Predicting 1,3,5,7-tetrakis(4-aminophenyl)adamantine based covalent-organic frameworks as hydrogen storage materials

Abstract

Four types of 1,3,5,7-tetrakis(4-aminophenyl)adamantine based covalent-organic frameworks (tapa-COFs) have been designed with diamond, ctn and bor net topologies using the methods of molecular mechanics and density function theory. Their low density (0.096–0.258 g cm⁻³), high porosity (90–96%) and large H₂ accessible surface area (5511–6810 m² g⁻¹) forecast excellent hydrogen uptake capacities. The grand canonical Monte Carlo (GCMC) simulation revealed that at 77 K tapa-COF-1 possesses the highest gravimetric hydrogen storage capacity at 49.10 wt%, while tapa-COF-3 has the highest volumetric hydrogen storage capacity at 58.66 g L⁻¹. Impressively, at 298 K, tapa-COF-1 and tapa-COF-2 possess rather high gravimetric hydrogen uptake capacities, which exceed both the U.S. Department of Energy’s goal (5.5 wt%) for onboard light-duty vehicles for 2020 and the criterion of 6 wt% for commercial use of hydrogen at room temperature. In addition, the possible schemes are also proposed to synthesize the tapa-COFs.

---

1. Introduction

The continued growth in worldwide demand for energy sources and a better environment has led to increasing pursuit of clean energy. Hydrogen has long been identified as a promising candidate due to its inherent advantages in comparison with other possible fuels.^1,2 However, the safe and effective storage of hydrogen is one of the main challenges in a hydrogen economy because of the drawbacks of traditional high pressure gas or cryogenic liquefaction storage methods.^2,3 In recent decades, material-based hydrogen storage has attracted more and more attention both experimentally and theoretically. Consequently, a great number of materials such as porous frameworks,^4–7 hydrides,^8,9 hydrogen clathrates,^10,11 and so on,^12,13 have been proposed as hydrogen storage media. The U.S. Department of Energy (DOE) has set a series of goals for commercialized on-board hydrogen storage systems.¹⁴ The goal for onboard light-duty vehicles for 2020 is a hydrogen gravimetric density of 4.5 wt% and volumetric capacity of 40 g L⁻¹ at near ambient temperature and applicable pressure less than 100 bar. Unfortunately, no existing material has met all the application requirements set by the U. S. DOE. Therefore, the search for novel materials with better hydrogen storage properties is still an urgent priority in the field of hydrogen storage.

Currently, porous covalent-organic materials (COMs) including covalent-organic frameworks (COFs),^15–18 porous aromatic frameworks (PAFs),^19,20 polymers of intrinsic microporosity (PIMs),^21,22 conjugated microporous polymers (CMPs)^23,24 and so on, have been considered as new star materials with promising prospects in the hydrogen storage field. Constructed by strong covalent bonds between light elements such as C, H, O, B, N, Si, etc.,^15,25 versatile COMs possess low density, tunable pore size, high porosity, and large specific surface area which are advantages for hydrogen storage. Therefore, many investigations have been performed to study the hydrogen storage properties of pristine COMs or their modified counterparts.^25–29 For example, Furukawa and Yaghi measured the H₂ adsorption isotherms of a series of COFs at pressures of 1–85 bar and temperatures of 77–298 K in their experiments.²⁶ Garberoglio simulated the hydrogen adsorption isotherms of several COFs with the GCMC method and found that these COFs might attain a 30% increase in hydrogen uptake compared with the analogous simulations performed for metal organic frameworks at 77 K and 298 K.²⁷ Cao et al. studied the hydrogen storage capacities of lithium doped COFs and demonstrated that the gravimetric adsorption capacities for hydrogen in Li-doped COF-105 and COF-108 can reach 6.84 wt% and 6.73 wt% at 298 K and 100 bar, meeting the requirements for commercial use set by the U. S. DOE.²⁹ Zhu et al. designed a series of PAFs and investigated their advanced applications in hydrogen storage, carbon dioxide capture, drug release and so on.^25,30 Moreover, Xiang et al. first synthesized a covalent organic polymer (COP) experimentally, and then they proposed a novel lithium-decorating approach to enhance its H₂ adsorption properties.²⁸ Surprisingly, the H₂ uptake of the lithium-modified COPs increased by 70.4% compared with the unmodified compounds. All these vivid examples indicate the potential of COMs as practical hydrogen storage media. However, none of the existing COMs can be used as practical hydrogen storage media for commercial use and hence the study on COMs as hydrogen storage materials still has a long way to go.

