Ammonia Synthesis via Chemical Looping Using Nano-Confined Lithium Hydride in Alloy Matrix

Abstract

Recently, the kinetic improvement of the nitrogenation reaction of lithium hydride (LiH) to form lithium imide (Li₂NH) by adding a scaffold was reported. The scaffold prevents agglomeration of Li₂NH and maintains the activity of LiH, achieving a reduction in reaction temperature and an increase in reaction rate. In this work, a Li–Si alloy, Li₂₂Si₅, was used as a starting material to form nano-sized LiH dispersed in a Li alloy matrix. Lithium nitride (Li₃N) is generated by the reaction between Li₂₂Si₅ and N₂ to form Li₇Si₃, and then Li₃N is converted to LiH with ammonia (NH₃) generation during heat treatment under H₂ flow conditions. Since Li₃N is formed at the nano-scale on the surface of alloy particles, LiH generated from the above nano-Li₃N is also nano-scale. The differential scanning calorimetry results indicate that direct nitrogenation of LiH in the alloy matrix occurred from around 280 °C, which is much lower than that of the LiH powder itself. Such a highly active state might be achieved due to the nano-crystalline LiH confined by the Li alloy as a self-transformed scaffold. From the above experimental results, the nano-confined LiH in the alloy matrix was recognized as a potential NH₃ synthesis technique based on the LiH-Li₂NH type chemical looping process.

1. Introduction

To utilize unstable and distributed renewable resources, efficient and cost-effective techniques for energy storage and transportation are being strongly demanded. H₂ is one of the most promising energy carriers for long-term storage, with its high gravimetric energy density and many advantages. On the other hand, the unfavorable properties of H₂ itself, such as low volumetric energy density, high flammability, and harsh liquification conditions, are still challenging. Therefore, various kinds of H₂ carriers, including H₂ storage alloys and metal hydrides, are being investigated [1,2,3,4,5]. Ammonia (NH₃) is an outstanding H₂ and/or energy carrier because of its high volumetric and gravimetric H₂/energy density and mild liquification conditions [6,7].

For the establishment of a system that converts renewable energies into NH₃ onsite, small-scale and distributed types of NH₃ synthesis techniques are required. The Haber–Bosch process is conventionally used for industrial NH₃ mass production, in which NH₃ is synthesized under high pressure (10–35 MPa) and temperature (400–600 °C) [8,9]. Considering the utilization of renewable energies, this method is not suitable for downsizing because it lowers the scale merits. Thus, the novel NH₃ synthesis techniques adapted with unstable renewable energy sources, which can be operated under lower temperature and pressure conditions, are required. The dissociation of stable nitrogen (N₂) molecules is considered the bottleneck in the NH₃ synthesis reaction, and the key to the development of innovative techniques is how to efficiently dissociate N₂ under mild conditions. One approach is the improvement of conventional catalysts by adding promoters to enhance N₂ dissociation activity [10,11,12,13,14,15,16]. Ru is well known as an active metal because of its high N₂ dissociation properties. Ru-based catalysts exhibit high activity even at low temperatures and have become a trend in catalyst development [17,18,19,20,21,22,23]. On the other hand, since Ru is an expensive precious metal, research on methods using less costly elements continues [24,25,26,27,28,29,30,31,32,33,34]. Li is a promising element because it can dissociate N₂ even at room temperature. Various Li-based NH₃ synthesis methods, including electrosynthesis [35,36,37,38], light-driven methods [39,40,41], liquid alloy catalysts [42,43], and heterogeneous catalysts [44,45,46,47,48,49,50], were reported. Many reports focus on chemical looping processes, in which N₂ and H₂ interact with Li-based materials at different reaction steps. The NH₃ production process is not limited by the equilibrium of the direct NH₃ formation reaction, different from conventional catalytic synthesis; therefore, the chemical looping processes are expected to achieve efficient NH₃ synthesis even under moderate conditions.

