Tailoring the pore geometry and chemistry in microporous metal–organic frameworks for high methane storage working capacity

Abstract

We realized that tailoring the pore size/geometry and chemistry, by virtue of alkynyl or naphthalene replacing phenyl within a series of isomorphic MOFs, can optimize methane storage working capacities, affording an exceptionally high working capacity of 203 cm³ (STP) cm⁻³ at 298 K and 5–80 bar.

The exponential increase in global energy usage over the past 50 years has pushed the implementation of a cleaner fuel for replacing conventional petroleum fuels to a level of utmost importance. Natural gas (NG), primarily composed of methane, holds bright promise as a viable alternative fuel due to its abundant reserves, low CO₂ emission and high research octane number (RON = 107). However, its widespread use is limited by the relatively low volumetric energy density. Although liquefied NG (LNG) at 113 K and compressed NG (CNG) at 250 bar are already in use, cryogenic and high-pressure storage conditions have severely hampered practical applications. For instance, overhigh pressures required for liquefaction can only be achieved by using complex and expensive multi-stage compressors, leading to some cost, space, and safety issues for use in passenger vehicles. In this regard, adsorbed natural gas (ANG) systems have aroused considerable interest to overcome these issues, which involves filling storage tanks with porous materials to store a high density of methane at modest pressures. To guide the research of the ANG technology, the U.S. Department of Energy (DOE) has set the ambitious volumetric NG storage target of 263 cm³ (STP) cm⁻³ at room temperature (RT) assuming the adsorbent packing loss is ignored, equivalent to that of CNG at 250 bar and 298 K.¹

Extensive research efforts have been devoted to developing new materials with high methane storage capacity in order to meet the challenging storage target. Microporous metal–organic frameworks (MOFs) are known to be very promising for methane storage because of their high surface areas, tunable pores, and versatile chemistry.^2–4 A large number of MOFs, such as HKUST-1, UTSA-76, Co(BDP), MOF-5, NU-125, Al-soc-MOF-1, MAF-38, MOF-905, NJU-Bai43, and MFM-115 have been reported as outstanding storage adsorbents, showing some of the highest gravimetric or volumetric CH₄ storage capacities.^5–11 It is notable that the volumetric working capacity is a key parameter to evaluate the performance of methane storage materials, because it determines the driving range of vehicles powered by natural gas. Since the automobile industry requires an engine inlet pressure of 5 bar, the working capacity (also called deliverable capacity) is defined as the difference of the amount of methane adsorbed between the target storage pressure (generally 35–80 bar) and 5 bar. Prominently, there are two pathways to improve the methane working capacity of a given porous material. On one hand, optimizing pore sizes or functional sites in MOFs can enforce the strength of host–methane and methane–methane interactions (Q_st), contributing to high volumetric methane uptake at high pressure.^5,6 Nevertheless, overhigh Q_st in MOFs leads to a rapid increase in the low-pressure uptake, sometimes delimiting their high working capacity, as exemplified by MOF-74 and PCN-14.⁶ On the other hand, large-pore MOFs (e.g., MOF-905, Al-soc-MOF-1 and NU-111) commonly show moderate Q_st for methane, resulting in a low CH₄ adsorption at 5 bar;^9,10 however, too weak CH₄ affinity delimits the total CH₄ uptake at high pressure, thus sacrificing the working capacity in some cases.¹⁰ Evidently, it is highly desired to optimize the porous structures with suitable methane binding affinity for balancing the CH₄ adsorption amounts between the high and low pressure, thus targeting higher volumetric working capacity.

With this in mind, we herein used a crystal-engineering strategy to finely modulate the pore size and chemistry in a family of isomorphic MOFs for methane storage. We chose NOTT-102 as the fundamental framework backbone because of its high methane storage capacity and easily adjustable functionality on the center phenyl rings.¹⁴ The use of alkynyl or naphthalene groups to replace the phenyl rings in NOTT-102 afforded two porous materials (namely PCN-46 and ZJU-105), in which the pore size and pore-chemistry can be well modulated (Fig. 1). It was found that the expanded pore size and reduced aromatic CH₄-affinity sites in PCN-46a can cause a slight decrease in the low-pressure CH₄ uptake while having a negligible effect on the adsorption at 65/80 bar, leading to an exceptionally high volumetric working capacity of 203 cm³ (STP) cm⁻³ between 5 and 80 bar at 298 K. This value is comparable to the best-performing MOFs reported, such as MFM-115a (208 cm³ (STP) cm⁻³)¹² and MOF-905 (203 cm³ (STP) cm⁻³).¹³

Fig. 1.

