1 Introduction

Hydrogen is the ideal energy carrier because it burns completely free of pollutants and has a potentially high energy content per mass. The biggest challenge for the use of hydrogen as an energy carrier for mobile applications is a safe and cost-effective storage system. So far two commercially available methods have been developed: Hydrogen gas in high-pressure vessels and liquid hydrogen in cryogenic tanks. Both methods have serious disadvantages, e.g., the large volume occupied by gaseous hydrogen at room temperature even at high pressure, or extremely low temperatures of around 20 K that are required to liquefy H2.

One alternative is hydrogen storage in solids. Here, two kinds of interactions have to be distinguished, strong chemical binding (chemisorption), which forms hydrides or complex hydrides, and weak binding on the surface by adsorption (physisorption). Compared to chemical storage, the physical adsorption of H2 has several advantages. Hydrogen is reversibly stored by physisorption, and since no energy barrier exists between the gas phase and the adsorbed state, fast adsorption and desorption kinetics are guaranteed. Furthermore, owing to the low heat of adsorption involved in the physisorption of hydrogen, an operation is possible without additional thermal control systems or requiring only small heat management. However, the low adsorption enthalpy involved in molecular adsorption on porous materials requires low temperatures of about 80 K for achieving technologically relevant storage densities.

Physical adsorption of a gas on a solid can be described as the accumulation of molecules at the interface between the surface of the solid and the gas phase. Responsible for this phenomenon are attractive van der Waals or London dispersion forces. This kind of interaction originates in the attraction between fluctuating and induced dipoles between molecules in the gas phase and atoms on the surface of the solid. Since the available surface is dominating this phenomenon, highly porous materials with large internal surface areas are of particular interest for hydrogen storage by physisorption.

In this mini-review, we summarize the historical development of the past 25 years, with a special focus on the development of the laboratory at MPI MF now MPI IS.

1.1 Early investigations on activated carbons

Hydrogen storage by cryo-adsorption on high surface area activated carbons started already in the 1960s and have been continued by only a few laboratories in the 1980s and 1990s [1, 2], for a brief history see, e.g., Dillon and Heben [3]. The possibility of hydrogen storage by physisorption on activated carbons has been demonstrated. At cryogenic temperatures (mainly near the boiling point of nitrogen), cryo-adsorption can provide a viable method of storing hydrogen at lower pressure and therefore economically. Chahine and Bose formulated the goal The key to achieving this level of performance is the use of a densification process which can produce activated carbons with a high surface area, coupled with a high bulk density [4].

The linear correlation between the hydrogen uptake at 77 K and the specific surface area (SSA) is now called Chahine’s rule [5]. Activated carbons are typically produced by burning organic waste, e.g., coconut shells. The porosity of the resulting activated carbons can be increased by KOH treatments and specific surface areas of over 3000 m2/g have been reached. However, the structural characterization of these disordered materials is difficult, making a well-defined optimization difficult or impossible.

1.2 The hope and hype on carbon nanotubes

In 1991, Iijima’s discovery of carbon microtubules with an extension of a few nanometers in two dimensions opened the field of carbon nanostructures with well-defined geometry and long extensions in one dimension [6]. Especially, single-walled carbon nanotubes (SWCNTs) showed exceptional electric and thermal conductivities due to their long-range order. Their enormous mechanical strength led to inspirations such as the space elevator, a satellite connected to earth with a cable of an SWCNT bundle, see the title page of American Scientist 1997. Nobel Laureate Rick Smalley stated “This means one might expect for such ropes a real-life strength of 130 gigapascals, almost a hundred times stronger than steel but one-sixth its weight. This may be a useful combination.” [7]. The hopes for large-scale production of well-defined, purified SWCNTs have been pushed by many laboratories worldwide [8].

Inspired by these “tantalizing” properties, researchers started to look at hydrogen storage properties. The first results on hydrogen uptake in SWCNTs have been based on an extrapolation of a small amount of hydrogen assigned to a small content of SWCNTs in the carbon soot. Nevertheless, this promising result at room temperature and ambient pressure created an enormous research activity, even an economic analysis predicted: "still a long road from science to commerce" [9]. In contrast, a “Review of theoretical calculations of hydrogen storage” by Merregalli and Parinello [10] indicated a hydrogen uptake smaller than some of the more optimistic experimental results.

