CALF-20 Isoreticular MOF Series

• CALF-20 isoreticular series comprises six zinc triazolate MOFs with varied linkers that modulate pore volume, pore limiting diameter, and cavity size.
• The series shows that balanced CO2 adsorption and suppressed CH4 uptake—achieved via precise linker substitution—are key for effective PVSA biogas upgrading.
• Process-level evaluations demonstrate that FumCALF-20 outperforms its analogues by achieving >90% CH4 purity and recovery, underscoring the impact of isoreticular tuning.

The CALF-20 isoreticular series denotes, in the explicit sense used in recent process-screening literature, a six-member family of zinc triazolate metal-organic frameworks comprising CALF-20 and five linker-substituted derivatives—SquCALF-20, FumCALF-20, BdcCALF-20, TtdcCALF-20, and CubCALF-20—in which the CALF-20 framework motif is retained while the oxalate pillar/linker is replaced (Shin et al., 20 Jul 2025). The parent framework is a zinc triazolate MOF made of 1,2,4-triazolate-bridged Zn(II) layers pillared by oxalate ligand, and the broader CALF-20 literature establishes it as a benchmark CO$_2$-capture sorbent with unusual adsorption, transport, and mechanical behavior. At the same time, earlier atomistic studies on CALF-20(Zn) were focused on a single exemplar and explicitly did not define a formal isoreticular family, so the phrase “CALF-20 isoreticular series” is best understood as a recent, application-oriented framing rather than an older, fully systematized reticular taxonomy (Fan et al., 2023).

1. Definition, membership, and scope

The six materials examined as the CALF-20 isoreticular series are the parent CALF-20 and five derivatives obtained by replacing the oxalate pillar/linker with squarate, fumarate, benzenedicarboxylate, thieno[3,2-b]thiophene-2,5-dicarboxylate, or cubanedicarboxylate (Shin et al., 20 Jul 2025). In that study, the parent material was included because it is already known as the benchmark CALF-20 adsorbent, with high CO$_2$ selectivity, thermal and chemical stability, and pilot-scale deployment for CO$_2$ capture. The derivatives were previously proposed by Gopalsamy et al. as linker-engineered analogues intended to tune pore structure and adsorption energetics.

This usage of “series” is narrower than a general family genealogy. CALF-20(Zn), previously reported by Shimizu and co-workers, is treated in earlier simulation work as a single zinc-based framework rather than as one entry in a systematically compared reticular family (Yann et al., 2023). Those earlier papers are therefore foundational for understanding the parent framework’s architecture and physics, but not for defining the series itself.

The rationale for studying the series is application-specific. The series paper evaluates whether isoreticular tuning of CALF-20 through linker substitution can improve pressure/vacuum swing adsorption for biogas upgrading, specifically CO$_2$/CH$_4$ separation. In that sense, the CALF-20 isoreticular series is a materials-design construct organized around a conserved Zn–triazolate framework motif and chemically varied pillars.

2. Framework motif and structural variation

Across the series, the conserved architectural motif is the CALF-20 framework: a zinc triazolate MOF composed of 1,2,4-triazolate-bridged Zn(II) layers pillared by a dicarboxylate-type linker into a three-dimensional lattice (Shin et al., 20 Jul 2025). For the parent CALF-20, more detailed single-material analysis describes a three-dimensional pillared framework in which 1,2,4-triazolate-bridged Zn ions form 2D layers and oxalate acts as a pillar/hinge connecting these layers. The framework flexibility is attributed to a crossed 3D configuration with zinc triazolate grids and oxalate pillars, with the zinc triazolate grids lying parallel to the ac plane (Fan et al., 2023).

The principal structural differences within the series were quantified by pore volume, pore limiting diameter (PLD), largest cavity diameter (LCD), and crystal density.

