appreciable loss of C₂H₄ uptake capacity and C₂H₄/C₂H₆ adsorption selectivity over three consecutive DCB cycles. These results highlight the utility of post‐synthetic incorporation of metal salts into MOFs for olefin‐selective separation [288].

In 2018, an example of molecular sieving was detailed by Lin et al. in a rigid ultra‐microporous MOF (UTSA‐280, Figure 9c) [287]. UTSA‐280 features 1D pore channels with two aperture sizes of 3.2 × 4.5 Ų and 3.8 × 3.8 Ų, whose cross‐sectional area (14.4 Ų) is midway between those of C₂H₄ (13.7 Ų) and C₂H₆ (15.5 Ų) molecules. As a result, UTSA‐280 exhibited a high C₂H₄ uptake of 2.5 mmol/g at 298 K and 1 bar, with almost complete exclusion of C₂H₆ (0.09 mmol/g). The calculated IAST selectivity of UTSA‐280 for an equimolar C₂H₄/C₂H₆ mixture is over 10,000 at 298 K, setting a new benchmark. SCXRD analysis and DFT calculations revealed that C₂H₄ interacts with UTSA‐280 through multiple weak C─H···O hydrogen bonds and van der Waals interactions. In contrast, C₂H₆ is prevented from passing into the channels due to its larger molecular size. Breakthrough experiments confirmed that UTSA‐280 separated C₂H₄ from a C₂H₄/C₂H₆ mixture with a high productivity of 1.86 mol/kg and effluent purity >99.2%.

In the same year, another family of reticular sorbents for C₂H₄/C₂H₆ separation, M‐gallate (M = Ni, Mg, Co), was reported by Bao et al. (Figure 9d) [291]. The 3D interconnected zig‐zag channels and the aperture sizes of 3.47 × 4.85, 3.56 × 4.84, and 3.69 × 4.95 Ų for Ni, Mg, and Co‐gallate, respectively, enabled molecular sieving of C₂H₄ (3.28 × 4.18 Ų) over C₂H₆ (3.81 × 4.08 Ų). Co‐gallate exhibited a much higher uptake of C₂H₄ than C₂H₆ (3.37 mmol/g vs. 0.31 mmol/g at 298 K and 1 bar) with a high IAST selectivity of 52 at 298 K. High‐resolution neutron powder diffraction experiments revealed strong binding affinity between C₂D₄ molecules and the ligands with multiple C─D···O hydrogen bonding interactions.

In 2019, UMCs were created through immobilization of metal ions within MOFs for highly efficient C₂H₄/C₂H₆ separation. Qian et al. reported two tailor‐made copper(I)‐chelated adsorbents (CuI@UiO‐66‐COOH and CuI@UiO‐66‐(COOH)₂ ). The introduction of Cu(I) ions cannot only present multiple strong binding sites but also fine‐tune the pore size to better fit C₂H₄. CuI@UiO‐66‐(COOH)₂ was found to exhibit optimal pore size (4.1 Å) and open Cu(I) sites to selectively bind C₂H₄ through π‐complexation. The IAST selectivity of CuI@UiO‐66‐(COOH)₂ for an equimolar C₂H₄/C₂H₆ mixture was 80.8 at 298 K and 1.0 bar. The separation performance of CuI@UiO‐66‐(COOH)₂ was confirmed by the breakthrough experiments [290].

Overall, the top‐performing reticular sorbents for C₂H₄/C₂H₆ separation are enabled by (A) difference in molecular sizes (Table 2) and/or (B) strong noncovalent interactions (e.g., UMCs and a combination of multiple hydrogen bonds). Both factors can play roles in C₂H₄‐selective separation.

3.2.4. C₂H₆‐Selective Separation

3.2.4.1. C₂H₆/C₂H₄ Separation

As mentioned above, removal of impurities is required for the production of polymer‐grade C₂H₄. However, for solid sorbents, C₂H₄‐selective separation requires a second step of sorbent regeneration to obtain purified C₂H₄. One‐step purification with selective removal of C₂H₆ from C₂H₄ is therefore a more desirable approach. However, because C₂H₄ features a larger quadrupole moment (Table 2; 1.50 × 10²⁶ esu cm²) and lower polarizability (Table 2; 42.52 × 10²⁵ esu cm³) compared to C₂H₆ (Table 2; 0.65 × 10²⁶ esu cm² and 44.7 × 10²⁵ esu cm³, respectively), C₂H₄/C₂H₆ selectivity is typically observed. C₂H₆‐selective “reverse” sorption is challenging, but recent reports have achieved such selectivity by utilizing multiple C─H···O/N/F hydrogen bonds, C─H···π bonds, van der Waals interactions, or combinations thereof. Each of these interactions could be potentially augmented by structural flexibility that results in induced fit binding. C₂H₆‐selective sorbents are a topical subject [85], and Table 9 presents the leading reticular sorbents reported to date for C₂H₆/C₂H₄ separation, arranged in decreasing order of adsorption selectivity. As revealed by Table 9, compared to C₂H₄‐selective sorption, the selectivity values for C₂H₆/C₂H₄ tend to be lower, with the highest reported selectivity being 15.4. For this type of “reverse” selective sorption, pore size cannot lead to sieving as C₂H₆ is larger than C₂H₄. Rather, non‐covalent interactions, which can differ between C₂H₆ and C₂H₄, are key to their separation. As detailed in Table 9, H‐bonding has been consistently asserted as being involved in the separation mechanism. Flexibility can also play a role in C₂H₆/C₂H₄ separation, as it can enhance the generally low Q _st values observed for C₂H₆. Five studies reporting reticular sorbents with high selectivity are discussed below.

X. M. Chen and co‐workers reported a metal‐azolate framework (MAF) in 2015, MAF‐49 [Zn(batz)] _n (H₂batz = bis(5‐amino‐1H‐1,2,4‐triazol‐3‐yl)methane, Figure 10a), with equimolar C₂H₆/C₂H₄ IAST selectivity of 9 and C₂H₆ uptake of 2.5 mmol g⁻¹ at 316 K [314]. The combination of (a) multiple C−H···N hydrogen bonds between C₂H₆ molecules and the MAF‐49 pore surface and (b) dipolar repulsion groups located at specific positions across the narrow pore channels (3.3 × 3.0 Ų) enabled preferential adsorption of C₂H₆ over C₂H₄.

FIGURE 10.

Examples of top‐performing reticular sorbents for C₂H₆/C₂H₄ separation. (a) MAF‐49 [314], (b) Fe₂(O₂)(dobdc) [315], (c) Cu(Qc)₂ [317], and (d) X‐dia‐1‐Ni_0.89Co_0.11 [316], with feed ratio, temperature, and total pressure set at 1:1, 298 K, and 1 bar, respectively. (Reprinted with permission from ref. [314]: Copyright 2015, Springer Nature; ref. [315]: Copyright 2017, American Association for the Advancement of Science; ref. [317]: Copyright 2018, American Chemical Society; ref. [316]: Copyright 2024, American Chemical Society.).

In 2018, B. Chen et al. reported the microporous MOF Fe₂(O₂)(dobdc) (dobdc⁴⁻: 2,5‐dioxido‐1,4‐benzenedicarboxylate), which displayed highly selective separation of C₂H₆/C₂H₄ (Figure 10b) [315]. At 298 K and 1 bar, the C₂H₆ uptake and the selectivity for an equimolar C₂H₆/C₂H₄ mixture were 3.32 mmol/g and 4.4, respectively. Despite the relatively large pore dimensions (14 × 14 Ų), the Fe‐peroxo sites on the pore surface of Fe₂(O₂)(dobdc) enabled recognition of C₂H₆ and adsorption of a larger amount of C₂H₆ than C₂H₄.

In the same year, another report from B. Chen's group on controlling pore apertures and chemistry enabled the material Cu(Qc)₂ (Qc⁻ = quinolone‐5‐carboxylate) to feature a weak polar pore surface (Figure 10c). Ultramicroporous Cu(Qc)₂ , with a pore size of 3.3 × 3.3 Ų exhibited self‐adaptive sorption behavior for C₂H₆ and thus higher binding affinity (C─H···π interactions) towards C₂H₆ over C₂H₄ [317]. At 298 K and 1 bar, it presented an IAST selectivity of 3.4 for equimolar C₂H₆/C₂H₄ mixtures and a C₂H₆ uptake of 1.85 mmol/g.

As reported in 2020, introducing HC pendant groups in a family of NIIC‐20 MOFs reduced pore dimensions and increased hydrophobicity, whereas glycerol introduced hydrophilicity with OH‐groups that facilitated hydrogen bond formation [313]. Specifically, NIIC‐20‐Bu (1,2‐butane‐diol) exhibited a record‐high equimolar C₂H₆/C₂H₄ IAST selectivity of 15.4 with C₂H₆ uptake of 2.5 mmol/g at 298 K. A tight pore aperture (3.5 Å) and three adsorption sites were identified within the nanocages of NIIC‐20‐Bu. Multiple C─H···π and C─H···O interactions between C₂H₆ and two phenyl rings were observed; in parallel, carboxyl moieties enabled high inverse adsorption affinity C₂H₆ over C₂H₄.

Recently, S. M. Wang et al. reported a family of flexible MOFs, X‐dia‐1‐Ni_0.89Co_0.11 and X‐dia‐1‐Ni, that exhibit high equimolar C₂H₆/C₂H₄ IAST selectivities of 5.47 and 3.51 with respective C₂H₆ uptakes of 4.96 and 5.54 mmol g⁻¹ at 273 K (Figures 10d and 11) [316]. Insights from structural analysis revealed that C₂H₆ induced phase transformation abruptly with pressure increase, switching the sorbent from its narrow pore (NP) phase to a large pore (LP) phase. Multiple C‐H···O interactions between C₂H₆ and X‐dia‐1‐Ni‐LP were found to contribute toward the higher binding energy, −23.4 and −20.5 kJ/mol for C₂H₆ and C₂H₄, respectively, of X‐dia‐1‐Ni‐LP.

FIGURE 11.

Illustration of rigid and flexible adsorbents for C₂H₆/C₂H₄‐selective adsorption. (A) Rigid MOFs tend to show little adsorption difference between C₂H₄ and C₂H₆. (B) Flexible C₂H₄‐selective adsorbents require an additional desorption step to release the adsorbed C₂H₄ molecules. (C) Flexible C₂H₆‐selective porous materials can thereby be optimal for C₂H₆/C₂H₄ separation. (Reprinted with permission from ref. [316]: Copyright 2024, American Chemical Society.).

As summarized in Figure 12, six mechanisms have been asserted in the 15 top‐performing reticular sorbents that demonstrate C₂H₄/C₂H₆ and C₂H₆/C₂H₄ adsorption selectivities: pore size (e.g., ultramicropores inducing partial and/or complete size sieving of C₂H₄ over C₂H₆ or the opposite); π‐π interactions (including π···π and C─H···π interactions); framework flexibility (gate opening or switching selective to either C₂H₄ or C₂H₆); hydrogen bonding (H‐bonding); electrostatics; UMCs. Occurrences of these six mechanisms revealed the dominating factors behind both types of separation (C₂H₄/C₂H₆ is the “normal” selectivity, whereas C₂H₆/C₂H₄ represents “inverse” selectivity).

FIGURE 12.

Radar plot presenting the contributions of six separation mechanisms (and their combinations) among the 15 top‐performing reticular sorbents that exhibit binary C₂H₄/C₂H₆ and C₂H₆/C₂H₄ adsorption selectivities. Every concentric hexagon denotes an individual sorbent, with the central hexagon symbolizing one sorbent.

With regard to C₂H₄/C₂H₆ separation, UMCs are the most typical feature. UMCs can result from unsaturated metal centers on the pore walls or post‐synthetic modification by the introduction of salts. Specific binding interactions can be generated between vacant orbitals of the UMCs and π electrons of the olefinic C₂H₄ bond. Interactions between π electrons of C₂H₄ and aromatic rings from the linker ligands can also enable C₂H₄‐selective binding. Coadsorption is generally unavoidable and hinders the harvesting of high‐purity C₂H₄ streams. Thanks to the different molecular sizes of C₂H₄ and C₂H₆ (Table 2), pore size control can result in sieving and ultrahigh C₂H₄ selectivity. C₂H₆/C₂H₄ separations predicated upon superior C₂H₆ binding are primarily enabled by noncovalent interactions between C₂H₆ and pore walls. For instance, H‐bonding interactions between the alkane hydrogens of C₂H₆ and H‐bond acceptor sites, along with C─H···π interactions between C₂H₆ and aromatic rings, can play a significant role in C₂H₆/C₂H₄ separation.

To produce high‐purity C₂H₄, a one‐step C₂H₄ production process that removes C₂H₆ from a C₂H₆/C₂H₄ mixture is ideal, especially when considering the low energy footprint of sorbent recycling. Unfortunately, unlike C₂H₄/C₂H₆ separation, molecular sieving is inapplicable for selective capture of C₂H₆ over C₂H₄. Also, limited by the relatively low difference in binding between C₂H₆ and C₂H₄, the selectivity for C₂H₆/C₂H₄ (highest value = 15.4) is much lower than that for C₂H₄/C₂H₆ (highest value = 106). Physisorbents that feature C₂H₆‐induced structural flexibility and selective binding of C₂H₆ over C₂H₄ possess high upside potential to further boost C₂H₆/C₂H₄ selectivity.

Overall, for the binary separation of C₂H₆/C₂H₄, the IAST selectivity trends among the top‐performing reticular sorbents indicate that three approaches tend to enhance performance: (A) strengthening binding affinity towards C₂H₆; (B) enhancing host (sorbent)–guest (C₂H₆) binding strength; (C) decreasing sorbent‐C₂H₄ binding strength.

3.3. C3 Separations

C3 HCs, particularly propylene (C₃H₆) and propane (C₃H₈), are high‐volume commodity chemicals used in the production of fuels, polymers, and other chemical products. The demand for C3 HCs is on the rise. However, achieving the required purity (>99.95%) for polymer production and other chemical processes currently relies on costly and energy‐intensive methods such as cryogenic distillation, which remains the industry standard. Adsorptive separations using porous materials could provide a more energy‐efficient alternative. Notably, after C₂H₄, C₃H₆ is the second most valuable olefin feedstock in the petrochemical industry, and serves as a feedstock for synthesising various products, including polypropylene. Industrially, C₃H₆ is primarily produced through the steam cracking of naphtha, a process that inevitably introduces impurities such as C₃H₈, methylacetylene (MA, C₃H₄), and trace amounts of propadiene (PD, C₃H₄).

3.3.1. C₃H₄‐Selective Separation

3.3.1.1. C₃H₄/C₃H₆ Separation

C₃H₆ feeds are inevitably mixed with trace amounts of C₃H₄, which is undesirable because of catalyst poisoning. Polymer‐grade C₃H₆ requires that the concentration of C₃H₄ must be <5 ppm. Compared to the traditional method, cryogenic distillation, adsorptive separation of C₃H₄ from C₃H₆ using reticular porous materials promise to be more cost‐effective and energy‐efficient (Table 10). Table 10 presents the leading reticular sorbents reported to date for C₃H₄/C₃H₆ separation, arranged in decreasing order of adsorption selectivity. The sorbents with the highest selectivity were found to have pore sizes ranging from 3.4 to 8 Å. However, pore size was not observed to play a significant role in C₃H₄‐selective sorption. Compared to C1‐C2 separations, the Q _st values between C₃H₄ and the frameworks are generally higher, leading to increased energy requirements for sorbent regeneration. H‐bonding between C₃H₄ and the frameworks has been frequently observed as a key separation mechanism. Four studies reporting reticular sorbents with high C₃H₄ selectivity are discussed below.

In 2017, B. Chen and colleagues accomplished C₃H₄/C₃H₆ separation for the first time using a flexible yet robust MOF, ELM‐12 [Cu(bpy)₂(OTf)₂] (OTf⁻ = trifluoromethanesulfonate) [438]. The strong binding affinity and pore confinement for C₃H₄ result in high uptake capacity and selectivity, as demonstrated by neutron powder diffraction studies and density functional theory calculations, to achieve C₃H₆ purity of >99.9998%.