Probing into the structures of crystalline COMs, one can easily find that COMs are combined by ever-changing organic linkers with regular geometric shapes under various net topologies. At the same time, different structures can be formed by the same organic linkers under different net topologies. Typically, COF-102, COF-103, COF-105 have ctn topology while COF-108 has a bor topology, and, furthermore, COF-105 and COF-108 contain the same building blocks but different topology nets.¹⁷ Also, PAFs are mainly constructed by various aromatic organic linkers under different net topologies.^20,25,31,32 Relying on this feature, many novel COMs with targeted properties have been predicted by theoretical design. Huang et al. proposed a series of PAFs by replacing the silicon in zeolites with tetrakis(4-bromophenyl)methane. Marvelously, hydrogen uptake can reach up to 5.9 wt% at 100 bar and 298 K, exceeding the DOE 2017 target (4.5 wt%).²⁰ Also, Mendoza-Cortes designed 14 COFs which can adsorb large amounts of methane at 298 K and up to 300 bar.³² Similarly, Farha et al. first used computational design to predict and characterize a metal–organic framework. Then, they synthesized and tested it for its hydrogen storage properties experimentally.³³ All these investigations indicate that the design of new COMs with computational methods prior to experimental synthesis can not only save time and effort but also make the experimental synthesis goal-directed.

Motivated and inspired by these studies, four types of novel covalent-organic frameworks were designed as hydrogen storage medium based on 1,3,5,7-tetrakis(4-aminophenyl)adamantine (TAPA) and three anhydrides. Adamantane is a cycloaliphatic hydrocarbon with a diamond-like cage structure that contains six secondary carbons and four tertiary carbons. Its bulky molecular volume, rigidity and highly symmetrical tetrahedral shape render it an ideal building block for the construction of porous materials. The four hydrogen atoms attached to the tertiary carbons can be substituted by phenyls through arylation to form 1,3,5,7-tetraphenyladamantane (TPA).³⁴ Moreover, 1,3,5,7-tetrakis(4-aminophenyl)adamantine (TAPA) can be generated from the nitration of TPA with a sequence of reaction procedures.^35,36 The amino in TAPA is chemically active for polymerization to generate different kinds of function materials. However, polyimides constructed from TAPA have rarely appeared in the literature up to now. Also, there are few investigations on the hydrogen storage properties of polyimides. Therefore, four types of TAPA based polyimides were theoretically designed and their hydrogen storage properties studied in the present work. We hope our design can provide some new thinking for the experiment synthesis of novel high-capacity hydrogen storage materials.

2. Design schemes and computation details

The structures of 1,3,5,7-tetrakis(4-aminophenyl)adamantine (TAPA) and three other anhydride monomers A, B and C are depicted in Fig. 1. As can be seen in Fig. 1, TAPA possesses a tetrahedral structure and four amino groups sit at the vertices of the tetrahedral. The amino group is reactive and can react with the anhydrides A, B and C. A and B are linear monomers while C has a planar triangular shape. Four types of polyimides can be formed between TAPA and the three chemical monomers and the possible synthesis schemes are depicted in Fig. 1. Here, for convenience, we term the TAPA based polyimides as tapa-COFs and the four types of tapa-COF networks are respectively named tapa-COF-1, tapa-COF-2, tapa-COF-3 and tapa-COF-4. Here, the possible schemes to synthesize these tapa-COFs are proposed and shown in Fig. 1. As depicted in Fig. 1, tapa-COF-1 and tapa-COF-2 are possibly generated through polycondensation between A or B and TAPA while tapa-COF-3 and tapa-COF-4 can be both formed through polycondensation between TAPA and C but possess different topological structures. Similar schemes to synthesize other porous materials have been performed in experiments in previous work.^36–38 As shown in Fig. 1, tapa-COF-1 and tapa-COF-2 are designed with the topological structure of diamond (dia). For these two networks, TAPA is equivalent to the carbon atom and the linear monomers A and B are equivalent to the C–C bond in diamond. As we all know, for design frameworks with tetrahedral and triangular units, the ctn (a hypothetical C₃N₄ structure, I4̄3m space group) and bor (boracite, P4̄3m space group) topology nets are considered to be the most stable.^17,32,39 Hence, we designed tapa-COF-3 and tapa-COF-4 with the same chemical monomers with the corresponding ctn and bor net topologies. The optimized structures of the four tapa-COFs are shown in Fig. 2.

Fig. 1 The possible schemes to synthesize four 1,3,5,7-tetrakis(4-aminophenyl)adamantine (TAPA) based polyimide (tapa-COF) networks.

Fig. 2 The optimized structures of (a) tapa-COF-1, (b) tapa-COF-2, (c) tapa-COF-3 and (d) tapa-COF-4.

To construct the tapa-COF networks, first, we optimized four chemical monomers using a generalized gradient approximation within the framework of density function theory. Then we assembled the four tapa-COFs by adding irreducible representations of the ligands into the topological structures of the corresponding space groups. Based on the preliminary design, subsequently, we optimized the tapa-COF networks using the classical molecular mechanics method with a COMPASS^40,41 force field. Most parameters of the COMPASS force field were derived from ab initio results and the COMPASS force field was parameterized to predict various properties for a variety of materials including the most common organics, small inorganic molecules, and polymers in isolation and in condensed phases. During the optimized process, the conjugate gradient method was implemented to begin with the steepest descent method as the first step. To get the best geometrical structures, no space group symmetry constraints were imposed on the networks and all the bond lengths, angles, and cell parameters were optimized to get the best framework structures. The geometrical structures were optimized until the remaining atomic forces were less than 0.001 kcal (mol⁻¹ Å⁻¹) on each atom and the energy convergence criterion was chosen as 1.0 × 10⁻⁷ kcal mol⁻¹ between two steps.

