De-hydrogenation/Rehydrogenation Properties and Reaction Mechanism of A m Zn(NH 2 ) n -2 n LiH Systems (A = Li, K, Na, and Rb)

: With the aim to ﬁnd suitable hydrogen storage materials for stationary and mobile applications, multi-cation amide-based systems have attracted considerable attention, due to their unique hydrogenation kinetics. In this work, A m Zn(NH 2 ) n (with A = Li, K, Na, and Rb) were synthesized via an ammonothermal method. The synthesized phases were mixed via ball milling with LiH to form the systems A m Zn(NH 2 ) n -2 n LiH (with m = 2, 4 and n = 4, 6), as well as Na 2 Zn(NH 2 ) 4 · 0.5NH 3 -8LiH. The hydrogen storage properties of the obtained materials were investigated via a combination of calorimetric, spectroscopic, and diffraction methods. As a result of the performed analyses, Rb 2 Zn(NH 2 ) 4 -8LiH appears as the most appealing system. This composite, after de-hydrogenation, can be fully rehydrogenated within 30 s at a temperature between 190 ◦ C and 200 ◦ C under a pressure of 50 bar of hydrogen.


Introduction
Hydrogen is a renewable and environmentally friendly energy carrier, regarded as a potential candidate for supporting the epochal transition from fossil to renewable energy sources. Compared to common fossil fuels such as methane, diesel, etc., hydrogen possesses a higher gravimetric energy density [1]. However, under ambient conditions, its volumetric energy density is lower than fossil fuels. In order to increase this value, compression up to 750 bar, and liquefaction at cryogenic temperatures are commonly used methods. An additional method that enables achieving extremely high volumetric energy density is to store hydrogen in a solid-state using metal hydride and/or complex metal hydride-based systems. Compared to compression and liquefaction, solid-state storage allows operating the storage system at lower pressures (e.g., ≤100 bar) and without losses of hydrogen over time [2]. Due to their availability and high gravimetric and volumetric storage densities, metal hydrides such as magnesium and magnesium-based compounds were extensively explored as prospective hydrogen storage materials prior to the 2000 s [3][4][5][6]. However, in the present century, the scientific community's focus has been shifting to the development of sophisticated complex metal hydride-based hydrogen storage systems. Among the several classes of these hydrides that are potentially suitable for hydrogen storage applications, in the last decade, amide-based systems have received increasing attention due to their high gravimetric and volumetric hydrogen capacities [7][8][9][10][11][12]. The first example of a fully reversible amide-based system for hydrogen storage applications was reported in 2002 by Chen et al. [13]. In this work, it was demonstrated that the system LiNH 2 -2LiH could store reversibly about 9.5 wt.% of hydrogen.
Zinc possesses chemical properties similar to those of nickel and titanium, which is known to be an effective additive for these systems, therefore the possibility to synthesize ternary amides containing alkali metals and zinc might be considered a good strategy to develop novel amide-based hydrogen storage systems with improved thermodynamic and kinetic properties. In this regard, previous works [47][48][49] on the systems Li 4 Zn(NH 2 ) 6 -12LiH and K 2 Zn(NH 2 ) 4 -8LiH showed extremely promising results, especially for the latter. In fact, the desorbed state of the system K 2 Zn(NH 2 ) 4 -8LiH can be fully rehydrogenated at ca. 230 • C within 30 s under a hydrogen pressure of 50 bar, whereas under the same conditions the de-hydrogenated state of LiNH 2 -2LiH can absorb only a small amount of H 2 . It must be mentioned that these excellent absorption properties are maintained upon cycling.