A chemical looping process based on lithium hydride (LiH) can generate NH₃ below 500 °C under 0.1 MPa [51]. In this process, LiH reacts with N₂ to form lithium imide (Li₂NH), and then NH₃ is produced by heating Li₂NH under a H₂ flow with the re-formation of LiH. Both nitrogenation and hydrogenation in the LiH system, expressed by reactions (1) and (2), are exothermic reactions with an enthalpy change of ΔH = −79 kJ/mol H₂ and −13 kJ/mol NH₃, respectively [52]. It was also reported that the nitrogenation temperature and reaction speed of LiH can be drastically improved by mixing Li₂O as a scaffold, which prevents the agglomeration of reaction intermediates. Similar scaffold effects were observed for the chemical looping process of sodium hydride [33]. Namely, controlling the reaction environment and sample geometry using scaffolds is effective for improving the reaction properties.

$$4\mathrm{L}\mathrm{i}\mathrm{H}+{\mathrm{N}}_2\to 2{\mathrm{L}\mathrm{i}}_2\mathrm{N}\mathrm{H}+{\mathrm{H}}_2$$

(1)

$${\mathrm{L}\mathrm{i}}_2\mathrm{N}\mathrm{H}+{2\mathrm{H}}_2\to {\mathrm{N}\mathrm{H}}_3+2\mathrm{L}\mathrm{i}\mathrm{H}$$

(2)

The chemical looping NH₃ synthesis using Li and other group 14 element alloys is operated below 500 °C under 0.1 MPa [45]. This process consists of three steps: nitrogenation, NH₃ formation, and initial alloy regeneration, as in the following reactions. Li₂₂Si₅ reacts with N₂ to form Li₃N and Li₇Si₃, and then Li₃N reacts with H₂ and is converted to NH₃ and LiH. At the third step, the initial alloy phase, Li₂₂Si₅, is regenerated by releasing H₂.

$${\displaystyle \frac{18}{31}}{\mathrm{L}\mathrm{i}}_{22}{\mathrm{S}\mathrm{i}}_5+{\mathrm{N}}_2\to 2{\mathrm{L}\mathrm{i}}_3\mathrm{N}+{\displaystyle \frac{30}{31}}{\mathrm{L}\mathrm{i}}_7{\mathrm{S}\mathrm{i}}_3$$

(3)

$${\mathrm{L}\mathrm{i}}_3\mathrm{N}+3{\mathrm{H}}_2\to 3\mathrm{L}\mathrm{i}\mathrm{H}+{\mathrm{N}\mathrm{H}}_3$$

(4)

$${\displaystyle \frac{30}{31}}{\mathrm{L}\mathrm{i}}_7{\mathrm{S}\mathrm{i}}_3+6\mathrm{L}\mathrm{i}\mathrm{H}\to {\displaystyle \frac{18}{31}}{\mathrm{L}\mathrm{i}}_{22}{\mathrm{S}\mathrm{i}}_5+3{\mathrm{H}}_2$$

(5)

In the process of N₂ dissociation by Li alloys, Li₃N was formed with a nano-size at around the particle surface [48,49], and therefore LiH could be formed in nano-scale during the reaction (4). Therefore, we expected that this nano-sized LiH formed in the alloy matrix to possess high reactivity with N₂, lowering the reaction temperature for the exothermic process. In this case, the host Li alloys would act as the self-scaffold. Since there are no reports about direct nitrogenation of the products after the NH₃ generation step for the Li–S system, we investigated the chemical looping process based on Li₂₂Si₅ without a regeneration step to know whether nano-scaled LiH can be formed as speculated above or not in this work. Furthermore, the reactivity of the generated LiH was examined. As a result, the nano-confined LiH in the Li alloy matrix showed high reactivity toward N₂ dissociation, and the reaction temperature was much lower than that of LiH powder. Thus, this system is recognized as a potential small-scale NH₃ synthesis technique operated below 300 °C under 0.1 MPa.

2. Materials and Methods

Li (99%) and Si (99.999%) were purchased from Sigma-Aldrich Co. LLC (Saint Louis, MO, USA) and used as raw materials, where surface oxide layers of Li granules were roughly removed. A Li alloy, Li₂₂Si₅, was prepared by heating a stoichiometric mixture of Li and Si at 500 °C for 10 h in an air-tight stainless-steel reactor under argon (Ar) atmosphere.