Flexible MOFs that exhibit stepwise adsorption isotherm profiles can theoretically offer high (even infinite) separation selectivity. However, their practical utility in separating gas mixtures often falls significantly short of the theoretical values derived from pure‐component sorption isotherms. This discrepancy is due to the unpredictable effects of exposure to gas mixtures. In 2018, H. Xing's group reported a flexible HUM, TIFSIX‐14‐Cu‐i (also termed as ZU‐13, TIFSIX = TiF₆ ²⁻, 14 = 4,4′‐azobipyridine, i = interpenetrated) with an anion‐pillared structure and ultramicroporous pore diameter of 3.4 Å [432]. TIFSIX‐14‐Cu‐i featured an ultra‐low C₃H₄ threshold pressure of 500 ppm, an important requirement to produce polymer‐grade C₃H₆ as the effluent. TIFSIX‐14‐Cu‐i demonstrated exceptionally high C₃H₄/C₃H₆ IAST selectivity of 355 for separating 1/99 C₃H₄/C₃H₆ mixtures, achieving an adsorption capacity of 1280 mL/g and C₃H₆ purity exceeding 99.9999%, as evidenced by breakthrough results. Density functional theory calculations revealed that the underlying mechanism for C₃H₄/C₃H₆ separation in TIFSIX‐14‐Cu‐i involves hydrogen bonding and Van der Waals interactions with the C3 sorbates (Figure 13a).

FIGURE 13.

Examples of top‐performing reticular sorbents for C₃H₄/C₃H₆ separation with high selectivity: (a) TIFSIX‐14‐Cu‐i [432] with feed ratio, temperature, and total pressure set at 1:1, 298 K, and 1 bar, respectively. (b) NKMOF‐1‐Cu [428] with feed ratio, temperature, and total pressure set at 0.5:99, 298 K, and 1 bar, respectively. (c) FJI‐W1 [431] with feed ratio, temperature, and total pressure set at 1:99, 298 K, and 1 bar, respectively. (d) BUT‐306 [430] with feed ratio, temperature, and total pressure set at 1:1, 298 K, and 1 bar, respectively. (Reprinted with permission from ref. [432]: Copyright 2018, The Royal Society of Chemistry; ref. [428]: Copyright 2019, Wiley‐VCH GmbH; ref. [431]: Copyright 2022, American Chemical Society; ref. [430]: Copyright 2020, The Royal Society of Chemistry.).

In 2019, Y. Peng et al. reported two isostructural HUMs, NKMOF‐1‐Ni and NKMOF‐1‐Cu (Cu [M(pdt)₂] (pdt = pyrazine‐2,3‐dithiol, M = Cu, Ni)) with pore dimensions of ca. 5.7 × 5.7 Ų [428]. High IAST selectivities of 1217.8 and 859.5 for the separation of C₃H₄/C₃H₆ at a v/v ratio of 0.5/99 were observed in NKMOF‐1‐Ni and NKMOF‐1‐Cu, respectively. Gas‐loaded single‐crystal structures provided insight into the separation mechanism, which involves hydrogen bonding and π···π interactions (Figure 13b). In 2021, the same research group reported an isostructural material, NKMOF‐11 (Ni [Cu(pdt)₂], pdt = pyrazine‐2,3‐dithiol). Half of the nickel (Ni) atoms in NKMOF‐1‐Ni were replaced by copper (Cu) atoms in NKMOF‐11, resulting in the same pore dimensions and high IAST selectivity of 1074 for C₃H₄/C₃H₆ at a v/v ratio of 1/99. Due to the high affinity for C₃H₄, the selectivity for C₃H₄/C₃H₆ at a v/v ratio of 1/99 can reach as high as 1388. Structural analysis indicated that. once again, hydrogen bonding and π···π interactions are crucial to the C₃H₄‐selective binding.

In 2022, M. Hong and colleagues reported a hybrid microporous material, FJI‐W1 ([Cu₃(TMTPB)₄(SiF₆)₃] _n , TMTPB = 1,3,5‐trimethyl‐2,4,6‐tris(4‐pyridyl)benzene) with pore dimensions of 8 × 8 Ų [431]. The IAST selectivity for a binary C₃H₄/C₃H₆ mixture at a v/v ratio of 1/99 in FJI‐W1 was calculated to be 593, with dynamic breakthrough experiments lasting up to 600 minutes per gram. Computational studies revealed that the separation of C₃H₄/C₃H₆ was facilitated by hydrogen bonding with C₃H₄ forming strong C─H···F hydrogen bonds with SiF₆ ²⁻ in the irregular hexagonal channels (Figure 13c).

In 2022, J. R. Li and co‐workers reported two isoreticular ultramicroporous Zn(II)‐MOFs, BUT‐305 (Zn₂(ATZ)₂(TPDC), H₂TPDC = [1,1′:3′,1′′‐terphenyl]‐4,4′′‐dicarboxylic acid, HATZ = 3‐amino‐1,2,4‐triazole) and BUT‐306 (Zn₂(ATZ)₂(MeTPDC), H₂MeTPDC = 5′‐methyl‐ [1,1′:3′,1′′‐terphenyl]‐4,4′′‐dicarboxylic acid) [430]. A pore diameter of 3 Å was observed for BUT‐306. High hydrolytic stability and hydrophobicity enabled breakthrough experiments on a binary C₃H₄/C₃H₆ gas mixture (v/v = 1/1) to be conducted with pre‐saturated water vapor, resulting in an IAST selectivity of 636. The single‐crystal structure of C₃H₄‐loaded BUT‐306 revealed that adsorbed C₃H₄ molecules lie in the center of the pores and interact with pore walls through multiple weak C ^{δ −}···C ^{δ +} dipole–dipole interactions and C─H···π interactions (Figure 13d).

3.3.2. C₃H₆‐Selective Separation

3.3.2.1. C₃H₆/C₃H₈ Separation

C₃H₆ is predominantly produced through steam cracking of naphtha or dry gas refining, and it is invariably mixed with minor impurities such as C₃H₈. Currently, the petrochemical industry relies almost entirely on cryogenic distillation for C₃H₆ purification. However, because C₃H₆ and C₃H₈ have very similar physicochemical properties (Table 2), cryogenic distillation is both inefficient and energy‐intensive. Adsorptive separation using reticular porous materials has been explored to achieve C₃H₆ purification from C₃H₆/C₃H₈ mixtures under ambient conditions. Table 11 presents the leading reticular sorbents reported to date for C₃H₆/C₃H₈ separation, listed in decreasing order of adsorption selectivity. The sorbents with high selectivity were found to exhibit pore sizes ranging from 3 to 7 Å. Pore size was observed to influence C₃H₆‐selective sorption. Interestingly, the uptake at 298 K and 1 bar for C₃H₆ did not correlate with pore size/surface area. Six reports of reticular sorbents with high selectivity are discussed below.

In 2009, C₃H₆/C₃H₈ separation was reported by Lamia et al. using a simulated moving bed (SMB) process conducted on the prototypal high surface area MOF HKUST‐1 [500]. The observed C₃H₆/C₃H₈ separation was attributed to UMCs and π‐complexation between the vacant s‐orbital of Cu(II) cations and the electron‐rich π orbitals in C₃H₆. In the same year, J. Li et al. reported the first example of the kinetic separation of C₃H₆/C₃H₈, based on their different diffusion rates in Zn(2‐cim)₂ (2‐cim = 2‐chloroimidazole), which has a pore diameter of 3.26 Å [468].

In 2017, J. Li and co‐workers reported two structurally related microporous MOFs, Zn(ox)_0.5(trz) (ox = oxalate, trz = 1,2,4‐triazole) and Zn(ox)_0.5(atrz) (atrz = 3‐amino‐1,2,4‐triazole) with pore sizes of only 2.9 and 3.3 Å, respectively [458]. Single crystal X‐ray data revealed that the pore channels in both MOFs exhibit a zig‐zag shape. As a result, due to the optimal pore size and the zig‐zag‐shaped ultramicroporous 1D channels, a significant difference in the diffusivity of C₃H₆ and C₃H₈ in Zn(ox)_0.5(trz) at 323 K was achieved, leading to a kinetic separation factor of 1565.

In 2020, B. Chen, W. Zhou, and co‐workers reported an ultramicroporous cobalt gallate MOF (Co‐gallate, [Co(gallate)] _n ) for highly C₃H₆‐selective sieving of C₃H₆ from C₃H₈ at 298 K [462]. The optimal pore structure of Co‐gallate enabled confinement of C₃H₆ while excluding the slightly larger C₃H₈, as demonstrated in the neutron diffraction crystal structure of Co‐gallate⊃0.38C₃D₆ (Figure 14a). The IAST selectivity for equimolar C₃H₆/C₃H₈ separation was calculated to be 330 at 298 K. This high separation performance was subsequently supported by gas sorption isotherms and column breakthrough experiments, resulting in high purity C₃H₆ (97.7%). Structural analysis indicated that hydrogen bonding, O‐H···π interactions, and van der Waals forces resulted in selective binding.

FIGURE 14.

Top‐performing reticular sorbents for C₃H₆/C₃H₈ separation: (a) Co‐gallate [462], (b) JNU‐3a [461], (c) MIP‐203 [460], and (d) NTU‐85‐WNT [457] with feed ratio, temperature, and total pressure set at 1:1, 298 K, and 1 bar, respectively. (Reprinted with permission from ref. [462]: Copyright 2020, American Chemical Society; ref. [461]: Copyright 2021, Springer Nature; ref. [460]: Copyright 2022, The Chemical Industry and Engineering Society of China, and Chemical Industry Press Co., Ltd. All rights reserved; ref. [457]: Copyright 2023, American Chemical Society.).

The potential separation ability of flexible MOFs has also been reported, but the rational design of flexibility in MOFs for dynamic molecular sieving remains a challenge. In 2021, D. Li et al. reported a flexible MOF, JNU‐3a (JNU = Jinan University, Co(MPTBDC)₂, MPTBDC = 5‐(3‐methyl‐5‐(pyridin‐4‐yl)‐4H‐1,2,4‐triazol‐4‐yl)‐1,3‐benzenedicarboxylate), exhibiting a molecular sieving effect for C₃H₆/C₃H₈ separation [461]. JNU‐3a features one‐dimensional channels with embedded molecular pockets that open to C₃H₆ and C₃H₈ at significantly different pressures (Figure 14b). The dynamic nature of these pockets is demonstrated by the single‐crystal‐to‐single‐crystal transformation of JNU‐3a when exposed to an atmosphere of C₃H₆ or C₃H₈. JNU‐3a exhibited a high IAST selectivity of 513 at 298 K and produced high‐purity C₃H₆ (>99.5%). The underlying separation mechanism, involving orthogonal array‐like dynamic molecular sieving, enabled both a large separation capacity and rapid adsorption‐desorption kinetics. In 2022, two flexible Zr‐based MOFs (MIP‐202 and MIP‐203) were prepared with C4 linkers and studied for C₃H₆/C₃H₈ separation [460]. The temperature‐induced and guest‐induced flexibility of C₃H₆‐selective MIP‐203 resulted in an IAST selectivity of 551.4 at 303 K. Computational studies provided insight into binding sites, which were found to involve H‐bonding and electrostatic interactions (Figure 14c). O‐H···π H‐bonding formed by the ‐OH group in the MBB metal cluster and sp² carbon atoms from C₃H₆ play a key role in the experimentally observed high C₃H₆/C₃H₈ selectivity of MIP‐203.

In 2023, J. Bai and co‐workers reported a super tetrahedral‐cluster Cu₁₀O₁₃‐based MOF (NTU‐85‐WNT) and studied its C₃H₆/C₃H₈ separation performance. With pore dimensions of 4.57 × 4.57 Ų [457], NTU‐85‐WNT recorded IAST selectivity of 1570 at 298 K for C₃H₆/C₃H₈. The C₃H₆ selectivity was credited to a new mechanism involving initial framework expansion and subsequent contraction of the confined water nanotubes (∼4.5 Å) exclusively triggered by C₃H₆ adsorption (Figure 14d). Thanks to the robust nature of the framework, the water nanotubes were restored by soaking the MOF in water.

3.3.3. C₃H₈‐Selective Separation (Inverse Separation)

3.3.3.1. C₃H₈/C₃H₆ Separation

Compared to C₃H₆‐selective adsorption from a C₃H₆/C₃H₈ mixture, the preferential removal of C₃H₈ impurities could be more practically useful, as there would be no need for a C₃H₆ desorption process. The subtle structural differences between C₃H₆ and C₃H₈ mean only slight differences in size and shape, rendering the development of C₃H₈‐selective materials a challenge. The use of C₃H₈‐selective CNs as adsorbents for the one‐step purification of C₃H₆ from C₃H₈/C₃H₆ binary mixtures is still in its early stages, with only a limited amount of literature available. In 2013, Bohme et al. reported separation of C₃H₈/C₃H₆ by CPO‐27‐Co (Co₂(dhtp), dhtp = dihydroxyterephthalate) and ZIF‐8 (Zn(MIM)₂, MIM = 2‐methylimidazole) [716]. Preferential sorption of C₃H₈ over C₃H₆ was observed in ZIF‐8, thanks to its nonpolar pore environment, which lacks any specific interaction with C₃H₆. Although these initial experiments indicated promising potential for C₃H₈‐selective adsorbents, there is room to improve capacities and selectivities. Table 12 presents the leading reticular sorbents reported for C₃H₈/C₃H₆ separation, arranged in decreasing order of adsorption selectivity. Similar to C₂H₆/C₂H₄ separation, C₃H₈/C₃H₆ selectivity is generally not as high as the reverse pair (C₃H₆/C₃H₈), and pore size does not seem to have a significant influence on C₃H₈ selectivity. The pore sizes of the highly C₃H₈‐selective sorbents are mostly within the range of 4–12 Å. The relatively large pore sizes mean that the uptake of C₃H₈ can be high (up to 8.73 mmolg⁻¹). The separation mechanism relies on non‐covalent interactions, with Q _st values between C₃H₈ and the frameworks ranging from 31 to 66 kJ/mol. Seven studies reporting reticular sorbents with high selectivity are discussed below.

In 2019, H. Xing and colleagues reported a microporous HUM, [Ni(bpe)₂(WO₄)] _n (bpe = 1,2‐bis(4‐pyridyl)ethylene), featuring a polycatenated molecular cage. This HUM exhibited IAST selectivity of 1.75 for an equimolar (v/v) C₃H₈/C₃H₆ mixture. Structural analysis revealed that the cage provided dense electronegative binding sites, facilitating multiple C^δ−─H^δ+···C^δ− interactions with C₃H₈ molecules, thereby conferring higher affinity for C₃H₈ over C₃H₆ (Figure 15a). Further, the cage exhibited shape selectivity for the oblate C₃H₈, while being less favourable to the relatively planar C₃H₆.

FIGURE 15.

Top‐performing reticular sorbents for C₃H₆/C₃H₈ separation with high selectivity: (a) Ni(bpe)₂(WO₄) [505], (b) PCP‐IPA [325], (c) FDMOF‐2 [507], and (d) JNU‐90 [506] with feed ratio, temperature, and total pressure set at 1:1, 298 K, and 1 bar, respectively. (Reprinted with permission from ref. [505]: Copyright 2021, American Chemical Society; ref. [325]: Copyright 2021, Springer Nature; ref. [507]: Copyright 2023, Wiley‐VCH GmbH; ref. [506]: Copyright 2024, American Chemical Society.).

In 2021, a robust MOF, Ni(ADC)(TED)_0.5 (ADC = 9,10‐anthracenedicarboxyl, TED = 1,4‐diazabicyclo [2.2.2] octane or triethylenediamine) featuring sorbate‐induced pore confinement for C₃H₈ trappin, was reported to purify C₃H₆ [504]. Ni(ADC)(TED)_0.5 demonstrated excellent separation performance for C₃H₈/C₃H₆ (v/v = 1/1), with a record‐high IAST selectivity of 6.4. Structural analysis revealed that the C─H···π interactions between the aromatic rings of the MOF and the C₃H₈ molecules enable the observed high C₃H₈‐selective sorption.

In 2022, H. Xing and co‐workers reported an ultramicroporous material PCP‐IPA, ([Co(IPA)(DPG)] _n (PCP‐IPA; PCP  =  porous coordination polymer, IPA  =  isophthalic acid, DPG  =  meso‐α,β‐di(4‐pyridyl) glycol)) with parallel‐aligned linearly extending isophthalic acid units along a one‐dimensional channel with dimensions of 4.7 × 5.6 Ų. PCP‐IPA was found to produce high‐purity C₃H₆ (99.99%) [325]. PCP‐IPA also achieved a separation selectivity of 2.48 and a separation potential of 1.20 mol/L for an equimolar C₃H₈/C₃H₆ binary mixture. Computational studies indicated that the periodically expanded and parallel‐aligned aromatic moieties acted as a paraffin nano‐trap, engaging the exposed hydrogen atoms of C₃H₈ through hydrogen bonding and van der Waals interactions. (Figure 15b).