On the basis of the optimized structures, the hydrogen adsorption properties of these tapa-COF networks were investigated by the method of grand canonical Monte Carlo (GCMC) simulation. The 12-6 Lennard-Jones (L-J) potential shown in eqn (1) was used to describe the van der Waals (VDW) interactions between the H₂ molecules and the networks. The potential parameters for the network atoms were from the DREIDING force field of Mayo et al.⁴² The potential parameters of the H₂ molecule were from the work of Buch,⁴³ where a united-atom model for H₂ was employed. All the potential parameters used in the present work are listed in Table 1. The cross-interaction parameters between two different atoms were treated using the Lorentz–Berthelot mixing rules shown in eqn (2) and (3).

σ_ij = (σ_ii + σ_jj)/2 (2)

Table 1 The L-J potential parameters of the H₂ and He molecules and all the network atoms used in the present work

	H2	He	C	H	O	N
σ (Å)	2.956	2.64	3.473	2.846	3.033	3.263
ε/kB (K)	36.7	10.9	47.856	7.649	48.156	38.948

---

The GCMC simulations were carried out using the code MuSic.⁴⁴ 2 × 2 × 2 simulation cells of frameworks were adopted in the simulations and periodic boundary conditions were applied in all three dimensions. The frameworks were kept rigid with frozen atoms during the simulation processes. The L-J interactions were evaluated with a spherical cut off of 12.9 Å. The number of trial moves in a typical GCMC simulation was 2 × 10⁷. The first 1 × 10⁷ steps were used for equilibration and the subsequent 1 × 10⁷ steps were used for ensemble averages. Three different trial moves were attempted in the GCMC simulation, namely, translation, random insertion, and deletion. In many studies, the excess adsorbed amounts were calculated from absolute adsorbed amounts to compare with the experiments. In this work, the excess adsorbed amounts (N_exc) were computed by the following eqn (4)

N_exc = N_abs − ρV_p (4)

where N_abs denotes the amount of absolute H₂ molecules adsorbed, ρ refers to the density of the hydrogen under the thermodynamics condition studied, and V_p represents the pore volume of adsorbent. The density of the hydrogen is calculated by the Peng–Robinson equation of state⁴⁵ at a given temperature and pressure. The pore volumes of these materials were estimated with the method proposed by Talu and Myers.⁴⁶ They suggested that the pore volume could be calculated by the amounts of helium molecules contained in a unit mass adsorbent (N_a) from GCMC simulation at low pressures (P) and room temperature (T₀) with eqn (5)

V_p = N_ak_BT₀/P (5)

where k_B refers to the Boltzmann constant.

3. Results and discussions

Based on geometry optimization, the final structures of the four tapa-COFs were obtained. Under the tolerance of 10⁻² Å, the optimized tapa-COF-1 and tapa-COF-2 belong to cubic lattices without more symmetry, while tapa-COF-3 retains the I4̄3m space group symmetry and tapa-COF-4 retains the P4̄3m space group symmetry. The cell and physical parameters of the four tapa-COF networks are listed in Table 2 and the structure models are shown in Fig. 2. As shown in Table 2, tapa-COF-1 and tapa-COF-2 possess almost the same cell length due to the same dia topology structures and nearly equally sized building units. On the other hand, although having the same build unit, the cell length of tapa-COF-3 is much larger than that of tapa-COF-4, which is attributed to their different space group symmetry. It can be found in Table 2 that all four tapa-COF networks possess very low densities (0.096–0.258 g cm⁻³), especially tapa-COF-1 and tapa-COF-2 whose densities are nearly the same as water. It is well known that COFs are some of the lowest density COMs. To the best of our knowledge, the lowest densities reported for three-dimensional COFs are COF-108 (0.17 g cm⁻³) and COF-105 (0.18 g cm⁻³). Here, the densities of tapa-COF-1 and tapa-COF-2 are even lower than that of COF-108 and COF-105, which shows that the tapa-COFs designed in this work can be ranked amongst the members of COMs with the lowest density.

