Exploring the capability of mayenite (12CaO·7Al2O3) as hydrogen storage material

We utilized nanoporous mayenite (12CaO·7Al2O3), a cost-effective material, in the hydride state (H−) to explore the possibility of its use for hydrogen storage and transportation. Hydrogen desorption occurs by a simple reaction of mayenite with water, and the nanocage structure transforms into a calcium aluminate hydrate. This reaction enables easy desorption of H− ions trapped in the structure, which could allow the use of this material in future portable applications. Additionally, this material is 100% recyclable because the cage structure can be recovered by heat treatment after hydrogen desorption. The presence of hydrogen molecules as H− ions was confirmed by 1H-NMR, gas chromatography, and neutron diffraction analyses. We confirmed the hydrogen state stability inside the mayenite cage by the first-principles calculations to understand the adsorption mechanism and storage capacity and to provide a key for the use of mayenite as a portable hydrogen storage material. Further, we succeeded in introducing H− directly from OH− by a simple process compared with previous studies that used long treatment durations and required careful control of humidity and oxygen gas to form O2 species before the introduction of H−.

www.nature.com/scientificreports/ (0.08988 g/L) at 1 atm. Additionally, the transportation of high-pressure gas is not widespread because of safety risks and added costs. On the other hand, in the case of cryogenic systems, the low temperature requirements of insulated containers render the process very expensive 20,25 . In particular, safe, cost-effective, and stable storage materials featuring efficient physical or chemical adsorption-desorption of hydrogen are needed for widespread applications of hydrogen, such as in portable electronics.
We propose the use of the nanocage structure of mayenite in the hydride state (H − ) for the storage and safe transportation of hydrogen. Mayenite ceramics react with water, during which the ions trapped in the structure are easily desorbed. The easy desorption of hydride ions in water can allow the use of this material in portable applications. Additionally, this material is 100% recyclable because the cage structure can be recovered by the removal of water and subsequent heat treatment (1250 ℃ in air). We confirmed the presence of hydrogen as hydride by 1 H-NMR spectroscopy, gas chromatography (GC), and neutron diffraction analyses. Further, we confirmed the hydrogen state stability inside the mayenite cage by first-principles calculations to better understand the adsorption mechanism and storage capacity and to provide a key to the development of mayenite as a hydrogen storage vehicle.

Experimental
Sample preparation. The mayenite 12CaO·7Al 2 O 3 samples were prepared by the citrate gel technique using Ca(NO 3 ) 2 ・4H 2 O (Nacalai, 99.5%), Al(NO 3 ) 3 ・9H 2 O (Nacalai, 98.9%), and citric acid (C 6 H 8 O 7 ) (Nacalai, 99%). The detailed preparation method is reported elsewhere 26 . Briefly, the citrate-nitrate was heated and stirred at 90 °C until a gel was formed and then heated for 2 h to evaporate excess water. The powder was then crushed and calcinated at 1250 °C for 3 h in air atmosphere. Thereafter, hydrogen treatment was conducted in a tubular furnace at 1250 °C for 2 h in a 100% hydrogen atmosphere. Additionally, we studied the effect of sintering time, hydrogen treatment time and temperature (for sintering at hydrogen treatment). We have reported the best conditions for hydrogen generation and omitted the details.
Structure characterization. The crystal structures of the samples were analyzed using an X-ray diffractometer (XRD; Rigaku Corporation, RINT 2500HF) operated at 50 kV and 300 mA with a scanning rate of 0.02 s −1 . The XRD analysis was carried out at room temperature, and Cu Kα radiation of 1.5406 Å wavelength was used. In addition, the diffraction angle (2θ) range was 10°-70°. The powder diffraction data were analyzed using JADE software to identify the phases present. The microstructures of the samples were analyzed using a field emission scanning electron microscope (FE-SEM; Nippon Electronics Co., Ltd., JSM-6705F) with an acceleration voltage of 3 kV. Neutron powder diffraction profiles were measured using a high-throughput diffractometer iMATERIA installed at the Japanese Particle Accelerator Research Complex (J-PARC). Rietveld refinements were performed using the program RIETAN-FP Version 2.32 27 for XRD and Z-Rietveld Version 1.0.4. 28 , and 3D visualizer VESTA was used to demonstrate the crystal structures 29 .
Cage characterization. 1 H-NMR spectroscopic measurements were performed using a Bruker AVANCE III 800 MHz US plus spectrometer equipped with a 2.5 mm MAS probe and operated at a resonance frequency of 800 MHz. Each sample was weighed to obtain quantitative results and sealed in a zirconia rotor. The MAS frequency was 30 kHz and the 1H 90 pulse length was 1.3 μs. Fully relaxed spectra were obtained with the recycle delay of 20 s. The chemical shifts were expressed as values relative to tetramethylsilane using the resonance line at 1.91 ppm for adamantane as an external reference.
Additionally, we measured ESR to verify the presence of O 2− in the cage structure.
Hydrogen desorption. The desorption of hydrogen was verified by the reaction of the mayenite sample with water as follows: the sample (0.05 g) was added to distilled water (1 ml) at 60 °C in a head space recipient for 1 h. The sample was naturally cooled to room temperature and the gas inside the recipient was then analyzed by a gas chromatograph equipped with a thermal conductivity detector (GC-TCD) (GC-8A, Shimadzu Corporation) and a molecular sieve/5A column. To determinate the activation energy from Arrhenius plot, we plotted desorption reaction temperature (RT ~ 80 °C) of dissolved mayenite in water versus the amount of hydrogen detected by the GC. The amount of hydrogen detected by the GC represented the desorbed hydride ions in the cage.
Density functional theory calculation. DFT calculations were performed using OpenMx (an opensource package for Material eXplorer) 30 . The exchange correlation energy was approximated using the generalized gradient approximation 31 . An energy cutoff of 300 Ry was employed with a 2 × 2 × 2 k-point grid in the 124 and 122 atoms unit cell for structural optimization. We used the following base functions: s4p3d3 for Ca, s3p3d2 for Al, s2p2d1 for O, and s2p1 for H. The cutoff radii were chosen as 11.0, 8.0, 6.0, and 6.0 au for Ca, Al, O, and H, respectively. The convergence criteria were set to 2.0 × 10 −4 Hartree/Bohr or 1.0 × 10 −5 Hartree for structural optimization. The structures are visualized using Materials Studio Visualizer 8.0 32 .