One hypothesis has been that SWCNTs have to be opened for access of hydrogen molecules to the inside; therefore, ultrasonication in acidic environments seemed to be the choice. Under these harsh conditions, unfortunately, metal nanoparticles are eroded from the ultrasonication horn, which is typical of a titanium alloy [11]. In addition, Ti is a well-known hydride former and therefore, caused an additional hydrogen uptake of the sample. A collaborative project funded by the Federal Ministry for Education and Research in Germany (BMBF) reported a hydrogen uptake below 1 wt% for SWCNTs at ambient temperatures. This opened a controversy which has been critically discussed in Nature in 2001 [12] and one year later in a cover story of Chemical and Engineering News [13]. Finally, initial results on high hydrogen uptake at ambient conditions could not be independently reproduced [14]. In 2009, Harris described this in his book on carbon nanotube science, as ‘‘the most controversial episode in nanotube science.’’ [15]. At cryogenic conditions, i.e., 77 K, hydrogen uptake in SWCNTs depends linearly on the specific surface area following Chahine’s rule as for any carbonaceous material [16, 17]. Since purified SWCNTs have a specific surface area of 1000 m2/g, the hydrogen uptake at 77 K was just around 2 wt% and below most commercial activated carbons. Therefore, the long-range order responsible for the "tantalizing" properties as electric/thermal conductivity or mechanical strength play no role in hydrogen adsorption since physisorption is a local phenomenon influenced by only the next nearest neighbor atoms of the adsorbent.

1.3 The race to high surface area and gravimetric capacity

The development of coordination polymers with permanent porosity opened a new route to nanoporous materials with ultra-high internal surfaces. Mainly metal–organic frameworks (MOFs) have been attracting a great deal of attention. In 2002, the first publication on MOF-5 as a potential material for methane storage demonstrated the high permanent porosity and accessibility to gas molecules [18]. The first investigation of hydrogen storage in MOF-5 was reported, showing a sharp increase in H2 uptake at very low pressures and a maximum of 4.5 wt% below 1 bar at 77 K [19]. In the same year 2003, a hydrogen storage capacity of 3.8 and 3.1 wt% loaded on MIL-53 with Al3+ and Cr3+ was reported at 77 K under 1.6 MPa [20]. Soon after, the exceptional hydrogen uptake and isotherm shape at 77 K in MOF-5 were corrected to 1.3 wt% at 1 bar by the same group [21], and the correct isotherm shape and a lower uptake at room temperature were observed by Panella and Hirscher [22]. In a follow-up, a maximum uptake in MOF-5 has been determined by B. Panella to be 5.1 wt% at 77 K and high pressure [23]. Hydrogen adsorption measurements confirmed an excess uptake of about 5 wt%, therefore reaching already values comparable to the best activated carbons, i.e., 4.5 wt% on activated carbon [24].

At 77 K and high pressures above 20 bar, the maximum storage capacities are closely related to the specific surface area accessible to H2 molecules. Gravimetric capacities, therefore, tend to follow Chahine's rule, which is a linear correlation stating approximately 1 wt% H2 uptake per 500 m2 g−1 of SSA, as shown in Fig. 1. This result started a race to synthesize new MOFs with higher and higher surface area and porosity, and therefore optimizing gravimetric hydrogen capacity. As a result, in 2006 MOF-177 made a huge step in increasing porosity and specific surface area to approximately 4700 m2 g−1, showing an excess H2 adsorption capacity of 7.5 wt% at 77 K and 70 bar [25]. Composed of {Zn4O(CO2)6} units and 4,4′,4′′-(benzene-1,3,5-triyl-tris(ethyne-2,1-diyl)) tribenzoate (BTE) and biphenyl-4,4′-dicarboxylate (BPDC) linkers, MOF-210 has shown an H2 storage capacity of 8.6 wt.% at 77 K and 70 bar [26], while an even higher value of 9.95 wt% has been reported for NU-100 at 56 bar [27]. The Brunauer–Emmett–Teller (BET) specific surface areas have been reported of 6250 m2 g−1 and 6143 m2 g−1, respectively for MOF-210 and NU-100. These extremely high surface areas may be very close to the ultimate limit possible for porous structures and promote the highest gravimetric H2 storage capacity.

Fig. 1
figure 1

Excess gravimetric hydrogen adsorption capacity at 77 K versus BET specific surface area for various of metal–organic frameworks measured in different laboratories [28]

Full size image

However, the materials possessing the highest surface area tend to have large pores. This will increase the free volume in their open framework and in turn lower volumetric H2 capacity, because, in large pores, H2 is more likely to form a gas-like phase. A trade-off between volumetric and gravimetric H2 capacity has been identified, indicating that MOFs with the best gravimetric performance will generally exhibit relatively modest volumetric capacities. To be specific, high gravimetric capacity will result in a lighter tank, but moderate volumetric capacity will increase the occupied space for the tank vessel. Therefore, both factors need to be considered for technical applications.