Table 2 Unit cell parameters, chemical formula, molar mass (M), density, pore volume (V_p) and H₂ accessible surface (S) of four tapa-COF networks

Materials	a = b = c (Å)	Chemical formula	M (g mol^−1)	Density (g cm^−3)	Vp (cm^3 g^−1)	S (m^2 g^−1)
tapa-COF-1	49.35	C432H256N32O64	6918.93	0.096	10.05	6810.19
tapa-COF-2	50.34	C496H288N32O64	7719.89	0.100	9.51	6755.06
tapa-COF-3	39.74	C600H336N48O96	9753.49	0.258	3.49	5654.44
tapa-COF-4	25.91	C150H84N12O24	2438.37	0.233	3.95	5511.10

---

Surface area and pore size are two important indices determining the gas adsorption capacity of porous materials. The surface areas of the four tapa-COFs were evaluated by using a numerical Monte Carlo integration technique proposed by Frost et al.⁴⁷ It was performed by “rolling” a probe molecule with a diameter equal to the L-J σ parameter for H₂ (2.958 Å) over the network surface. The probe was inserted randomly around the surface of each network atom with a diameter equal to the LJ σ parameters one by one. The ratio of probes that did not overlap with other network atoms was used to calculate the accessible surface area. It is obvious that surface area calculated in this way is highly dependent on the probe size used for measurement and calculating the surface area with a H₂ probe provides the amount of area accessible to H₂ molecules. The calculated H₂ accessible surface areas of the four tapa-COFs are listed in Table 2. As shown in Table 2, for tapa-COF-1 and tapa-COF-2 with diamond net topologies, tapa-COF-1 exhibits a larger H₂ accessible surface area than tapa-COF-2, attributed to the different linear building units depicted in Fig. 1. For tapa-COF-3 and tapa-COF-4 with the same building units, tapa-COF-3 has a larger H₂ accessible surface area than tapa-COF-4, since the network with the ctn net topology has a larger surface area than that with the bor net topology. Nevertheless, all tapa-COFs have quite large surface areas. Up to now, the highest surface area reported for COF materials was COF-103 with a Brunauer–Emmett–Teller (BET) surface area of 4210 m² g⁻¹.¹⁷ Moreover, PAF-1 has a very high BET surface area of 5600 m² g⁻¹ and a Langmuir surface area of 7100 m² g⁻¹.³⁰ The accessible surface areas of the four tapa-COFs in Table 2 reflect that all the designed tapa-COFs here are on a par with porous materials with the highest surface area. Using the method mentioned before, the pore volumes of the four tapa-COFs were estimated by the amounts of helium molecules adsorbed at the temperature of 298 K and the pressure of 0–1 bar through eqn (5). As revealed in Table 2, the sequence for pore volume is opposite that of density, that is, tapa-COF-1 (tapa-COF-3) has a larger pore volume than tapa-COF-2 (tapa-COF-4). To gain a concrete concept of the pore size, we also calculated the percent pore volumes of the four tapa-COFs and they are 96%, 95%, 90% and 92% for tapa-COF-1 to tapa-COF-4, respectively. To our knowledge, the percent pore volumes of COF-105 (88.22%), COF-108 (88.84%) and PAF-1 (77.60%)^17,29,30 are among the highest ones reported for COMs experimentally. Obviously, the percent pore volume of the tapa-COFs means they can join the rank of porous materials with the highest porosity up to now. The high porosity and the large surface area implies an excellent hydrogen storage capacity of these tapa-COFs.

In order to verify the rationality of the GCMC simulation and the employed force field based on the L-J potential mode, the hydrogen adsorption isotherm of COF-102 was simulated at 77 K and the result was compared with the experimental data by Furukawa and Yaghi²⁶ and the simulated results by Han et al.⁴⁸ and Klontzas et al.⁴⁹ with different parameters, as depicted in Fig. 3. As displayed in Fig. 3, the trend of our simulated results is consistent with the experimental data but the hydrogen uptake is a little larger. Therefore, the force field based GCMC simulation can accurately predict the hydrogen storage capacities of the tapa-COF networks in the present work.

Fig. 3 Comparison of the simulated absolute H₂ adsorption isotherm in COF-102 at 77 K with experimental data by Furukawa et al.²⁶ and other simulated results by Han et al.⁴⁸ and Klontzas et al.⁴⁹ with different methods or parameters.