In 2023, the incorporation of fluorinated functional groups into the confined pore space of a MOF was reported in FDMOF‐2, Zn₂(BDC‐(CF₃)₂)₂(DABCO) (BDC‐(CF₃)₂ = 2,5‐bis(trifluoromethyl) terephthalate, DABCO = 1,4‐diazabicyclo [2.2.2]octane)). FDMOF‐2 effected C₃H₈‐selective separation for an equimolar C₃H₈/C₃H₆ mixture [507]. IAST selectivity of 2.18 for C₃H₈/C₃H₆ (v/v = 1/1) separation was measured for FDMOF‐2. SCXRD studies revealed that the tailored pore confinement in FDMOF‐2 facilitates stronger and multiple attractive interactions (C─H···π interactions and hydrogen bonding) between the ‐CF₃ moieties and C₃H₈ (Figure 15c), enabling high C₃H₆ purity (99.99%). The separation performance of FDMOF‐2 was maintained even at 70% relative humidity.

Machine learning is increasingly being used fpr prediction of materials properties. In 2024, D. Li and colleagues employed computational screening methods of existing MOFs for C₃H₈/C₃H₆ separation, identifying JNU‐90 (Zn(4,4′‐(hexafluoroisopropylidene) bis(benzoic acid)) as a promising candidate [506]. IAST selectivity of 2.7 was obtained, a number consistent with the computational results. H‐bonding and C─H···π interactions between C₃H₈ and the ‐CF₃ functional groups in JNU‐90 are thought to play a role in facilitating the observed C₃H₈/C₃H₆ selectivity. (Figure 15d).

For the three C3 HC binary pair separations discussed above, the distinct structural and physicochemical characteristics (of C₃H₄, C₃H₆, and C₃H₈) resulted in different primary separation mechanism(s) for each pair. In the case of C₃H₄‐selective separation over C₃H₆, hydrogen bonding was found to play the key role, as revealed by Figure 16. Similar to C₂H₂, C₃H₄ has relatively acidic alkyne hydrogen atoms, which can act as hydrogen bond donors to heteroatoms that can serve as hydrogen bond acceptors. For C₃H₆ separation over C₃H₈, it was found that reticular sorbents with hydrogen acceptors and appropriate pore sizes generate stronger noncovalent interactions with C₃H₆ than C₃H₈. As for C2 HCs, one‐step C₃H₈‐selective separation over C₃H₆ is more challenging. Designing sorbents with electron‐rich groups that facilitate the formation of C─H···π, π···π, and van der Waals interactions between C₃H₈ and the sorbents was found to be the most effective strategy.

FIGURE 16.

Radar plot presenting the contribution of nine separation mechanisms (and their combinations) among the 15 top‐performing reticular sorbents that exhibit binary C₃H₄/C₃H₆, C₃H₆/C₃H_8, and C₃H₈/C₃H₆ adsorption selectivities. Each concentric enneagon represents one sorbent, while the central enneagon represents two.

3.4. C1‐C3 Multicomponent Separation

Industrial production of HCs through steam cracking of crude petroleum generates unwanted by‐products. As detailed above, the separation of binary mixtures using HC‐selective adsorbents has been addressed. However, from a practical perspective, the separation of a more complex multicomponent gas mixture is more reflective of real‐world scenarios. This necessitates a more nuanced level of molecular recognition capability in physisorbents (Table 13). To design reticular physisorbents with optimal separation abilities for multicomponent mixtures, an understanding of structure‐function relationships is needed. Table 13 presents the leading reticular sorbents reported to date for C1‐C3 multicomponent hydrocarbon separations. Sorbents for multicomponent mixtures are typically reported with single‐component selective sorption from mixtures, with one binary pair playing a dominant role. The mechanisms of C1‐C3 separation are similar to those presented for binary separations, for three types of reticular sorbents, HUMs, MOFs, and COFs, in this context.

In 2012, B. Chen, Krishna, and co‐workers reported a MOF, [Cu₃(H₂L)(H₂O)₃ ], UTSA‐34 (UTSA = University of Texas at San Antonio; H₈L = 1,2,4,5‐tetra(5‐isophthalate)benzene), with high separation capacity and selectivity for C2 HCs over CH₄ [425]. The activated form of UTSA‐34b possesses pore cavities of 12.8 Å and high‐density Cu(II) UMCs, which facilitate a high separation capacity of 3.0 mol kg⁻¹ and a selectivity of 35 for the separation of C2 HCs from CH₄ at 298 K. In the same year, B. Chen and co‐workers utilized H₃BTN (6,6′,6′′‐benzene‐1,3,5‐triyl‐2,2′,2′′‐trinaphthoic acid and Cd(NO₃)₂·4H₂O to synthesize Cd₃(BTN)₂(H₂O)₃(DMF)₆ (UTSA‐35a) [411]. With pore dimensions of 7.7 × 5.8 Å, UTSA‐35a demonstrated high selectivities for multiple components, with selectivities for C₃H₆/CH₄ (1/1) of 80 and C₃H₈/CH₄ (1/1) of 75. Van der Waals interactions were found to enable the observed selective adsorption of higher HCs.

In 2019, Zaworotko, Chen, and colleagues reported three ultramicroporous CN sorbents that collectively achieved one‐step purification of C₂H₄ from a mixture of C₂H₂/C₂H₄/C₂H₆/CO₂, introducing the concept of synergistic sorbent separation technology (SSST) for multicomponent separations [70]. SIFSIX‐3‐Ni (SIFSIX = SiF₆ ²–, 3 = pyrazine), Zn‐atz‐ipa (atz = 3‐amino‐1,2,4‐triazolate; ipa = isophthalate), and TIFSIX‐2‐Cu‐i (TIFSIX = TiF₆ ²–, 2 = 4,4′‐dipyridylacetylene, i = interpenetrated) were packed in tandem in a column, with SIFSIX‐3‐Ni selectively removing CO₂ from a C₂H₂/C₂H₄/C₂H₆/CO₂ mixture, TIFSIX‐2‐Cu‐i capturing C₂H₂, and Zn‐atz‐ipa offering high inverse selectivity for C₂H₆. Breakthrough experiments confirmed that the SSST approach produced high‐purity C₂H₄ in one step from a four‐component gas mixture. Computational studies revealed that hydrogen bonding (C‐H···F) between the hydrogen atoms of HCs and TiF₆ ²⁻ anions, along with electrostatic interactions between CO₂ molecules and electronegative fluorine atoms, enabled the observed separation (Figure 17a).

FIGURE 17.

Examples of top‐performing reticular sorbents for separating multicomponent C1‐C3 gas mixtures with high selectivity: (a) SIFSIX‐3‐Ni [70], (b) MOF‐303 [533], (c) BFFOUR‐Cu‐dpds [525], and (d) NKCOF‐62 [413] with feed ratio, temperature, and total pressure set at 1:1, 298 K, and 1 bar, respectively. (Reprinted with permission from ref. [70]: Copyright 2019, the Authors; ref. [533]: Copyright 2022, the Authors; ref. [525]: Copyright 2024, the Authors; ref. [413]: Copyright 2023, American Chemical Society.).

J. Li and colleagues reported in 2023 a record‐high IAST selectivity of 5114 for the separation of C₃H₈/CH₄ (v/v = 5:85) using MOF‐303 (Al(OH)(PZDC), PZDC = 1‐H‐pyrazole‐3,5‐dicarboxylate) [533]. High surface polarity, combined with a high density of heteroatoms and an appropriate pore diameter (5‐7 Å), is the key factor that was attributed to the observed separation performance (Figure 17b). MOF‐303 was shown to efficiently separate C₃H₈/C₂H₆/CH₄ through ternary breakthrough experiments. Most recently, in 2024, S. S. Chen, X. Z. Guo, and colleagues reported on a paddle‐wheel Co(II)‐MOF, FNU‐2 (Co₂(NTA)(bpy)₂, NTA = 4,4′,4″‐tricarboxylate triphenylamine), that purified CH₄ from a C₃H₈/C₂H₆/CH₄ ternary mixture, achieving a selectivity of 638.9 for C₃H₈/CH₄ (v/v = 5:85) [519]. Computational studies revealed that van der Waals interactions between guest molecules and FNU‐2 played a crucial role in influencing its separation performance. J. Wang and colleagues, also in 2024, reported that BFFOUR‐Cu‐dpds (Cu(dpds)₂(BF₄)₂, BFFOUR  =  BF₄ ⁻, dpds = 4,4′‐bipyridinedisulfide) simultaneously sieved C₂‐C₄ olefins from their corresponding paraffins [525]. Interlayer spaces were opened through hydrogen bonding and C─H···π sorbent‐sorbate interactions induced by the unsaturated C = C bonds in olefins (Figure 17c). BFFOUR‐Cu‐dpds simultaneously separated olefins from paraffins in an equimolar six‐component mixture of C₂H₄/C₂H₆/C₃H₆/C₃H₈/n‐C₄H₈/n‐C₄H₁₀. High‐purity C₂H₄ (> 99.99%) was obtained in a second column using granular porous carbons.

In 2023, Z. Zhang's group reported a robust olefin‐linked COF, NKCOF‐62, synthesized through the melt polymerization method using tetramethylpyrazine and terephthalaldehyde as monomers [413]. NKCOF‐62 ushers in cost‐effective, kilogram‐scale fabrication of olefin‐linked COFs in a single pot reaction. With a pore size of 8 × 8 Ų, NKCOF‐62 was found to facilitate selective adsorption of C₂H₂ and C₂H₆ over C₂H₄, allowing polymer‐grade C₂H₄ to be obtained directly from C₂H₂/C₂H₆/C₂H₄ (1/1/1) mixtures through a one‐step process. GCMC simulations indicated that noncovalent interactions between C2 HCs and the pore walls of NKCOF‐62 contributed to the separation of C₂H₂ and C₂H₆ over C₂H₄ (Figure 17d).

For the separation of multicomponent C1–C3 mixtures, seven mechanisms were identified among the 15 top‐performing reticular sorbents: hydrogen bonding; electrostatics; kinetics, pore size; polarizability; π···π and C─H···π interactions; and van der Waals interactions (Figure 18). This finding is consistent with the results in the sections on binary C1, C2, and C3 separations. Overall, hydrogen bonding, π···π and interactions, and van der Waals interactions are the three primary mechanisms behind the top‐performing C1‐C3 reticular sorbents.

FIGURE 18.

Radar plot illustrating the contribution of seven separation mechanisms (and their combinations) across the 15 top‐performing reticular sorbents, each of which demonstrates C1–C3 HC adsorption selectivities. Every concentric heptagon denotes an individual sorbent, with the central heptagon symbolizing one sorbent.

3.5. C4 Separation

C4 HCs, including 1,3‐butadiene (C₄H₆), 1‐butene (n‐C₄H₈), isobutene (i‐C₄H₈), cis‐2‐butene (2‐C₄H₈), isobutane (i‐C₄H₁₀), and n‐butane (n‐C₄H₁₀) serve as raw materials for a range of products such as synthetic rubbers and elastomers, resins, detergents, fuels, plastics, and more [717]. Generally, C4 olefins are produced through fluidized catalytic cracking (FCC), steam cracking of petroleum or methanol‐to‐olefins (MTO) processes [718]. C4 olefins produced from these processes usually comprise 30%–60% C₄H₆, 10–30% i‐C₄H₈, 10–20% n‐C₄H₈, 2–6% 2‐C₄H₈, 1–3% i‐C₄H₁₀, and 3–10% n‐C₄H₁₀ [719]. However, due to their similar molecular shapes and physicochemical properties, developing adsorbents for C4 olefin separation is a challenge. Extractive distillation is the established method for obtaining high‐purity C₄H₆ (greater than 99.5%) and operates at high pressure (3 bar) and elevated temperatures (323 to 393 K) with tall towers (over 110 trays). This process is energy‐intensive, environmentally unfriendly, and carries the risk of C₄H₆ polymerization at high temperatures [720]. Reticular adsorbents with tailored pore size, shape, and controllable surface functionalization offer an opportunity to address these handicaps and develop an economically viable separation process without the need for heat (Table 14). Table 14 presents the leading reticular sorbents reported to date for C4 isomer separation, arranged in decreasing order of adsorption selectivity. The pore sizes of sorbents with high selectivity for C4 isomer separation (4–12 Å) are generally larger than those for C1–C3 sorbents, due to the larger molecular size of C4 isomers. However, an analysis of studies on sorbents for C4 separation reveals that pore size is not a key parameter in achieving the separation. Instead, noncovalent interactions, primarily hydrogen bonding, tend to play a role in the separation process. The regeneration temperatures for the foregoing C4 sorbents typically fall within the range of 323 K to 423 K. Seven studies reporting high selectivity are discussed below.

Assen et al. reported an fcu‐MOF with precisely tuned pore size to achieve efficient molecular sieving of n‐C₄H₁₀ from i‐C₄H₁₀ in 2015 [569]. This rare‐earth (RE) fcu‐MOF was built from 12‐connected RE hexanuclear clusters and formed well‐defined triangular windows as well as interconnected octahedral and tetrahedral cages (Figure 19a). The Y‐fcu‐MOF analogue ((CH₃)₂NH₂)₂ [Y₆(μ₃‐OH)₈(fum)₆(H₂O)₆], fum = fumarate) excluded branched paraffins from normal paraffins. Its window of approximately 4.7 Å diameter, considered a sorbate‐size cut‐off, enabled sieving of branched paraffins from normal paraffins, with n‐C₄H₁₀ and i‐C₄H₁₀ capacities of 1.97 and 0.05 mmol/g, respectively.

FIGURE 19.

Examples of top‐performing reticular sorbents for C4 isomer separation with high selectivity: (a) fcu‐MOF [569], (b) Zn₂(btm)₂ [82], (c) MnINA [577], (d) ZU‐601 [578], and (e) COF‐300 [583] with feed ratio, temperature, and total pressure set at 1:1, 298 K, and 1 bar, respectively. (Reprinted with permission from ref. [569]: Copyright 2015, Wiley‐VCH GmbH; ref. [82]: Copyright 2017, The American Association for the Advancement of Science; ref. [577]: Copyright 2019, Wiley‐VCH GmbH; ref. [578]: copyright 2024, American Chemical Society; ref. [583]: Copyright 2022, Institute of Process Engineering, Chinese Academy of Sciences.).

Conventional adsorbents typically favour adsorption of smaller, higher‐polarity C₄H₆ over other C4 HCs. In 2017, X. M. Chen and colleagues reported inverse selectivity for C4 HCs by Zn₂(btm)₂ (H₂btm = bis(5‐methyl‐1H‐1,2,4‐triazol‐3‐yl)methane), which is comprised of quasi‐discrete pores (Figure 19b) [82]. As revealed by SCXRD and computational studies, the pores of Zn₂(btm)₂ can induce conformational changes in the flexible guest molecules, weakening C₄H₆ adsorption through a significant bending energy penalty. This allows C₄H₆ to be purified at 298 K (>99.5%) using a guest conformation‐controlling adsorbent that elutes C₄H₆ first, followed by C₄H₁₀, n‐C₄H₈, and i‐C₄H₈.

In 2019, H. Xing and colleagues reported a flexible MOF, MnINA (INA = isonicotinate), which features one‐dimensional pore channels with periodic cavities connected by narrow windows. MnINA was found to exhibit n‐/i‐C₄H₈ selectivity [577]. With an optimal pore size of 4.62 Å, MnINA exhibits steep adsorption isotherms and a high capacity of 1.79 mmol g⁻¹ for n‐C₄H₈ (4.46 Å) enabled by C−H⋅⋅⋅π interactions (Figure 19c). At the same time, bottlenecks create barriers for larger i‐C₄H₈ molecules (4.84 Å) in the gate‐opening pressure range of 0‐0.1 bar. MnINA achieved selectivity of 327.7 (based on the initial slope ratio, n‐/i‐C₄H₈ v/v = 1/1) for n‐/i‐C₄H₈. In 2024, the same group reported two flexible sulfonate‐functionalized organic frameworks, ED‐6‐Cu (ZU‐601) and 13‐6‐Cu (ZU‐602) (ED = 1,2‐ethanedisulfonate, 13 = benzene‐1,3‐disulfonate, 6 = 1,3‐di(4‐pyridyl)propane, ZU = Zhejiang University) for C4 separation [578]. Thanks to the right aperture size enabled by the flexibility of coordinated organic ligands, ZU‐601 (6.7 × 2.2 Ų) and ZU‐602 (6.8 × 2.5 Ų), which are pillared with different sulfonate anions, can discriminate between C4 olefin isomers. Their layered structure allows the utilization of both intra‐ and interlayer space so that ZU‐601 exhibits a record‐high C₄H₆ adsorption capacity of 2.90 mmol/g (0.5 bar, 298 K) and a high uptake ratio for C₄H₆/n‐C₄H₈ (207), C₄H₆/trans‐C₄H₈ (10.1). DFT‐D calculations revealed that the C₄H₆ molecule within ZU‐601 binds to the oxygen atoms of three 1,2‐ethanedisulfonate anions from adjacent layers at shorter distances than n‐C₄H₈ (Figure 19d).