1.4 Volumetric capacity and packing density

Since MOFs are crystalline materials, in contrast to activated carbons, their single-crystal volume can be easily determined by analyzing XRD spectra. The single-crystal density allows us to convert directly the gravimetric hydrogen capacity into the volumetric uptake of a single crystal, being equivalent to the maximum that could be reached if the tank vessel is fully occupied by a single crystal.

Based on experimental adsorption data and single-crystal densities, B. Streppel’s thesis shows one of the first correlations of volumetric versus gravimetric hydrogen storage capacity [29], as shown in Fig. 2. MFU-4 has a volumetric hydrogen uptake comparable to MOF-177, even though the gravimetric hydrogen uptake is only one-fourth. Therefore, the optimum value combining high volumetric and gravimetric capacity was identified for Cu-BTC (HKUST-1). A similar correlation between gravimetric and volumetric capacity was later illustrated and confirmed by theoretical calculations by Goldsmith et al. [30], as shown in Fig. 3. They employed high-throughput computational screening through over 20,000 MOFs within the Cambridge Structural Database (CSD) and predicted the hydrogen storage properties of these porous absorbers. The screening reveals that the relationship between gravimetric and volumetric H2 density is concave downward, denoted from ~ 4000 compounds, with maximal volumetric performance occurring for surface areas of 3100–4800 m2 g−1.

Fig. 2
figure 2

Taken from ref. [29]

Total volumetric and gravimetric hydrogen uptake for different MOFs at 2 MPa and 77 K. For MFU-4 and Mg-formate excess hydrogen uptake under these conditions is shown.

Full size image
Fig. 3
figure 3

Theoretical correlation between total volumetric and gravimetric density of stored H2 in ∼4000 MOFs screening from the CSD. For comparison, the region bounded by the dashed lines represents the DOE 2017 targets for H2 storage systems. Crossed circles represent common MOFs with incomplete or disordered crystal data in the CSD, from ref. [30]

Full size image

Therefore, H2 storage capacity in MOFs will not benefit from further improvements in the surface area alone. To increase volumetric capacity, different approaches have been developed. One strategy is to increase the available volumetric surface area by interpenetrating MOFs, denoting the intergrowth of two or more frameworks. Even though often seen as a problem in MOF synthesis, interpenetration has been found to reduce the pore volume of the frameworks and thus, the volumetric surface area may increase. R. Balderas-Xicohténcatl reported a volumetric surface area of 2697 m2 mL−1 for the interpenetrated network of the MFU-4 family, CFA-7, while the non-interpenetrated MFU-4 l was possessing only 1670 m2 mL−1. Therefore, the volumetric H2 storage capacity at 77 K increases from 25 g H2 L−1 to 50 g H2 L−1 at 20 bar in MFU-4 l and CFA-7, respectively [31].

The volumetric storage density calculated is based on the single-crystal density; however, in technical applications, the MOFs are used as powder materials, and the storage density, thus, depends on the packing density of the powder. In most cases, powders can be compressed to increase their bulk density and enhance their volumetric H2 storage capacity. One way improving bulk density is pelletization. This has been demonstrated by several groups for a number of materials that could improve the volumetric hydrogen uptake. MOF-5 powder was reported to be processed into cylindrical tablets by J. J. Purewal et al. [32]. An optimal hydrogen storage property is achieved for ρ ∼ 0.5 g/cm3, yielding a 350% increase in volumetric H2 density, reaching up to 42 g H2 L−1. A total volumetric H2 storage capacity of MIL-101 has been reported by G. Blanita et al. to reach 46.5 g H2 L−1 at a pressure of 15 MPa [33]. To enhance the volumetric capacity, R. Zacharia et al. found that compressing MOF-177, for example, can increase its volumetric excess H2 capacity by up to 80%, compared to MOF-177 loose powder. The total volumetric capacity of monoliths prepared from MOF-177 is reported to be 48.0 ± 2.1 g L−1 at 13 MPa and 77 K [34].