Comparison of the organic linkers and crystal structures of ZJU-105, NOTT-102, and PCN-46.

The linkers of H₄L1 and H₄L2 were elaborately designed by using two naphthalene or alkynyl groups replacing the phenyl rings on the linker of NOTT-102 (Fig. 1). Both of the organic linkers were synthesized through a multistep reaction procedure (Schemes S1 and S2, ESI†). A solvothermal reaction of H₄L1 with Cu(NO₃)₂·2.5H₂O in a mixture of DMF/H₂O (in the presence of HCl) at 80 °C for 3 days afforded blue block crystals of a novel MOF (ZJU-105) with a framework formula of [Cu₂(L1)(H₂O)₂]_n. High-quality crystals of PCN-46, [Cu₂(L2)(H₂O)₂]_n, were synthesized by a solvothermal reaction of H₄L2 with Cu(NO₃)₂·2.5H₂O in a mixture of DMF/EtOH/H₂O (in the presence of HNO₃) at 85 °C for 1 day. It is worth noting that the sample of PCN-46 in the literature was directly synthesized by using a precursor compound (5-ethynylisophthalic acid) of H₄L2 to react with copper salts, which cannot guarantee the high quality of the bulk samples.¹⁵ In this work, the phase purity of bulk materials for ZJU-105 and PCN-46 was confirmed by powder X-ray diffraction (PXRD) and the following surface area/pore volume measured (Fig. S10–S15, ESI†).

The crystal structures of the three MOFs investigated are schematically shown in Fig. 1. Structural analysis reveals that both of the ZJU-105 and PCN-46 frameworks are isoreticular to NOTT-102, which consist of paddle-wheel dinuclear Cu₂(COO)₄ units linked by the carboxylates of L⁴⁻ linkers to form a 3D NbO-type network.^14d As expected, the use of alkynyl or naphthalene groups to replace the phenyl rings on NOTT-102 can finely modulate the pore size/geometry and chemistry. As shown in Fig. 1, the framework of NOTT-102 has two types of cages: one cuboctahedral cage of about 12.5 × 14.7 Ų and another large irregular elongated cage of ≈9.6 × 28.6 Ų. When the bulky naphthalene rings were incorporated, the pores of ZJU-105 were significantly decreased, wherein the large irregular cage was divided into three small cages with the diameters of 3.8 and 8.0 Å. In contrast, we slim down the organic ligands in PCN-46 by virtue of triple-bond spacers replacing the phenyls in NOTT-102. Relative to NOTT-102, PCN-46 exhibits a similar pore geometry with slightly expanded pore spaces. In addition, the incorporation of alkynyl or naphthalene groups in PCN-46 and ZJU-105 can reduce or increase the aromatic moieties that provide π···CH₄ interactions.¹⁶ Therefore, the fine-tuned pore size/geometry and aromatic moieties in this series of MOFs can systematically affect the strength of CH₄ binding affinities, thus optimizing the target working capacities.

Before methane adsorption measurements, we first measured the N₂ adsorption isotherms at 77 K of the activated ZJU-105a and PCN-46a. As shown in Fig. 2a, the N₂ isotherms of ZJU-105a and PCN-46a exhibit a reversible type-I character with saturated uptake of 670 and 803 cm³ g⁻¹, respectively. The Brunauer–Emmett–Teller (BET) surface area and pore volume of ZJU-105a were determined to be 2608 m² g⁻¹ and 1.037 cm³ g⁻¹, respectively, notably lower than those of NOTT-102a (3342 m² g⁻¹ and 1.268 cm³ g⁻¹) due to the incorporated bulky naphthalene groups. In contrast, PCN-46a with the slimmed alkynyl groups shows a comparable BET surface area and pore volume (3224 m² g⁻¹ and 1.243 cm³ g⁻¹). We note that these values of PCN-46a are significantly higher than the reported values (2500 m² g⁻¹ and 1.012 cm³ g⁻¹) in the previous literature,¹⁵ and agree well with the theoretical ones (3634 m² g⁻¹ and 1.266 cm³ g⁻¹) calculated from the crystal structure. This further confirms the high quality of the bulk samples that we synthesized, encouraging us to retest their methane storage capacities.