The hydrogen uptake capacities of the four tapa-COF networks at 77 K were calculated using the GCMC simulation and the results are shown in Fig. 4. As shown by the hydrogen adsorption isotherms in Fig. 4, the absolute gravimetric and volumetric H₂ uptake capacities increase gradually with increasing H₂ gas pressure from 0.1 bar to 100 bar. However, the situation is not the same for excess H₂ uptake capacities. For tapa-COF-1 and tapa-COF-2, the excess H₂ uptake capacities also increase gradually in the range of pressure studied while for tapa-COF-3 and tapa-COF-4, the excess H₂ uptake capacities reach a maximum at a moderate H₂ gas pressure. This implies that tapa-COF-1 and tapa-COF-2 possess a larger pore volume and surface area than tapa-COF-3 and tapa-COF-4, which is consistent with the results displayed in Table 2. Fig. 4(a) shows that the H₂ gravimetric adsorption capacity of tapa-COF-1 (tapa-COF-4) is higher than that of tapa-COF-2 (tapa-COF-3) at all H₂ gas pressures studied. The maximum gravimetric H₂ uptake capacities are 49.10 wt%, 46.30 wt%, 22.90 wt% and 24.80 wt% for tapa-COF-1, tapa-COF-2, tapa-COF-3 and tapa-COF-4, respectively. In addition, the excess gravimetric H₂ uptake capacities are 14.80 wt% at 100 bar for tapa-COF-1, 13.90 wt% at 100 bar for tapa-COF-2, 12.50 wt% at 50 bar for tapa-COF-3 and 12.60 wt% at 50 bar for tapa-COF-4. From Fig. 4(b), we can see that tapa-COF-1 and tapa-COF-2 possess almost the same absolute and excess volumetric H₂ uptake capacities. At the same time, tapa-COF-3 has higher absolute and excess volumetric H₂ uptake capacities than tapa-COF-4, which is contrary to the gravimetric H₂ uptake capacities. The maximum absolute (excess) volumetric H₂ uptake capacities are 46.60 g L⁻¹ (23.80 g L⁻¹ at 100 bar), 46.20 g L⁻¹ (13.60 g L⁻¹ at 40 bar), 58.66 g L⁻¹ (31.80 g L⁻¹ at 50 bar) and 57.30 g L⁻¹ (28.90 g L⁻¹ at 50 bar) for tapa-COF-1 to tapa-COF-4, respectively. On the whole, tapa-COF-1 and tapa-COF-2 have much higher gravimetric H₂ uptake capacities than tapa-COF-3 and tapa-COF-4, while tapa-COF-3 and tapa-COF-4 have higher volumetric H₂ uptake capacities than tapa-COF-1 and tapa-COF-2. Nevertheless, the simulation results show the excellent H₂ storage capacities of these tapa-COFs at 77 K.

Fig. 4 The computed absolute (abs) and excess (exc) H₂ adsorption isotherms in four tapa-COFs at 77 K. (a) Gravimetric H₂ adsorption isotherms and (b) volumetric H₂ adsorption isotherms.

Since the practical application of hydrogen should be at room temperature, we also simulated the hydrogen adsorption isotherms of the four tapa-COFs at 298 K and the results are depicted in Fig. 5. As shown by the hydrogen adsorption isotherms, both the gravimetric and volumetric H₂ uptake capacities increase linearly with rising H₂ gas pressure at room temperature. This tendency conforms to the characteristic of Henry’s linear isotherm equation.^50,51 The Henry law constant can be expressed by the slope of the linear isotherm. The Henry’s linear isotherm equation is n = KP, where n is the absolute adsorbed amount per unit weight of adsorbent (wt%), P is the adsorbate H₂ gas pressure at equilibrium (bar), and K is the Henry’s law constant (wt% per bar). The linear fits were done for the absolute gravimetric hydrogen adsorption isotherms and the fitted equations are listed in Fig. 5(a), including the respective degree of linear correlation (r). The subscripts 1 to 4 in these equations denote the corresponding equations and degrees of linear correlation for tapa-COF-1 to tapa-COF-4. From these equations, the isotherms have good linearity, which reveals that the H₂ uptake capacity of these tapa-COFs is mainly related to the pore volume and is virtually independent of binding energy or surface area at room temperature. As can be seen from Fig. 5 and Table 2, tapa-COF-1 (tapa-COF-4) does possess higher gravimetric H₂ uptake capacities than that of tapa-COF-2 (tapa-COF-3) at all H₂ pressure studied, which is consistent with the size of their pore volume. The highest absolute gravimetric (volumetric) H₂ uptake capacities of tapa-COF-1 to tapa-COF-4 are 8.06 wt% (7.64 g L⁻¹), 7.53 wt% (7.51 g L⁻¹), 3.13 wt% (8.03 g L⁻¹), and 3.49 wt% (8.05 g L⁻¹), respectively. Surprisingly, the maximum gravimetric H₂ uptake capacities for tapa-COF-1 and tapa-COF-2 exceed the U.S. Department of Energy’s goal (4.5 wt%) for 2017 and also the capacity (6 wt%) for practical application of hydrogen storage at room temperature.¹⁴ The only disadvantage is that the four tapa-COFs have low absolute volumetric H₂ uptake capacities and the capacities are almost the same at 298 K, which also indicates that the amount of hydrogen adsorbed is mainly dependent on pore volume. In order to use the tapa-COFs as practical hydrogen storage media, some modification methods such as doping metals or introducing functionalized groups^15,28,52 can be adopted to improve the volumetric H₂ uptake capacity at room temperature. For most investigations, excess H₂ uptake capacity is rarely mentioned at room temperature, since it is always very low. To make the studies more comprehensive, we also calculated the excess H₂ adsorption capacities at 298 K. As shown in Fig. 4, the maximum excess gravimetric (volumetric) H₂ uptake capacities of tapa-COF-1 to tapa-COF-4 are 0.21 wt% (0.16 g L⁻¹), 0.11 wt% (0.08 g L⁻¹), 0.39 wt% (0.95 g L⁻¹) and 0.39 wt% (0.84 g L⁻¹), respectively. Here, the excess H₂ adsorption capacities of tapa-COF-3 and tapa-COF-4 are higher than those of tapa-COF-1 and tapa-COF-2, which implies that tapa-COF-3 and tapa-COF-4 can adsorb H₂ molecules more strongly than tapa-COF-1 and tapa-COF-2.