In 2022, C. Zhong's team reported a flexible MOF, Mn‐bpdc (bpdc = 2,2′‐bipyridine‐4,4′‐dicarboxylate), with a gate‐opening effect for C₄H₆ separation from C4 HCs (n‐C₄H₈, i‐C₄H₈, n‐C₄H₁₀, i‐C₄H₁₀) [582]. In this study, C₄H₆ triggers the gate‐opening of Mn‐bpdc at 0.13 bar and 298 K, allowing C₄H₆ uptake, whereas other C4 HCs did not induce gate‐opening even at 1 bar. The selectivities of Mn‐bpdc for C₄H₆/n‐C₄H₈ and C₄H₆/i‐C₄H₈ reach 40.0 and 45.0, respectively, at 298 K and 1 bar, respectively. DFT calculations indicated multiple H‐bonding interactions between the one‐dimensional channel of Mn‐bpdc and C₄H₆, with a static adsorption energy of C₄H₆ of up to 74.33 kJ mol⁻¹.

In 2020, H. Xing and co‐workers reported ZU‐36‐Co (GeFSIX‐3‐Co, GeFSIX = GeF₆ ²⁻, 3 = pyrazine) for separating both C4 linear/branched olefins (n‐/i‐C₄H₈) and paraffin isomers (n−/i‐C₄H₁₀) [572]. With a pore size ranging between 3.82‐5.25 Å, ZU‐36‐Co adsorbs n‐C₄H₈ (2.35 mmol/g) and n‐C₄H₁₀ (2.20 mmol/g) while excluding i‐C₄H₈ and i‐C₄H₁₀. This sieving effect highlights the importance of pore window size for effective C4 gas separation. ZU‐36‐Co demonstrated n‐/i‐C₄H₈ uptake ratio of 13.8, an n‐/i‐C₄H₁₀ uptake ratio of 18.3, and IAST selectivity values of 2050 for n‐/i‐C₄H₈ (v/v = 1/1) and 140.4 for n‐/i‐C₄H₁₀ (v/v = 1/1). Computational studies revealed that the H‐bonding interactions between F atoms from GeF₆ ²⁻ anions and n‐C₄H₈ (or n‐C₄H₁₀) are key to trapping linear C4 isomers while excluding branched C4 isomers.

In 2022, A. Zheng's group reported a series of COFs, COF‐300 (dia‐c5), COF‐320 (dia‐c9), and azo‐COP‐2 (dia‐c5), for the separation of C4 HCs. Among these, COF‐300 (dia‐c5) demonstrated the highest selectivity (i‐C₄H₈/trans‐C₄H₈, 38.4; i‐C₄H₈/cis‐C₄H₈, 35.8) [583]. The exceptional separation performance of COF‐300 (dia‐c5) is attributed to its optimal interpenetration isomerism‐controlled interlayer space and the pedal motion‐altered shape of the 1D channel (Figure 19e).

For the separation of C4 HCs, seven mechanisms were identified among the 15 top‐performing reticular sorbents for selective C4 adsorption. These mechanisms include UMCs, hydrogen bonding, kinetics, pore size (molecular sieving), π···π & C─H···π interactions, van der Waals interactions, and flexibility (Figure 20). Overall, pore size and hydrogen bonding were found to play prominent roles in C4 separation. The best‐performing molecular sieve exhibited nearly infinite selectivity for n‐C₄H₁₀/i‐C₄H₁₀ separation. However, achieving the precise design of an ideal pore size in sorbents remains a challenge. Hydrogen bonding interactions between C4 molecules (e.g., C₄H₆ and n‐C₄H₁₀) and reticular sorbents also enable C4 separation.

FIGURE 20.

Radar plot presenting the contribution of seven separation mechanisms (and their combinations) among the 15 top‐performing reticular sorbents that exhibit binary or multiple C4 HCs adsorption selectivities. Each concentric heptagon denotes one sorbent, while the central heptagon represents two.

4. HC Separation in Vapor and/or Liquid Phase

HCs with higher molecular weights and bulkier molecular volumes typically exist as liquids at 298 K and 1 bar. While much of the focus on physisorbents has been on adsorption and purification of lower molecular weight gaseous HCs, separation of HCs in the vapor/liquid phases has also been addressed, even though it remains understudied. C6 and C8 HC isomers are the most widely studied sorbates as detailed below [721].

4.1. C6 Separation

Separation of C6 HCs has focused upon two subsets of industrial relevance: C6 alkane isomers (n‐hexane (n‐HEX), 3‐methylpentane (3MP), 2‐methylpentane (2MP), 2,2‐dimethylbutane (22DMB), and 2,3‐dimethylbutane (23DMB)); cyclic C6 HCs (benzene, Bz, and cyclohexane, Cy). Table 15 presents the leading reticular sorbents reported to date for the separation of these two subsets arranged in decreasing order of adsorption selectivity within each category. For C6 alkane isomers, most sorbents thus far studied favour n‐HEX selective sorption, presumably because of the more linear shape and smaller cross‐section of n‐HEX compared to its branched isomers (2MP, 3MP, 22DMB, 23DMB). The pore sizes of reticular sorbents studied for C6 alkane isomer separation range from 3 to 12 Å, with pore size being key to separation performance. For C6 cyclic HC separation, the reticular sorbents studied also exhibit pore sizes between 3 and 12 Å, but the separation mechanism primarily relies on π interactions between the sorbent and sorbate. Six studies on C6 alkane isomers and ten studies on C6 cyclic HCs have reported reticular sorbents with high selectivity and are discussed below.

4.1.1. C6 Alkane HCs

C6 alkane HCs are commodities in the petrochemical industry. The quality of gasoline is related to its combustion efficiency and proclivity to detonate. To obtain high‐quality gasoline, an effective approach is to increase the degree of branching [5, 722, 723, 724, 725, 726, 727, 728]. Performance is measured by the research octane number (RON); isomers with a higher degree of branching usually possess higher RONs. The five isomers of C6 alkane HCs produced as primary products from the light naphtha isomerization process, n‐HEX, 3MP, 2MP, 22DMB, and 23DMB, have RONs of 24.8, 73.4, 74.5, 91.8, and 101.7, respectively [623]. Octane enhancement can be achieved by effectively separating alkane isomers with varying degrees of branching. However, the similar physicochemical properties of hexane isomers present significant challenges for C6 alkane separation. In particular, the similar boiling points of C6 alkane HCs (Table 2) make separation by conventional fractionation energy‐intensive. In this context, adsorptive separation using porous materials offers an energy‐efficient alternative to the traditional methods. The distinct linear and branched shapes of C6 alkane isomers mean that control over pore size and shape is likely impact separation performance.

Early work on C6 alkane separation was reported in 2006 by B. Chen, S. Dai, and colleagues. They studied a flexible MOF‐508, [(Zn(BDC)(4,4′‐bpy)_0.5] (BDC = 1,4‐benzenedicarboxylic acid), a primitive cubic framework with one‐dimensional channels of approximately 4 × 4 Ų in cross‐section [658]. Selective separation of n‐HEX from n‐HEX/2MP/22DMB mixtures was achieved using MOF‐508 as a GC column. In addition, other HC mixtures were separated by GC, and the guest‐loaded/activated structures of MOF‐508 were determined through SCXRD. The authors’ analysis indicated that the GC separation of C6 alkanes was based on van der Waals interactions enabled by subtle size‐ and shape‐selective matching.

In 2013, Long and co‐workers reported that self‐assembly of a rigid and nearly planar dipyrazolate linker ligand, 1,4‐benzenedipyrazolate (BDP²⁻), with Fe(III), afforded a stable 3D MOF, Fe₂(BDP)₃ . Fe₂(BDP)₃ features triangular channels with sharp‐angled crevices, which enabled the separation of hexane isomers according to their degree of branching (Figure 21a) [71]. The linear shape of n‐HEX enabled it to be selectively adsorbed from n‐HEX/2MP/3MP/22DMB/23DMB mixtures, driven by stronger van der Waals contacts than branched alkanes.

FIGURE 21.

Examples of top‐performing reticular sorbents for C6 isomers’ separation with high selectivity: (a) Fe₂(BDP)₃ [71] with feed ratio and temperature set at 1:1 and 298 K, respectively. (b) UU‐200 [615] with feed ratio and temperature set at 1:1 and 298 K, respectively. (c) CopzNi [617] with feed ratio and temperature set at 1:1 and 303 K, respectively. (d) MIP‐214 [651] with feed ratio and temperature set at 1:1 and 298 K, respectively. (e) BTAPa‐CF₃ [655] with feed ratio and temperature set at 1:1 and 298 K, respectively. (Reprinted with permission from ref. [71]: Copyright 2013, The American Association for the Advancement of Science; ref. [615]: Copyright 2022, Wiley‐VCH GmbH; ref. [617]: Copyright 2023, The Elsevier B.V.; ref. [651]: Copyright 2024, Wiley‐VCH GmbH; ref. [655]: Copyright 2024, American Chemical Society.).

In 2022, D. Zhao and co‐workers reported a bismuth‐MOF synthesized from Bi³⁺ and benzene‐1,3,5‐tricarboxylic acid (H₃BTC), UU‐200, that was effective for the isolation of di‐branched alkanes from their isomers by combining two mechanisms, molecular recognition and molecular sieving [615]. The auxetic structure of UU‐200, featuring honeycomb‐like cavities connected by narrow pore sizes (5 × 5 Ų) enabled rejection of dibranched alkanes while offering high capacities for linear and monobranched isomers (Figure 21b). IAST calculations highlighted UU‐200’s separation ability, with equimolar selectivities for n‐HEX/22DMB of 1 × 10⁵, for 3MP/22DMB of 3 × 10⁴, and for n‐HEX/3MP at 160. This was confirmed by breakthrough tests at 298 K, which produced an HC mixture with a RON >96. DFT calculations revealed that UU‐200 binds to n‐HEX through multiple H‐bonding and C─H···π interactions.

In 2023, Z. Bao and colleagues reported that M(pz) [Ni(CN)₄], MpzNi; M = Co, Ni, pz = pyrazine, can separate ternary mixtures into individual components through molecular sieving and temperature programming [617]. The multiple UMCs and electron‐rich environment of the pore channel enabled CopzNi to exhibit high IAST selectivities: 420.0 for 2MP/22DMB (50/50, v/v) at 303 K; 171.7 for n‐HEX/2MP (50/50, v/v) at 363 K (Figure 21c).

In 2024, Serre and colleagues reported a multi‐cage microporous Fe(III)‐MOF, MIP‐214 (MIP: Materials of the Institute of Porous Materials of Paris, Fe₃(OH)(H₂O)₂(μ₃‐O)(PyC)₃, PyC = 4‐pyrazolecarboxylate) that achieved separation of a mixture of high RON di‐branched hexane and mono‐branched pentane isomers from their low RON counterparts, resulting in a product with a RON of 92 [651]. The narrow windows and large cages in MIP‐214 were generated by using a short ditopic ligand, H₂PyC, to replace the bdc linker in MCF‐48, Fe₃(μ₃‐O)(bdc)₃(L^T)₃, bdc = 1,4‐benzenedicaboxylate (Figure 21d). Experimental data revealed that MIP‐214 facilitates separation based on the degree of branching: linear (n‐HEX) ≫ mono‐branched (2MP, 3MP) ≫ di‐branched (22DMB, 23DMB). Also in 2024, X. P. Yan and colleagues reported two 2D COFs, BTAPa‐CF₃ and TbPa‐CF₃ , and studied their HC separation properties [655]. Compared to TbPa‐CF₃ , which has a uniform pore dimension of 14.3 × 14.3 Ų, BTAPa‐CF₃ possesses two pores of dimensions 6.3 × 6.3 and 14.5 × 14.5 Å. BTAPa‐CF₃ not only has a richer pore structure but also a higher density of trifluoromethyl functional groups, giving it superior performance in a bonded capillary column for n‐HEX‐selective separation from C6 alkane mixtures (Figure 21e).

4.1.2. C6 Cyclic HCs

Separation of Bz and Cy is another challenge for the petrochemical industry, with their binary mixtures having industrial utility. Their effective separation is difficult due to their similar molecular geometries, close boiling points (353.25 K and 353.85 K, respectively), and their tendency to form azeotropic mixtures [4, 729]. Cy is a feedstock for producing cyclohexanone and cyclohexanol, which are, in turn, used in the production of nylon‐66 and nylon‐6 [5, 724]. Industrially, Cy is primarily produced by the catalytic hydrogenation of Bz, resulting in a product that is a Bz/Cy binary mixture [730, 731]. Moreover, volatile organic compounds (VOCs) like Bz are a class of toxic pollutants that contribute to both indoor and outdoor air pollution, causing environmental and health issues even at trace concentrations [732, 733, 734, 735]. Instead of using heat‐driven distillation to separate Bz/Cy mixtures, adsorptive separation with porous reticular sorbents offers an alternative approach for selectively adsorbing either Bz over Cy (Bz/Cy), or Cy over Bz (Cy/Bz). Cy adopts either a boat or chair conformation and is bulkier than Bz, a planar molecule with a narrow cross‐section and a π‐electron cloud. MOF adsorbents typically adsorb Bz over Cy due to stronger binding affinities driven by host‐guest interactions, including π···complexation, π···π interactions, and H···π interactions [62, 679, 687]. Examples of Cy/Bz separation are relatively rare, their separation mechanisms primarily relying on van der Waals interactions between the reticular sorbents and Cy as well as appropriate pore size or shape [85].

In 2007, the first example of Bz‐selective adsorption in MOFs was reported by J. S. Chang and co‐workers with MIL‐101(Cr), (Cr₃F(H₂O)₂O [bdc)₂]₃, bdc = 1,4‐benzene dicarboxylate) [736]. High Bz uptake of 16.7 mmol·g^{− 1} at P/P₀ = 0.5 was observed, surpassing that of porous materials like mesoporous silica SBA‐15 (3.0 mmol·g^{− 1}), HZSM‐5 zeolite (1.9 mmol·g^{− 1}), and commercial activated carbon (8.0 mmol/g). In the same year, S. Kitagawa's group reported Bz‐selective sorption over Cy by a flexible [Zn(μ₄‐TCNQ‐TCNQ)bpy] (TCNQ = 7,7,8,8‐Tetracyano‐p‐quinodimethane) [62]. The undulating channel in the reticular sorbent enabled CH···π interactions between the host framework and the guest molecules.

In 2008, X. M. Chen reported a flexible MOF [Cu(etz)]_n , MAF‐2 (Hetz = 3,5‐diethyl‐1,2,4‐triazole) that exhibited efficient Bz/Cy separation [682]. Structural analysis revealed that its flexible framework distorted to allow Bz molecules to diffuse through the altered apertures while Cy did not. As a result, MAF‐2 readily adsorbed high amounts of Bz (206 mg/g) but only surface adsorption of Cy (9 mg/g at P/P₀ = 0.96). The strong interaction between the MAF‐2 framework and Bz was attributed to the π system and C(sp ²)‐H moieties of Bz (Figure 22a). In 2015, Ghosh and colleagues reported another flexible MOF, Cu‐1‐NO₃ (CuL₂(NO₃)₂), which achieved Bz/Cy separation enabled by a structural transformation from a 1D porous phase to a 2D non‐porous phase [683]. Cu‐1‐NO₃ was synthesized based on a flexible neutral amide‐based N‐donor ligand. This guest‐induced structural transformation enabled Bz‐selective adsorption via C─H⋯π and π⋯π interactions generated between aromatic rings of Cu‐1‐NO₃ and Bz (Figure 22b).

FIGURE 22.