Fig. 2.

(a) Nitrogen isotherms of ZJU-105a, NOTT-102a and PCN-46a at 77 K. (b) Total volumetric CH₄ adsorption isotherms of PCN-46a and (c) ZJU-105a. (d) Comparison of the CH₄ adsorption amount of the indicated MOFs at 5 and 65 bar.

Next, we collected the high-pressure methane adsorption isotherms for ZJU-105a and PCN-46a at 273 and 298 K over the pressure range of 0–100 bar. As presented in Fig. 2b and c, ZJU-105a and PCN-46a exhibit high total volumetric methane uptake values of 228 and 233 cm³ (STP) cm⁻³ at 65 bar and 298 K, respectively. These values are comparable to that of NOTT-102a (233 cm³ (STP) cm⁻³), indicating that tuning the pore size/geometry in these isomorphic MOFs shows a negligible effect on the total volumetric CH₄ uptake at high pressure. This might be attributed to the mutual compensation between the porosity, crystal density and CH₄ binding strength. With the pressure increasing to 80 bar, the volumetric uptake of ZJU-105a and PCN-46a can be further increased to 243 and 246 cm³ (STP) cm⁻³, respectively, achieving 93% of the new DOE’s volumetric target assuming the packing density loss is ignored. These values are higher than those of most of the promising MOFs, such as MOF-905 (228 cm³ (STP) cm⁻³)¹³ and Al-soc-MOF-1 (221 cm³ (STP) cm⁻³),^9a and approach those of the best-performing MFM-115 (256 cm³ (STP) cm⁻³)¹² and HKUST-1 (272 cm³ (STP) cm⁻³).^4b

In practical applications, the driving range of an ANG vehicle is determined primarily by the working capacity of the adsorbent. We first evaluated the volumetric working capacities of ZJU-105a and PCN-46a to compare with NOTT-102a and some other indicated MOFs (Table 1). As shown in Fig. 2d, due to the smaller pore spaces and more aromatic moieties, ZJU-105a takes up more CH₄ (53 cm³ (STP) cm⁻³) at 5 bar than NOTT-102a (48 cm³ (STP) cm⁻³). In contrast, PCN-46a shows a lower adsorption at 5 bar (43 cm³ (STP) cm⁻³) than NOTT-102a. Therefore, PCN-46a exhibits a higher volumetric working capacity of 190 cm³ (STP) cm⁻³ than NOTT-102a (185 cm³ (STP) cm⁻³) and ZJU-105a (175 cm³ (STP) cm⁻³) between 65 and 5 bar at 298 K. It is worth noting that this working capacity of PCN-46a outperforms that of most of the promising MOFs (e.g., NU-125, NU-111, UTSA-20, Al-soc-MOF-1, and MOF-5, Table 1), and is comparable to that of the benchmark HKUST-1 (190 cm³ (STP) cm⁻³),^4a MFM-115 (191 cm³ (STP) cm⁻³),¹² and UTSA-76a (197 cm³ (STP) cm⁻³).¹¹ When the storage pressure is increased to 80 bar, PCN-46a exhibits an exceptionally high CH₄ working capacity of 203 cm³ (STP) cm⁻³ between 80 and 5 bar at RT (Fig. 3), much higher than that of ZJU-105a (190 cm³ (STP) cm⁻³) and rivaling the state-of-the-art materials reported, such as MFM-115a (208 cm³ (STP) cm⁻³)¹² and MOF-905 (203 cm³ (STP) cm⁻³).¹³

Table 1.

Summary of some promising MOFs for high-pressure CH₄ storage

MOFs	SBETa (m2 g−1)	Total uptakeb cm3/cm3	Working capacityc cm3/cm3	Qst (kJ mol−1)
PCN-46a	3224	246 (233)	203 (190)	14.2
ZJU-105a	2608	243 (228)	190 (175)	16.58
MFM-115a12	3394	256 (238)	208 (191)	16.3
MOF-90513	3490	228 (206)	203 (181)	11.7
NOTT-102a14a	3342	−(233)	−(185)	14.90
Al-soc-MOF-19a	5585	221 (197)	201 (176)	11.0
HKUST-14a,b	1850	272 (267)	200 (190)	17.0
UTSA-7611	2820	−(257)	−(197)	15.5
MAF-3816a	2022	273 (263)	197 (187)	21.6
FDM-89b	3643	215 (193)	193 (171)	10.4
MOF-17710a	4500	205 (187)	185 (167)	9.9
NU-1254a	3120	−(232)	−(183)	15.1
PCN-144a,b	2000	250 (230)	178 (157)	18.7
Cu-tbo-MOF-516b	3971	216 (199)	175 (158)	20.4
Ni-MOF-744a,b	1350	267 (251)	152 (129)	21.4

^aCalculated from N₂ isotherms at 77 K.