Results and discussion
Structure and cage characterization. Figure 1a show the presence of two peaks for the all the analyzed samples after hydrogen treatment. The peak around 6.1 ppm corresponds to H − and that at − 0.75 ppm corresponds to OH − . The assignment was carried out based on a previous study 33 .  www.nature.com/scientificreports/ It further, the NMR results suggest that a part of H − was introduced into the cage, while OH − remained on the sample despite hydrogen treatment. Assuming that 4 ions can be introduced in the free cages, we calculated the fraction amount of OH − and H − from the NMR data. The calculations were based on the integration of the peak area for H − and OH − . Based on these fractions, the amount of H − inside the cage was calculated to be 7.3 × 10 -4 mol g −1 . This corresponds to 17.9 ml of hydrogen by grams of mayenite.
ESR analysis, shown in Fig. 4 did not show any presence of ESR signals due to O 2− (g z = 2.020) in the structure. In Fig. 4, the ESR spectra of O 2− and e − are shown as reference spectra.
Additionally, to verify the hydrogen state and possible adsorption on the surface, we analyzed the sample after hydrogen treatment by thermogravimetry differential thermal analysis photoionization mass spectrometry (TG-DTA-PIMS). The DTA-TG-PIMS data were collected under He flow and the result is presented in Fig. 5. The observed temperature versus gas evolution profile of the mayenite sample hydrogen treated at 1250 °C for 2 h shows a strong evolution peak of H 2 centered at approximately 600 °C (from IC m/z = 2 band of MS). These results indicate that hydrogen was not present at the surface and all the hydrogen was stored in the cage. Additionally, Fig. 5 shows that the H 2 peak is absent for the sample before hydrogen treatment.
Crystallography. Hayashi     Extraction mechanism. Hydrogen desorption. The storage hydrogen amount was confirmed using the GC-TCD. Figure 7 shows the GC-TCD results for mayenite ceramics before and after hydrogen treatment, in pure water at 60 °C. Retention time of 0.58, 1.25, and 1.8 min corresponds to hydrogen, oxygen, and nitrogen gases, respectively. The sample before hydrogen treatment only showed the peaks corresponding to oxygen and nitrogen. Oxygen and nitrogen peaks originate due to the presence of air in the head space. The amount detected from GC results is 18.14 ml of hydrogen per gram of mayenite. These results are in good agreement with the theoretical amount (17.9 ml g −1 ) calculated from the NMR results, which is listed in Table 1.
However, these results correspond to a storage density of < 1 mass%, which is a very low energy density to be useful for real-world applications. Further studies are needed to improve the amount of H − in the cage to use mayenite as a possible hydrogen storage material.
The mayenite, 12CaO·7Al 2 O 3, samples completely decomposed in water. Therefore, the hydride species trapped inside the cage were released and generated hydrogen. This reaction was almost independent of temperature as shown by Fig. 8 (Arrhenius plot). The activation energy of hydrogen released from mayenite cage in water was calculated to be 2.6 kJ mol −1 from the slope of the line in the graph by using the following equation: This very low value of activation energy is because almost no energy is required to dissolve mayenite in water. Mayenite with H − ions dissolves in water according to the following reaction.
Ln(k) = −315.77(1/T) + 3.7981  www.nature.com/scientificreports/ The cage structure can be recovered by removal of water and then applying heat treatment (1250 ℃ in air as follows: A schematic representation of the possible cycle life of mayenite is shown in Fig. 9. After dissolution in water and hydrogen release, mayenite can be recovered completely (100%) by 2 h heat treatment in air at 1250℃. We verified the structure by XRD analysis and the hydrogen treatment was performed. The amount of hydrogen released was verified by GC-TDC. The results are showed in S1.