Examples of top‐performing reticular sorbents for C6 isomers separation with high selectivity: (a) MAF‐2 [682] (b) Cu‐1‐NO₃ [683] (c) Mn‐MOF‐74 [679] (d) BUT‐55 [695] (e) Zn₁₂(tdc)₆(glycolate)₆(dabco)₃ [737] (f) CUB‐30 [680] with feed ratio and temperature set at 1:1 and 298 K, respectively. (Reprinted with permission from ref. [682]: Copyright 2008, American Chemical Society; ref. [683]: Copyright 2015, Wiley‐VCH GmbH; ref. [679]: Copyright 2016, The Royal Society of Chemistry; ref. [695]: Copyright 2022, The Author(s); ref. [737]: Copyright 2019, American Chemical Society; ref. [680]: Copyright 2024, Wiley‐VCH GmbH.).

In 2011, Y. Chen and co‐workers reported a 9‐connected trinuclear cluster‐based MOF, Ni₃(OH)(Ina)₃(BDC)_1.5 (Ina = isonicotinate and BDC = 1,4‐benzenedicarboxylate), for Bz/Cy separation [676] that adsorbs more Bz (22.60%) than Cy (1.40%). Bz selectivity was attributed to π⋯π interactions between Bz and the phenyl ring of the BDC linker, as well as the two types of cages in Ni₃(OH)(Ina)₃(BDC)_1.5 .

In 2016, S. Mukherjee et al. reported the sorption properties of a family of MOFs with UMCs, M‐MOF‐74 (M₂(dobdc) (dobdc⁴⁻ = 2,5‐dioxido‐1,4‐benzenedicarboxylate) or CPO‐27‐M or M₂(dhtp) (dhtp = 2,5‐dihydroxyterephthalate), where M = Mg, Mn, Fe, Co, Ni, Cu, Zn) [679]. In this work, π‐complexation triggered Lewis acid‐base interactions between the UMCs of MOFs and Bz. M‐MOF‐74 marked the first report of utilizing UMCs to enable Bz/Cy separation (Figure 22c).

In 2022, J. R. Li, Zaworotko, and colleagues reported a family of double‐walled MOFs, BUT‐53 to BUT‐58 (BUT = Beijing University of Technology) comprised of Co²⁺ or Zn²⁺ ions and ditopic pyrazolate (dipyrazolate) ligands. Bz uptakes at 298 K ranged from 2.47 to 3.28 mmol g^{− 1} at pressures below 10 Pa [695]. In particular, BUT‐55 (Co(BDP), BDP = H₂BDP  =  1,4‐di(1H‐pyrazol‐4‐yl)benzene), a supramolecular isomer of the MOF Co(BDP), captured trace levels of Bz, producing an air stream with Bz content below acceptable limits. Importantly, BUT‐55 also performed under humid conditions. An exceptionally high breakthrough time of 8000 h/g and Bz uptake capacity of 2.14 mmol/g were recorded for BUT‐55 under both dry conditions and 50% RH. Its breakthrough time and capacity were reduced to 6000 h/g and 1.61 mmol/g at 80% RH, respectively. The performance of BUT‐55 can be understood through the crystal structure of the Bz‐loaded phase (Bz@BUT‐55) and DFT calculations, which revealed that C─H⋯X interactions are responsible for the strong binding of Bz in a confined space suitable for Bz (Figure 22d).

In 2019, Cy/Bz separation was reported by Macreadie and co‐workers through employing the ditopic cubane‐1,4‐dicarboxylate linker (cdc) to prepare CUB‐5 [Zn₄O(cdc)₃, cdc = cubane‐1,4‐dicarboxylate], an analogue of MOF‐5 [73]. The high Bz/Cy selectivity of CUB‐5 was found to be enabled by non‐covalent interactions between the methine groups of the cubane linker and Bz. MOF‐5 was observed to exhibit Cy/Bz selectivity. DFT‐D3 calculations were employed to elucidate the driving factor behind this Cy/Bz selectivity and showed multiple interactions between the MOF‐5 walls and Cy, compared to Bz driven by Cy altering its conformation to maximize interactions with the pore.

Another example of Cy/Bz separation was reported by Dybtsev and co‐workers in 2019. [Zn₁₂(tdc)₆(glycolate)₆(dabco)₃] (H₂tdc = thiophene‐2,5‐dicarboxylic acid, glycolate = 1,2‐pentanediol, dabco = 1,4‐diazo [2.2.2] bicyclooctane), which features an alkyl chain, demonstrated adsorption selectivity for Cy/Bz of 2.5:1 in liquids and 5:1 in vapors [737]. van der Waals interactions between Cy and alkyl chains were found to contribute to this selective adsorption of Cy over Bz (Figure 22e).

In 2020, Hill and colleagues reported separation of Cy over Bz using CUB‐30 ([Zn₄O(hmtt)_4/3(bpdc)_1/2(cdc)_1/2], hmtt = 5,5′,10,10′,15,15′‐hexamethyltruxene‐2,7,12‐tricarboxylate; bpdc = biphenyl‐4,4′‐dicarboxylate; cdc = cubane‐1,4‐dicarboxylate) and MUF‐77 ([Zn₄O(hmtt)_4/3(bpdc)_1/2(bdc)_1/2], bdc = benzene‐1,4‐dicarboxylate) [680]. The preference for Cy adsorption in CUB‐30 compared to MUF‐77 at low partial pressures was attributed to the pore environment of the smaller dodecahedral pore (Figure 22f). Based on N₂ sorption isotherms, the theoretical pore size distribution of MUF‐77 and CUB‐30are 8, 18, 21 Å and 8.2, 17.5, 20.5 Å, respectively. Additionally, breakthrough curves were simulated to assess the capability of CUB‐30 to separate Cy and Bz under dynamic conditions [196, 738]. The results indicated that CUB‐30 can efficiently trap Cy under dynamic conditions.

For the separation of C6 HCs, seven mechanisms were identified among the top‐performing reticular sorbents. These mechanisms include UMCs, hydrogen bonding, pore size (molecular sieving), polarizability, π···π and C─H···π interactions, van der Waals interactions, and structural flexibility, all of which contribute to C6 adsorption selectivities (Figure 23). As might have been expected, pore size plays a crucial role in C6 alkane separation, thanks to the molecular size and shape differences among C6 HCs. Specifically, n‐hexane, with its linear molecular shape, tends to fit better within the channels of sorbents, whereas the branched C6 alkanes tend to be excluded. For the separation of C6 cyclic HCs, π···π and C─H···π sorbate–sorbent interactions are the predominant mechanism, exploiting the difference between Bz and Cy in terms of their electronic characteristics and shapes. Specifically, Bz is planar and π‐electron‐rich, Cy is of non‐planar shape and π‐electron‐deficient.

FIGURE 23.

Radar plot illustrating the contribution of seven separation mechanisms (and their combinations) among the ten top‐performing reticular sorbents that exhibit binary or multiple C6 HC adsorption selectivities. Each concentric heptagon corresponds to a different sorbent, with the central heptagon representing one sorbent.

4.2. C8 Aromatics Separation

C8 aromatics, which tend to be produced as mixtures which include the three xylene isomers, para‐xylene (PX), ortho‐xylene (OX) and meta‐xylene (MX), and ethylbenzene (EB), are each feedstocks for the synthesis of several important chemicals and polymers [4, 739]. PX is the most widely produced of the xylene isomers, accounting for 50%–60% of global xylenes production [740]. As the starting material for synthesis of terephthalic acid or dimethyl terephthalate, PX must have a purity greater than 99% for the subsequent production of polyester fiber and polyethylene terephthalate resin (PET) [741]. MX is primarily used to synthesize isophthalic acid as a co‐monomer in the production of PET‐based resin blends [742]. OX is used in the production of phthalic anhydride, which is then converted into a plasticizer [742]. EB is used in the production of polystyrene (PS) after undergoing catalytic dehydrogenation to styrene. Industrially, C8 aromatics are primarily produced through catalytic reforming of crude oil, which invariably results in the presence of other isomers [743]. Current technologies for separating C8 aromatics include distillation, crystallization, and adsorptive separation. Due to the similar boiling points and molecular sizes of the C8 aromatics, separation by distillation is costly and energy‐intensive (Table 2) [744]. Separation of aromatics through selective adsorption should be comparatively more energy‐efficient. Table 16 presents the leading reticular sorbents reported to date for the separation of the C8 aromatic isomers (OX, PX, MX, EB), arranged in decreasing order of adsorption selectivity within each category. For C8 isomer separation, more examples are reported for OX‐ or PX‐selective separation than for MX‐ or EB‐selective separation. The pore sizes for OX‐ or PX‐selective separation range between 3 and 21 Å, with pore size influencing performance. Pore size has had less influence on MX‐ or EB‐selective separation. π interactions and Van der Waals forces also play a role. Additionally, some molecular compounds with structural flexibility have shown exceptional performance, achieving record high selectivity. Eight studies reporting reticular sorbents with high selectivity are discussed below.

TABLE 16.

C8 aromatic (EB, PX, MX, and OX) isomers binary or multicomponent separations. The following parameters are listed for comparison: pore size; BET surface area (S_BET); single‐component gas uptakes; adsorption enthalpies (Q _st); separation factor; analyzation method, regeneration temperature; and attributed mechanisms. Reticular sorbents are listed in decreasing order of binary or multicomponent separation factor for C8 isomers.