Differential thermal analyses (DTA), and mass spectroscopy (MS) analyses were carried out using a Netzsch STA 409 C instrument, coupled with a Hiden Analytical HAL 201 mass spectrometer. The samples were investigated upon heating in the temperature range of 20-500 • C, using a heating rate of 5 • C/min. Each measurement was carried out under an argon flow of 50 mL/min. For the MS characterization of the gaseous species, the evolution of H 2 (m/z = 2) and NH 3 (m/z = 17) was monitored. Considering that the purity of hydrogen is a key requirement for PEM fuel cells and that even the presence of a few ppm of NH 3 has a detrimental effect on their performance, [52,53] this characterization aims to understand if upon heating, only hydrogen is released from the investigated materials or if traces of ammonia are also present. It must be noted, however, that the release of extremely small amounts of NH 3 along with H 2 could be easily sequestered using targeted hydrides and/or ammonia adsorbents.
De-/rehydrogenation experiments were performed using a Sievert's type apparatus (Hera, QC J4G 1A1, Canada). In this machine, the values of desorbed and absorbed hydrogen are evaluated based on the differential pressure method, i.e., two holders of equal volume, one containing the investigated sample and the other empty, are connected by a differential pressure gauge. Upon heating and cooling, the temperature-induced pressure changes on the two sides of the gauge compensate for each other, whereas pressure differences due to desorption/absorption processes are recorded. The hydrogen absorption experiments were carried out in dynamic temperature conditions, heating the specimens from room temperature (RT) up to the final temperature of 300 • C using a heating rate of 3 • C/min, under a hydrogen pressure of 50 bar. The hydrogen desorption experiments were also carried out in dynamic temperature conditions, with the specimens heated from RT to 400 • C at a rate of 3 • C/min under static vacuum. Isothermal de-/rehydrogenation experiments were performed using the same Sievert's type apparatus at temperatures of 192 • C and 222 • C, in static vacuum (starting pressure of 10 −2 bar) and 50 bar of hydrogen, respectively. The decisions to perform the de-hydrogenation measurement under static vacuum conditions and the hydrogenation reaction at 50 bar were taken to increase the likelihood of NH 3 release and to match the hydrogen pressure that is provided by a state-of-the-art commercial electrolyzer.
In-situ Synchrotron Radiation Powder X-ray Diffraction (in-situ SR-PXD) experiments were performed at the PETRA III Synchrotron facility at DESY beamline P.02.1, Germany. The X-ray wavelength (λ) utilized was fixed at 0.20745 Å and a Perkin Elmer plate image detector (2048 × 2048 pixels, each of size 200 × 200 µm 2 ) was used to acquire the 2D patterns, with a sample-to-detector distance of 1403 mm. Within an argon-filled glovebox GP (Campus) from Jacomex (H 2 O and O 2 levels below 1 ppm), the samples were loaded into sapphire capillaries and then placed into a custom-designed sample cell, which allows controlling the temperatures and applying gases. SR-PXD experiments of the dehydrogenation processes were performed by heating the specimens from RT to 350 • C, with a heating rate of 3 • C/min, under a static vacuum environment (starting pressure of 10 −2 bar). SR-PXD experiments on the rehydrogenation processes were performed by heating the samples from RT to 300 • C with a heating rate of 10 • C/min, under an atmosphere of 50 bar of H 2 . The software FIT2D was used to convert the acquired 2D images into 1D files.
Ex-situ PXD experiments were performed with a Bruker D8 Discover diffractometer from Bruker AXS GmbH, which was equipped with a copper X-ray source (Cu K radiation, λ = 1.54184 Å) and a VANTEC-500 area detector. The diffraction patterns were acquired in steps of 15 • (2θ) with an exposure time per step of 400 • s under rotation conditions (1 rotation/min). In order to avoid detrimental oxidation phenomena, prior to performing the PXD investigation, the investigated specimens were loaded under an argon atmosphere into an especially designed polymethyl methacrylate (PMMA) airtight sample holder.
Although a comprehensive examination of the microstructural features of the researched materials is theoretically achievable using Rietveld refinement or the Debye Scherrer equation, this paper focuses on the refinement of only a few diffraction patterns made with the MAUD software [54][55][56].