R. Balderas-Xicohténcatl et al. [35] developed a correlation of the inverse of the packing density plotted against SSA for a range of porous materials, as shown in Fig. 4, which has been updated by powder and monolith data reported in the literature [33, 34, 36]. The specific volume shows a linear correlation with SSA calculated for many MOFs using the single-crystal volume (blue) and the packing volume of the powder (red). For the single-crystal case, the intercept corresponds to the skeletal volume, while for powders, the intercept corresponds to the skeletal volume plus the interparticle void volume. This gap between the loose powder packing and single-crystal densities can be closed by compacting powder to form monoliths or pellets and reducing the interparticle void volume [35, 37].

Fig. 4
figure 4

Adapted from ref. [37]

Specific volume using the packing density and single-crystal density (red circles and blue triangles, respectively) as a function of the specific surface area for porous materials. Data from Bambalaza et al [36]. obtained for UiO-66 powder and pellets are included as gray squares.

Full size image

Most recently, nanoporous materials formed into monoliths offer an alternative to enhance the volumetric capacity. Recently, D. Madden et al. [38] reported on a high-throughput screening and deep analysis using a database of MOFs and identified HKUST-1 as the optimal structure for optimized hydrogen storage performance in accordance with Streppel’s experimental finding 11 years earlier [29]. After synthesis of an optimal monolithic monoHKUST-1 a volumetric storage capacity of 46 g L–1 H2 at 50 bar and 77 K and deliverable of 41 and 42 g L–1 H2 with a temperature–pressure swing 25 − 50 bar/77 K → 5 bar/160 K, respectively.

1.5 Usable capacity and operating temperature

For technical applications, the key parameter is usable or working capacity, which is the amount of hydrogen that can be delivered between the maximum tank pressure and the back pressure required by the end user [39]. So far, most of the reported hydrogen storage capacities are still given as the maximum uptake at the upper measurement pressure, whereas for methane storage in porous materials the working or usable capacity is often reported. Already in 2011, B. Streppel reported the usable capacity of hydrogen in MOF-177 which was determined by the difference between maximum and minimum uptake in dependence of the pressure for different temperatures. The maximum uptake depends on the specific surface area and the pore volume, while the minimum hydrogen uptake at low pressure is determined by the enthalpy of adsorption. Increasing the temperature during unloading of the tank was found to yield a lower amount of hydrogen adsorbed at 0.1 MPa which in turn resulted in a higher usable capacity.

In 2012, M. Schlichtenmayer and M. Hirscher studied a series of nanoporous materials and found a correlation between the average enthalpy of adsorption and the excess H2 uptake at 77 K and 20 bar, and a tendency of declining saturation uptake with increasing enthalpy was reported [40]. Another detailed analysis based on experimental data of several MOFs was later reported by M. Schlichtenmayer and M. Hirscher [39]. Depending on the heat of adsorption, an optimum operating temperature was reported, which maximized the usable capacity for a particular material, between 2 and 20 bar. Using the same analysis method, S. Glante et al. [41] recently investigated the correlation between the optimal operating temperature and usable capacity. A series of MOFs were investigated and compared to zeolite Ca–A. The optimal operating temperature for most of the MOFs was below 90 K. The reported experimental data have been summarized in Fig. 5 on the correlation of usable fraction and the optimum operating temperature for a range of porous materials. As the materials have different SSAs (and thus uptakes), the usable fraction, defined as the usable capacity at the optimum operating temperature normalized to the uptake at 77 K, has been compared to eliminate the effect of different SSAs of various materials. The higher the enthalpy of adsorption is, the higher working temperature is, however, at the expense of usable capacity. This phenomenon has also been confirmed by a computational study by Sun et al., [42] in which the maximum working capacity was predicted to decrease with increasing optimal temperatures, after a high-throughput screening of 64 state points.

Fig. 5
figure 5

The usable fraction (usable capacity at the optimum operating temperature normalized to the uptake at 77 K and 25 bar) of all materials versus their optimum operating temperature. The materials PAF-1, DUT-6, DUT-8(Cu), DUT-9, and IRMOF-1 are inserted at 77 K since their optimum temperature could not be identified within the measured temperature range [37]

Full size image

Recent work by A. Ahmed et al. [43] screened nearly half a million MOFs in a theoretical study and found an upper value of about 40 g H2 L−1 for the usable volumetric capacity. To overcome this limitation, on strategy has been recently demonstrated by J. Mason et al. for methane storage through the use of flexible frameworks, exhibiting ‘S-shaped’ or stepped adsorption isotherms (see Fig. 6), for which the amount of adsorption at low pressure is small but rises significantly above the delivery pressure and below the maximum storage capacity [44].

