Rapid Microwave Synthesis, Characterization and Reactivity of Lithium Nitride Hydride, Li₄NH

Abstract

Lithium nitride hydride, Li₄NH, was synthesised from lithium nitride and lithium hydride over minute timescales, using microwave synthesis methods in the solid state for the first time. The structure of the microwave-synthesised powders was confirmed by powder X-ray diffraction [tetragonal space group I4₁/a; a = 4.8864(1) Å, c = 9.9183(2) Å] and the nitride hydride reacts with moist air under ambient conditions to produce lithium hydroxide and subsequently lithium carbonate. Li₄NH undergoes no dehydrogenation or decomposition [under Ar_(g)] below 773 K. A tetragonal–cubic phase transition, however, occurs for the compound at ca. 770 K. The new high temperature (HT) phase adopts an anti-fluorite structure (space group Fm$\overline{3}$m; a = 4.9462(3) Å) with N³⁻ and H⁻ ions disordered on the 4a sites. Thermal treatment of Li₄NH under nitrogen yields a stoichiometric mixture of lithium nitride and lithium imide (Li₃N and Li₂NH respectively).

1. Introduction

The Li–N–H system is a promising hydrogen storage candidate, with the ability to store 11.5 wt % of H₂ reversibly [1]. This process occurs via two exothermic steps (Equations 1 and 2):

Li₃N + H₂ → Li₂NH + LiH   ΔH = −165 kJ/mol H₂

(1)

Li₂NH + H₂ → LiNH₂ + LiH  ΔH = −44.5 kJ/mol H₂

(2)

However, it has been demonstrated that the reaction pathway may be more complex than originally indicated. In-situ powder neutron diffraction (PND) [2,3] revealed the possibility of a reaction pathway involving the formation of the lithium nitride hydride, Li₄NH [4,5] from lithium nitride in addition to the hydrogenated phase, lithium imide, Li₂NH. At a low partial pressure of hydrogen, the formation of LiH appears to be suppressed, leading to the overall reaction shown in Equation (3). The dehydrogenation behaviour of Li₄NH itself, however, remains essentially unknown and Li₄NH is the only nitride hydride currently known in the Li–N–H system.

2Li₃N + H₂ → Li₄NH + Li₂NH

(3)

Further, non-stoichiometric phases can be formed at 723 K from the reaction between the hydride and imide products in Equation (3). These complex non-stoichiometric phases thus contain N³⁻, H⁻ and (NH)²⁻ anions and form a solid solution [Equation (4)] [4]:

(1−x)Li₄NH + xLi₂NH → Li_4−2xN_1−xH_1−x(NH)_x

(4)

A full understanding of the structure and reactivity of Li₄NH is thus required in order to determine its role in the Li–N–H system and the process of hydrogen uptake and release. One of the problems in developing such an understanding centres on the reliable synthesis of single phase Li₄NH. Preparation of the phase requires the solid state reaction of Li₃N and LiH at high temperature under strictly anaerobic conditions while preventing side reactions with container materials.

In this work we demonstrate how microwave synthesis of Li₄NH using both commercial multi-mode and single-mode microwave (MW) cavities can provide a solution to this problem. The result is a reproducible route for the synthesis of phase-pure Li₄NH over timescales orders of magnitude shorter than those required for conventional heating methods, which are less energy-efficient and more difficult to control. This facile synthesis approach has allowed us to produce bulk powders of Li₄NH for a subsequent comprehensive study of structure, stability and reactivity. This synthesis method may well be extrapolated successfully to other hydrogen storage materials.

2. Experimental Section

2.1. Synthesis of Li₄NH

All manipulations were performed in an N₂-filled glovebox (5 ppm O₂; 10 ppm H₂O). Lithium nitride, Li₃N (Alfa Aesar, Heysham, Lancashire, UK, 300 mesh, 99.95%; ca. 0.1 g) and lithium hydride, LiH (Sigma Aldrich, Gillingham, UK, 30 mesh, 95%) (1:1.1 molar ratio) were ground manually with an agate mortar and pestle, pressed into pellets (13 mm internal diameter, 1–2 mm thickness) for 30 min using a hand press and placed in an N₂-filled silica tube (25 cm × 8 mm × 2 mm) sealed with parafilm. The silica tube was transferred from the glove box and sealed under vacuum (10⁻² m∙bar). Reactions were conducted in either a multi-mode microwave reactor (Panasonic 4697 NN-TS53W, Panasonic UK Ltd., West Berkshire, UK, 900 Wmax. output, 2.45 GHz) or a single-mode microwave reactor (CEM Discover^®, CEM corporation, Matthews, NC, USA, 300 Wmax. output, 2.45 GHz). The resultant products were collected in the glovebox.

2.2. Characterization

Powder X-ray diffraction (PXD) was conducted using a Bruker D8 diffractometer (Bruker Corporation, Billerica, MA, USA, Cu Kα source) or a PANalytical X’Pert Pro MPD powder diffractometer (Cu Kα₁ source) in capillary mode. The air-sensitive samples were ground into fine powders and placed in 0.5 mm diameter sealed glass capillaries for data collection. Data were collected in the range 5° ≤ 2θ ≤ 85° using a 0.0168° 2θ step size for 1 h for phase identification or 10° ≤ 2θ ≤ 110° for 12 h for structure refinement. PXD data were indexed and refined by least squares fitting using the CELREF software package [6]. Structural refinements were conducted via the Rietveld method using the GSAS and EXPGUI packages [7,8] The scale factor, zero point and background were refined in initial cycles, A shifted Chebyschev polynomial function (background function 1 in GSAS) was employed to model the background. The unit cell parameters, peak profile parameters and atomic parameters were refined subsequently. The peak shape was modelled using the pseudo-Voigt function (profile function 2 in GSAS). Constraints were applied to the thermal parameters of the N and H atoms within both the LT- and HT-Li₄NH phases.

Simultaneous thermal analysis (thermogravimetric and differential thermal analysis; TG-DTA) was performed using a NETZSCH STA 409PC thermobalance coupled to a HIDEN HPR20 mass spectrometer (MS). Approximately 30 mg of Li₄NH was placed in an alumina pan and heated from ambient temperature to either 773 K or 873 K at 5 K∙min⁻¹ under a flow of Ar or N₂ (60 mL∙min⁻¹), respectively. The maximum temperature was held for 1 h before cooling (5 K∙min⁻¹). Simultaneously, mass spectra for nitrogen, hydrogen, ammonia and water were recorded during heating.

IR spectra were collected at room temperature (20 scans/sample, 8 cm⁻¹ resolution) using a Shimadzu FTIR 8400S instrument with a Pike MIRacle ATR sampling accessory. Raman spectra were collected at room temperature using a Horiba LabRAM HR confocal microscope system (Horiba Itd., Kyoto, Japan) with a 532 nm laser, 1200 gr∙mm⁻¹ grating and a Synapse CCD detector. A hole aperture of 50 μm and a 25 times reduced laser intensity were used in order to minimise sample decomposition.

3. Results and Discussion

3.1. Li₄NH Synthesis Using a Multimode Microwave Reactor

MW synthesis in a commercial multimode cavity (MMC) reactor offers faster processing (over times of the order of minutes), increased energy efficiency and lower cost [9] than conventional high temperature approaches. To date, MW heating experiments with solid-state hydrogen storage materials have been limited to the study of the dehydrogenation properties of a small number of alkali, alkaline-earth and transition metal hydrides and of the alkali metal borohydrides, LiBH₄, NaBH₄ and KBH₄ [10,11,12]. Nevertheless, given the difficulties in mapping the microwave field distributions in MMCs, there are some drawbacks related to this synthetic approach such as variable reaction reproducibility and poor material homogeneity.

Table 1 shows representative experiments performed to ascertain the appropriate reaction conditions for the successful synthesis of Li₄NH (samples 1–5). A schematic of the reaction set-up for the synthesis of the samples described in this section is shown in Figure 1. Several reaction parameters such as microwave power, reaction times, atmosphere and the use of graphite (G) as an external microwave susceptor were considered. It should be noted that the reaction times in Table 1 were not continuous; each reaction was stopped at regular intervals (ca. 1 min steps) to allow for cooling in an attempt to avoid overheating of the silica reaction ampoule.

Figure 1. Reaction set-up using an multimode cavity (MMC) reactor.

Figure 1. Reaction set-up using an multimode cavity (MMC) reactor.

Table 1. Summary of Li₃N + LiH reactions using an MMC reactor.

Reaction	Power/W	Time/s	Other reaction conditions	PXD results
1	900	960	Ar(g) atm	LiH, α-Li3N
2	900	240	Vacuum	LiH, α-Li3N, Li4NH, Li2O and SiO2
3	600	70	Ar(g) atm, G	LiH, α-Li3N and Li4NH
4	600	300	Vacuum, G	LiH, α-Li3N, Li4NH and Li2O
5	600	390	Vacuum, G	Li4NH, Li2O

Table 1. Summary of Li₃N + LiH reactions using an MMC reactor.

Reaction	Power/W	Time/s	Other reaction conditions	PXD results
1	900	960	Ar(g) atm	LiH, α-Li3N
2	900	240	Vacuum	LiH, α-Li3N, Li4NH, Li2O and SiO2
3	600	70	Ar(g) atm, G	LiH, α-Li3N and Li4NH
4	600	300	Vacuum, G	LiH, α-Li3N, Li4NH and Li2O
5	600	390	Vacuum, G	Li4NH, Li2O

From these results it was shown that Li₄NH could be synthesised under the conditions highlighted for sample 5. Figure 2 shows the PXD data obtained from sample 5; additional PXD of a partially reacted sample 2 was plotted for comparison. Silica peaks from the reaction vessel were also observed (marked with asterisks in the PXD pattern). For sample 5, the indexed cell parameters for Li₄NH of a = 4.893(2) Å and c = 9.936(8) Å match well to those previously reported by Marx [Tetragonal space group I4₁/a, a = 4.8918(1) Å, c = 9.9321(3) Å] [4]. Neither N–H nor O–H bands were observed in Raman spectra collected for sample 5. The use of an external microwave susceptor, i.e., graphite, in these reactions appears essential in order to drive the reaction to completion. This can be rationalised in terms of the necessity to raise the reaction temperature to a point where the reactants exhibit a loss tangent (tan δ) that is sufficient to couple effectively with microwaves [13]. In fact, it has been reported that graphite achieves a temperature of 1345 K in ca. 2 min of microwave irradiation [13]. Reactions performed under argon gas were not observed to proceed to completion, even in the presence of a graphite susceptor.

Figure 2. PXD data for the as-formed products obtained from MW irradiation in an MMC reactor (a) at 600 W for 390 s sample 5; and (b) at 900 W for 240 s sample 2. Indexed peaks correspond to the tetragonal Li₄NH phase.

Figure 2. PXD data for the as-formed products obtained from MW irradiation in an MMC reactor (a) at 600 W for 390 s sample 5; and (b) at 900 W for 240 s sample 2. Indexed peaks correspond to the tetragonal Li₄NH phase.

Despite successfully preparing high purity Li₄NH using an MMC, concerns over repeatability across experiments from uncertainties in the field distribution led us to consider the use of an single mode cavity (SMC) reactor to improve homogeneity, reproducibility and potentially further reduce reaction times. This cavity offers better control over reactions given the opportunity to accurately position the sample in a well-defined electric field [14]. Moreover, the use of an external susceptor in these reactions is avoided and therefore any risk of product contamination can be minimised.

3.2. Li₄NH Synthesis Using a Single Mode Microwave Reactor

A summary of the reactions conducted in SMC system (Figure 3) is shown in Table 2. As for the MMC syntheses described in Section 3.1, cooling intervals were introduced between irradiation periods to avoid melting of the silica reaction ampoule (i.e., melting point 2073 K). Indeed, heating at 300 W for t > 240 s led to the destruction of the SiO₂ reaction vessel. It is evident from PXD data collected for sample 8 that single phase Li₄NH could be successfully synthesized at 300 W in 180 s; no reflections from the starting materials α-Li₃N and LiH were observed (Figure 4). The final product had the appearance of a yellow/beige pellet. Previously, Li₄NH was synthesised from the reaction between Li₃N and LiH at 763 K for 6 h under Ar [4] and thus with the synthetic approach described here, reaction times could be reduced by a factor of 100 and performed without the need for an inert cover gas.

Table 2. Summary of Li₃N + LiH reactions using a SMC reactor.

Reaction	Power/W	Time/s	Other reaction conditions	PXD results
6	150	270	Vacuum	LiH, α-Li3N and Li4NH
7	250	270	Vacuum	LiH, α-Li3N and Li4NH
8	300	180	Vacuum	Li4NH

Table 2. Summary of Li₃N + LiH reactions using a SMC reactor.

Reaction	Power/W	Time/s	Other reaction conditions	PXD results
6	150	270	Vacuum	LiH, α-Li3N and Li4NH
7	250	270	Vacuum	LiH, α-Li3N and Li4NH
8	300	180	Vacuum	Li4NH

Figure 3. Reaction set-up using a SMC reactor.

Figure 3. Reaction set-up using a SMC reactor.

The MW synthesis of lithium nitride hydride is possible due to the ability of the starting materials to absorb microwave energy and convert this into heat (as reflected in the loss tangent, tan δ). The ability of Li₃N to produce heat in a microwave field may be attributed to its inherent fast ionic conductivity and semiconducting behavior [15,16]. In fact, it is well established that microwaves couple directly to charge carriers leading to extremely rapid reactions in many ionic conductors and semiconductors [17]. Conversely, LiH does not generate significant heat under a microwave field and, for example, no changes in temperature were observed when LiH was placed within SMC (400 W; 20 min) or MMC (500 W; 30 min) reactors [7,9]. In fact, in these previous studies among NaH, MgH₂, CaH₂, TiH₂, VH_0.81, ZrH₂ and LaH_2.48 only the transition metal and lanthanide hydrides showed a rapid increase in temperature, which even then only led to the desorption of a small percentage of hydrogen (< 0.5 wt %).

Figure 4. PXD data for tetragonal Li₄NH obtained in an SMC reactor at 300 W for 180 s under vacuum (sample 8).

Figure 4. PXD data for tetragonal Li₄NH obtained in an SMC reactor at 300 W for 180 s under vacuum (sample 8).

During the reactions described here (samples 1–8), a purple plasma was observed along the length of the silica reaction tube. The purple plasma was followed on most occasions by yellow/orange flashes (Additional Supplementary Information). The observation of these plasmas/flashes provides evidence for the high local temperatures achieved in the reaction vessel (i.e., Li evaporation occurs at ca. 1573 K) [18].

3.3. Thermal Stability of Li₄NH

The thermal stability of the nitride hydride was investigated by TG-DTA under flowing argon. TG-DTA of sample 8 showed no evidence of mass change and hence decomposition or dehydrogenation when the sample was heated to 773 K (Figure 5a). Moreover, it was also evident from mass spectra collected simultaneously while heating that no hydrogen or other gases were evolved over the entire m/z range (1 ≤ m/z ≤ 200) (Figure 5b). These results corroborate previous investigations conducted to 698 K under argon [4]. Indexing of the PXD pattern from sample 8 following the TG-DTA experiment (Figure 5c) yielded cell parameters for Li₄NH of a = 4.891(2) Å and c = 9.9252(8) Å. These lattice parameters are within 2σ of those obtained for this sample prior to the TG-DTA and therefore no significant changes were noted. An Li₂O impurity was noted in the post-TG-DTA diffractogram and was attributed to the presence of moisture in the Ar_(g) and/or a reaction between a small amount of Li₄NH and the alumina sample holder.

The DTA profile for sample 8 however reveals an interesting feature above 700 K with no corresponding simultaneous weight change. This endothermic peak at 770 K can thus be attributed to a structural phase transition in Li₄NH. An equivalent exothermic peak in the DTA was observed at 755.6 K on cooling, demonstrating that the phase transition is reversible (and as corroborated by PXD where the tetragonal Li₄NH is observed as discussed above).

Figure 5. (a) Thermogravimetric and differential thermal analysis (TG-DTA) profiles for sample 8 under Ar_(g). The sample was heated to 773 K at 5 K min⁻¹ held for 1 h and cooled at 5 K min⁻¹. TG, DTA and temperature curves are represented in black, blue and red respectively; (b) Mass spectra obtained under the conditions shown in (a). NH_3(g), N_2(g), H_2(g) and H₂O were monitored; and (c) PXD pattern showing the experimental data for sample 8 after heating and cooling under Ar. The open squares denote a Li₂O minor impurity phase.

Figure 5. (a) Thermogravimetric and differential thermal analysis (TG-DTA) profiles for sample 8 under Ar_(g). The sample was heated to 773 K at 5 K min⁻¹ held for 1 h and cooled at 5 K min⁻¹. TG, DTA and temperature curves are represented in black, blue and red respectively; (b) Mass spectra obtained under the conditions shown in (a). NH_3(g), N_2(g), H_2(g) and H₂O were monitored; and (c) PXD pattern showing the experimental data for sample 8 after heating and cooling under Ar. The open squares denote a Li₂O minor impurity phase.

In light of the evidence for a high temperature phase transition from the TG-DTA data, attempts were made to isolate the high-temperature (HT) Li₄NH phase by heating as-prepared Li₄NH (sample 8) to 798 K at 5 K min⁻¹ under flowing argon using a conventional furnace and quenching in liquid nitrogen. Figure 6 shows the PXD pattern of the reaction products collected after quenching (sample 9). The diffractogram comprises reflections for the tetragonal low temperature (LT) phase and Li₂O but is notable for the appearance of a small number of new peaks corresponding to an HT-phase. The reflections for the HT-Li₄NH phase could be indexed to a cube with a = 4.915(1) Å. The presence of Li₂O in the reaction products is attributed to a possible reaction between Li₄NH and the SiO₂ reaction tube. Hence, it is quite likely that the unknown peaks observed in the powder pattern for sample 9 might correspond to Li–Si–(N–O) impurities from a side reaction with the reaction ampoule. Raman spectra collected for sample 9 showed the complete absence of either N–H or O–H bands.

Figure 6. PXD pattern of sample 8 (a) after quenching Li₄NH from high temperature (to form sample 9); and (b) prior to heating.

Figure 6. PXD pattern of sample 8 (a) after quenching Li₄NH from high temperature (to form sample 9); and (b) prior to heating.

3.4. Structure Determination

Structure refinements performed against PXD data were conducted for sample 8. Selected Rietveld refinement data collected at room temperature are displayed in Table 3 and additional results are included in the Supplementary Information (Tables S1 and S2). Although a satisfactory fit for the data could be obtained using a single phase tetragonal model for LT-Li₄NH, a marked improvement in the residuals was achieved when including the second HT-phase identified by the quenching experiments (see above). Traces of silica (assumed to originate from the reaction ampoule) were also found in the diffraction profile of sample 8. The observed-calculated-difference (OCD) profile plot is shown in Figure 7.

The presence of the HT-phase in sample 8 can be rationalised by the relatively fast cooling rate from the SMC MW reaction (as compared to conventional heating), which allows some of the kinetically stable HT-Li₄NH phase to remain in the sample at room temperature.

Table 3. Selected Rietveld refinement data from the lab X-ray refinement of sample 8 at 298 K.

Empirical formula	LT-Li4NH	HT-Li4NH
Collection temperature/K	298	298
Crystal system	Tetragonal	Cubic
Space group	I41/a	Fm$\overline{3}$m
Lattice parameters/Å	a = 4.8864(1)	a = 4.9462(3)
	c = 9.9183(2)
V/Å3	236.82(1)	121.01(2)
Z	4	4
Unit cell formula weight, Mw	171.116	85.558
Density/g cm−3	1.200	1.174
Phase fraction/wt%	98.1(5)	1.8(5)
No. of observations, parameters	12,117, 35
Rwp, Rp	0.0373, 0.0273
χ2	1.962

Table 3. Selected Rietveld refinement data from the lab X-ray refinement of sample 8 at 298 K.

Empirical formula	LT-Li4NH	HT-Li4NH
Collection temperature/K	298	298
Crystal system	Tetragonal	Cubic
Space group	I41/a	Fm$\overline{3}$m
Lattice parameters/Å	a = 4.8864(1)	a = 4.9462(3)
	c = 9.9183(2)
V/Å3	236.82(1)	121.01(2)
Z	4	4
Unit cell formula weight, Mw	171.116	85.558
Density/g cm−3	1.200	1.174
Phase fraction/wt%	98.1(5)	1.8(5)
No. of observations, parameters	12,117, 35
Rwp, Rp	0.0373, 0.0273
χ2	1.962

Figure 7. Observed-calculated-difference (OCD) profile plot from the room temperature Rietveld refinement for sample 8. Observed data are shown in red, calculated data are shown in green and the difference between the two profiles is shown in pink. Black tickmarks correspond to tetragonal Li₄NH, red tickmarks correspond to cubic Li₄NH and blue tickmarks correspond to SiO₂.

Figure 7. Observed-calculated-difference (OCD) profile plot from the room temperature Rietveld refinement for sample 8. Observed data are shown in red, calculated data are shown in green and the difference between the two profiles is shown in pink. Black tickmarks correspond to tetragonal Li₄NH, red tickmarks correspond to cubic Li₄NH and blue tickmarks correspond to SiO₂.

The LT-Li₄NH phase was modeled in tetragonal space group I4₁/a with lattice parameters a = 4.8864(1) Å and c = 9.9183(2) Å (Figure 8a) and corresponds to the structure originally reported by Marx and re-determined by Niewa et al. [4,5].from PXD and PND data and single crystal X-ray diffraction data respectively. Initial attempts were made to fit HT-Li₄NH in tetragonal space groups but following the elucidation of the second phase from the quenching experiments described in Section 3.3, a better structural model was chosen. The structure of HT-Li₄NH was refined in the cubic space group Fm$\overline{3}$m with a = 4.9462(3) Å (Figure 8b) using a model based on a modified Li₂NH-type anti-fluorite structure in which N³⁻ and H⁻ equally occupy the 4a site occupied by the imido N atom in Li₂NH [19]. The structure thus corresponds to a regular Li₂(N,H) cube derived from the tetragonal LT-Li₄NH structure where a_cubic ~ a_tetragonal and a_cubic ~ c_tetragonal/2.

Figure 8. Crystal structures of (a) the LT-Li₄NH phase (space group I4₁/a), where Li atoms are represented in pink; N1/H1 are represented in orange; and N2/H2 atoms are represented in blue; and (b) the HT-Li₄NH phase (space group Fm$\overline{3}$m). Lithium atoms are represented in pink and N/H atoms are represented in blue. Polyhedra showing the coordination environment of the Li and N/H atoms are represented in yellow.

Figure 8. Crystal structures of (a) the LT-Li₄NH phase (space group I4₁/a), where Li atoms are represented in pink; N1/H1 are represented in orange; and N2/H2 atoms are represented in blue; and (b) the HT-Li₄NH phase (space group Fm$\overline{3}$m). Lithium atoms are represented in pink and N/H atoms are represented in blue. Polyhedra showing the coordination environment of the Li and N/H atoms are represented in yellow.

Atomic positions and isotropic thermal parameters for each Li₄NH phase are presented in Table 4 and Table 5. Given that PXD is not able to locate hydride accurately, the occupancies of the N³⁻ and H⁻ anions in the LT-Li₄NH phase were fixed at values reported by Marx [4]. Attempts to fix the site occupancies at alternative values representing different distributions of anions led to poorer quality fits. The LT-phase therefore exhibits anion ordering with the 4a site predominantly occupied by N³⁻ (at 95%) and the 4b site similarly almost fully occupied by H⁻. The cubic HT-Li₄NH phase displays a disordered N³⁻/H⁻ anion distribution over the 4a site. The occupancies for the N³⁻ and H⁻ atoms in the HT-Li₄NH phase were both fixed to 50% in order to achieve charge balance. Given the structural relationship between the LT- and HT-Li₄NH phases, it is not surprising that the coordination environments around the cations and anions in each phase are very similar. Whereas in the LT- structure the essentially fully ordered N³⁻ and H⁻ anions are in a distorted cubic coordination with Li⁺ (with distances ranging between 2.044(2)–2.082(2) Å and 1.949(2)–2.556(3) Å respectively), in the HT-structure fully disordered N³⁻/H⁻ anions are in a regular cubic coordination, with an Li–(N³⁻/H⁻ distance of 2.1418(1) Å (Supplementary Information; Table S1). Li atoms are tetrahedrally coordinated to N/H atoms in the HT-phase. There are also strong similarities between the anti-fluorite structures of HT-Li₄NH (Li₂N_0.5H_0.5) and Li₂NH (Li₂(NH)). The Li–N³⁻/H⁻ bond lengths are shorter than the lithium-imido Li–N distances reported by Balogh et al. [19] in Li₂NH. (2.205 Å).

Nitride hydrides are relatively rare but N³⁻/H⁻ anion ordering similar to that in the LT-Li₄NH phase has also been observed in alkaline earth metal nitride hydrides such as Ca₂NH(D) (cubic space group Fd$\overline{3}$m) [20,21], Ba₂NH(D) and Sr₂NH (both hexagonal space group R$\overline{3}$m) [22,23]. Although there are no previously reported examples of complete N³⁻/H⁻ disorder in the solid state, the anion disorder in HT-Li₄NH is paralleled by the N³⁻/F⁻ distribution in nitride fluorides such as Ba₂NF [24,25]. Further studies on deuterated LT-Li₄NH and HT-Li₄NH using powder neutron diffraction will be performed to elucidate the crystal structures more fully (i.e., determine accurate hydrogen (deuterium) occupancies and anisotropic thermal parameters).

Table 4. Atom positions and isotropic thermal parameters generated by Rietveld refinement against lab X-ray data for LT-Li₄NH (sample 8) at 298 K.

Atom	N1	H1	N2	H2	Li1
Site	4a	4a	4b	4b	16f
x	0	0	0	0	0.1959(5)
y	0.25	0.25	0.25	0.25	0.4618(4)
z	0.125	0.125	0.625	0.625	0.2794(2)
100 × Uiso/Å2	3.9(1)	3.9(1)	3.1(4)	3.1(4)	6.93(9)
Site occupancy	0.95	0.05	0.05	0.95	1.00

Table 4. Atom positions and isotropic thermal parameters generated by Rietveld refinement against lab X-ray data for LT-Li₄NH (sample 8) at 298 K.

Atom	N1	H1	N2	H2	Li1
Site	4a	4a	4b	4b	16f
x	0	0	0	0	0.1959(5)
y	0.25	0.25	0.25	0.25	0.4618(4)
z	0.125	0.125	0.625	0.625	0.2794(2)
100 × Uiso/Å2	3.9(1)	3.9(1)	3.1(4)	3.1(4)	6.93(9)
Site occupancy	0.95	0.05	0.05	0.95	1.00

Table 5. Atom positions used for Rietveld refinement against PXD data for HT-Li₄NH (in sample 8) at 298 K. (Thermal parameters were fixed for this minority phase).

Atom	Li1	N1	H1
Site	8c	4a	4a
x	0.25	0	0
y	0.25	0	0
z	0.25	0	0
100 × Uiso/Å2	2.5	2.5	2.5
Occupancy	1.00	0.50	0.50

Table 5. Atom positions used for Rietveld refinement against PXD data for HT-Li₄NH (in sample 8) at 298 K. (Thermal parameters were fixed for this minority phase).

Atom	Li1	N1	H1
Site	8c	4a	4a
x	0.25	0	0
y	0.25	0	0
z	0.25	0	0
100 × Uiso/Å2	2.5	2.5	2.5
Occupancy	1.00	0.50	0.50

3.5. Reactivity of Li₄NH with Air and Nitrogen

3.5.1. Li₄NH in Air

To determine the reactivity of Li₄NH in air, a freshly made sample was exposed to the ambient atmosphere for different times and the as-formed products were analysed by PXD (Figure 9). After 4 h of air exposure, Li₄NH had completely reacted to form crystalline phases of Li₂CO₃, LiOH·H₂O and LiOH. Prolonged (e.g., 24 h) exposure of Li₄NH to air led predominantly to Li₂CO₃ with some Li₂OH·H₂O still present. Given this experimental evidence and by analogy to Li₃N [26], the hydrolysis of the nitride hydride is likely to proceed via LiOH formation (with evolution of ammonia and hydrogen; Equation 5) followed by further hydration to obtain the monohydrated hydroxide [Equation (6)]:

Li₄NH + 4H₂O → 4LiOH + NH₃ + H₂  ΔH_298K= −793.117 kJ∙mol⁻¹ [27]

(5)

LiOH + H₂O → LiOH·H₂O   ΔH_298K= −60.757 kJ∙mol⁻¹ [16]

(6)

Although there is the possibility that both LiOH and LiOH·H₂O react with CO₂ to form lithium carbonate [Equations (7) and (8) respectively], from the calculated reaction enthalpies we can establish that the former reaction is more favourable. In fact, it was not possible to identify any LiOH in the PXD pattern obtained after 24 h [although as previously stated, this also reacts with air moisture from the air to form the monohydrated LiOH (Equation 6)].

2LiOH + CO₂ → Li₂CO₃ + H₂O   ΔH_298K = −89.487 kJ kJ∙mol⁻¹ [16]

(7)

2LiOH·H₂O + CO₂ → 3H₂O + Li₂CO₃  ΔH_298K = +32.027 kJ∙mol⁻¹ [16]

(8)

Figure 9. PXD patterns of the products after Li₄NH air exposure for (a) 24 h; and (b) 4 h.

Figure 9. PXD patterns of the products after Li₄NH air exposure for (a) 24 h; and (b) 4 h.

3.5.2. Li₄NH under N_2(gas) at Variable Temperature

The TG-DTA profile of the reaction between Li₄NH and N_2(g) on heating to 873 K and cooling (both at 5 K∙min⁻¹) (Figure 10) shows an increase in weight (with onset at ca. 573 K) that stabilises by 873 K. Overall, a weight gain of 19.2 wt % was achieved over the course of the reaction. Mass spectra taken during the analysis did not show any gas evolution (Supplementary Information, Figure S1). PXD of the post-STA products revealed the presence of α-Li₃N and cubic-Li₂NH, as the main products with no evidence of remaining nitride hydride starting material. Some evidence of LiOH and Li₂O was also found in the PXD pattern. By contrast to spectra for Li₄NH, further characterisation using Raman spectroscopy showed two characteristic bands at 3162.6 cm⁻¹ and 3227.8 cm⁻¹ (Figure 10, inset).These bands correspond to the symmetric and asymmetric N–H vibrational modes in Li₂NH [28,29].

The mass change in the TGA profile can therefore be rationalised in terms of the reaction:

3Li₄NH + N₂ → 2Li₃N + 3Li₂NH

(9)

The formation of lithium nitride and lithium imide from the nitride hydride corresponds to a theoretical gain of 21.8 wt %, which is in reasonable agreement with the experimental values obtained by thermal analysis (given also the observation of LiOH and Li₂O as noted above). By contrast to previous suggestions, TG-DTA would thus indicate that the reaction of Li₄NH requires a temperature in excess of 823 K to reach completion [4].

Figure 10. (a) TG-DTA plot of Li₄NH heated to 873 K and cooled at 5 K∙min⁻¹ under an N₂ flow; and (b) PXD pattern of the reaction products obtained after thermal treatment under nitrogen; Inset: Raman spectra of the reaction products in the range 2500–4000 cm⁻¹, showing the characteristic N-H bands from Li₂NH.

Figure 10. (a) TG-DTA plot of Li₄NH heated to 873 K and cooled at 5 K∙min⁻¹ under an N₂ flow; and (b) PXD pattern of the reaction products obtained after thermal treatment under nitrogen; Inset: Raman spectra of the reaction products in the range 2500–4000 cm⁻¹, showing the characteristic N-H bands from Li₂NH.

4. Conclusions

In summary, Li₄NH has been synthesized in multi-mode and single-mode cavity microwave reactors over unprecedented timescales. Single-mode microwave reactions demonstrate several advantages over multi-mode approaches, such as increased efficiency and higher reproducibility. This new synthetic approach can reduce reaction times by a factor of 100 compared to conventional synthesis methods. Diffraction data served to confirm the purity of the as-formed product and to provide a structural model for Li₄NH by means of Rietveld refinement. Thermal treatment under argon showed that a phase transition to a high temperature cubic anti-fluorite phase occurs at ca. 770 K. HT-Li₄NH contains disordered nitride and hydride anions. In addition, results on the reactivity of Li₄NH under air and N₂ were also shown. In the former case, the nitride hydride reacts to form hydroxides (anhydrous and monohydrated) and subsequently lithium carbonate, under ambient conditions. In the latter case, Li₄NH reacts to produce Li₃N and Li₂NH at high temperature.