Adsorption preference	Adsorbent, network dimensionality (nD)	Pore size (Å)	Components	Separation factor a	C8 capacity (mmol/g)	Analyzation method	Regeneration temperature (K)	SBET (m2/g)	Mechanism	Refs.
OX‐selective	ZUL‐C3, 3D (303 K)	4.5 × 5.2, 3.5 × 3.2 3.4 × 3.4, 3.6 × 4.2	OX/PX	21.8	PX: 0.17 MX: 0.29 OX: 0.68 EB: 0.47	NMR	423K with N2 flow	1200	C‐H···π interactions	[745]
			OX/MX	7.71
			OX/EB	7.81
	CD‐MOF‐1, 3D	7.8 × 7.8	OX/MX	6.73	—	HPLC	—	1220 [746]	Van der Waals interactions	[747]
			OX/PX	17.93
			MX/PX	2.67
	CD‐MOF‐2, 3D	7.8 × 7.8	OX/MX	4.76	—	HPLC	—	1030 [746]	Van der Waals interactions	[747]
			OX/PX	16.37
			MX/PX	3.44
	CD‐MOF‐1, 3D	21.4 b	OX/MX	3.4	OX: 0.59 MX: 0.29 PX: 0.19	Breakthrough, GC	—	391	Pore size	[748]
			MX/PX	11.2
			OX/PX	14.6
	MIL‐47, 3D	10.5 × 11.0 [749]	PX/EB	9.7	—	Breakthrough; HPLC	Hexane flow at 298 K	930	π···π interaction	[750]
			PX/MX	2.9
			OX/MX	2.0
			OX/EB	10.9
			OX/PX	1.4
			MX/EB	4.2
	MIL‐53(Al)ht , 3D	8.5 × 8.5	PX/EB	3.1	—	Breakthrough, GC	Hexane flow at 298 K	940	Pore size, π···π interactions	[751]
			MX/PX	1.2
			OX/MX	2.7
			OX/EB	10.9
			OX/PX	3.5
			MX/EB	3.8
	CAU‐23, 3D	6.3 × 7.6	OX/PX	10.4	OX: 1.90	GC	n‐Heptane	1243	H‐bonding	[752]
			OX/MX	7.6
	MIL‐53‐TDC, 3D	8.5 × 8.5	OX/PX	3.2	OX: 1.29	Breakthrough, GC	n‐Heptane	1101	C‐H···π, π···π interactions	[752]
			OX/MX	5.4
	Co2(dobdc), 3D (306 K)	8.1 × 8.1	OX/MX	2.5	—	Breakthrough, GC	—	1410 c	UMCs	[753]
			OX/PX	3.9
			OX/EB	1.21
			EB/MX	2.05
			EB/PX	3.21
			MX/PX	1.6
	MIL‐53(Fe), 3D (293 K)	6.8 × 21.3 (narrow phase) [754]	MX/PX	2.63	—	HPLC	—	23 [755]	π···π interaction, flexibility	[756]
			OX/MX	1.34
			OX/PX	3.52
	MIL‐53(Al), 3D	8.5 × 8.5	PX/EB	3.3	—	HPLC	—	904	C‐H···O interactions	[757]
			MX/PX	1.52
			OX/MX	1.93
			PX/OX	1.5
	CPO‐27‐Ni, 3D (398 K)	11 × 11	MX/PX	2	PX: 2.0 MX: 2.1 OX: 1.9	GC	—	1425	Pore shape, electrostatics	[758]
			OX/MX	1.7
			OX/PX	3.3
	CPO‐27‐Ni, 3D	11 × 11	OX/PX	3.3	PX: 0.86 OX: 2.93	Breakthrough, GC	—	1351 [759]	UMCs	[760]
	Cu‐BTC@Fe3O4 , 3D (453 K)	9 × 9	OX/MX	3.3	OX: 3.46 MX: 1.04	Breakthrough, GC	—	562	Electrostatics	[761]
			OX/PX	2.7
			MX/PX	1.1
	CMOMO‐7, 3D (303 K)	6.6 × 8.1	OX/PX	3.1	OX: 2.3 MX: 2.1 PX: 1.9	Breakthrough	N2 flow at 423 K	819	C‐H···π, interactions	[762]
			MX/PX	2.5
	UiO‐66, 3D (313 K)	8 × 8, 11 × 11	OX/MX	1.8	—	HPLC	—	885	Pore shape	[630]
			OX/PX	2.4
	SIFSIX‐1‐Cu, 3D	8 × 8 [32]	MX/PX	2.3	OX: 1.23 MX: 1.05 PX: 0.74	Breakthrough, GC	—	1012	H‐bonding	[763]
			OX/PX	2.7
			OX/MX	1.1
	COF1, 3D	7.8 × 12.3	PX/EB	1.4	—	HPLC	—	666	Polarizability, H‐bonding	[764]
			MX/PX	1.3
			OX/MX	2.0
	CAU‐13, 3D	—	OX/PX	1.5	OX: 0.34 mol mol−1 PX: 0.29 mol mol−1 MX: 0.28 mol mol−1	GC	—	—	Flexibility	[765]
			OX/MX	1.9
			PX/MX	1.3
	Zn(BDC)(dabco)0.5 , 3D(448 K)	7.5 × 7.5	MX/EB	1.15	—	Breakthrough, GC	—	—	Selective inclusion	[766]
			OX/EB	1.62
			OX/MX	1.12
			EB/PX	1.15
			MX/PX	1.25
			OX/PX	1.88
	MIL‐101(Cr), 3D	12, 16 × 14.5	PX/EB	1.64	—	HPLC	Hexane/DCM flow at 298 K	4000 [767]	π···π interaction, UMCs	[768]
			MX/PX	1.44
			OX/MX	1.86
			OX/PX	1.17
	RHO ZMOF, 3D	18 × 18, 9 × 9	OX/PX	1.3	PX: 0.63 OX: 0.92	Breakthrough, GC	—	871 [769]	UMCs	[760]
	HKUST‐1, 3D	6.5 × 6.5, 10, 12, 4.6 × 4.6	OX/PX	1.3	PX: 1.20 OX: 1.63	Breakthrough, GC	—	1055 [603]	UMCs	[760]
	HKUST‐1, 3D (398 K)	6.5 × 6.5	MX/PX	1.1	PX: 2.8 MX: 2.4 OX: 2.8	Breakthrough, GC	—	1718	Electrostatics	[758]
			OX/MX	1.1
			OX/PX	1.2
	ZIF‐76, 3D	11.6 × 11.6, 5.4 × 5.4, 18 × 18	OX/PX	1.1	PX: 0.18 OX: 0.26	Breakthrough, GC	—	827 [770]	Polarizability	[760]
	hybrid [3]arene H (Hα), 0D	—	OX/MX	97.5%	—	NMR	—	—	C‐H···π and C‐H···O interactions	[771]
			OX/PX	97.61%
	MIL‐47(V), 3D	10.5 × 10.5	—	—	OX selective	—	—	—	π···π interactions	[772]
	Tz‐DHTA, 2D	25.9 b	—	—	OX selective	GC	—	822	Pore‐filling	[773]
	Tph‐DHTA, 2D	17.8 b	—	—	OX selective	GC	—	528	Pore‐filling	[773]
	Tab‐DHTA, 2D	20 b	—	—	OX selective	GC	—	1137	Pore‐filling	[773]
	TpTFMB, 2D	15.5	OX/PX/MX/EB	—	OX selective	GC	—	964	H‐bonding	[774]
	TpBD, 2D	17.2	OX/PX/MX/EB	—	—	504	—	[774]
	TpPa‐CF3 , 2D	14.5	OX/PX/MX/EB	—	OX selective			1306	Polarizability, pore size	[774]
	TbTFMB, 2D	17.8			—			500	—	[774]
	MOF‐1, 3D (448 K)	4, 7.5	OX/(MX+PX+EB)	—	OX selective	Vapor sorption isotherms	—	2342.5	Pore size, shape	[775]
			EB/(OX+MX+PX)
	MIL‐47, 3D (448 K)	10.5 × 11.0	OX/(MX+PX+EB)	—	OX selective	Vapor sorption isotherms	—	1837.8	Pore size, shape	[775]
			EB/(OX+MX+PX)
	IRMOF‐1, 3D (448 K)	15, 11	OX/(MX+PX+EB)	—	OX selective	Vapor sorption isotherms	—	3821.5	Pore size, shape	[775]
			EB/(OX+MX+PX)
	BTAPa‐CF3 , 2D	6.3 × 6.3, 14.5 × 14.5	OX/MX/PX	—	OX selective	GC	—	953	Pore size	[655]
PX‐selective	MAF‐89, 3D	4.2 × 4.2, 5.5 × 5.5	PX/MX	19.4	PX: 2.00	Breakthrough, GC	—	562	Kinetics, van der Waals, electrostatics	[776]
			PX/OX	221
	HIAM‐203, 3D (393 K)	6 b	PX/OX	211.2 d	PX: 1.4 MX: 0.07 OX: 0.11	Breakthrough, GC	—	480	Kinetics, pore size, flexibility	[777]
			PX/MX	171.9 d
	TPBDB, 0D	5.4 × 9.1	PX/MX	50.6 (22.1)	—	GC, NMR	423 K	—	H‐bonding	[778]
			PX/OX	54.1 (76.1)
			PX/EB	47.3 (49.2)
			PX/OX/MX	60.4 (56.3)
			PX/OX/MX/EB	85.7 (37.5)
	β‐ [Cu2(pypz)2], 3D	—	PX/MX/OX	51	—	GC	DCM (423 K)	—	Flexibility	[779]
	sql‐1‐Co‐NCS, 2D	4.003, 4.172, 4.397 e	OX/MX	5.3	OX: 87.0 MX: 87.0 PX: 84.8 EB: 43.5	NMR	313 K under vacuum	—	C‐H···π interactions	[780]
			OX/PX	3.7
			PX/MX	1.5
			MX/EB	2.8
			PX/EB	28.9
	H/ZSM5, 3D (403‐443 K, 6–8 bar)	5.95 b	PX/MX	25.0	OX: 0.28 MX: 0.12 PX: 1.34 EB: 0.57	Breakthrough, GC	PDEB (403‐443 K)	—	Van der Waals interactions (Lewis acid, basic binding sites)	[781]
			PX/OX	16.8
			PX/EB	6.8
	NU‐2000, 3D	7.0 × 7.0	PX/OX	20	PX: 1.14 MX: 0.29 OX: 0.06	NMR	—	570	Molecular sieving	[782]
			PX/MX	3.9
	sql‐4,5‐Zn, 2D	—	PX/OX	13.69	PX: 20.7% f EB: 15.9% g	Breakthrough, GC	—	—	C–H···π, π···π interactions	[783]
			EB/OX	16.92
			MX/OX	4.23
	α‐ [Cu2(pypz)2], 3D	7.0 × 6.2	PX/MX/OX	16	—	GC	DCM (423 K)	—	Flexibility	[779]
	ZSM‐5(SiO2/Al2O3 = 600), 3D	—	PX/EB	5.52	EB: 0.21 PX: 1.2	GC	—	—	Pore shape	[784]
			PX/OX	12.42
			OX/EB	2.25
	MIL‐120(Al), 3D	5.4 × 4.7	PX/OX	11.5	PX: 0.5	GC	—	322	Pore shape, size	[785]
			PX/MX	8.3
			PX/EB	3.6
	MAF‐88, 3D	5.3 × 5.4, 6.4 × 9.6	PX/MX	6.7	PX: 2.00	Breakthrough, GC	—	—	—	[776]
			PX/OX	10.7
	Cu(CDC), 3D	5.4 × 5.4	PX/OX	10.0	PX: 1.1	GC	Methanol (423 K)	348	Pore shape	[786]
			PX/MX	7.0
			PX/EB	5.0
	Li/ZSM5, 3D (403‐443 K, 6–8 bar)	5.71 b	PX/MX	8.2	—	Breakthrough, GC	PDEB (403‐443 K)	—	Van der Waals interactions (Lewis acid, basic binding sites)	[781]
			PX/OX	6.0
			PX/EB	4.0
	DUT‐8(Cu), 3D	9.6 × 9.6	PX/OX	5.4	PX: 1.8	GC	—	2694	Pore size, π···alkyl interactions	[787]
			PX/MX	7.2
			PX/EB	5.9
	HOF‐PX‐a, 3D	—	PX/OX	7.2	—	NMR	—	—	C‐H···π, π···π interactions	[788]
			PX/MX	6.1
			PX/EB	4.1
	UIO‐66(Zr), 3D	4‐5	OX/MX	7.1	—	Breakthrough	—	—	π···π interactions	[789]
			OX/PX	2.3	—	Breakthrough	—	—	π···π interactions	[789]
			PX/MX	3.1	—	Breakthrough	—	—	π···π interactions	[789]
	Na/ZSM5, 3D (403‐443 K, 6–8 bar)	5.56 b	PX/MX	6.7	—	Breakthrough, GC	PDEB (403‐443 K)	—	Van der Waals interactions (Lewis acid, basic binding sites)	[781]
			PX/OX	5.5
			PX/EB	2.0
	Ce(HTCPB), 3D (383 K)	2.25 × 2.54, 2.04 × 2.33	PX/MX	4.5	—	GC	—	378.95	Flexibility, C–H···π interactions, H‐bonding	[790]
			PX/OX	5.6
			PX/EB	2.4
	nano‐zeolite K‐X, 3D (423 K)	7.1 b	PX/MX	5.36	—	Breakthrough, GC	Toluene at 423 K	688	Van der Waals interactions (Lewis acid, basic binding sites)	[791]
			PX/OX	2.43
			PX/EB	3.22
	MAF‐X8, 3D (433 K)	8.5 × 8.5	PX over all three	5.3	PX: 2.2	—	—	1465	Pore shape	[792]
	K/ZSM5, 3D (403‐443 K, 6–8 bar)	5.34 b	PX/MX	5.1	—	Breakthrough, GC	PDEB (403‐443 K)	—	Van der Waals interaction (Lewis acid, basic binding sites)	[781]
			PX/OX	3.9
			PX/EB	1.7
	MIL‐125(Ti), 3D	6.13 × 6.13, 12.55 × 12.55 [632]	PX/MX	4.4	—	HPLC	—	1446	Pore shape	[793]
	BaY zeolite, 3D (353 K)	7.0 b	PX/OX	4.0	—	HPLC	n‐octane (353 K)	313	Molecular sieving	[794]
			PX/MX	3.9
			PX/EB	1.8
	nano‐zeolite Na‐X, 3D (423 K)	7.2 b	PX/MX	3.73	—	Breakthrough, GC	Toluene at 423 K	697	Van der waals interactions (Lewis acid, basic binding sites)	[791]
			PX/OX	1.68
			PX/EB	2.78
	CAU‐1(Al)‐NH2 , 3D	10 × 10 [795]	PX/MX	3.5	—	HPLC	—	1495	Pore shape	[793]
	ZIF‐8, 3D	11.6 × 11.6, 3.4 × 3.4	PX/OX	3.3	PX: 0.47 OX: 0.15	GC	—	739 [796]	Molecular sieving	[760]
	MIL‐125(Ti)‐NH2 , 3D	12.5 × 12.5, 6.1 × 6.1	PX/MX	2.8	—	HPLC	—	1380	Pore shape	[793]
	nano‐zeolite Li‐X, 3D (423 K)	7.2 b	PX/MX	2.58	—	Breakthrough, GC	Toluene at 423 K	699	Van der Waals interactions (Lewis acid, basic binding sites)	[791]
			PX/OX	1.36
			PX/EB	1.32
	MIL‐140B, 3D (323 K)	3.8 × 3.8 [667]	PX/EB	2.1 d	—	Breakthrough, GC	323	460	Pore shape	[797]
			PX/MX	1.7 d
			PX/OX	1.8 d
	CAU‐1‐OH, 3D	6.0, 8.0 b	PX/MX	1.86	PX: 4.33	Breakthrough, GC	—	1366	C–H···π interactions	[798]
	MOF‐48, 3D (323 K)	5.2 × 5.2 [799]	PX/EB	1.5 d	—	Breakthrough, GC	323	195	Pore shape	[797]
			PX/MX	1.7 d
			PX/OX	1.7 d
	nano‐zeolite H‐X, 3D (423 K)	7.2 b	PX/MX	1.22	—	Breakthrough, GC	Toluene at 423 K	712	Van der waals interactions (Lewis acid, basic binding sites)	[791]
			PX/OX	1.20
			PX/EB	1.19
	EtP6, 0D	6.7 × 6.7	PX/OX/MX	99.1%	—	NMR	433 K	—	π···π interactions	[800]
	P [4]Q [1]L, 0D	—	PX/(MX+OX+EB)	97.5%	—	NMR	—	—	C‐H···π and C‐H···O interactions	[801]
	phenanthrene [2]arene, 0D	—	PX/MX	81%	—	GC, NMR	423 K	—	C‐H···π and C‐H···O interactions	[802]
	TEMPO@ZIF‐8, 3D	7.26 × 7.26	PX/MX	93%	—	NMR	—	—	Polarizability	[803]
			PX/OX	95%
			MX/OX	92%
	[Zn(Hpidba)] n , 3D	8.2‐8.7, 5.6 × 6.8	PX/MX/OX	—	PX selective	GC	—	1319	C‐H···O, π···π interactions	[804]
	JNU‐5, 2D	13	—	—	PX selective	GC	—	419	C‐H···π, π···π interactions	[805]
	SCOF‐303, 2D	8.31	—	—	PX selective	GC	—	490	π···π interactions	[806]
	DynaMOF‐100 ([Zn4O(L)3(DMF)2]), 3D	5.1 b	—	—	PX: 2.86 MX: 0.22 OX: 0.05	GC	—	—	Pore size	[807, 808]
	Co2L2(AzoD)2 , 3D	3.3 × 10.6	—	—	PX: 2.5 MX: 0 OX: 0.95	—	—	113	Pore size	[809]
	JUC‐77, 3D	10.8 × 7.3	—	—	PX: 3.3	—	—	976	Molecular sieving	[810]
	[Zn(μ4‐L)] n , 3D	14.8 × 18.7	—	—	PX:0.02 MX: 0.01	GC, NMR	CH3Cl	—	Pore size, shape	[811]
	Co‐CUK‐1, 3D	6.8 × 6.8	—	—	PX: 2.1 f MX: 0 f OX: 0 f EB: 0.2 f	—	—	730	Electrostatics, pore shape	[812]
MX‐selective	sql‐1,3‐Co‐NCS, 2D	7.5 × 11.5	PX/EB	9.8	OX: 37% MX: 37% PX: 37% EB: 18.7%	NMR	—	—	C–H···π, π···π stacking interactions	[813]
			MX/EB	10.8
			OX/EB	7.9
	MIL‐160, 3D	9.05 × 9.05	MX/PX	4.0	MX: 1.3	Breakthrough	—	1243	Electrostatics, π···π stacking interactions	[814]
			MX/OX	2.8
			OX/PX	1.4
	MFM‐300(Fe), 3D	6.6 × 6.8	MX/PX	3.9 d	—	Breakthrough, GC	—	924	Pore size, π‐π interactions, π‐O interactions	[815]
			MX/OX	1.6 d
			OX/PX	2.5 d
	MFM‐300(V), 3D	6.8 × 7.0	MX/PX	3.7 d	—	Breakthrough, GC	—	1091	Pore size, π‐π interactions, π···O interactions	[815]
			MX/OX	1.1 d
			OX/PX	3.0 d
	MFM‐300(In), 3D	7.2 × 7.4	MX/PX	3.5 d	—	Breakthrough, GC	—	903	Pore size, π‐π interactions, π‐O interactions	[815]
			MX/OX	2.3 d
			OX/PX	1.3 d
	NIIC‐30(Ph), 3D	2.9 × 9.6, 4 × 9.2	MX/OX	3.03 (1.94 h )	—	NMR	—	451	C–H···O, C–H···π interactions	[816]
			PX/OX	2.97 (2.12 g )
			PX/MX	1.05 (1.19 g )
	ZU‐61, 3D (333 K)	7.8 b	MX/PX	2.9	PX: 0.62 MX: 1.15 OX: 1.61	Breakthrough, GC	398K	1384	Pore shape, C–H···F interactions	[817]
	Zn2(aip)2(bpe), 3D	14.7	MX/OX/EB	—	MX selective	NMR	—	125	π···π interactions	[818]
EB‐selective	Zn‐ETTOB, 3D	7 × 7	PX/EB	1.2	EB: 8.08 PX: 8.00 MX: 7.54 OX: 6.51	—	—	2594.9	Van der Waals interactions	[819]
			PX/MX	3.9
			PX/OX	6.2
			EB/MX	3.3
			EB/OX	13.1
			MX/OX	3.5
	MOF‐monoclinic, 3D (393 K, 80kPa)	7 × 7	MX/PX	2.52	—	Breakthrough, GC	—	225	Polarizability	[820]
			EB/PX	5.17
			OX/PX	4.55
	MOF‐5, 3D (523 K, 120 kPa)	12 × 12	EB/OX	1.96	—	Breakthrough, GC	—	773	Polarizability	[820]
			EB/MX	2.34
			EB/PX	4.14
	MIL‐47, 3D (343 K)	10.5 × 11.0 [749]	EB/PX	1.83	—	Breakthrough, GC	—	—	Polarizability, π⋅⋅⋅π interactions	[821]
			EB/MX	1.41
			EB/OX	1.39
			MX/PX	2.07
			OX/MX	1.17
			OX/PX	1.01
	Zn2(aip)2(bpy), 3D	15.2	EB/MX/OX	—	EB selective	NMR	—	332	π···π interactions	[818]

^a Selectivity calculated from the uptake from mixtures of binary (v/v = 1/1) C8 aromatic compounds.

^b Pore size distribution, BET surface area calculated from the N₂ adsorption isotherm at 77 K.

^c Langmuir surface area calculated from the N₂ adsorption isotherm at 77 K.

^d Selectivities calculated from binary breakthrough curves.

^e Distances between adjacent layers.

^f IAST calculated uptakes for gas phases.

^g Gravimetric Uptake.

^h Vapor selectivity.

In 1989, T. Y. Yan reported the separation of PX and EB from C8 aromatics using the zeolite ZSM‐5 (SiO₂/Al₂O₃) [784]. The effective separation of PX/EB, with an uptake ratio of 5.5, was attributed to the optimal pore size and commensurate packing of PX in the cavities. In 2007, Alaerts et al. reported the first separation of xylene isomers and EB using MOFs (MIL‐47, VO(BDC)·(H₂BDC)_0.75 (BDC = 1,4‐benzenedicarboxylate)) [750]. As demonstrated by breakthrough and chromatographic experiments, MIL‐47 exhibited high selectivity for PX/EB, OX/EB, and MX/EB, with selectivity values of 9.7, 10.9, and 4.2, respectively. The difference in π···π interactions between the C8 aromatics and the MIL‐47 framework was determined to be key to the separation performance.

In 2016, Nair, Sholl, and co‐workers selected a series of MOFs for C8 aromatics separation using computational methods, and their performance was subsequently confirmed through experimental evaluation. Based on the results of the computational screening, four top‐performing MOF materials were identified: MIL‐47 [VO(BDC)·(H₂BDC)₀.₇₅ (BDC = 1,4‐benzenedicarboxylate)]; MIL‐125‐NH₂ [Ti₈O₈(OH)₄(BDC‐NH₂)₆]; MIL‐140B [ZrO(NDC)] _n (NDC = 2,6‐naphthalenedicarboxylate); and MOF‐48 [VO(DMBDC)·(H₂DMBDC)₀.₄ (DMBDC = 2,5‐dimethylbenzenedicarboxylate)] _n [797]. At 323 K, the selectivity of MOF‐48, as determined from breakthrough data, was PX/EB: 1.5, PX/MX: 1.7, and PX/OX: 1.7, respectively. Computational analysis of free energy and structural characteristics revealed that PX has the lowest free energy along the corrugated channels of MOF‐48. This work demonstrates an approach for identifying PX‐selective reticular sorbents from the MOF database using computational methods.