Powder X-ray diffraction (XRD) (Rigaku Holdings Corporation, Tokyo, Japan, SmartLab, Cu, Kα radiation) was used for phase identification of the synthesized alloy and the solid products after the reactions. The hand-milled samples were placed on a glass plate and covered with a polyimide sheet (Du Pont-Toray, Tokyo, Japan, Kapton) to prevent oxidation during measurements. Differential scanning calorimetry (DSC) (TA Instruments, Tokyo, Japan, Q10 PDSC), placed in an Ar-filled glovebox, was used for the thermal analysis of the nitrogenation reaction of the samples. A total of 2–3 mg of sample was placed into an aluminum sample pan (TA Instruments,, Tokyo, Japan, Tzero Lid) and placed in the heating chamber. Before starting DSC measurements, the gas inside the sample chamber was purged with pure N₂ and Ar several times, and the measurements were performed in a closed system under atmospheric pressure.

For the NH₃ synthesis, the sample powder was loaded into a home-made stainless-steel reactor with an inner diameter of 8 mm and held by stainless-steel filters (SMC Corporation, Tokyo, Japan). Weight was placed on top of the upper filter to prevent the moving and carrying away of the samples by the gas flow, as shown in Figure 1. The samples were heated under N₂ (>99.9999%) or H₂ (>99.9999%) gas flow from the bottom side of the sample. The gas flow rate was fixed to 50 cm³/min by mass flow controllers, and the total pressure was kept at 0.1 MPa. A total of 40.8 mg of Li₂₂Si₅ was nitrogenated at 500 °C for 10 h under N₂ flow. After cooling down to room temperature, the flow gas was switched to H₂, and the sample was heated at 300 °C for 15 h. The above experimental procedure was repeated 2 times. After the 1st cycle, the sample was taken out of the reactor. A part of the sample was used for the XRD and DSC measurements, and the remaining sample was used for the 2nd cycle without any treatment. Generated gases were analyzed using a quadrupole mass spectrometer (MS) (MKS Inc., Andover, MA, USA, Microvision 2) connected to the downstream pipeline via about 4 m of capillary tube. All the downstream side pipelines, including the capillary, were kept at 90 °C to prevent adsorption of NH₃ on the inner surface of pipelines [33,53]. All the sample handling was performed in a glovebox filled with Ar (>99.9999%) equipped with a circulation gas purifier system (Miwa Manufacturing Co., Ltd., Osaka, Japan, MDB-2BL) to avoid any undesired reaction such as oxidation.

3. Results and Discussion

Figure 2 shows the MS spectra of the 1st and 2nd cycles of NH₃ synthesis using Li₂₂Si₅, where m/z = 2, 17, and 28 correspond to H₂, NH₃, and N₂, respectively; other related signals are shown in Figure S1. Figure 3 shows the XRD patterns before and after each reaction step. In the 1st cycle, NH₃ was produced at around 250 °C during the H₂ flow experiment following nitrogenation, as reported in a previous paper [45]. This result implies that Li₂₂Si₅ reacted with N₂ to form Li₇Si₃ and Li₃N in the nitrogenation step, and NH₃ was produced from the reaction between Li₃N and H₂, forming LiH. This is also suggested by the phase change from Li₂₂Si₅ to Li₇Si₃ after the nitrogenation shown in Figure 3, in which a broad peak in the range of 15–35° is a diffraction by a SiO₂ amorphous glass plate. Here, the XRD pattern shows peaks assigned to Li₇Si₃; however, this phase does not produce any NH₃ under the H₂ flow conditions (see Figure S2). Namely, there are other products, even if they are undetectable in the XRD measurements. The existence of Li₃N and LiH was expected based on the experimental facts, which were the reaction temperature of NH₃ generation, thermal properties obtained by DSC shown later, and the results reported before [45,49]. The crystalline size of the formed Li₃N and LiH is probably nano-scale, where the formation of Li₃N with a nano-size by the reaction between the Li alloys and N₂ was clarified in previous works [48,49]. Different from the previous work, the formation of Li₁₂Si₇ by partial hydrogenation of Li₇Si₃ during the NH₃ generation step was not observed in this experiment, likely due to differences in experimental conditions. The sample amount used for the home-made gas flow system, as shown in Figure 1 in this work, is ten times larger than that used for the thermogravimetry system in previous work, leading to a difference in reaction progress due to the influence of temperature, route, and heat capacity (heat release) of the flow gas.