Results and Discussion
As previously stated, the operating temperature and purity of the released hydrogen are crucial factors to consider when choosing the right material for a storage system that will work in conjunction with a PEM fuel cell (even if only as an additive). Aiming at investigating the thermal stability of the prepared materials and the purity of the released hydrogen, the coupled DTA-MS analyses of the ball-milled LH, LZH, NZH, KZH, and RZH were carried out, and the results are summarized in Figure 1. The DTA de-hydrogenation trace of LH shows the presence of a single endothermic signal having an onset temperature of about 160 • C and a maximum peak temperature of about 250 • C. The associated MS analysis indicates that during the decomposition of LH only hydrogen is released. In contrast to the LH system, the DTA traces of the other materials indicate the occurrence of the decomposition over several steps. For the LZH system, the first small and exothermic signal is visible, having an onset temperature of 70 • C, and it is accompanied by the release of a small amount of hydrogen. This event is followed by a broad endothermic signal, with an onset temperature of about 160 • C. For this signal, it is difficult to assign a unique value for the peak maximum, as it appears to be composed of overlapping endothermic events having peak maxima at about 210 • C, 240 • C, 280 • C, and 315 • C, respectively. Associated with these events, the release of pure hydrogen is observed. The DTA trace of the NZH system shows the presence of a broad endothermic signal having a temperature of about 160 • C, which can be separated into three different signals having peak maxima of about 205 • C, 240 • C, and 260 • C, respectively. This main event is followed by an additional endothermic signal having a peak maximum of 365 • C. The MS analysis of the gaseous phases released during the decomposition of NZH indicates that the release of hydrogen occurs in three main events having signal maxima of 205 • C, 365 • C, and 425 • C. Interestingly, this hydrogen desorption is accompanied by the release of small quantities of NH 3 in the range between 160 • C and 240 • C. The amount of released NH 3 increases significantly above 240 • C, reaching a maximum at 290 • C, before decreasing monotonically. The DTA trace of KZH shows the presence of two main endothermic signals. In addition, a small endothermic signal precedes the second main event. The first signal has an onset temperature of 160 • C and a peak maximum of 240 • C, whereas the second signal has an onset temperature of about 305 • C and a peak maximum of 365 • C. Both signals are associated with the release of hydrogen only. The DTA trace of the system RZH shows the presence of four main endothermic signals, having onset temperatures of 160 • C, m, 290 • C, and 350 • C, whereas the peak maxima are 240 • C, 270 • C, 315 • C, and 430 • C, respectively. The MS traces measured for the RZH system indicate that the system decomposes releasing only hydrogen. It must be noticed that the release of hydrogen associated with the endothermic signal at 350 • C is relatively small if compared with the DTA signal. This DTA peak is most likely connected with the sublimation of a fraction of the sample. In order to shed further light on the hydrogen storage properties of the investigated systems, the kinetics of the first de-hydrogenation/rehydrogenation cycle was characterized by means of volumetric technique. The results of this investigation are displayed in Figure 2. Upon heating, the LH system starts to release hydrogen at about 160 • C, however, significant quantities of hydrogen are released only at temperatures higher than 200 • C. The overall de-hydrogenation process occurs in a single step and the final amount of released hydrogen at 390 • C is 4.2 wt.%. As previously observed in the coupled DTA-MS analyses, the hydrogen release from the system LZH starts at around 100 • C and increases considerably at temperatures higher than 180 • C, reaching a value of 5.3 wt.% at 390 • C. The system NZH, similarly to LH, starts to desorb hydrogen at around 160 • C. As observed in the DTA analysis of Figure 1, the decomposition reaction appears to occur in two separate steps. The measured amount of released hydrogen at 390 • C (i.e., 4.8 wt.%) is slightly overestimated, due to the simultaneous release of NH 3 . The de-hydrogenation reaction of KZH starts at about 160 • C and, through a three steps kinetics, reaches a final amount of released hydrogen of 3.9 wt.