Fig. 6
figure 6

The usable capacity is compared for an idealized adsorbent exhibiting a classical Langmuir-type adsorption isotherm a and an ‘S-shaped’ or ‘stepped’ adsorption isotherm b, with the minimum desorption pressure Pdes and the maximum adsorption pressure Pads indicated by the vertical dashed gray lines. c, d, Total CH4 adsorption isotherms for Co(bdp) c and Fe(bdp) d at 25 °C [44]

Full size image

Owing to the lower isosteric heat of hydrogen than methane, it will be more difficult to design and tailor a flexible MOF for which H2 adsorption at RT will, for example, induce a structural phase transition to an open structure with higher gas storage capacity. J. Mota et al. reported that MIL-53(Al) exhibited flexibility during H2 adsorption, revealing an increase in usable capacity, suggesting flexible MOFs as a potential candidate [45]. Although promising initial results have been obtained, the challenge still remains to improve the usable capacity with more systematic studies.

1.6 Hydrogen storage at subcritical temperatures

The alternative way for storing massive amounts of hydrogen on nanoporous materials is by cooling it down even lower than 77 K (Cryogenic adsorption at subcritical temperatures) to make the adsorbed phase of hydrogen denser. Of course, cooling to subcritical temperatures requires much energy, but storing hydrogen at low temperatures (e.g., LH2) for long-term usage at any scale or efficient long-distance transportation is still beneficial. Thus, despite its energy-inefficient cooling processes, the fundamental interest in the microscopic nature (understanding of origin for denser adsorption) in adsorbed phase at (sub-)critical temperature has remained substantially challenging to characterize, requiring more in-depth investigation.

In 2011, B. Streppel et al. [46] reported the first fully-automated high-resolution cryo-adsorption device, which can measure hydrogen at (sub-)critical temperature of 20 K (Fig. 7). Such hydrogen isotherm measured at critical temperature enables to reveal of the direct textural properties such as specific surface area, total pore volume, and full saturation uptake. In particular, for the case of non-flat high surface area porous materials, as the accessible surface area strongly depends on the kinetic diameter of the probe gas, the smaller size of the probe gas, such as H2 (2.89 Å) measured at 20 K, will provide a more accurate surface area than the typical probe gas of N2 (3.68 Å) measured at 77 K [46]. Please note that the inner surfaces are determined to be too small, and the outer ones are too large when the larger probe molecule is used (Fig. 8).

Fig. 7
figure 7

Schematic view (&real photo) of the fully-automated high-resolution 20 K cryo-adsorption device

Full size image
Fig. 8
figure 8

Probe molecule size-dependent surface area calculation [47]

Full size image

The pore size is another critical textural property of the nanoporous material, as the smaller pore size provides stronger host–guest interaction due to the potential overlap between pore walls. Hence, the hydrogen-filling behavior at different pore sizes exhibits a step-like adsorption isotherm at subcritical temperatures. As shown in Fig. 9, B. Streppel et al. [46] measured the 20 K isotherm of MIL-101, which possessed a trimodal pore size distribution. In this (log-scale) isotherm at 20 K, a clear step sorption isotherm is observed, implying the existence of energetically different sorption sites (or pore size). Furthermore, the successive hydrogen filling in the different pore sizes can be associated with hydrogen uptake so that this measurement at the boiling temperature of hydrogen can roughly estimate an individual hydrogen-filling amount (and, thus, volume) at each pore. Another advantage of this cryo-adsorption measurement at critical temperatures up to P/P0 ~ 1 is that a physical upper limit of hydrogen storage capacity can be determined experimentally. H. Oh et al. [48] provided experimental evidence that the high hydrogen saturation uptake in a large cavity in MOF is closely related to the complete pore filling. They investigate the physical upper limit of hydrogen uptake for compressed pellet MIL-101 samples, determined to be 9.6 wt% and 42.3 g L−1, which exceeds the 2025 U.S. DOE Target. These results implied that assessing any nanoporous materials' potential for a promising adsorbent candidate is possible.

Fig. 9
figure 9

Taken from ref. [29]

Hydrogen adsorption (filled triangles) and desorption (open triangles) of MIL-101 at a critical temperature of 20 K.