In the 2nd cycle, H₂ was detected at around 250 °C in the nitrogenation step, and NH₃ was produced again in the H₂ flow process, indicating that the products after the 1st NH₃ generation, which might be composed of Li₇Si₃ and nano-LiH, successfully react under N₂ flow, releasing H₂. The main phase after the 2nd nitrogenation is also Li₇Si₃. Here, the ternary nitride phase, Li₅N₃Si, was observed as a byproduct, which did not contribute to the NH₃ generation and decomposed under H₂ flow below 300 °C (Figures S3 and S4). Although Li₅N₃Si is recognized as a stable impurity without any influence on the main reactions, the Li₅N₃Si formed by consuming reactive Li in the sample possibly affects the durability in long cycles. The optimization of reaction conditions to suppress the formation of Li₅N₃Si phase and further investigation of the detailed reaction pathway would be required for improvement of the system efficiency. Although quantitative analyses were difficult in this study due to experimental limitations, it was reported that the NH₃ generation rates are about 100 and 700 μmol g⁻¹ h⁻¹ for LiH and palladium-doped LiH, respectively [44]. Considering the temperature range, it is expected that the NH₃ formation rate would be similar to that of LiH itself.

There are two possible reaction pathways in the 2nd nitrogenation step, which are direct nitrogenation of LiH and continuous reactions of the Li₂₂Si₅ regeneration and its nitrogenation. In the former case, LiH reacts with N₂ directly by the exothermic process to form Li₂NH, and NH₃ can be produced via the hydrogenation of Li₂NH, as shown in reactions (1) and (2), as reported before [44,51]. In another case, Li₇Si₃ reacts with LiH to form Li₂₂Si₅ with H₂ release by the endothermic process following reaction (5), and then Li₂₂Si₅ successively reacts with N₂ to produce Li₇Si₃ and Li₃N by the exothermic process following reaction (3). The total reaction is described in reaction (6), and it is an endothermic process. The standard heat of formation ($H_{\mathrm{f}}^0$) of Li₇Si₃, Li₂₂Si₅, LiH, and Li₃N is about −250, −640, −91, and −160 kJ/mol, respectively [54,55,56,57,58,59,60,61]. Thus, the ΔH of reactions (5), (3), and (6) is estimated to be around 416, −202, and 214 kJ, respectively. If the nitrogenation process of the cycled samples could be dominated by the regeneration process (reaction (5)), an external heat supply is required. Therefore, the direct nitrogenation of LiH to form Li₂NH (reaction (1)) is thermodynamically preferable; however, more than 400 °C is required for activation of this process, considering previous work using the powder state samples.

$$6\mathrm{L}\mathrm{i}\mathrm{H}+{\mathrm{N}}_2\to 2{\mathrm{L}\mathrm{i}}_3\mathrm{N}+3{\mathrm{H}}_2$$

(6)

To determine the reaction process during the reaction with N₂ of the as-synthesized and cycled samples, the DSC experiments were carried out under N₂ and Ar atmospheres. As shown in Figure 4, the synthesized Li₂₂Si₅ showed clear exothermic peaks below 200 °C and around 350 °C, and a small endothermic peak appeared around 450 °C under N₂ atmosphere (solid black line). A small sharp endothermic peak around 180 °C and small broad peaks above 300 °C were observed under an Ar atmosphere (dashed black line). These peaks obtained from the DSC results under an Ar atmosphere would correspond to the melting of the remaining Li metal used for the sample synthesis and decomposition of tiny impurity alloy phases, such as Li₁₃Si₄. The nitrogenation reaction of Li₂₂Si₅ is an exothermic and weight-increasing reaction, and these exothermic peaks obtained by DSC are consistent with the successive weight change in thermogravimetric analysis conducted in a previous report [45]. For the sample after one cycle, the heat absorption due to the regeneration process by reaction (5) proceeded at a higher temperature region than 350 °C under Ar. On the other hand, the gradual heat release was observed in the temperature range of 280–400 °C under N₂, and this exothermic process probably corresponds to the nitrogenation of LiH expressed by reaction (1). At higher temperatures than 450 °C, thermal reactions were observed. This would be caused by the regeneration of the initial alloy from the kinetically remaining LiH with Li₇Si₃ (reaction (5)) and the immediate nitrogenation of the alloy to form Li₃N (reaction (3)). In the previous experimental studies, it was clarified that the regeneration and nitrogenation reaction of Li₂₂Si₅ proceeds below 380 °C and around 200 °C, respectively [45]. Thus, more than 380 °C would be necessary for the regeneration process, and the temperature range was consistent with that observed in this work. The above results revealed that the nitrogenation of LiH was prior to the regeneration reaction of the Li₂₂Si₅ phase. The exothermic reaction proceeded from around 280 °C under N₂ and might be assigned to be the nitrogenation of LiH; however, the nitrogenation of LiH powders with a 50–100 μm particle size requires around 400 °C, as reported before [52]. From detailed analyses of the nitrogenation process of alloys conducted in a previous study [49], it was revealed that Li₃N generated from the alloys possessed a nano-crystalline size, which was characterized by transmission electron microscopy (TEM) with energy dispersive X-ray spectroscopy (EDS) techniques. Thus, it is expected that LiH generated from nano-Li₃N on the alloy surface during the NH₃ generation process would also be nano-sized. In addition to the scaffold effects reported before [52], the surface of LiH nanoparticles may be quite active, in which the nitrogenation reaction at lower temperatures than that of LiH powder would be achieved. The nano-confinement may affect the electronic states on the particle surface, resulting in a decrease in activation energy for the nitrogenation reaction. Further accurate analyses of the chemical state of Li in LiH and Li₃N by in situ Fourier transform infrared spectroscopy (FT-IR) and nuclear magnetic resonance (NMR) are required to understand the detailed reaction mechanism.