% at 390 • C. The de-hydrogenation reaction of the system RZH starts at about 160 • C and, in two-step kinetics, reaches a maximum amount of desorbed hydrogen equal to 3.2 wt.% at 390 • C. The de-hydrogenated materials were subsequently rehydrogenated under dynamic conditions in the temperature range between RT and 300 • C. Upon heating, the LZH system starts to absorb hydrogen already at 140 • C, reaching a final H 2 amount equal to 5.2 wt.%, while the LH system starts to absorb hydrogen at about 180 • C, and the maximum amount is equal to 4.0 wt.% at 300 • C. The NZH system starts to absorb hydrogen at 180 • C and, in two steps, reaches a final amount equal to 3.8 wt.%. The hydrogen absorption of the KZH system starts at about 140 • C. As already reported in ref. [48], the hydrogen absorption of this system is characterized by the presence of an extremely fast absorption step occurring at about 230 • C. In this step, within less than a minute, more than 50% of the overall hydrogen is absorbed. A final absorption step between 230 • C and 300 • C resulted in an overall capacity of 3.0 wt.%. The RZH system starts to absorb small quantities of hydrogen already at temperatures lower than 100 • C. A sensible increase of the hydrogenation rate is observed at about 140 • C, leading to a total hydrogen capacity of about 0.6 wt.% at 190 • C. Interestingly, this system, like KZH, shows an extremely fast absorption step. At 195 • C, the material absorbs more than half of the total capacity (i.e., 1.4 wt.%) in less than a minute. A final absorption step at temperatures ranging from 210 to 300 • C results in a final hydrogen storage capacity of 2.6 wt.%.
Among the investigated materials, the KZH and RZH systems are the only ones that show an extremely fast hydrogen uptake at a well-defined temperature. Considering that the properties and reaction mechanism of KZH were already partially studied in a previous publication [47], the focus of the present work will be on the RZH system. The results obtained by the volumetric investigation of the first three de-hydrogenation/rehydrogenation cycles of the system RZH are reported in Figure 3. The second and third de-hydrogenation/ rehydrogenation cycles present similar features. The de-hydrogenation occurs always in two consecutive steps that lead to a final amount of released hydrogen of about 2.8 wt.%. The de-hydrogenation onset temperature for the first step shifts (to 180 • C and 225 • C for the second and third cycles, respectively). The second and third rehydrogenation cycles trace out the first one in the range of temperatures between RT and 190 • C. However, the fast hydrogenation step, although leading to the same amount of stored hydrogen, is shifted to higher temperatures, i.e., about 210 • C, and appears to result from two different hydrogenation events, closely following one another. As observed also for the de-hydrogenation measurement, the final hydrogen storage capacity of the material remains about 2.6 wt.%. Figure 3. Volumetric analysis of the first three de-/rehydrogenation cycles of RZH carried out under static vacuum (initial pressure value 10 −2 bar) and 50 bar of H 2 , respectively. The material was heated from RT up to 400 • C for the de-hydrogenation measurement, whereas for the rehydrogenation measurement was heated from RT up to 300 • C. The used heating rate was 3 • C/min, in both cases.
The de-hydrogenation reaction of the RZH system was further characterized by in-situ SR-PXD measurements. The results of this investigation are shown in Figure 4. The ball-milled RZH powders are composed of LiNH 2 , LiH, and Rb 2 ZnH 4 . Upon heating at ca. 310 • C, the diffraction peaks of Rb 3 ZnH 5 appear along with those of Li 2 NH and Li 0.105 Zn 0.985 ; finally, at 340 • C, only the signals of Li 2 NH and Li 0.105 Zn 0.985 are visible. Unexpectedly, no diffraction peaks related to Rb containing phases were observed even during the cooling period. In a previous work on the de-hydrogenation mechanism of KZH, in-situ SR-PXD showed the formation of KH [47]. Figure 3 indicates that the hydrogen storage properties and reaction mechanisms of KZH and RZH are similar [47,57]. Thus, the formation of phases such as RbH, Rb-based alloys, or Rb (depending on the applied hydrogen backpressure) during the de-hydrogenation of RZH is expected. Figure 4. In-situ SR-PXD analysis of the de-hydrogenation process of the system RZH. The measurement was carried out heating the material from RT to 350 • C with a heating rate of 3 • C/min under dynamic vacuum (pressure value 10 −2 mbar). Utilized X-ray wavelength (λ) = 0.20745 Å.