In 2019, Zaworotko and co‐workers reported a flexible sql topology MOF that is effective for C8 aromatics separation [780]. The non‐porous layered material [Co(bpy)₂(NCS)₂]_n (sql‐1‐Co‐NCS) was found to reversibly switch to large pore phases to accommodate C8 aromatics, with varying switching pressures and kinetics. High OX selectivity (S_OX/EB ≈ 60) and saturation capacity (>80 wt%) were observed. SCXRD analysis of xylene‐included single crystals revealed additional C─H···π interactions between the phenyl hydrogens of OX and sql‐1‐Co‐NCS, compared to those seen for the other isomers. These supplementary C─H···π interactions enabled the high OX‐selectivity over PX, MX, and EB and high OX uptake capacity (Figure 24a). Also in 2019, J. P. Zhang and co‐workers reported another example of a flexible CN for C8 aromatics separation, the metal‐azolate framework, MAF‐36. The open phase of MAF‐36, ([Cu₂(pypz)₂]·0.5p‐xylene, Hpypz = 4‐(1H‐pyrazol‐4‐yl)pyridine) features one‐dimensional channels and discrete small cavities, while a non‐porous phase was obtained upon removal of the guest molecules [779]. This guest‐induced structural flexibility plays a crucial role in C8 aromatics separation. The open phase, α‐ [Cu₂(pypz)₂], with a pore size of 7.0 × 6.2 Ų, exhibited a PX/MX/OX uptake ratio of 16, while the closed phase, β‐ [Cu₂(pypz)₂], exhibited a higher PX/MX/OX uptake ratio of 51.

FIGURE 24.

Examples of top‐performing sorbents for C8 aromatics separation with high selectivity: (a) sql‐1‐Co‐NCS [780], (b) MAF‐89 [776], (c) TPBD [778], and (d) HIAM‐203 [777] with feed ratio and temperature set at 1:1and 298 K, respectively. (Reprinted with permission from ref. [780]: Copyright 2019, Wiley‐VCH GmbH; ref. [776]: Copyright 2022, Science China Press and Springer‐Verlag GmbH Germany; ref. [778]: Copyright 2023, the Authors; ref. [777]: Copyright 2023, Wiley‐VCH GmbH.).

In 2022, J. P. Zhang and C. T. He reported on the C8 sorption properties of two isostructural MOFs α‐ [Zn(pba)] (MAF‐88, H₂pba = 4‐(1H‐pyrazol‐4‐yl) benzoic acid) and β‐ [Zn(pba)] (MAF‐89) [776]. MAF‐88 and MAF‐89 were found to possess similar pillared‐column structures, porosities, and high PX capacities of 2.0 mmol/g. However, due to differences in pore size and pore topology, their PX selectivities were quite different. For PX/OX and PX/MX separation, MAF‐88, with its narrow 1D channels (5.3 × 5.4 and 6.4 × 9.6 Ų), exhibited selectivities of 10.7 and 6.7, respectively. In contrast, MAF‐89, featuring 3D‐connected quasi‐discrete pores (4.2 × 4.2 and 5.5 × 5.5 Ų), exhibited selectivities of 221 and 19.4, respectively (Figure 24b). MD simulations indicated that the mechanism involved diffusion‐controlled kinetic separation. The movement trajectories showed guest molecules trapped in the 0D pores. These guest xylene molecules were found to occasionally jump to neighbouring pores through the connecting apertures. Consequently, selectivity was governed by kinetics, and attributed to the gating actions of the flexible narrow apertures.

In 2023, H. Wang and co‐workers reported another example of a flexible MOF with a temperature‐driven kinetic separation mechanisms [777]. HIAM‐203 (Ca(chloranilate)), was found to separate PX from OX and MX isomers driven by its flexible structure. At 303 K, all three isomers are accommodated, but the adsorption kinetics of OX and MX were found to be significantly slower than that of PX (Figure 24d). At the elevated temperature of 393 K, OX and MX were excluded while PX was adsorbed.

In addition to using porous materials, molecular compounds can also be applied for C8 aromatics separation. In 2018, K. Jie et al. reported a flexible molecular compound, EtP6 (perethylated pillar [6]arene), which separated PX with a purity of 99.1% [800]. Selectivity was found to be an intrinsic property of the EtP6 host, resulting from its size and shape as well as flexible cavities that adapt during adsorption to trap PX in the solid state.

In 2023, Zaworotko and co‐workers reported another flexible molecular compound (TPBD, 4‐(1H‐1,2,4‐triazol‐1‐yl)‐phenyl‐1H‐benzo [de]isoquinoline‐1,3(2H)‐dione) that exhibited record‐high PX‐selective separation from PX/OX/MX/EB mixtures [778]. Binary selectivity values for PX, as determined by ¹H NMR spectroscopy and gas chromatography, ranged from 22.4 to 108.4, setting new benchmarks for PX/MX (70.3) and PX/EB (59.9) selectivity, and achieving near‐benchmark selectivity for PX/OX (108.4) (feed ratio = 1:1, temperature: 298 K). Analysis of the C8 aromatics‐induced flexibility and PX‐loaded structure of TPBD revealed that it formed channels that are a particularly good shape match for PX. Multiple noncovalent interactions (C─H···π, C‐H···N, C─H···O) were observed between PX and TPBD. TPBD was the first sorbent of any class to demonstrate high across‐the‐board PX selectivity from quaternary mixtures of C8 aromatics under ambient conditions (Figure 24c).

For the separation of C8 HCs, six mechanisms were identified among the 15 top‐performing sorbents. The mechanisms include electrostatics, kinetics, pore size, π···π and C─H···π interactions, van der Waals interactions, and flexibility (Figure 25). Pore size was found to be the leading driver for C8 separation enabled by the smaller molecular sizes of PX and OX (Table 2). OX and PX selective sorbents are the most commonly reported, thanks to size exclusion. van der Waals and π···π and C─H···π interactions are the second most reported mechanism to drive selectivity.

FIGURE 25.

Radar plot presenting the contribution of six separation mechanisms (and their combinations) among the 15 top‐performing sorbents that exhibit binary or multiple C8 HCs adsorption selectivities. Each concentric hexagon denotes one sorbent each, while the central hexagon represents one.

5. Conclusions and future directions

5.1. We Have Come an Awfully Long Way

Properties: This article details the emergence over the past two decades of reticular crystalline materials as physisorbents, exemplified by MOFs and COFs, for separating and/or purifying industrially relevant HCs, including separations involving CO₂. Overall, exceptional enhancements in performance have been reported. Whereas trace separations (removal of minor impurities at levels <1%) were generally infeasible for physisorbents as recently as 2010, as detailed herein, this is no longer the case. Today, benchmark selectivity values for several industrially relevant separations go far beyond those exhibited by traditional adsorbents like zeolites and silica. These advances are summarized in Table 17 as exemplified by seven adsorbate pairs and three entries: the existing benchmark material; the sorbent that represented a significant leap forward; the sorbent that is the current state‐of‐the‐art.

TABLE 17.

Summary of binary selectivities and binding mechanisms (towards the more preferred sorbate) for a select few reticular sorbents, as listed below.

Adsorbent	Year	Selectivity a	Mechanism	Refs.
Adsorbate pair (CO2/N2)
Zeolite 13X	2002	420 b /562 c	Cation···π interactions	[822]
SIFSIX‐3‐Zn	2013	1818†	Electrostatics (F‐CO2); pore size	[823]
NbOFFIVE‐1‐Ni	2016	601 b /6528 c	Electrostatics (F‐CO2); pore size	[824]
Adsorbate pair (CO2/CH4)
Zeolite 13X	2004	790	Electrostatics	[825]
TIFSIX‐3‐Ni	2019	3501	Electrostatics (F‐CO2); pore size	[134]
Cu‐F‐pymo	2022	107 c	Molecular sieving	[128]
Adsorbate pair (C2H2/C2H4)
Fe‐MOF‐74	2012	2.1	UMCs	[66]
SIFSIX‐2‐Cu‐i	2016	44.5	Electrostatics (C‐H∙∙∙F); pore size	[35]
UTSA‐300a	2017	104 c	Flexibility; H‐bonding; pore size	[159]
Adsorbate pair (C2H4/C2H6)
Zeolite 5A	2013	4.5	Cation···π interactions	[307]
Co‐gallate	2018	52	Molecular sieving	[291]
UTSA‐280	2018	104 c	Molecular sieving; H‐bonding; π···π stacking; van der Waals interactions	[287]
Adsorbate pair (C2H6/C2H4)
ZIF‐7	2010	1.6	Flexibility	[387]
MAF‐49	2015	9	H‐bonding; electrostatics	[314]
NIIC‐20‐Bu	2020	15.4	H‐bonding; C‐H···π interactions	[313]
Adsorbate pair (C3H6/C3H8)
Co‐MOF‐74	2012	45	UMCs	[472]
UTSA‐400	2023	107 c	H‐bonding; van der Waals interactions	[456]
HAF‐1	2025	1.67 × 107 c	Molecular sieving; H‐bonding; C‐H···π interactions	[455]
Adsorbate pair (PX/OX)
nano‐zeolite K‐X	2015	2.43	Van der Waals interactions	[791]
HIAM‐203	2023	211.2	Kinetics; pore size; flexibility	[777]
MAF‐89	2022	221	Kinetics; van der Waals interactions; electrostatics	[776]

^a IAST selectivity at 298 K at 1 bar.

^b For N₂ mole fraction of 0.9 (CO₂/N₂ = 1/9).

^c For N₂ mole fraction of 0.9995 (CO₂/N₂ = 500 ppm/0.9995 bar).

As revealed by Table 17, in some cases performance enhancements are driven by selective binding but in others molecular sieving. Highlights include the prototypal HUM, SIFSIX‐3‐Zn, which was the first physisorbent to offer ultrahigh CO₂/N₂ selectivity (∼2500:1), roughly an order of magnitude beyond the prior benchmarks zeolite 13X and Mg‐MOF‐74 [33]. Second generation variants established that narrow pores with strongly electrostatic binding sites can capture CO₂ at partial pressures relevant to direct air capture of CO₂. The rigid CN UTSA‐280 exhibited equimolar C₂H₄/C₂H₆ adsorption selectivity exceeding 4,100, producing >99% pure ethylene in the effluent stream [287]. Flexible HUMs can produce polymer‐grade propylene from propyne, with TIFSIX‐14‐Cu‐i, adsorbing C₃H₄ at a very low threshold pressure (∼500 ppm) and with an IAST selectivity of ∼355 for a 1/99 C₃H₄/C₃H₆ mixture [432].

Reticular sorbents can also capture trace gas and vapor components from more complex mixtures, including benzene removal from moist streams. Benzene is a toxic VOC and is often encountered as a minor impurity (ppm‐level) in petrochemical processes (e.g., in cyclohexane) or in air. The double‐walled CN BUT‐55 was found to adsorb benzene at partial pressures below 10 Pa (∼100 ppm) even under humid conditions with effluent benzene levels below acceptable exposure limits. These examples underscore the extraordinary progress that has been made in HC separations with selectivity values and working capacities that could render energy‐intensive techniques (cryogenic distillation, solvent scrubbing, etc.) obsolete.

Insight: There is now much more understanding about the underlying structure‐function relationships that drive these strong separation performances. This insight comes in part from in situ experiments that have provided experimental observation of sorbent‐sorbate binding and also from advances in molecular modelling. Further, the inherent modularity of PCNs and COFs enables crystal engineering studies to systematically study families of related reticular sorbents to control structure/properties and further enhance performance. Several recent in situ studies have demonstrated how XRD and IR, typically coupled with computational experiments (DFT and/or GCMC), can reveal the fine details of those sorbent binding sites that enable high selectivity for a specific sorbate. For example, single‐crystal X‐ray diffraction studies have pinpointed the C₂H₂ binding site in the Cu(II) CN, CPL‐1. In CPL‐1, each adsorbed C₂H₂ molecule forms an H‐bond with a non‐coordinated framework O atom [60]. The acidic C–H on C₂H₂ can engage these O‐donors whereas CO₂ (which lacks an H donor) cannot, thereby explaining CPL‐1’s strong C₂H₂/CO₂ and C₂H₂/C₂H₄ selectivity. Similarly, Al(HCOO)₃ (ALF) achieves excellent CO₂ selectivity over C₂H₂ (≈6.5 × 10⁵) via directed C─H⋯O bonding [266]. In ALF’s ∼4.1×5.3 Å² channels, in situ FTIR of CO₂‐loaded ALF revealed that CO₂ molecules bind at the electropositive C–H moiety of the formate linkers. This specific C─H⋯O interaction (along with the tight pore size) is key to the aforesaid ultrahigh CO₂/C₂H₂ IAST selectivity. These examples illustrate that precise pore dimensions together with localized polar binding motifs (H‐bond donors or charged anions) can enable single sorbate discrimination.

Systematic crystal engineering studies can also reveal how even subtle composition changes, such as adding a fluorine atom to a linker and/or metal substitution, including doping, can profoundly impact sorbent performance. This is exemplified by the dramatic changes in the 195 K CO₂ sorption profiles between [Co(bimpy)(bdc)] _n (X‐dia‐4‐Co) and [Co(bimbz)(bdc)] _n (X‐dia‐5‐Co) (H₂bdc = 1,4‐benzendicarboxylic acid; bimpy = 2,5‐bis(1H‐imidazole‐1‐yl)pyridine; bimbz = 1,4‐bis(1H‐imidazole‐1‐yl)benzene) [826]. Specifically, X‐dia‐4‐Co showed a gradual phase transformation accompanied by a steady increase in CO₂ uptake, whereas X‐dia‐5‐Co displayed a sharp switching isotherm, with the phase transition occurring at a relative pressure (P/P ₀) of 0.008 at 195 K, and at an absolute pressure of 3 bar at 298 K [826]. In another recent study, a Zn(II) CN, ZnFPCP, with fluorinated aromatic nanotraps was shown via DFTB/GCMC and operando FTIR to preferentially bind alkanes, yielding high C₃H₈/C₃H₆ and C₂H₆/C₂H₄ selectivities to produce ultra‐high purity (99.99%) C₃H₆ and C₂H₄. Likewise, pore tuning in Fe‐MOF‐74 via pore‐space partitioning sharply improved C₂H₆/C₂H₄ selectivity [72]. A flexible Ni‐based diamondoid net (X‐dia‐1‐Ni) [54], was partially doped with Co to enhance its inverse C₂H₆/C₂H₄ IAST selectivity from 3.5 to 5.5 while maintaining high C₂H₆ uptake [316]. Even more extreme changes in properties can occur in interpenetrated vs. non‐interpenetrated variants of a CN, e.g., SIFSIX‐2‐Cu vs. SIFSIX‐2‐Cu‐i, where interpenetration means tight C₂H₂ binding sites with greatly enhanced performance for trace C₂H₂/C₂H₄ separation [35].

These examples span rigid ultramicroporous PCNs to soft flexible sorbents, and underscore not just the feasibility of fine‐tuning of composition, but how even small changes, typically guided by in situ insight and modelling, can significantly enhance performance. In short, there has not just been remarkable progress in terms of design (crystal engineering) and properties of reticular sorbents for HC separation, there is now insight into why these benchmark sorbents exhibit their performance.

5.2. How Did We Get Here?

That crystal engineering approaches can enable chemists to fine‐tune pore size, shape, and chemistry to achieve optimal sorbate binding has enabled a focus upon the mechanisms of sorbent‐sorbate binding to guide the next generation of reticular sorbents. As detailed herein, we and others have identified nine such mechanisms (and combinations thereof), including molecular sieving, UMCs, H‐bonding, electrostatic interactions, π···π and C─H···π interactions, van der Waals forces, framework flexibility, and kinetic sieving effects (Figure 26). Several of these are summarized below.

FIGURE 26.

Nine mechanisms have been identified as factors that drive adsorptive HC separation by reticular materials: UMCs, H‐bonding, electrostatics, kinetics, flexibility, van der Waals interactions, C─H···π and π···π interactions, polarizability, and pore size.