Fig. 5 The computed absolute (abs) and excess (exc) H₂ adsorption isotherms of the four tapa-COFs at 298 K. (a) Gravimetric H₂ adsorption isotherms and (b) volumetric H₂ adsorption isotherms.

In order to study the energetics of the adsorbed H₂ molecules, we calculated the isosteric heat of adsorption (Q_st) as a function of H₂ loading from eqn (6).⁵³

Here, R is the ideal gas constant, T is the temperature, 〈ν〉 is the average potential energy of the adsorbed phase, and 〈N〉 is the average number of molecules in the adsorbed phase. Table 3 lists the isosteric heat of adsorption for H₂ of each tapa-COF at 77 K and 298 K. Also, the average isosteric heat of adsorption for H₂ of each tapa-COF at 77 K and 298 K was calculated. As shown in Table 3, the isosteric heats of adsorption for H₂ at 298 K is larger than that of 77 K for the each tapa-COF as a whole. This is related to the fact that much more H₂ is adsorbed on the same tapa-COF at 77 K than at 298 K. Hence, at 298 K the H₂ molecules are mainly adsorbed near the surface of the tapa-COFs, while more H₂ molecules fill the pore cavities far from the network surface at 77 K. Since the isosteric heat of adsorption for H₂ mainly comes from the Van der Waals interactions between the H₂ molecules and the network surface, it was smaller at 77 K than at 298 K. On the other hand, tapa-COF-3 and tapa-COF-4 possess a little larger isosteric heat of adsorption for H₂ than that of tapa-COF-1 and tapa-COF-2. This can be explained by the size of the pore volume of the networks. Since tapa-COF-1 and tapa-COF-2 have larger pore volumes than tapa-COF-3 and tapa-COF-4, the average distance between the adsorbed H₂ molecule and the network surface in tapa-COF-1 and tapa-COF-2 is longer than that in tapa-COF-3 and tapa-COF-4 and hence larger energetics is attained. Unfortunately, the isosteric heats of adsorption for H₂ in the four tapa-COFs are all rather small. A previous study has pointed out that the value of the isosteric heat of adsorption needs to be larger than 15 kJ mol⁻¹ for practical hydrogen storage application.⁵⁴ The low isosteric heats of adsorption for H₂ in COFs is a common problem. In a previous feature article, Xiang and Cao summarized the hydrogen adsorption properties of porous COFs and the isosteric heats of adsorption for H₂ in COFs lies between 3.9–9.5 kJ mol⁻¹.¹⁵ Therefore, to enhance the isosteric heats of adsorption for H₂ in the tapa-COFs, some modified methods can be adopted, which was mentioned in the previous paragraph to improve the volumetric H₂ uptake capacity at room temperature.

Table 3 The isosteric heat of H₂ adsorption (kJ mol⁻¹) in four tapa-COF networks under pressures from 0.1–100 bar and their average values at both 77 K and 298 K

Pressure (bar)	tapa-COF-1		tapa-COF-2		tapa-COF-3		tapa-COF-4
	77 K	298 K	77 K	298 K	77 K	298 K	77 K	298 K
0.1	2.88	2.67	2.16	2.48	4.15	3.47	3.43	3.29
0.5	3.05	2.94	2.89	2.90	3.48	3.43	3.21	4.06
1	3.07	3.24	3.31	2.88	3.84	3.98	3.08	2.99
2.5	2.92	3.63	2.69	2.54	3.28	3.72	3.28	3.85
5	2.63	2.47	3.68	2.40	2.26	2.54	2.56	3.58
7.5	2.55	2.61	2.87	2.89	3.22	4.32	2.46	3.95
10	1.53	2.85	2.27	3.19	2.64	3.83	3.34	4.22
20	2.16	3.31	2.83	3.74	3.06	3.16	3.00	3.33
30	2.31	2.59	1.65	2.96	3.24	3.85	2.92	4.59
40	1.80	2.80	1.89	3.60	3.15	3.46	3.39	3.96
50	1.26	2.89	2.27	2.32	2.25	3.46	3.36	4.12
60	1.01	2.53	2.38	3.81	2.89	4.29	2.85	4.33
70	2.58	3.68	2.01	3.14	3.05	3.78	2.91	4.00
80	2.11	3.44	2.06	3.02	3.66	3.96	2.58	3.47
90	2.04	2.96	1.98	3.29	3.19	3.58	2.96	4.84
100	2.17	3.47	2.22	3.01	3.56	4.19	2.69	3.77
Average	3.18	3.69	3.00	3.90	2.25	3.01	2.45	3.01