Material	Linker	Key pore metrics
CALF-20	oxalate	$0.35\ \mathrm{cm^3/g}$; PLD $3.0\ \text{Å}$; LCD $4.4\ \text{Å}$
SquCALF-20	squarate	$0.40\ \mathrm{cm^3/g}$; PLD $2.9\ \text{Å}$; LCD $_2$0
FumCALF-20	fumarate	$_2$1; PLD $_2$2; LCD $_2$3
BdcCALF-20	benzenedicarboxylate	$_2$4; PLD $_2$5; LCD $_2$6
CubCALF-20	cubanedicarboxylate	$_2$7; PLD $_2$8; LCD $_2$9
TtdcCALF-20	thieno[3,2-b]thiophene-2,5-dicarboxylate	$_2$0; PLD $_2$1; LCD $_2$2

CALF-20 is the most compact member, with the smallest pore volume and the highest density, $_2$3. SquCALF-20 has the narrowest PLD, $_2$4, and the series paper explicitly attributes its restricted molecular accessibility to that feature. FumCALF-20 combines a relatively large pore volume with the largest PLD in the series, $_2$5, and is described as having excellent pore connectivity and high surface area, although no numerical surface area value is given in that paper. TtdcCALF-20 has the largest pore volume, $_2$6, and the most open pore architecture (Shin et al., 20 Jul 2025).

These structural differences are process-relevant because the series does not reward simple pore enlargement. The results show that both excessive restriction and excessive openness are detrimental under PVSA biogas-upgrading conditions. A plausible implication is that the series is governed by a narrow structural window in which pore accessibility, CH$_2$7 suppression, and regenerability can be balanced simultaneously.

3. Adsorption thermodynamics and equilibrium separation behavior

The equilibrium evaluation of the series was carried out by rigid-framework grand canonical Monte Carlo simulations in RASPA 2.0 from $_2$8 to $_2$9 at 273, 298, and 323 K, using DREIDING for framework Lennard-Jones parameters, TraPPE for adsorbates, PACMAN DDEC06 charges, Lorentz-Berthelot mixing rules, and Widom particle insertion for adsorption enthalpies (Shin et al., 20 Jul 2025). For the parent CALF-20, the simulated adsorption isotherm at 298 K was reported to agree satisfactorily with experiment from Gopalsamy et al. All six materials show steep low-pressure CO$_2$0 uptake followed by saturation, while CH$_2$1 uptake remains lower but varies enough to dominate process ranking.

The adsorption selectivity and working-capacity metrics used in the series study were

$_2$2

and

$_2$3

with binary isotherms predicted by the extended dual-site Langmuir model at 298 K for a 50:50 CO$_2$4:CH$_2$5 mixture and working capacities defined between 1 bar adsorption and 0.1 bar desorption (Shin et al., 20 Jul 2025).

A central result is that equilibrium quality is not captured by CO$_2$6 uptake alone. CALF-20 has the strongest CO$_2$7 adsorption enthalpy in the series, $_2$8, and also the strongest CH$_2$9 affinity except for CubCALF-20 in the CH$_4$0 ranking, with $_4$1. Yet its working capacities are poor for the target application: $_4$2 and $_4$3, with adsorption selectivity $_4$4. By contrast, FumCALF-20 has more moderate adsorption enthalpies, $_4$5 for CO$_4$6 and $_4$7 for CH$_4$8, but the best equilibrium separation profile in the series: $_4$9, $0.35\ \mathrm{cm^3/g}$0, and adsorption selectivity $0.35\ \mathrm{cm^3/g}$1.

The comparison between FumCALF-20 and TtdcCALF-20 is especially revealing. TtdcCALF-20 has the highest reported single-component CO$0.35\ \mathrm{cm^3/g}$2 capacity, approximately $0.35\ \mathrm{cm^3/g}$3, and a CO$0.35\ \mathrm{cm^3/g}$4 working capacity of $0.35\ \mathrm{cm^3/g}$5, essentially equal to FumCALF-20. However, it also has a much larger CH$0.35\ \mathrm{cm^3/g}$6 working capacity, $0.35\ \mathrm{cm^3/g}$7, and a lower adsorption selectivity, $0.35\ \mathrm{cm^3/g}$8. SquCALF-20 lies at the opposite extreme: it suppresses CH$0.35\ \mathrm{cm^3/g}$9 well, with $3.0\ \text{Å}$0, but its restricted accessibility limits $3.0\ \text{Å}$1 to $3.0\ \text{Å}$2. BdcCALF-20 and CubCALF-20 occupy intermediate positions, with respectable CO$3.0\ \text{Å}$3 capacities but too much CH$3.0\ \text{Å}$4 uptake for optimal PVSA operation.

The series therefore establishes a design rule stated explicitly in the source study: moderate CO$3.0\ \text{Å}$5 binding plus low CH$3.0\ \text{Å}$6 binding is better than simply maximizing CO$3.0\ \text{Å}$7 affinity or raw CO$3.0\ \text{Å}$8 capacity (Shin et al., 20 Jul 2025).