Pore size really matters as it can be the key to enabling molecular sieving, generally regarded as the most energy‐efficient approach for molecular separations [39, 130, 287]. Controlling pore size until it matches the kinetic diameter of a target molecule can enable molecular sieving, a striking example being ethylene/ethane separation by UTSA‐280. Two narrow channels (3.2 Å × 4.5 Å and 3.8 Å × 3.8 Å) lie between the cross‐sectional areas of C₂H₄ and C₂H₆, allowing C₂H₄ to enter while excluding C₂H₆ [287]. Consequently, UTSA‐280 adsorbs 2.5 mmol g⁻¹ of ethylene at 1 bar and 298 K but almost no ethane, translating to exceptional selectivity and >99% ethylene purity. SIFSIX‐2‐Cu‐i (pore ∼5.15 Å) and SIFSIX‐3‐Zn (∼3.84 Å) offer single‐site binding energies of ca. 45–60 kJmol⁻¹ for C₂H₂ and CO₂, respectively. These energies are in the sweet spot when they are strong enough to capture the target gases (C₂H₂ and CO₂, respectively) when present in trace amounts, yet are not nearly as energy‐intensive as chemisorption [33, 35]. This balance can overcome the oft‐encountered Catch‐22 of physisorption, where strong binding improves selectivity but hinders regeneration [827]. Indeed, tight binding thanks to ultramicropores enabled SIFSIX‐3‐Zn and its derivatives to become the first class of PCN physisorbents to exhibit selectivity consistent with that needed for DAC [33, 828]. Conversely, wider pore SIFSIX variants with the same pore chemistry exhibit much weaker binding for CO₂ [36].

Pore chemistry also really matters as incorporation of open metal sites, hydrogen bonding moieties, or strong electrostatics can drive affinity for one HC over another. For instance, incorporation of Ag(I) cations into NUS‐6(Hf)‐Ag enables ethylene molecules to selectively bind to Ag⁺ via π‐electron donation and backbonding, affording an IAST selectivity of ∼106 for C₂H₄ over C₂H₆ [288]. Notably, this chemically modified CN maintained separation performance over repeated cycles. Likewise, UMCs in the M‐MOF‐74 family (M = Mg, Mn, Fe, Co, Ni, Cu, Zn) were preferentially benzene over cyclohexane, marking the first use of Lewis acid–base interactions (π‐complexation) to separate aromatic/aliphatic HC isomers, viz., Bz over Cy [679]. In the same vein, electronegative SiF₆ ²–sites favour strong binding with the more acidic C₂H₂ molecules (pK _a 25) vs. the less acidic C₂H₄ (pK _a 44) as demonstrated by SIFSIX‐2‐Cu‐i [35]. It is now evident that designing pore surfaces with the right functional group(s) can “programme” a reticular sorbent to favour one HC over another.

Pore shape and topology can sometimes matter as the spatial arrangement of pores (1D channels vs 3D cages, interpenetrated vs open frameworks) can influence diffusion rates and coadsorption, introducing kinetic selectivity in addition to thermodynamic selectivity. A good example is the separation of xylene isomers by two isostructural Zn(pba) CNs, MAF‐88 and MAF‐89 [776]. These CNs exhibit different pore architectures, narrow 1D channels in MAF‐88 versus quasi‐discrete 0D cavities in MAF‐89. Consequently, MAF‐89 achieves p‐xylene selectivities 10–20 times higher than MAF‐88 for the same p/o/m‐xylene mixture. Molecular dynamics simulations support a diffusion‐controlled mechanism with the flexible narrow apertures in MAF‐89 acting as gated traps that slow or exclude the larger isomers, whereas the more open channels of MAF‐88 provide less discrimination. As discussed earlier (section 4.2, and Table 16), the molecular sorbent TPBD exhibits benchmark PX selectivity from C8 aromatic mixtures can be attributed to its structural adaptability and conformational polymorphism [778]. Despite being non‐porous, conformational polymorphs of TPBD (TPBD‐αI, TPBD‐αII) differ by ring orientations, each transforming to a PX–bound inclusion phase (TPBD–PX) upon PX exposure. Only p‐xylene was found to trigger the necessary induced fit pore opening while excluding o‐/m‐xylene and ethylbenzene. TPBD illustrates that pore shape can critically dictate separation outcomes. These are just two examples that illustrate how pore shape/topology can drive separation outcomes, including the harnessing of kinetic effects (diffusional “sieving”) [829].

Flexible frameworks have recently emerged in the context of separations, with several studies revealing that they can exhibit induced‐fit guest binding [159]. Specifically, flexible CNs can undergo guest‐triggered structural transformations that selectively accommodate certain molecules by in effect adapting its structure to a preferred sorbate. For example, the flexible CN JNU‐1 was found to “breathe” and reposition its UMCs to bind acetylene, achieving a C₂H₂/CO₂ selectivity of ∼285 at 298 K [160]. Similarly, the HUM TIFSIX‐14‐Cu‐i remains essentially nonporous until C₃H₄ induces a structural change that enables its capture and production of C₃H₆ with 99.9999% purity [432]. The ability of a PCN to distort or expand only in the presence of the target sorbate confers a type of “exclusive” selectivity. Although the design rules for flexible sorbents are not yet mature, there are indications that they can offer the best of both worlds: high working capacity (since they can open up when fully loaded); high selectivity (since they remain closed to undesired sorbates) [830]. In this context, approaches like “sorbate‐selective induced fit” [237], and “double sieving” (multi‐step exclusion mechanisms) offer potential to target even the most challenging of trace separations. Nevertheless, the potential of flexible, stimuli‐responsive [831], transiently porous [832], and traditionally non‐porous sorbents, as well as derived composites, remains largely untested for HC separations. Given the distinct characteristics of rigid and flexible sorbents in adsorptive separations of light hydrocarbons, we present a decision framework to guide the choice between rigid and flexible sorbents for HC‐selective separations (Figure 27).

FIGURE 27.

Decision framework: selecting rigid versus flexible hosts for physisorption‐based separations (Ex. = example(s)).

In summary, while a single mechanism may serve as the primary driver for the effective separation of HC mixtures, one must recognize that high selectivity and capacity are most often achieved through the synergistic combination of multiple noncovalent interactions. Rather than relying on a solitary interaction between sorbate and sorbent, the cooperative effects of van der Waals forces, hydrogen bonding, π‐π interactions, and host‐guest complementarity within porous frameworks can work in concert to enhance recognition and separation efficiency.

5.3. The Future in Focus: Remaining Challenges

Overall, future sorbents must satisfy a spectrum of performance parameters beyond selectivity alone (Figure 28). Key requirements include:

1. High adsorption capacity (i.e., working capacity) for the target HCs, to handle large throughputs.

2. Fast kinetics for both adsorption and desorption, enabling rapid cycling and high productivity.

3. High selectivity over competing species (ideally at relevant trace concentrations).

4. Tolerance to humidity and other impurities.

5. Mild regeneration conditions, i.e., low energy input for sorbent recycling.

6. Processability into industrial forms (uniform pellets, durable monoliths, etc.).

7. Chemical, thermal, and mechanical stability over several cycles.

8. Scalability of synthesis and deployment (feasible cost and volume).

9. Multicomponent mixtures separation including the use of tandem beds.

FIGURE 28.

The spectrum of performance parameters that must be exhibited by a sorbent with regard to adsorptive separations and purifications, especially for commercial adoption.

All of the above must be balanced when moving from laboratory demonstrations to commercial adoption. These performance parameters must be addressed from a chemist's lens in that composition can be fine‐tuned from the ground up since, as is the case for many top‐performing sorbents, this requires controlling pore chemistry and pore sizes with sub‐angstrom precision [815, 833]. In addition, it remains a challenge to predict many of the above requirements in advance of conducting an extensive range of experiments, which creates a bottleneck in that the number of PCNs being reported far outpaces our ability to test them comprehensively. This raises the issue of whether AI and predictive modelling are primed to reliably screen for functional properties and accelerate the commercial development of PCNs. It is the opinion of these authors that there is some way to go before this bottleneck is addressed. Some of the above criteria are discussed individually below.

Interference from humidity presents a major hurdle. In real‐world conditions, HC streams (especially gases from air or combustion processes) often contain water vapor, which tends to compete for adsorption sites and/or degrade some PCNs s. Further, some of the highest‐performing CNs are hydrophilic and thus prone to water adsorption. Mitigating this issue is essential unless the application itself is water capture. One solution is to design sorbents with inherent hydrophobicity and water stability – for example, by using hydrophobic linkers or high‐valence metal clusters that resist hydrolysis [834, 835]. Some robust PCNs, like CALF‐20 and BUT‐55, retain their selectivity under humid conditions [74, 695], and some can even exploit pre‐adsorbed water to enhance hydrocarbon uptake via host–guest cooperation [836, 837, 838]. Still, adsorbents that are hydrophobic and offer the high selectivity values needed for trace separations remain rare for some separations and non‐existent for others [839].

Beyond binary separations, the highest volume gas mixtures (e.g., biogas, syngas, air, natural gas) are complex mixtures of varying compositions. That bespoke physisorbents have setting binary selectivity benchmarks for multiple impurities that are present in important gas mixtures (air, natural gas, C₂ and C₃ gases) suggests that a tandem packed fixed‐bed configuration comprising multiple sorbents, each tailored to target a specific impurity, might enable single‐step purification. Such an approach, SSST, was introduced in 2019 to produce polymer‐grade C₂H₄ from ternary (C₂H₂/C₂H₆/C₂H₄) or quaternary (CO₂/C₂H₂/C₂H₆/C₂H₄) gas mixtures, using three ultramicroporous metal‐organic physisorbents, SIFSIX‐3‐Ni [69], TIFSIX‐2‐Cu‐i [36], and Zn‐atz‐ipa [840], in a tandem packed‐bed configuration in the right order [70]. In addition, there are now examples of single sorbents that can address multicomponent HC mixtures. Zn‐atz‐oba (oba = 4,4′‐oxobisbenzoate) was found to produce polymer‐grade purity (>99.95%) C₂H₄ from binary (1:1 for C₂H₄/C₂H₆), ternary (1:1:1 for C₂H₂/C₂H₄/C₂H₆) and quaternary (1:1:1:1 for C₂H₂/C₂H₄/C₂H₆/CO₂) gas mixtures [419]. The Mn(II) CN, NPU‐1 delivered C₂H₄ in >99.9% purity from a 1:1:1 C₂H₂/C₂H₄/C₂H₆ mixture in a single step [840]. Bringing this approach to a new level of complexity, CALF‐20 was recently shown to achieve >99.99% C₂H₄ recovery from a seven‐component HC mixture, C₂H₂/C₂H₄/C₂H₆/CO₂/C₃H₄/C₃H₆/C₃H₈ (0.6/62/10/0.3/0.6/26/0.5, v/v/v/v/v/v/v), even at 74% RH [841], whereas a tandem‐packed connection of two ultramicroporous PCNs, MFM‐300(In) (pore size ≈ 7.2–7.4 Å) and MFM‐300(V) (pore size ≈ 6.8–7 Å) enabled one‐step separation of p‐ and m‐xylene isomers from a ternary equimolar xylene mixture [815].

Another overarching challenge is scalability. To transition from lab discoveries to industrially adopted sorbents, materials must be economical to produce and compatible with standard process engineering. Unfortunately, many CNs and COFs rely on more expensive metals or complex organic linker ligands that can be difficult to synthesize on scale. In addition, shaping materials into usable forms (beads, pellets, or monoliths) without losing performance is non‐trivial. Several studies have explored shaping reticular sorbents into pellets or monoliths. In 2018, Xing's group pelletized several PCNs (SIFSIX‐3‐Ni, SIFSIX‐2‐Cu‐i, GEFSIX‐2‐Cu‐i, TIFSIX‐2‐Cu‐i) for C₂H₂/C₂H₄ separation, with GEFSIX‐2‐Cu‐i pellets showing an IAST selectivity of 116.4 for a 1:99 C₂H₂/C₂H₄ mixture at 298 K and 1 bar [842]. In 2021, Fairen‐Jimenez's group developed monolithic MOFs (monoUiO‐66, monoUiO‐66‐NH₂ , monoHKUST‐1) for CO₂/N₂ and CO₂/CH₄ separations, with monoUiO‐66‐NH₂ achieving a CO₂/CH₄ IAST selectivity of 54 for an equimolar mixture [843]. These studies underscore the main challenge in shaping MOFs for industrial use: balancing adsorption capacity with mechanical strength. Conventional binders often block pores, lowering performance. Recent strategies include ultrahigh‐loading pellets with minimal binder and binder‐free monoliths offering fast kinetics. While monoliths outperform pellets, their scalability is limited. Pelletization therefore remains the more practical route for applications such as carbon capture, provided the trade‐off between capacity and strength can be resolved by mitigating pore blockage.

Thus, an important future direction is the development of adsorbents from low‐cost, abundant, earth‐friendly precursors and simplifying their synthesis. A blueprint in this context is CALF‐20 [74], which is comprised from low‐cost linker ligands, exhibits exceptional steam stability, and performance has been demonstrated over nearly half a million adsorption/desorption cycles. CALF‐20’s success in flue gas CO₂ capture makes it a promising proof‐of‐concept for high TRL development of PCNs.

Finally, there are design limitations to confront. While the first generation of ultramicroporous coordination networks (Gen‐1 HUMs and related PCNs) has delivered several prototypes, they mostly belong to just a few families (e.g., Gen‐2 versions of SIFSIX‐3‐Zn, Mg‐MOF‐74). The development of new Gen‐1 materials with benchmark performance is still largely based on empirical discovery rather than design. A deeper understanding of structure–function relationships is needed to target new families of sorbents. This includes exploring understudied classes of potential sorbents that are already archived in the CSD. In addition, most research has thus far focused on rigid porous frameworks, yet flexible, stimuli‐responsive, and even nonporous or closed pore sorbents (such as molecular organic cage crystals with discrete voids, or composite materials lacking continuous porous channels but with potential to exhibit transient porosity) remain almost unstudied for HC separations. These unconventional sorbents could offer new mechanisms (e.g., gate‐opening, phase change, selective swelling). Additionally, hybrid approaches—like mixed‐matrix membranes or adsorbents integrated with catalytic sites—might address the constraints of pure adsorbents. In parallel, computational screening and machine learning may accelerate the identification of candidate sorbents that meet all of the desired criteria.

In conclusion, the strides made thus far mark what might be considered “the end of the beginning” for HC sorbents. Reticular chemistry and crystal engineering have brought us to the point where physisorbents can viably perform trace separations thought to be infeasible as recently as 15 years ago. The next decade will likely witness an even greater expansion of what is possible, as we leverage the insights gained so far to reach beyond this strong foundation toward more sustainable industrial separation technologies.

Funding

Research Ireland, 16/IA/4624, IRCLA/2019/167, 21/PATH‐S/9454; European Research Council, ADG 885695); Synthesis and Solid State Pharmaceutical Centre, 12/RC/2275_P2.

Conflicts of Interest

The authors declare no conflicts of interest.

Biographies

Xia Li is a postdoctoral researcher in the College of Chemistry at Nankai University (NKU). She received her Ph.D. from the Department of Chemical Sciences at the University of Limerick (UL) in 2024, followed by postdoctoral study at UL (2024‐2025) and NKU (2025‐present). Her research addresses structural design and synthesis of novel azo MOFs using solvothermal/green solvent‐free methods, aiming to create efficient and low‐cost next‐generation materials for water sorption and gas sorption applications. Her work has been published in journals such as JACS and Angewandte Chemie, and she is a recipient of the China Postdoctoral Science Foundation (CPSF) grant.

Soumya Mukherjee is an associate professor of materials chemistry in the Department of Chemical Sciences at University of Limerick (UL). Soumya's recent recognitions include the UL President's Research Output Award 2025/Research Excellence Impact Award 2022, the Thieme Chemistry Journals Award 2025, the Research Ireland–Wales Innovation Network Research Alliance Award 2025, the Bernal Institute Mid‐Career Researcher Award 2024, the SSPC Research Article of the Year Award 2024, and the SSPC Academic Collaboration of the Year Award 2024. Led by Soumya, the Mukherjee Group hosted at the Bernal Institute, UL, focuses upon reticular porous materials that address global challenges in chemical purification, air, and water treatment.

Michael J. Zaworotko (“Dr. Z”) was born in Wales in 1956 and received his B.Sc. and Ph.D. degrees from Imperial College (1977) and the University of Alabama (1982), respectively. He served as a faculty member at Saint Mary's University (1985‐1998), University of Winnipeg (1998‐99), both in Canada, and the University of South Florida, USA, (1999‐2013). In 2013, he joined the University of Limerick, Ireland, where he serves as the Bernal Chair of Crystal Engineering. Research activities focus upon fundamental and applied aspects of crystal engineering. Design and properties of hybrid ultramicroporous materials, flexible sorbents, and multi‐component pharmaceutical materials such as cocrystals are of particular interest.