---

To understand the adsorption process of H₂ molecules in these tapa-COF frameworks, we investigated their adsorption behavior at a molecular level by detecting snapshots of the four tapa-COFs with H₂ molecules adsorbed at both 77 K and 298 K. Fig. 6 shows typical snapshots of the four tapa-COFs with adsorbed H₂ molecules at 77 K and 100 bar. By studying the adsorption behavior of the H₂ molecules in these tapa-COFs, we found that the monomers A, B and C (seen in Fig. 1) are the preferential adsorption sites for H₂ molecules, which is consistent with studies in previous work by ab initio calculations. During the process, the H₂ molecules are first adsorbed near the monomers A, B and C at low H₂ pressure. With rising H₂ pressure, the H₂ molecules begin to occupy the corner sites near the TAPA units. Finally, the H₂ molecules accommodate the pore space far from the framework surface at high H₂ pressure.

Fig. 6 Equilibrium snapshots of the four tapa-COFs with H₂ molecules adsorbed under the pressure of 100 bar at 77 K. (a) tapa-COF-1, (b) tapa-COF-2, (c) tapa-COF-3 and (d) tapa-COF-4.

4. Conclusions

In this work, four novel covalent-organic frameworks have been designed based on 1,3,5,7-tetrakis(4-aminophenyl)adamantine and proposed as hydrogen storage materials. The simulated results reveal that the tapa-COFs have high pore volumes, large H₂ accessible surface areas and very low densities. The GCMC simulations show that tapa-COF-1 has the highest gravimetric hydrogen uptake while tapa-COF-3 possesses the highest volumetric hydrogen uptake at both 77 K and 298 K. The good hydrogen uptake capacities of these tapa-COFs implies their potential as hydrogen storage materials. At the same time, possible schemes to synthesize the tapa-COFs were also proposed. Although the experimental synthesis of these tapa-COFs is still to be achieved, we expect that the ideas gained from this work may motivate inspiration for the corresponding experiments in future.

Acknowledgements

This work is supported by the National Natural Science Foundation of China (Grant No. 11447142, 11504088, 11304079 and 51201059). X.-D. Li also acknowledges financial support from the Talent Introduction Fund (No. 2013BS039) and the Plan of Nature Science Fundamental Research (No. 2013JCYJ10) in Henan University of Technology and the support from the Fundamental Research Funds for the Henan Provincial Colleges and Universities (No. 2014YWQN07, 16A140005 and 14A140027).