4. PVSA process evaluation and ranking

Process-level performance was assessed with a modified five-step Skarstrom PVSA cycle comprising pressurization, adsorption, heavy reflux, counter-current depressurization, and light reflux, using a 1D dynamic adsorption-column model that is non-isothermal and non-isobaric (Shin et al., 20 Jul 2025). The feed was $3.0\ \text{Å}$9 CO$4.4\ \text{Å}$0:CH$4.4\ \text{Å}$1 at $4.4\ \text{Å}$2. The model assumed ideal-gas behavior, axially dispersed plug flow, gas-solid thermal equilibrium, no radial gradients, an LDF solid-phase mass-transfer law,

$4.4\ \text{Å}$3

Ergun pressure drop, and no heat transfer across the column wall. Optimization was performed with the Thompson Sampling Efficient Multi-objective Optimization algorithm using 90 initial evaluations followed by 150 consecutive iterations.

The ranking is unambiguous: FumCALF-20 is the only material in the series that can reach CH$4.4\ \text{Å}$4 purity $4.4\ \text{Å}$5 while maintaining high recovery (Shin et al., 20 Jul 2025). The source study further states that FumCALF-20 achieved CH$4.4\ \text{Å}$6 purity and recovery simultaneously above 0.90 and was the only member to meet that target. All other materials failed for distinct reasons. CALF-20 underperformed because its CO$4.4\ \text{Å}$7 working capacity was by far the lowest in the series and regeneration remained problematic even at $4.4\ \text{Å}$8. SquCALF-20 failed because narrow accessibility limited CO$4.4\ \text{Å}$9 swing capacity. BdcCALF-20 and CubCALF-20 admitted too much CH$0.40\ \mathrm{cm^3/g}$0. TtdcCALF-20 most clearly demonstrated the inadequacy of equilibrium CO$0.40\ \mathrm{cm^3/g}$1 capacity as a sole screening criterion: despite near-top CO$0.40\ \mathrm{cm^3/g}$2 working capacity, its process-level CH$0.40\ \mathrm{cm^3/g}$3 purity reached only about 60%.

For FumCALF-20, the paper reports four example PVSA operating points where both purity and recovery exceed 90%: $0.40\ \mathrm{cm^3/g}$4, $0.40\ \mathrm{cm^3/g}$5, $0.40\ \mathrm{cm^3/g}$6, and $0.40\ \mathrm{cm^3/g}$7 (Shin et al., 20 Jul 2025). Its economic Pareto front spans a tradeoff between energy consumption of about $0.40\ \mathrm{cm^3/g}$8–$0.40\ \mathrm{cm^3/g}$9 and productivity up to about $2.9\ \text{Å}$0.

The process study’s broader significance is methodological. Several members that appear attractive from equilibrium adsorption alone fail when regenerability, CH$2.9\ \text{Å}$1 co-adsorption, and cycle-level purity-recovery tradeoffs are taken into account. The series is therefore a direct example of why MOF screening for separations cannot be reduced to uptake or selectivity measured at a single state point.

5. Mechanistic descriptors from CALF-20(Zn) case studies

Although earlier CALF-20 papers do not define the isoreticular series, they establish the parent framework’s microscopic behavior in ways that are highly informative for series-level comparison. A detailed transport study of CALF-20(Zn) describes a three-dimensional zinc–triazolate–oxalate framework with cage-like subnanometer pores, repeatedly termed “angstropores” or sub-nanopores, with pore size and cage diameters of $2.9\ \text{Å}$2–$2.9\ \text{Å}$3 and surface area about $2.9\ \text{Å}$4 (Yann et al., 2023). In that framework, CO$2.9\ \text{Å}$5 initially adsorbs near the center of cages rather than at obvious wall sites, while H$2.9\ \text{Å}$6O forms quasi-1D water wires rather than conventional clusters. The diffusion of both CO$2.9\ \text{Å}$7 and H$2.9\ \text{Å}$8O is non-monotonic with loading, with a minimum in corrected CO$2.9\ \text{Å}$9 diffusivity around $_2$00 and a minimum in corrected H$_2$01O diffusivity around $_2$02. For CO$_2$03, a flexible framework increases diffusivity by about one order of magnitude relative to a rigid framework. The same paper explicitly notes that it is not a comparative study across multiple CALF-20 analogues, but it suggests that changes in pore size, polarity, linker chemistry, or metal substitution could strongly reshape both uptake and microscopic transport.