From the nitrogenation and hydrogenation process of Li₂₂Si₅, “nano-confined LiH” might be generated in the Li alloy matrix, as shown in the schematic image in Figure 5. In fact, the Li₃N and LiH phases were not observed in the XRD analyses, as shown in Figure 3, while the existence of these materials is indicated from the MS and DSC results, suggesting that these phases may be nano-sized. Likely due to the small crystalline size and highly dispersed state, it is speculated that the nano-LiH is quite reactive, resulting in the progress of the exothermic reaction with N₂ to form Li₂NH at a lower temperature than that of LiH powder. The H₂ desorption observed from around 280 °C during the 2nd nitrogenation step in the MS profile would originate from the nitrogenation of the nano-confined LiH. In addition, similar to the Li₂O case, the agglomeration of Li₂NH would be suppressed by the Li₇Si₃ matrix as expected. In fact, this phenomenon can be explained by the absence of peaks corresponding to Li₂NH in the XRD pattern after the 2nd nitrogenation [52].

From 22. Si₅. The reactivity of LiH is enhanced due to the nano-size and the scaffold effect of Li₇Si₃, resulting in a lower reaction temperature than 300 °C. Here, it is expected that the LiH particles are possibly agglomerated during long-term utilization. However, for the proposed nano-confined LiH supported by the alloy, re-activation (re-formation of nano-confined state) would be easy via the regeneration of Li₂₂Si₅ at high temperatures under flow conditions of inert gases, as reported in previous works, and its nitrogenation.

4. Conclusions

In this work, kinetic improvement for the nitrogenation of LiH using Li alloy as a starting material was investigated. Li₂₂Si₅ was heated up to 500 °C under a N₂ flow to form Li₃N, and then the products were heated to 300 °C under a H₂ flow. After that, the same reaction procedure was conducted without the alloy regeneration step. The formation of LiH on the Li alloy surface was suggested based on the NH₃ generation, its temperature, and comparison with previous studies. The DSC measurement under a N₂ and Ar atmosphere revealed that direct nitrogenation of LiH to form Li₂NH would occur prior to the regeneration and nitrogenation of Li₂₂Si₅ in the 2nd nitrogenation process. Namely, the nitrogenation temperature of LiH was clearly reduced to around 280 °C, which is lower than the nitrogenation temperature (400 °C) of the typical LiH powder of a 50–100 μm particle size. The extraction of Li from Li₂₂Si₅ during the nitrogenation leads to the formation of nano-Li₃N on the particle surface, and it would result in the formation of LiH at the nano-scale in the Li alloy matrix during the NH₃ generation process. This “nano-confined” state of LiH and the contribution of Li alloy as a self-transformed scaffold would realize significant improvement in kinetics and enable LiH to dissociate N₂ at such a low temperature. Although further investigation of the reaction mechanism and state of the LiH particles is required, the proposed chemical looping process by the nano-confined LiH in the alloy matrix is recognized as the potential NH₃ synthesis technique, which can be operated below 300 °C under 0.1 MPa. In future works, further research on the investigation/improvement of the reaction rate, cyclic properties, NH₃ synthesis behaviors with unstable H₂ supply from renewables, and design of a practical system including gas purification is required and will be performed.