^bAt 80(65) bar and 298 K.

^cThe working capacity is defined as the difference in the total uptake values between 80(65) and 5 bar.

Fig. 3.

Comparison of the volumetric CH₄ working capacities of PCN-46a and ZJU-105a with some best-performing MOFs reported at 298 K.

To gain better insight into the origin of different volumetric low-pressure uptake values and working capacities of ZJU-105a and PCN-46a, we first calculated their isosteric heats of adsorption (Q_st) from the temperature-dependent isotherms using the virial method (Fig. S22–S24, ESI†). The initial Q_st of ZJU-105a was calculated to be 16.58 kJ mol⁻¹, much higher than those of NOTT-102a (14.90 kJ mol⁻¹) and PCN-46a (14.2 kJ mol⁻¹). Due to the same framework backbone and topology, these MOFs have a similar density of open metal sites. The incorporation of naphthalene groups in ZJU-105a can not only notably reduce the pore sizes but also offer more aromatic moieties to interact with CH₄ molecules via π···CH₄ interactions,¹⁶ contributing to the enhanced CH₄ affinity and thus a higher CH₄ adsorption at low pressure of 5 bar than that of NOTT-102a. In contrast, the use of alkynyl groups to replace the benzene rings in PCN-46a can slightly expand the pore sizes with no aromatic moieties, resulting in lower CH₄ affinity and uptake at 5 bar. Therefore, PCN-46a shows the lowest low-pressure CH₄ uptake among these three MOFs, thus affording the highest volumetric working capacity.

We have also studied the CO₂ and H₂ uptake properties of ZJU-105a and PCN-46a over wide temperature and pressure ranges (Fig. S18–S21, ESI†). ZJU-105a and PCN-46a exhibit high total volumetric H₂ storage capacities of 48 g L⁻¹ (6.78 wt%) and 50 g L⁻¹ (7.47 wt%) at 65 bar and 77 K, meeting the DOE’s targets of 40 g L⁻¹ (5.5 wt%), albeit at 77 K rather than at ambient temperature. These volumetric values are comparable to those of some of the benchmark MOFs (Table S4, ESI†). The volumetric and gravimetric working capacities of PCN-46a, calculated from (T_min = 77 K, P_max = 65 bar) to (T_max = 160 K, P_min = 5 bar), are 47.4 g L⁻¹ and 7.1 wt%, rivaling the best-performing material IRMOF-20 (47 g L⁻¹ and 8.4 wt%).¹⁷ In addition, ZJU-105a and PCN-46a also exhibit high volumetric CO₂ storage capacities of 328 and 309 cm³ (STP) cm⁻³ (Table S5, ESI†). Finally, we calculated the Q_st values of H₂ and CO₂ for both MOFs. As shown in Fig. S25–S30 (ESI†), the Q_st values of ZJU-105a for both H₂ and CO₂ are much higher than those of PCN-46a, consistent with those for CH₄, further confirming that PCN-46a has a weaker interaction with various gas molecules.

In summary, we herein realized that the systematical tuning of the pore size and chemistry in a series of isomorphic MOFs based on NOTT-102 can effectively regulate the methane storage and working capacities. The use of alkynyl and naphthalene replacing phenyl in the NOTT-102 backbone has enabled a gradually decreased low-pressure CH₄ uptake from ZJU-105a to PCN-46a, combined with a comparable total storage capacity at high pressure. PCN-46a thus exhibits the highest working capacity of 203 cm³ (STP) cm⁻³ (at 5–80 bar and 298 K) among the isomorphic MOFs, making it among the best performing porous materials for CH₄ storage. This work may provide some guidance for the development of new MOFs with common alkynyl connectivity instead of phenyl to reduce low-pressure CH₄ absorption and thus to improve the volumetric working capacity.