Storage mechanism. The mechanism of H − formation in mayenite has been discussed in a previous study 34 . Here, we briefly discuss the mechanism from the viewpoint of comparison with our experiments. After calcina-  Mainly two types of hydrogen doping mechanisms have been reported 35 . The first is the adsorption of a H 2 molecule from the gas phase onto the mayenite surface with subsequent dissociation into a pair of either H 0 or H + and H − ions. Then, the H atoms or ions diffuse into the bulk with a concentration gradient. This process involves long treatment times for the hydrogen to diffuse and dissociate 2,3,13 . However, in this study, the annealing duration in hydrogen atmosphere is very short (2 h) compared to the annealing duration (> 24 h) 2,3,13 reported in literature. Therefore, it is hard to assume this mechanism as a possible route for the H − formation in the cage.
Another proposed mechanism is the diffusion of H 2 molecules into the mayenite bulk and their participation in the chemical reaction 34 . This mechanism if more feasible in this study, assuming that H 2 rapidly diffuses into the cages of mayenite and then undergoes chemical reactions with OH − inside the cage as follows: Another possibility is that at elevated temperatures, dehydroxylation of the surface forms O 2− surface sites followed by exchange with H − ions as represented by the following equation: However, as discussed previously, O 2− was not present our samples, which was confirmed by the ESR results. To introduce O 2− in the cage, the sintering environment should be carefully controlled (generally, control of humidity in the oxygen atmosphere) 2,3,13. However, our experiments were simplified to explore the possibility of using mayenite as a hydrogen carrier in real-world applications. Thus, we sintered the samples in air without controlling the humidity and/or oxygen gas atmosphere. This was the reason for the absence of O 2− in our samples.
Density functional theory calculations. We studied the hydrogen states, viz., H + , H − , and/or H 2 , in the mayenite treated by H 2 gas using the DFT calculations. The calculated models of the crystal structures with P 1 obtained from the results of the Rietveld analysis, Ca 24 Al 28 O 64 + 4OH and Ca 24 Al 28 O 64 + 2OH + 2H, were optimized and evaluated under the constraints of lattice parameters, viz., a = b = c and α = β = γ = 90° (Figs. 10, 11). The structure defined by the Rietveld analysis using the TOF neutron diffractions were used for the initial structure model in DFT calculations, as shown in Tables S2 and S3. In this study, we performed DFT calculations for three models: Ca 24

Conclusions
In summary, we successfully demonstrated the application of mayenite (12CaO·7Al 2 O 3 ceramic) as a potential hydrogen storage material, There are no requirements of high temperatures or pressures for desorption, because mayenite has the advantage of hydrogen desorption by dissolution of mayenite in water through a reaction at a relatively low temperature (60 °C at 1 h). After the reaction with water, the cage structure of mayenite is transformed into a calcium aluminate hydrate and this transformation enables hydrogen desorption at a low temperature. The mayenite can be recovered by applying heat treatment to calcium aluminate hydrate and the subproducts generated in the reaction with water. The activation energy for hydrogen desorption in water was calculated to be 2.6 kJ mol −1 . Additionally, this material is highly stable in air and water vapor environments at low temperatures 33 , which is an advantage for its possible use as hydrogen carrier. However, the energy density is very low (less than 1 mass%) to be useful in real-world applications. There is a need to improve the amount of hydride adsorption sites in mayenite by surface treatments or other techniques.