Aiming at probing this eventuality, following the experimental conditions applied for the in-situ analysis of Figure 4, a small batch of RHZ was de-hydrogenated in a Sievert's type apparatus and the product characterized by XRD ( Figure 5). The presence of diffraction peaks ascribable to Rb, RbZn 13 , and RbZnH 5 in the diffraction patterns of the de-hydrogenated sample confirms the previously mentioned supposition. In addition, the presence of Li 2 NH, ZnO, Li 2 O, and Rb 2 O is also observed. The absence of any diffraction peak related to Rb containing phases in the cooling part of the SR-PXD analysis of Figure 4 is most likely due to their displacement (segregation), with respect to the X-ray beam position. Indeed, because the capillary of the SR-PXD cell is exposed to a significant temperature gradient between the central area (the area hit by the X-ray beam) and the peripheral areas (the areas of the connection between the capillary and the cell), molten phases can move to the colder areas of the capillary. This eventuality is supported by a picture of the capillary taken after the de-hydrogenation measurement and reported in Figure A1, where it is possible to notice the presence of a phase of shiny metallic grey color (presumably metallic Rb) at the extreme left of the capillary, next to the position where the sample was located at the beginning of the measurement (a part of the sample still occupies this area).
Based on the SR-PXD analysis of Figure 4 and the above discussion, a possible reaction scheme for the de-hydrogenation process can be envisaged as depicted in Figure 6.
Due to the displacement of a portion of the sample under operating conditions ( Figure A1), the in-situ SR-PXD analysis of the rehydrogenation process was not attempted. Instead, the PXD analysis of the RZH system after the third rehydrogenation cycle was carried out and the obtained diffraction pattern was analyzed by Rietveld's method ( Figure 7). As for the as-milled material (Figure 4), the crystalline phases composing the RZH system cycled three times are LiNH 2 , LiH, and Rb 2 ZnH 4 in the amounts of 26.9 wt.%, 11.9 wt.%, and 61.2 wt.%, respectively. Thus, as suggested by the volumetric analysis of Figure 3, the result of this PXD analysis confirms that the RZH system is fully reversible.  Even though no in-situ SR-PXD analysis is available on the rehydrogenation process of the RZH system, some speculations on the reaction mechanism can be made. The hydrogenation step, starting at about 100 • C, can be attributed to the hydrogenation of the Rb formed during the de-hydrogenation process. Then, considering that after this initial hydrogenation step, the system is composed most likely of Li 2 NH, Li 0.105 Zn 0.985 , LiH, and RbH/Rb, whereas the hydrogenation products are LiNH 2 , LiH, and Rb 2 ZnH 4 , the rehydrogenation process can be divided into two sub-processes: (1) the hydrogenation of Li 2 NH to form LiNH 2 -LiH (similarly to the system LH) and (2) the hydrogenation of Li 0.105 Zn 0.985 -RbH/Rb to form Rb 2 ZnH 4 -LiH. Observing the hydrogenation kinetics of the system LH (Figure 2), it is clear that the reaction occurs in a single step, which leads to reaching the system's final hydrogen storage capacity monotonously. Thus, it is unlikely that this sub-process is the one responsible for the fast hydrogenation behavior observed in the RZH system. In the opinion here advocated, it is more likely that this peculiar behavior is due to the hydrogenation of solid Li 0.105 Zn 0.985 and molten RbH/Rb (RbH melting point = 170 • C) to form Rb 2 ZnH 4 and LiH. The observation that this extremely quick phenomenon occurs in two different steps (visible in the 2nd and 3rd rehydrogenation cycles of the insert in Figure 3) might be justified by a possible inhomogeneous phase distribution in the cycled RZH system. Based on these assumptions, a possible reaction scheme for the