Full size image

Liquid hydrogen (LH2) is considered highly efficient for hydrogen storage, and advanced nanoporous materials could make LH2 transportation even more efficient, economical, and safe by reducing boil-off losses (evaporative losses after a short dormancy), which is a major unsolved technical issue up to date. In this regard, J. Park et al. [49] reported a novel LH2 storage/transportation approach based on physisorption at a critical temperature. Owing to the robust host–guest interaction, adsorbed hydrogen at 20 K required more energy to release the hydrogen. Hence, the hydrogen desorption started above a critical point (~ 35 K), leading to the improved operating temperature (enhanced dormancy). Moreover, the pressure profiles gradually developed into a sigmoid shape at an even higher temperature, providing the maximum temperature tolerance of phase transition. Therefore, the use of physisorption on MOF at critical temperature is noted to be an affirmative effect on long-term LH2 storage. Nevertheless, the delivery capacity must also be considered because of the volume loss by sample occupation (Fig. 10). Hence, they investigated the representative 5 MOFs (Flexible MIL-53, MOF-74 with open metal site, microporous MOF-177, hierarchically microporous MFU-4 l, micro- and mesoporous DUT-6) for their total storage capacity and correlated desorption on-set temperature. Their results indirectly revealed the denser adsorbed phase on MOFs than LH2, which implied that it possibly compensated for the space occupied by the adsorbent in storage vessels (up to 96%, compared to LH2 capacity) and increased operating temperature (Fig. 10). Although their report well demonstrated a potentially viable and dependable option for use in LH2 transportation, unfortunately, it failed to provide solid evidence of the dense adsorbed phase on MOF at subcritical temperatures.

Fig. 10
figure 10

Taken from ref. [47]

20 K H2 Physisorption on Metal–Organic Frameworks with Enhanced Dormancy Compared to Liquid Hydrogen Storage.

Full size image

Indeed, the formation of the super-dense hydrogen layer at a boiling temperature of 20 K is of fundamental interest. Hence recently, R Balderas-Xicohténcatl et al. [50] experimentally discovered the microscopic nature of the formation of a super-dense monolayer of hydrogen on highly ordered mesoporous silica (KIT-6) at critical temperatures. They reported that unusually high monolayer capacities surprisingly exceeded the density of LH2 by a factor of almost three, which was confirmed by a high-resolution cryo-adsorption device, in-situ inelastic neutron scattering, and first-principles calculations. At the boiling temperature, adsorbed hydrogen on well-ordered mesoporous silica forms a two-dimensional super-dense monolayer rather than a bilayer, resulting in the molecules squeezing closely together on the surface layer. This exceptionally high layer density can be also explained as a consequence of the high compressibility of the surface layer due to the absence of core electrons, resulting in the easy compression of hydrogen. Thus, the first layer sorption energy substantially exceeds the intermolecular interaction, leading to three times more compressed than uncompressed LH2 density (Fig. 11).

Fig. 11
figure 11

Taken from ref. [50]

Formation mechanism of a super-dense hydrogen monolayer on mesoporous silica.

Full size image

To sum up, the use of this super-dense monolayer capacity on the highly porous adsorbents can be exploited as a potentially viable option for volumetric-high-density hydrogen transportation and may also open new possibilities for large-scale energy storage by cryogenics.

1.7 Summary

The past 25 years of research in novel hydrogen storage materials has been dominated by the discovery of new materials with well-defined nanostructures. For hydrogen storage by physisorption on nanoporous adsorbents, firstly, long-range ordered carbon structures created new hopes and initiated wide attention and activity in this field. Unfortunately, the initial prospects could not be confirmed and the excess hydrogen storage capacity based on physisorption is limited to 2 wt% at 77 K, owing to the low BET area of SWNTs of about 1000 m2 g−1. Shortly after, a new class of porous and crystalline framework materials emerged as adsorbents. These metal–organic frameworks (MOFs) conceivably show extremely high internal surface areas, ranging from 3000 to 7000 m2 g−1, which surpass the BET areas of the best activated carbons (3300 m2 g−1) and boost the excess hydrogen storage capacity to more than 9 wt% at 77 K. After a race to higher and higher surface areas, now the focus moved to balance the gravimetric and volumetric hydrogen uptake and further optimize the usable capacity. The development of MOF monoliths led to total volumetric capacities of over 40 g L−1 at 77 K and similar values for the usable capacity by pressure–temperature swing operation (100 bar/77 K–5 bar/160 K). A very recent development is the use of adsorbents at subcritical temperatures of hydrogen. The first results show a reduced boil-off, meaning more extended dormancy, compared to liquid hydrogen storage. Furthermore, fundamental investigations on subcritical hydrogen adsorption in mesoporous silica show the formation of a super-dense hydrogen monolayer, which may open a new route to high-density hydrogen storage/transportation by physisorption.