References

1. C.-J. Winter, Int. J. Hydrogen Energy, 2009, 34, S1–S52.
2. L. Schlapbach and A. Zuttel, Nature, 2001, 414, 353–358.
3. K. Mazloomi and C. Gomes, Renewable Sustainable Energy Rev., 2012, 16, 3024–3033.
4. V. C. Menon and S. Komarneni, J. Porous Mater., 1998, 5, 43–58.
5. S. Ma and H. C. Zhou, Chem. Commun., 2010, 46, 44–53.
6. B. Assfour and G. Seifert, Microporous Mesoporous Mater., 2010, 133, 59–65.
7. E.-Y. Chen, Y.-C. Liu, M. Zhou, L. Zhang and Q. Wang, Chem. Eng. Sci., 2012, 71, 178–184.
8. S. Orimo, Y. Nakamori, J. R. Eliseo, A. Zuttel and C. M. Jensen, Chem. Rev., 2007, 107, 4111–4132.
9. B. Sakintuna, F. Lamari-Darkrim and M. Hirscher, Int. J. Hydrogen Energy, 2007, 32, 1121–1140.
10. P. K. Chattaraj, S. Bandaru and S. Mondal, J. Phys. Chem. A, 2011, 115, 187–193.
11. V. V. Struzhkin, B. Militzer, W. L. Mao, H.-k. Mao and R. J. Hemley, Chem. Rev., 2007, 107, 4133–4151.
12. Y. Zhao, Y.-H. Kim, A. C. Dillon, M. J. Heben and S. B. Zhang, Phys. Rev. Lett., 2005, 94, 155504.
13. J. Carrete, R. C. Longo, L. J. Gallego, A. Vega and L. C. Balbás, Phys. Rev. B: Condens. Matter Mater. Phys., 2012, 85, 125435.
14. http://www.hydrogen.energy.gov/roadmaps_vision.html.
15. Z. Xiang and D. Cao, J. Mater. Chem. A, 2013, 1, 2691–2718.
16. E. Tylianakis, E. Klontzas and G. E. Froudakis, Nanoscale, 2011, 3, 856–869.
17. H. M. El-Kaderi, J. R. Hunt, J. L. Mendoza-Cortes, A. P. Cote, R. E. Taylor, M. O’Keeffe and O. M. Yaghi, Science, 2007, 316, 268–272.
18. A. P. Cote, A. I. Benin, N. W. Ockwig, M. O’Keeffe, A. J. Matzger and O. M. Yaghi, Science, 2005, 310, 1166–1170.
19. T. Ben and S. Qiu, CrystEngComm, 2013, 15, 17–26.
20. L. Huang, X. Yang and D. Cao, J. Phys. Chem. C, 2015, 119, 3260–3267.
21. N. B. McKeown and P. M. Budd, Macromolecules, 2010, 43, 5163–5176.
22. N. B. McKeown and P. M. Budd, Chem. Soc. Rev., 2006, 35, 675–683.
23. A. I. Cooper, Adv. Mater., 2009, 21, 1291–1295.
24. J. X. Jiang, F. Su, A. Trewin, C. D. Wood, H. Niu, J. T. Jones, Y. Z. Khimyak and A. I. Cooper, J. Am. Chem. Soc., 2008, 130, 7710–7720.
25. X. Zou, H. Ren and G. Zhu, Chem. Commun., 2013, 49, 3925–3936.
26. H. Furukawa and O. M. Yaghi, J. Am. Chem. Soc., 2009, 131, 8875–8883.
27. G. Garberoglio, Langmuir, 2007, 23, 12154–12158.
28. Z. Xiang, D. Cao, W. Wang, W. Yang, B. Han and J. Lu, J. Phys. Chem. C, 2012, 116, 5974–5980.
29. D. Cao, J. Lan, W. Wang and B. Smit, Angew. Chem., Int. Ed., 2009, 48, 4730–4733.
30. T. Ben, H. Ren, S. Ma, D. Cao, J. Lan, X. Jing, W. Wang, J. Xu, F. Deng, J. M. Simmons, S. Qiu and G. Zhu, Angew. Chem., Int. Ed., 2009, 48, 9457–9460.
31. S. Amirjalayer, R. Q. Snurr and R. Schmid, J. Phys. Chem. C, 2012, 116, 4921–4929.
32. J. L. Mendoza-Cortes, T. A. Pascal and W. A. Goddard, J. Phys. Chem. A, 2011, 115, 13852–13857.
33. O. K. Farha, A. O. Yazaydin, I. Eryazici, C. D. Malliakas, B. G. Hauser, M. G. Kanatzidis, S. T. Nguyen, R. Q. Snurr and J. T. Hupp, Nat. Chem., 2010, 2, 944–948.
34. I. Boldog, A. B. Lysenko, E. B. Rusanov, A. N. Chernega and K. V. Domasevitch, Acta Crystallogr., Sect. C: Cryst. Struct. Commun., 2009, 65, O248–O252.
35. V. R. Reichert and L. J. Mathias, Macromolecules, 1994, 27, 7015–7023.
36. C. Shen, Y. Bao and Z. Wang, Chem. Commun., 2013, 49, 3321–3323.
37. Y. Luo, B. Li, L. Liang and B. Tan, Chem. Commun., 2011, 47, 7704–7706.
38. J. Weber, M. Antonietti and A. Thomas, Macromolecules, 2008, 41, 2880–2885.
39. O. Delgado-Friedrichs, M. O’Keeffe and O. M. Yaghi, Acta Crystallogr., Sect. A: Found. Crystallogr., 2006, 62, 350–355.
40. H. Sun, J. Phys. Chem. B, 1998, 102, 7338–7364.
41. D. Rigby, H. Sun and B. E. Eichinger, Polym. Int., 1997, 44, 311–330.
42. S. L. Mayo, B. D. Olafson and W. A. Goddard, J. Phys. Chem., 1990, 94, 8897–8909.
43. V. Buch, J. Chem. Phys., 1994, 100, 7610–7629.
44. A. Gupta, S. Chempath, M. J. Sanborn, L. A. Clark and R. Q. Snurr, Mol. Simul., 2003, 29, 29–46.
45. D. Y. Peng and D. B. Robinson, Ind. Eng. Chem. Fundam., 1976, 15, 59–64.
46. O. Talu and A. L. Myers, AIChE J., 2001, 47, 1160–1168.
47. H. Frost, T. Duren and R. Q. Snurr, J. Phys. Chem. B, 2006, 110, 9565–9570.
48. S. S. Han, H. Furukawa, O. M. Yaghi and W. A. Goddard, J. Am. Chem. Soc., 2008, 130, 11580–11581.
49. E. Klontzas, E. Tylianakis and G. E. Froudakis, J. Phys. Chem. C, 2008, 112, 9095–9098.
50. J. Palgunadi, H. S. Kim, J. M. Lee and S. Jung, Chem. Eng. Process., 2010, 49, 192–198.
51. L. Zhou and Y. P. Zhou, Ind. Eng. Chem. Res., 1996, 35, 4166–4168.
52. J. Lan, D. Cao, W. Wang and B. Smit, ACS Nano, 2010, 4, 4225–4237.
53. R. Q. Snurr, A. T. Bell and D. N. Theodorou, J. Phys. Chem., 1993, 97, 13742–13752.
54. A. W. C. van den Berg and C. O. Arean, Chem. Commun., 2008, 668–681.

---