A later theoretical treatment of CO$_2$04 in CALF-20 identifies a compact descriptor set that is especially suitable for series-level comparison, even though that paper also focuses on a single CALF-20 material (Gonçalves et al., 10 Jul 2025). Using an adsorption-energy-distribution approach extracted from one wide-range CO$_2$05 isotherm, the paper models CALF-20 as a two-site sub-nanoporous adsorbent with total saturation capacity $_2$06, site capacities $_2$07 and $_2$08, and $_2$09 site binding energies $_2$10 and $_2$11. The dominant diffusion pathway is along [011] with a barrier of about $_2$12, whereas [100] has a much larger barrier of about $_2$13. The favorable [011] barrier is approximately the difference between the two site binding energies, $_2$14. This suggests that, within a CALF-20-type architecture, the most transferable descriptors are the number of adsorption site classes, site-specific saturation capacities, site-specific binding energies, adsorbed-state vibrational frequencies, Henry constant, dominant transport-pathway barriers, and characteristic jump lengths.

Taken together, these single-member studies imply that the CALF-20 isoreticular series is not only a set of pore-size variants. It is also a platform in which adsorption-site hierarchy, guest–guest interactions, confinement, and transport topology are likely to remain tightly coupled.

6. Mechanical context, polymorphism, and unresolved boundaries

The parent CALF-20 framework also has an unusually rich mechanical and thermal phenomenology that informs how the series should be interpreted structurally. A first-principles and machine-learned-potential study describes CALF-20 as Zn$_2$15(1,2,4-triazolate)$_2$16(oxalate), built from 2D zinc–triazolate grids pillared by oxalate linkers into a 3D network, and repeatedly emphasizes the coexistence of a rhombic or lozenge-like zinc triazolate layer and oxalate pillars acting as hinges (Fan et al., 2023). The framework is strongly anisotropic: the a-axis is associated with the oxalate-pillar direction and greater stiffness, while the bc plane is the more flexible rhombic-shaped zinc triazolate grid. CALF-20 displays negative area compressibility, negative thermal expansion, a negative Poisson’s ratio reaching $_2$17, and a two-step elastic response under tension along [001]. At 0 K the first stress maximum along [001] occurs at $_2$18 strain and fracture at $_2$19; at $_2$20, the failure strain remains as high as $_2$21. The strain-induced metastable phase lies only $_2$22 per unit cell above the pristine form, with forward and reverse barriers of $_2$23 and $_2$24 per unit cell, and is explicitly noted to be similar to the humidity-responsive polymorph B-CALF-20.

This body of evidence clarifies a recurrent misconception. CALF-20 is a benchmark parent material, but it is not automatically the best member of the isoreticular series for every separation problem. In biogas PVSA, the parent framework is outperformed decisively by FumCALF-20 (Shin et al., 20 Jul 2025). A second misconception is terminological: the single-material CALF-20 papers do not establish a formal CALF-20 isoreticular family map, and the closest family-level contextualization in the mechanical study is the polymorphic relationship to B-CALF-20 rather than a linker-extended series.

Important boundaries remain unresolved. The series-level PVSA study used rigid-framework adsorption simulations, did not model humid feeds, applied fixed mass-transfer coefficients rather than material-specific diffusivities, and did not report material-by-material hydrolytic stability or thermal decomposition for the five derivatives. It also did not treat pelletization, binder effects, or mechanical strength across the series. In addition, the supporting-information equations for productivity and energy requirement appear CO$_2$25-based, whereas the main-text figure labels and discussion refer to CH$_2$26-based outputs, leaving an explicit metric inconsistency in the extracted text (Shin et al., 20 Jul 2025). These caveats do not alter the reported ranking, but they define the present limits of the term “CALF-20 isoreticular series” as a rigorously characterized materials family.

The current literature therefore supports a precise, limited conclusion. The CALF-20 isoreticular series is presently best understood as a six-member, linker-varied Zn–triazolate framework family whose separation performance is governed by a balance among pore accessibility, adsorption enthalpy, CH$_2$27 suppression, and regenerability. Within that family, FumCALF-20 is the standout candidate for PVSA biogas upgrading, while the parent CALF-20 remains the mechanistically best characterized member and the structural template from which series-level comparisons derive.