rehydrogenation process can be envisaged as depicted in Figure 8. The extremely fast kinetics of the formation of ternary transition metal hydrides, such as Rb 2 ZnH 4 , combined with their high degree of reversibility, makes them appear as extremely interesting compounds for hydrogen storage purposes. Unfortunately, their hydrogen storage capacity is insufficient to qualify them for use as hydrogen storage materials. For instance, despite the fact that their gravimetric capacity is comparable to that of compounds with a high potential for commercial use (e.g., TiFe [58,59], Ti 0.95 Zr 0.05 Mn 1.46 V 0.45 Fe 0.09 [60], etc.), their operational temperature is too high. However, because of the high catalytic activity that transition metal-based elements have on metal amide-based systems [53], A m Zn(NH 2 ) n could play an important role in the development of novel hydrogen storage materials with enhanced kinetic properties. In this regard, the beneficial effect of an A m Zn(NH 2 ) n compound on the hydrogen storage properties of the amide system 2LiH-Mg(NH 2 ) 2 has recently been demonstrated, and the mechanism of action has been partially explained [54]. One might object that the operating temperatures of the species responsible for the fast hydrogenation kinetics of the system studied in this work are still too high for the expected range of operation of a hydride-based system, which is supposed to work in combination with a PEM fuel cell. However, when it comes to the use of additive compounds similar to Rb 2 ZnH 4 for improving the hydrogen storage properties of amide-based systems, ZrCoH 3 is one of the most effective additives ever reported for the system 2LiNH 2 +1.1MgH 2 +0.1LiBH 4 [55][56][57][58][59] In fact, by adding 3 wt.% ZrCoH 3 , once desorbed, it is possible to store about 5.3 wt.% of hydrogen in 10 min at 150 • C and a hydrogen atmosphere of 70 bar, whereas at the same temperature, it is possible to desorb 3.8 wt.% of hydrogen in 60 min under 1 bar of hydrogen pressure. Under these operating conditions, the chemical environments of both Zr and Co atoms, as well as the crystal properties of ZrCoH 3 , appear to be unaltered during de-/rehydrogenation processes [56].

Conclusions
Ternary amides, containing alkali and transition metals, appear to be appealing compounds for the development of hydrogen storage systems. In previous works and in this manuscript, several different systems were synthesized and their hydrogen storage properties were investigated. Among them, K 2 Zn(NH 2 ) 4 -8LiH, K 2 Mn(NH 2 ) 4 -8LiH, and Rb 2 Zn(NH 2 ) 4 -8LiH display unique rehydrogenation features. In fact, the desorption products of these systems can reabsorb most of the released hydrogen within a short time frame. At 230 • C and 50 bar of hydrogen pressure, K 2 Zn(NH 2 ) 4 -8LiH fully rehydrogenates in 30 s, with a maximum hydrogenation rate of 6 wt.%/min. At 230 • C and 50 bar of hydrogen pressure, K 2 Mn(NH 2 ) 4 -8LiH fully rehydrogenates in 30 s, with a maximum hydrogenation rate of 3 wt.%/min. In this manuscript, it has been demonstrated that under a hydrogen pressure of 50 bar, Rb 2 Zn(NH 2 ) 4 -8LiH can reabsorb the majority of the released hydrogen between 190 and 200 • C in 30 s, reaching a maximum hydrogenation rate of 3 wt.%/min. For the other investigated systems, re-hydrogenation is slower and not fully reversible. Concerning desorption, the Rb system starts to release pure hydrogen at 160 • C and the reaction evolves in two steps. In the Na system, the hydrogen release starts at the same temperature, but some ammonia is released during the evolution of the process. Despite these unique hydrogen storage features, the use in commercial hydrogen storage applications of these systems is hampered by the fact that they do not meet the target requirements for working in combination with PEM fuel cells.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest.