LaNi₅-Assisted Hydrogenation of MgNi₂ in the Hybrid Structures of La_1.09Mg_1.91Ni₉D_9.5 and La_0.91Mg_2.09Ni₉D_9.4

Abstract

This work focused on the high pressure PCT and in situ neutron powder diffraction studies of the LaMg₂Ni₉-H₂ (D₂) system at pressures up to 1,000 bar. LaMg₂Ni₉ alloy was prepared by a powder metallurgy route from the LaNi₉ alloy precursor and Mg powder. Two La_3−xMg_xNi₉ samples with slightly different La/Mg ratios were studied, La_1.1Mg_1.9Ni₉ (sample 1) and La_0.9Mg_2.1Ni₉ (sample 2). In situ neutron powder diffraction studies of the La_1.09Mg_1.91Ni₉D_9.5 (1) and La_0.91Mg_2.09Ni₉D_9.4 (2) deuterides were performed at 25 bar D₂ (1) and 918 bar D₂ (2). The hydrogenation properties of the (1) and (2) are dramatically different from those for LaNi₃. The Mg-containing intermetallics reversibly form hydrides with ΔH_des = 24.0 kJ/mol_H2 and an equilibrium pressure of H₂ desorption of 18 bar at 20 °C (La_1.09Mg_1.91Ni₉). A pronounced hysteresis of H₂ absorption and desorption, ~100 bar, is observed. The studies showed that LaNi₅-assisted hydrogenation of MgNi₂ in the LaMg₂Ni₉ hybrid structure takes place. In the La_1.09Mg_1.91Ni₉D_9.5 (1) and La_0.91Mg_2.09Ni₉D_9.4 (2) (a = 5.263/5.212; c = 25.803/25.71 Å) D atoms are accommodated in both Laves and CaCu₅-type slabs. In the LaNi₅ CaCu₅-type layer, D atoms fill three types of interstices; a deformed octahedron [La₂Ni₄], and [La(Mg)₂Ni₂] and [Ni₄] tetrahedra. The overall chemical compositions can be presented as LaNi₅H_5.6/5.0 + 2*MgNi₂H_1.95/2.2 showing that the hydrogenation of the MgNi₂ slab proceeds at mild H₂/D₂ pressure of just 20 bar. A partial filling by D of the four types of the tetrahedral interstices in the MgNi₂ slab takes place, including [MgNi₃] and [Mg₂Ni₂] tetrahedra.

1. Introduction

Despite significant differences in chemistry between La and Mg, magnesium forms a very extensive solid solution in the LaNi₃ intermetallic alloy, crystallizing with a PuNi₃ type trigonal structure. Up to 67% of La atoms can be replaced by Mg to form a LaMg₂Ni₉ intermetallic compound. The LaNi₃ crystal structure is formed by a stacking of the LaNi₅ (Haucke CaCu₅ type) and MgNi₂ (Laves type) slabs along the trigonal 00z axis (LaNi₅ + 2MgNi₂ = LaMg₂Ni₉). Studies of hydrogen absorption–desorption properties of the LaMg₂Ni₉ [1,2] have shown that it forms a hydride containing up to 1.2 wt% H (~0.8 H/M; LaMg₂Ni₉H_9.6).

The building blocks of LaMg₂Ni₉—LaNi₅ and MgNi₂—are well characterized individually as hydride-forming intermetallic compounds. The thermodynamics and structural features of their interaction with hydrogen are quite different. At room temperature, LaNi₅ forms a saturated LaNi₅H_6.7 hydride and shows a reversible interaction with hydrogen at hydrogen pressures slightly exceeding atmospheric pressure. Hydrogen atoms fill tetrahedral La₂Ni₂, LaNi₃ and Ni₄ sites in the hydride crystal structure [3].

In contrast, hydrogenation of the Laves phase MgNi₂ compound is possible only at hydrogen pressures close to 30 kbar, while maintaining an interaction temperature of 300 °C. Formation of MgNi₂H₃ results in a complete rebuilding of the metal sublattice. Hydrogen atoms in the orthorhombic structure of trihydride fill two different sites, the Mg₄Ni₂ octahedra and the positions within the buckled Ni nets, consequently forming directional Ni-H bonds [4].

A gradual increase of Mg content in La_3−xMg_xNi₉ is accompanied by a linear decrease of the volumes of the unit cells. Interestingly, a substantial contraction takes place not only for the (La,Mg)₂Ni₄ slabs, but also for Mg-free CaCu₅-type LaNi₅ slabs. Hydrogen interaction with the La_3−xMg_xNi₉ alloys has been investigated by in situ synchrotron X-ray, neutron powder diffraction, theoretical modeling, electrochemical studies as metal hydride battery anode materials, rapid solidification and pressure–composition–temperature studies [1,2,5,6,7,8,9,10]. In the whole substitution range, La_3−xMg_xNi₉ alloys form intermetallic hydrides with H/M ratios ranging from 0.77 to 1.16. Magnesium influences structural features of the hydrogenation process and determines various aspects of the hydrogen interaction with intermetallics causing: (a) more than a 1,000-fold increase in the equilibrium pressures of hydrogen absorption and desorption for the Mg-rich LaMg₂Ni₉ as compared to the Mg-poor La_2.3Mg_0.7Ni₉ and a substantial modification of the thermodynamics of the formation–decomposition of the hydrides; (b) an increase of the reversible hydrogen storage capacities following increase of Mg content in the La_3−xMg_xNi₉ to ~1.5 wt% H for La₂MgNi₉; (c) improvement of the resistance against hydrogen-induced amorphisation and disproportionation and (d) change of the mechanism of the hydrogenation from anisotropic to isotropic. Thus, optimisation of the magnesium content provides different possibilities for improving properties of the studied alloys as hydrogen storage and battery electrode materials. Studies of the thermodynamics and crystal chemistry of the RE₂MgNi₉H₁₂₋₁₃ (RE = La and Nd) hydrides showed that La substitution by Pr or Nd causes destabilization of the formed hydrides without affecting their hydrogen storage capacities and leaves unchanged the most important features of their crystal structures [11].

Observed values of H capacities in the LaMg₂Ni₉-based hydride of 9.6 atoms H/f.u. cannot be explained by exclusive hydrogen insertion into the LaNi₅ slabs, and requires H incorporation into the MgNi₂ blocks of the structure to reach the experimentally observed H/M ratios. Thus, studies of the thermodynamics and crystal chemistry of La_3−xMg_xNi₉-H₂ systems are very interesting and important from the point of view of the effect of magnesium on the behaviours of the metal-hydrogen systems. The goal of the present study was to study two alloy compositions formed close to the limiting value of the magnesium solubility in LaNi₃, LaMg₂Ni₉, by performing in situ neutron powder diffraction studies of the deuterated La_0.91Mg_2.09Ni₉ and La_1.09Mg_1.91Ni₉ and by studying the thermodynamics of the metal-hydrogen interactions by measurements of the PCT diagrams.

2. Experimental

La_1.09Mg_1.91Ni₉ and La_0.91Mg_2.09Ni₉ alloys were prepared by a powder metallurgy route from LaNi₅ alloy precursor, Mg and Ni. Initial metals La, Mg and Ni with high purity exceeding 99.9% were used in the synthesis. LaNi₅ precursor was prepared by arc melting of a stoichiometric 1:5 mixture of La and Ni.

The powder mixture LaNi₅ + Mg + Ni was ball milled under protective atmosphere of argon gas in a SPEX 8000D mill for 8 h. After the milling process, the mixture was placed into a tantalum crucible and then annealed in Ar atmosphere in the sealed stainless steel containers at 600–1000 °C. Two samples with a slightly different stoichiometry were prepared. Their stoichiometric compositions were: sample 1: La_1.09(1)Mg_1.91(1)Ni₉; sample 2: La_0.91(1)Mg_2.09(1)Ni₉.

The first sample was annealed at 800 °C for 8 h and then at 600 °C for 8 h. The second sample was annealed at 1000 °C for 2 h and, later, at 800 °C for 12 h. The samples were quenched into a mixture of water and ice after the annealing. A small excess of Mg (5 wt%) was introduced into the initial mixtures to compensate for its sublimation at high temperatures.

The homogeneity of the prepared samples was characterized by XRD. Laboratory powder X-ray diffraction data were collected with a Siemens D5000 diffractometer (Oslo, Norway) equipped with a Ge primary monochromator giving Cu Kα₁ radiation. Initial phase-structural analysis was performed by X-ray powder diffraction using a Bruker D8 Advance diffractometer (Kjeller, Norway) with Cu-Kα radiation. High-resolution SR XRD data were collected at the Swiss-Norwegian Beamlines (SNBL, BM01B) at ESRF, Grenoble, France. A monochromatic beam with λ = 0.5009(1) Å was provided by a double Si monochromator. A 2θ angular range of 1°–50.5° was scanned with a detector bank consisting of six scintillation detectors mounted in series with 1.1° separation. The data were binned to the step size Δ2θ = 0.003°. The instrumental contribution to the line broadening was evaluated by refining the profile parameters for a standard Si sample.

In situ neutron powder diffraction studies were performed at HRPT diffractometer, SINQ, PSI, Switzerland using a wavelength of λ = 1.494 Å. The deuteride of sample 1 was synthesized at 25 bar D₂ and −30 °C (P_eq. for absorption ~20 bar); it was synthesized and studied by NPD using a thin walled stainless steel sample cell (6 mm OD). The deuteride of sample 2 was synthesized at 950 bar D₂ and measured at 912 bar D₂ at room temperature. The experimental setup for the in situ NPD study consisted of a high-pressure Sieverts’ manometric hydrogenator connected to a high-pressure sample cell made of a null matrix coherent scattering alloy (Zr–Ti) with a thin stainless steel inner liner.

Powder diffraction data were analysed by the Rietveld whole-profile refinement method using the General Structure Analysis System (GSAS) [12] and FULLPROF [13] software packages. Pressure-composition-temperature isotherms were measured at −40, −20, 0 and 20 °C.

3. Results and Discussion

3.1. XRD Characterization of the Initial Intermetallic Alloys La_0.91Mg_2.09Ni₉ and La_1.09Mg_1.91Ni₉

XRD characterization of two studied alloys La_0.91Mg_2.09Ni₉ and La_1.09Mg_1.91Ni₉ showed that they both contain PuNi₃ trigonal La_3−xMg_xNi₉ as the main phase constituents (80% for sample 1 and 75% for sample 2). The common secondary constituent was identified as a LaNi₅ binary intermetallic. Furthermore, sample 1 contained an admixture of the MgNi₂ Laves-type intermetallic phase, while sample 2 contained a cubic MgNi₃ intermetallic compound recently also observed during the studies of the MgNi₂-H₂ system [4]. MgNi₃ compound (sp.gr. Pm$\overline{3}$m; a = 3.7185(5) Å) has an AlCu₃-type structure and earlier it was synthesized by high-energy ball milling of a mixture of Mg and Ni metals [14]. We assume that in present study MgNi₃ was synthesised already during the reactive ball milling and remained stable during the consecutive annealing at 1,000 and 800 °C. As an example, Figure 1 shows an excellent fit of the experimental X-ray powder diffraction pattern collected for the sample 2, La_0.91Mg_2.09Ni₉.

Figure 1. XRD pattern of La_0.91Mg_2.09Ni₉ (sample 2) (Cu-Kα₁ radiation).

Figure 1. XRD pattern of La_0.91Mg_2.09Ni₉ (sample 2) (Cu-Kα₁ radiation).

Crystallographic data for the studied intermetallic samples obtained from the refinements of the XRD pattern are listed in Table 1.

Table 1. Crystal structure data for the La_0.91Mg_2.09Ni₉ and La_1.09Mg_1.91Ni₉ alloys from Rietveld refinements of the X-ray diffraction data. PuNi₃ type of structure, space group R$\overline{3}$m.

Alloy	Sample 1	Sample 2
Source of experimental data	SR XRD collected at BM01B, SNBL using a wavelength λ = 0.5009(1) Å	Siemens D5000 diffractometer, Cu Kα1 radiation
Composition of AB3 phase	La1.09(1)Mg1.91(1)Ni9	La0.91(1)Mg2.09(1)Ni9
Unit cell parameters:
a (Å)	4.94024(8)	4.8986(1)
c (Å)	23.8188(4)	23.957(1)
V (Å3)	503.44(1)	497.86(2)
Atomic parameters:
La1/Mg1 in 3a (0, 0, 0)
Uiso×100 (Å2)	0.43(5)	2.1(2)
nMg, (nLa = 1–nMg)	0.0(–)	0.09(1)
La2/Mg2 in 6c (0, 0, z)
z	0.1453(3)	0.1471(6)
Uiso×100 (Å2)	1.2(3)	0.5(3)
nMg, (nRE = 1–nMg)	0.954(5)	1.0(–)
Ni1 in 3b (0, 0, ½) Uiso × 100 (Å2)	0.7(1)	0.8(3)
Ni2 in 6c (0, 0, z)
z	0.3335(2)	0.3334(4)
Uiso×100 (Å2)	0.13(8)	1.8(3)
Ni3 in 18h (x, –x, z)
x	0.5009(3)	0.5014(6)
z	0.08529(8)	0.0854(2)
Uiso × 100 (Å2)	0.57(5)	1.4(2)
R-factors of refinements
Rp	8.9	7.4
Rwp	11.9	9.6
χ2	2.0	2.1
Impurity phases	LaNi5 7.8(2) wt% MgNi2 12.0(2) wt%	LaNi5 20.5(2) wt% MgNi3 4.2(3) wt%

Table 1. Crystal structure data for the La_0.91Mg_2.09Ni₉ and La_1.09Mg_1.91Ni₉ alloys from Rietveld refinements of the X-ray diffraction data. PuNi₃ type of structure, space group R$\overline{3}$m.

Alloy	Sample 1	Sample 2
Source of experimental data	SR XRD collected at BM01B, SNBL using a wavelength λ = 0.5009(1) Å	Siemens D5000 diffractometer, Cu Kα1 radiation
Composition of AB3 phase	La1.09(1)Mg1.91(1)Ni9	La0.91(1)Mg2.09(1)Ni9
Unit cell parameters:
a (Å)	4.94024(8)	4.8986(1)
c (Å)	23.8188(4)	23.957(1)
V (Å3)	503.44(1)	497.86(2)
Atomic parameters:
La1/Mg1 in 3a (0, 0, 0)
Uiso×100 (Å2)	0.43(5)	2.1(2)
nMg, (nLa = 1–nMg)	0.0(–)	0.09(1)
La2/Mg2 in 6c (0, 0, z)
z	0.1453(3)	0.1471(6)
Uiso×100 (Å2)	1.2(3)	0.5(3)
nMg, (nRE = 1–nMg)	0.954(5)	1.0(–)
Ni1 in 3b (0, 0, ½) Uiso × 100 (Å2)	0.7(1)	0.8(3)
Ni2 in 6c (0, 0, z)
z	0.3335(2)	0.3334(4)
Uiso×100 (Å2)	0.13(8)	1.8(3)
Ni3 in 18h (x, –x, z)
x	0.5009(3)	0.5014(6)
z	0.08529(8)	0.0854(2)
Uiso × 100 (Å2)	0.57(5)	1.4(2)
R-factors of refinements
Rp	8.9	7.4
Rwp	11.9	9.6
χ2	2.0	2.1
Impurity phases	LaNi5 7.8(2) wt% MgNi2 12.0(2) wt%	LaNi5 20.5(2) wt% MgNi3 4.2(3) wt%

The crystallographic characteristics of LaNi₃ change significantly on Mg → La substitution; a decrease in the unit cell parameters takes place from a = 5.0842(2); c = 25.106(1) Å (LaNi₃) to a = 4.8986(1) (sample 2)-4.94024(8) (sample 1); c = 23.8188(4) Å (sample 1)-23.957(1) (sample 2). Furthermore, comparison of the data shows that the studied intermetallic samples exhibit significant differences in the volumes of the unit cells and c/a ratios. A shrinkage along [001] appears to be more pronounced (Δc/c, −5.1%) as compared to Δa/a, −3.7%. The overall volume contraction is quite significant reaching 10.5%–11.5%. The measured dimensions of the unit cells well agree with the data reported for the stoichiometric LaMg₂Ni₉ alloy studied by single crystal XRD (a = 4.9241, c = 23.875 Å; V = 501.3 ų [15]), which shows intermediate values of a, c and V being in between the values for the samples 1 and 2, as it could be expected from comparison of their chemical compositions.

Refined volumes of the unit cells correlate with their chemical compositions and Mg/La ratios. Indeed, sample 1, La_1.09(1)Mg_1.91(1)Ni₉ with a larger unit cell has a higher content of lanthanum, while for sample 2, La_0.91(1)Mg_2.09(1)Ni₉ with a smaller unit cell, the content of lanthanum becomes smaller than 1 atom/f.u., and the content of Mg reaches overstoichiometric compositions with more than 2 Mg atoms/f.u. (La,Mg)₃Ni₉.

Comparison of the data presented in Table 1 with crystallographic data for the (La,Mg)₃Ni₉ intermetallics studied in [1] shows a linear dependence between the decrease of the unit cell volumes and the content of Mg in the alloys.

We note a very interesting feature of the crystal structure of La_0.91(1)Mg_2.09(1)Ni₉ where a partial substitution of La by Mg takes place within the CaCu₅ type layer in the position 6c. This contrasts with the behaviour of the alloys in the La-Mg-Ni system with compositions close to LaNi₅. In the latter case studies of phase equilibria showed no dissolution of an appreciable amount of Mg in LaNi₅ [16]. Thus, the present study demonstrates that the situation with Mg solubility in the LaNi₅ slabs of the LaNi₃ structure becomes different in the sample 2 La_0.91(1)Mg_2.09(1)Ni₉. Here LaNi₅, when influenced by the MgNi₂ slabs of the hybrid structure, becomes capable of forming solid solutions of such a type with experimentally refined composition of La_0.95Mg_0.05Ni₅. Thus, La_0.91(1)Mg_2.09(1)Ni₉ should be considered as the first reported case where a CaCu₅ type layer accommodates Mg atoms allowing a Mg content of 2.09 at./f.u. (La,Mg)₃Ni₉. Consequently, the limits of Mg solubility in LaNi₃ are not confined to LaMg₂Ni₉ and extend to the composition La_0.91Mg_2.09Ni₉.

3.2. Thermodynamics of the (La,Mg)₃Ni₉—H₂ systems

The hydrogenation/deuteration properties of the prepared La_1±0.1Mg_2±0.1Ni₉ intermetallics appear to be dramatically different from those for LaNi₃. While LaNi₃ is prone to the hydrogen-induced disproportionation, the Mg-containing intermetallics reversibly form hydrides with ΔH_des = 24.0 kJ/mol_H2and equilibrium pressure of H₂ desorption of 20 bar at room temperature for La_1.09Mg_1.91Ni₉ (see Figure 2). A pronounced hysteresis of H₂ absorption and desorption is evidenced by a high value of H₂ absorption pressure, more than 100 bar higher than that for desorption.

For La₂MgNi₉ [6] at room temperature the values of plateau pressures are 0.05 and 0.1 bar for hydrogen desorption and absorption, respectively, ΔH_des = 35.9 kJ/mol_H2. Equilibrium pressure of hydrogen desorption for La_0.91Mg_2.09Ni₉ is by more than 1000 times higher than that for La₂MgNi₉.

Figure 2. Room temperature isotherms of hydrogen absorption and desorption (a); and van’t Hoff plots (b) for La_1.09Mg_1.91Ni₉-based hydride. At room temperature equilibrium pressure of hydrogen absorption is ~120 bar D₂, while for the desorption P_eq. equals to ~20 bar D₂.

Figure 2. Room temperature isotherms of hydrogen absorption and desorption (a); and van’t Hoff plots (b) for La_1.09Mg_1.91Ni₉-based hydride. At room temperature equilibrium pressure of hydrogen absorption is ~120 bar D₂, while for the desorption P_eq. equals to ~20 bar D₂.

3.3. In situ NPD studies

In situ neutron powder diffraction studies of the La_1±0.1Mg_2±0.1Ni₉D_9.4−9.5 deuterides were performed at the Spallation Neutron Source SINQ accommodated at Paul Scherrer Institute (Villigen, Switzerland). Two samples, La_1.09Mg_1.91Ni₉D_9.5(3) (sample 1) and La_0.9Mg_2.1Ni₉D_9.4(6) (sample 2) were synthesised and studied under different conditions.

For the synthesis of La_1.09Mg_1.91Ni₉D_9.5, a 6 mm diameter stainless steel autoclave with a wall thickness of 0.2 mm was used. The synthesis was performed by saturating activated samples with deuterium gas (25 bar) at a sub–zero temperature of −30 °C. This was done in order to decrease the equilibrium pressure of hydrogen absorption-desorption in the La_1.09Mg_1.91Ni₉—D₂ system. The alloy absorbed deuterium to reach a composition La_1.09Mg_1.91Ni₉D_9.5 and was measured at 25 °C and deuterium pressure of 25 bar.

The second sample, La_0.9Mg_2.1Ni₉D_9.4, was synthesized at high pressure deuterium gas of 950 bar D₂. The studied sample was placed inside a TiZr sample cell with a stainless steel liner, which was used as a sample holder during the in situ NPD experiments (see Figure 3). The pressure during the NPD measurements performed at 20 °C was set to 912 bar D₂. No preliminary activation was applied prior to the synthesis.

Figure 3. High pressure synthesis setup for the in situ NPD measurements at pressures up to 1000 bar D₂.

Figure 3. High pressure synthesis setup for the in situ NPD measurements at pressures up to 1000 bar D₂.

For the La_1.09Mg_1.91Ni₉D_9.5(5) sample (No.1) at the highest applied deuterium pressure of 25 bar D₂, the deuteration resulted in the formation of a two-phase mixture of the α-solid solution of deuterium in the alloy and a corresponding β-deuteride. Such a mixture of the phase constituents was observed after allowing a deuteration time of ~20 h at interaction temperature of −30° C. Since applied temperature-pressure conditions were rather close to the equilibrium ones (see Figure 2), the transformation was slow and was not completed on the time scale of the measurements performed. The second sample with a slightly higher content of magnesium, La_0.91Mg_2.09Ni₉D_9.3(7) was saturated by deuterium at deuterium pressure of 950 bar and was equilibrated at 912 bar D₂ and 25 °C. Analysis of the diffraction pattern showed an excellent fit between the experimental data and calculated NPD profiles (Figure 4) and indicated a completeness of the transformation of the α-solid solution into the β-deuteride.

The results of the refinements of the NPD data for La_1.09Mg_1.91Ni₉D_9.5(5) and for La_0.91Mg_2.09Ni₉D_9.4(6) are summarized in Table 2. The data show a formation of very similar structures, with only minor differences in the occupancies of the specific D-sites of five various types. These sites are shown in Figure 5 and include four types of tetrahedral and one tetragonal bipyramid.

A partial filling by D atoms of the four types of the tetrahedral interstices takes place inside the MgNi₂ slab; these include two types of the [MgNi₃] (18h and 6c) tetrahedra and two types of the [Mg₂Ni₂] (36i and 18h) interstitial sites.

In addition, similar to the other studied La_3−xMg_xNi₉-based deuterides, the remaining 5.0 or 5.6 at. D/f.u. form a standard hydrogen sublattice within the LaNi₅ slab which are statistically distributed within the four types of the interstices; hydrogen atoms partially occupy [La₂Ni₄] octahedra, three types of [Ni₄] tetrahedra, and two types of the [LaMgNi₂] sites.

Figure 4. NPD pattern of La_0.9Mg_2.1Ni₉D_9.4(6) (912 bar D₂, 298 K). Note that the most significant contributions to the difference intensities are coming from the sample cell. R_p = 2.4%, R_wp = 3.2; χ² = 6.0.

Figure 4. NPD pattern of La_0.9Mg_2.1Ni₉D_9.4(6) (912 bar D₂, 298 K). Note that the most significant contributions to the difference intensities are coming from the sample cell. R_p = 2.4%, R_wp = 3.2; χ² = 6.0.

Table 2. Crystal structure data for the deuterated La_1±0.1Mg_2±0.1Ni₉ alloys (PuNi₃ type, sp.gr. R$\overline{3}$m) from the Rietveld refinements of in situ neutron diffraction data.

Deuteride	La1.09Mg1.91Ni9D9.5(5)	La0.91Mg2.09Ni9D9.4(6)
Conditions	25 bar at 25 °C (prepared at −30 °C)	912 bar at 25 °C
Unit cell parameters:
a (Å)	5.263(1)	5.212(1)
c (Å)	25.803(9)	25.71(1)
V (Å3)	618.9(3)	604.8(3)
Unit cell parameters:
Δa/a (%)	6.5	6.4
Δc/c (%)	8.3	7.3
ΔV/V (%)	23.0	21.6
ΔV/V[LaNi5] (%)	20.4	20.7
ΔV/V[MgNi2] (%)	25.4	22.2
Atomic parameters:
La1/Mg1 in 3a (0, 0, 0) nMg, (nLa = 1–nMg)	0.0(–)	0.09(–)
La2/Mg2 in 6c (0, 0, z) z Uiso × 100 (Å2) nMg, (nRE = 1–nMg)	1.0(–) 0.95(–)	1.0(–) 1.0(–)
Ni1 in 3b (0, 0, ½) Uiso × 100 (Å2)	1.0(–)	1.0(–)
Ni2 in 6c (0, 0, z) z Uiso × 100 (Å2)	0.3279(7) 1.0(–)	0.3220(6) 1.0(–)
Ni3 in 18h (x, –x, z) x z Uiso × 100 (Å2)	0.498(1) 0.0871(4) 1.0(–)	0.506(1) 0.0859(3) 1.0(–)
D1 in 18h (x, –x, z) x z n	0.484(4) 0.023(1) 0.33(1)	0.496(3) 0.023(1) 0.31(2)
D2 in 6c (0, 0, z) z n	0.390(1) 0.50(3)	0.385(1) 0.58(3)
D4’ in 18h (x, –x, z) x z n	0.814(3) 0.0626(9) 0.43(2)	0.792(2) 0.051(1) 0.33(3)
D5’ in 18h (x, –x, z) x z n	0.201(2) 0.120(1) 0.45(2)	0.192(3) 0.123(1) 0.35(2)
D6 in 18h (x, –x, z) x z n	0.819(4) 0.117(1) 0.20(2)	0.819(4) 0.117(1) 0.39(2)
Uiso × 100 (Å2) for D1-D6	2.0(–)	2.0(–)
Atomic parameters:
D distribution in the structure LaNi5 2 MgNi2	5.6(3) 3.9(2)	5.0(4) 4.4(2)
Shortest Metal—Hydrogen distances, Å La…D Mg…D Ni…D	2.34(3) 1.97(3) 1.56(3)	2.29(2) 1.93(2) 1.53(2)
R-factors of refinements Rp Rwp χ2	2.7 3.4 5.0	2.4 3.2 6.0
Secondary constituents	α-solid solution La0.9Mg2.1Ni9D0.9. Sp.gr. R$\overline{3}$m; a = 4.9459(2); c = 23.842(2) Å; V = 505.10(4). 0.3 D in D3 18h (0.15, 0.3, 0.085) and 0.6 D in D4 18h (0.3, 0.15, 0.085); 35.7(2) wt% LaNi5D7; Sp.gr. P63mc; a = 5.438(3), c= 8.598(5) Å; V = 220.3(2) Å3; 4.6(3) wt%. Atomic structure was taken from [3]. MgNi2; MgNi2 structure type; Sp.gr. P63/mmc; a = 4.8356(4), c = 15.850(3) Å; V = 320.97(5) Å3; 12.4(2) wt%. Atomic structure was taken from [4]. Sample holder: stainless steel; Sp.gr. Fm$\overline{ 3}m$; a = 3.598 Å.	LaNi5D7; Sp.gr. P63mc; a = 5.430(1), c = 8.606(4) Å; V = 219.8(2) Å3; 21.5(5) wt%. Atomic structure was taken from [3]. MgNi3; AuCu3 structure type; Sp.gr. Pm$\overline{3}$m; a = 3.7185 Å; 1 Mg in 1a: 0, 0, 0; 3 Ni in 3c: 1/2, 1/2, 0; 3.7(2) wt%. Sample holder: zero matrix TiZr alloy with Fe liner. The peaks from Fe liner are only observed. Sp.gr. Fm$\overline{ 3}m$; a = 3.5949(1) Å.

Table 2. Crystal structure data for the deuterated La_1±0.1Mg_2±0.1Ni₉ alloys (PuNi₃ type, sp.gr. R$\overline{3}$m) from the Rietveld refinements of in situ neutron diffraction data.

Deuteride	La1.09Mg1.91Ni9D9.5(5)	La0.91Mg2.09Ni9D9.4(6)
Conditions	25 bar at 25 °C (prepared at −30 °C)	912 bar at 25 °C
Unit cell parameters:
a (Å)	5.263(1)	5.212(1)
c (Å)	25.803(9)	25.71(1)
V (Å3)	618.9(3)	604.8(3)
Unit cell parameters:
Δa/a (%)	6.5	6.4
Δc/c (%)	8.3	7.3
ΔV/V (%)	23.0	21.6
ΔV/V[LaNi5] (%)	20.4	20.7
ΔV/V[MgNi2] (%)	25.4	22.2
Atomic parameters:
La1/Mg1 in 3a (0, 0, 0) nMg, (nLa = 1–nMg)	0.0(–)	0.09(–)
La2/Mg2 in 6c (0, 0, z) z Uiso × 100 (Å2) nMg, (nRE = 1–nMg)	1.0(–) 0.95(–)	1.0(–) 1.0(–)
Ni1 in 3b (0, 0, ½) Uiso × 100 (Å2)	1.0(–)	1.0(–)
Ni2 in 6c (0, 0, z) z Uiso × 100 (Å2)	0.3279(7) 1.0(–)	0.3220(6) 1.0(–)
Ni3 in 18h (x, –x, z) x z Uiso × 100 (Å2)	0.498(1) 0.0871(4) 1.0(–)	0.506(1) 0.0859(3) 1.0(–)
D1 in 18h (x, –x, z) x z n	0.484(4) 0.023(1) 0.33(1)	0.496(3) 0.023(1) 0.31(2)
D2 in 6c (0, 0, z) z n	0.390(1) 0.50(3)	0.385(1) 0.58(3)
D4’ in 18h (x, –x, z) x z n	0.814(3) 0.0626(9) 0.43(2)	0.792(2) 0.051(1) 0.33(3)
D5’ in 18h (x, –x, z) x z n	0.201(2) 0.120(1) 0.45(2)	0.192(3) 0.123(1) 0.35(2)
D6 in 18h (x, –x, z) x z n	0.819(4) 0.117(1) 0.20(2)	0.819(4) 0.117(1) 0.39(2)
Uiso × 100 (Å2) for D1-D6	2.0(–)	2.0(–)
Atomic parameters:
D distribution in the structure LaNi5 2 MgNi2	5.6(3) 3.9(2)	5.0(4) 4.4(2)
Shortest Metal—Hydrogen distances, Å La…D Mg…D Ni…D	2.34(3) 1.97(3) 1.56(3)	2.29(2) 1.93(2) 1.53(2)
R-factors of refinements Rp Rwp χ2	2.7 3.4 5.0	2.4 3.2 6.0
Secondary constituents	α-solid solution La0.9Mg2.1Ni9D0.9. Sp.gr. R$\overline{3}$m; a = 4.9459(2); c = 23.842(2) Å; V = 505.10(4). 0.3 D in D3 18h (0.15, 0.3, 0.085) and 0.6 D in D4 18h (0.3, 0.15, 0.085); 35.7(2) wt% LaNi5D7; Sp.gr. P63mc; a = 5.438(3), c= 8.598(5) Å; V = 220.3(2) Å3; 4.6(3) wt%. Atomic structure was taken from [3]. MgNi2; MgNi2 structure type; Sp.gr. P63/mmc; a = 4.8356(4), c = 15.850(3) Å; V = 320.97(5) Å3; 12.4(2) wt%. Atomic structure was taken from [4]. Sample holder: stainless steel; Sp.gr. Fm$\overline{ 3}m$; a = 3.598 Å.	LaNi5D7; Sp.gr. P63mc; a = 5.430(1), c = 8.606(4) Å; V = 219.8(2) Å3; 21.5(5) wt%. Atomic structure was taken from [3]. MgNi3; AuCu3 structure type; Sp.gr. Pm$\overline{3}$m; a = 3.7185 Å; 1 Mg in 1a: 0, 0, 0; 3 Ni in 3c: 1/2, 1/2, 0; 3.7(2) wt%. Sample holder: zero matrix TiZr alloy with Fe liner. The peaks from Fe liner are only observed. Sp.gr. Fm$\overline{ 3}m$; a = 3.5949(1) Å.

From the refinements of the NPD data we conclude that the overall chemical compositions La_1.09Mg_1.91Ni₉D_9.5/La_0.91Mg_2.09Ni₉D_9.4can be presented as LaNi₅H_5.6/LaNi₅H_5.0 + 2*MgNi₂H_1.95/MgNi₂H_2.2. Thus, in the hybrid La_1±0.1Mg_2±0.1Ni₉ structure, a LaNi₅-assisted hydrogenation of the MgNi₂ slab proceeds at rather mild H₂/D₂ pressure conditions; the equilibrium D₂ desorption pressure is just 20 bar D₂. In contrast, the parent MgNi₂ intermetallic remains inert with respect to hydrogenation even at much higher hydrogen pressures as well as the conditions applied in the present study of 912 bar D₂ for sample 2.

Figure 5. Crystal structure of La_1±0.1Mg_2±0.1Ni₉D_9.4−9.5 and types of the filled interstices.

Figure 5. Crystal structure of La_1±0.1Mg_2±0.1Ni₉D_9.4−9.5 and types of the filled interstices.

The shortest Me–D distances in the studied deuterides are listed in Table 2 and are within the regular values for the La–H, Mg–H and Ni–H distances in the structures of the metal and intermetallic hydrides.

The data of the present study clearly shows an influence of the LaNi₅ and MgNi₂ layers in the hybrid La_1±0.1Mg_2±0.1Ni₉ structures on the hydrogenation of the other buildings blocks of the structure. MgNi₂ slabs accommodate hydrogen up to a composition MgNi₂H_2.2 at much lower pressures as compared to those required to form a hydride by the pure MgNi₂ intermetallic. In contrast, the LaNi₅ block absorbs 5.0–5.6 at.H/f.u., which is quite close to the maximum hydrogenation capacity of the title intermetallic alloy, LaNi₅H₇; however, hydrogen desorption from the LaNi₅H_5.0/5.6 block proceeds much easier, at significantly higher pressures of H₂/D₂ as compared to the individual LaNi₅H₇ hydride—as a result of influence of the MgNi₂ slab.

4. Conclusions

LaNi₅-assisted hydrogenation of MgNi₂ is observed in the LaMg₂Ni₉ hybrid structure. Formation of LaMg₂Ni₉D_9.5 proceeds via an isotropic expansion of the trigonal unit cell. D atoms are accommodated in both Laves and CaCu₅-type slabs H atoms filling interstitial sites in both LaNi₅ and MgNi₂ structural fragments.

Limits of Mg solubility in LaNi₃ are not confined to LaMg₂Ni₉ and extend to the composition La_0.91Mg_2.09Ni₉ with a refined composition of the CaCu₅-type block of La_0.95Mg_0.05Ni₅.

Within the LaNi₅ CaCu₅-type layer, D atoms fill three types of interstices; a deformed octahedron [La₂Ni₄], and two types of tetrahedra, [LaNi₃] and [Ni₄], to yield LaNi₅D_5−5.6 composition. D distribution is very similar to that in the individual β-LaNi₅D₇ deuteride.

In the MgNi₂ slab hydrogen atoms fill two types of tetrahedra, [Mg₂Ni₂] and [MgNi₃]. The hydrogen sublattice formed is unique and is not formed in the studied structures of the Laves-type intermetallic hydrides.

A significant mutual influence of the LaNi₅ and MgNi₂ slabs causes a dramatic altering of their hydrogenation behaviours leading to:

(a)

significant decrease of the stability of the LaNi₅-type hydride;

(b)

much easier hydrogenation of the MgNi₂ slabs compared to the parent intermetallic compound;

(c)

increased hysteresis.