Structural and In Situ X-ray Diffraction Study of Hydrogenation of Ca_xMg_1−xNi₂ (0 ≤ x ≤ 1)

Abstract

In the quasi-binary system CaNi₂-MgNi₂ solid-solutions Ca_xMg_1−xNi₂ (0 ≤ x ≤ 1) were prepared from the elements. They crystallize in the hexagonal Laves phase type (MgNi₂, C36) for x ≤ 0.33 (P6₃/mmc, a = 482.51(7) pm, c = 1582.1(3) pm for x = 0, a = 482.59(3), c = 1583.1(1) for x = 0.33) and in the cubic Laves phase type (MgCu₂, C15) for 0.33 < x (Fd$\overline{3}$m, a = 697.12(3) pm for x = 0.5, a = 705.11(2) pm for x = 0.67, a = 724.80(2) pm for x = 1). After hydrogenation in an autoclave the X-ray diffraction patterns changed completely. Reflections assigned to CaNiH_3, and Ni and Rietveld refinement confirmed this. The hydrogenation properties of Ca_xMg_1−xNi₂ (0 ≤ x ≤ 1) compounds were also studied in situ by X-ray powder diffraction. In situ X-ray powder diffraction of Ca_xMg_1−xNi₂ (0 ≤ x ≤ 1) compounds under 0.3 MPa hydrogen gas flow (15 sccm), data collected on a Rigaku SmartLab diffractometer in an Anton Paar XRK 900 Reactor Chamber using Cu-Kα1 radiation. Scanning electron microscopy and EDX spectroscopy confirmed the entitled materials and elemental composition, respectively. From the Transmission electron microscopy and Selected area electron diffraction concluded that the Ca_xMg_1−xNi₂ (0 ≤ x ≤ 1) compounds were crystalline.

1. Introduction

The growing awareness of environmental protection and increasing demands for energy force us to find a new complementary solution. To replace traditional fossil fuels, hydrogen is receiving interest as a clean energy. Finding a new way to store hydrogen is the main problem in using hydrogen as a fuel in the future. To store hydrogen in the solid phase is the most promising solution because of the reversibility of the process, for safety reasons and, above all, its high capacity [1,2,3,4,5,6,7,8]. There are many materials that can be used as a medium for hydrogen storage. For hydrogen storage AB, AB₂, A₂B and AB₅ intermetallic phase-based alloys are the one basic group of materials [9]. In term of hydrogen storage, magnesium hydride is the mostly studied materials because of reversible absorption and desorption, high gravimetric density and relatively low cost [10]. The problem is that the high thermodynamics stability and slow kinetics of its absorption and desorption reactions complicate its uses. There are a number of publications regarding the possibility of improving these properties, mainly by doping with transition group elements or other chemical compounds and the formation of ternary metal hydrides [11,12,13,14]. Low-cost hydrogen storage materials include calcium-nickel and magnesium-calcium-nickel alloys. Many new hydrogen-absorbing materials have been developed to increase hydrogen storage capacity. Low-density Ca-Ni alloys are also good for hydrogen storage applications; for example, the CaNi₅ alloy absorbed 1.9 percent hydrogen [15]. With a PuNi₃ structure, the Ca_0.4Mg_0.6Ni₃ alloy has a hydrogen capacity of 1.56 percent at room temperature [16,17,18]. The MgCu₂ type (C15), MgZn₂ type (C14), and MgNi₂ type are the three main types of Laves phases (C36). For some subgroups, such as AM₂ and AM₃, both elements from the main groups, but occasionally M from the transition metal, valence electron concentration and size ratio are known [19,20,21]. Many Laves phases take up hydrogen reversibly at moderate pressure and temperature, making them a common example of hydrogen storage materials [22,23,24,25]. A statistical rule AMm + nH₂ = AMmH₂n was developed for the thermodynamic hydrogen uptake of an intermetallic phase [26]. CaNi₂ and MgNi₂ have cubic and hexagonal crystal structure, respectively. Ca_xMg_1−xNi₂ (x = 0.5, 0.67) also has cubic and C15 (MgCu₂)-type Laves phase structure while Ca_xMg_1−xNi₂ (x = 0.33) has hexagonal and C36 (MgNi₂)-type Laves phase structure (Figure 1).

Oesterreicher et al. [27] investigated the hydrogen storage properties and crystal structure of the Ca_1−xMg_xNi₂ alloy first, and Terashita et al. [28] investigated it further. CaNi₂ belongs to the C15 Laves phase, whereas MgNi₂ belongs to the C36 Laves phase. Terashita claims that the single C15 Laves phase exists in the CaNi₂ and Ca_0.32Mg_0.68Ni₂ composition ranges and was absorbed hydrogen 1.4 percent at 313 K. Basic structural units for the Hauke (AB5) and Laves (AB2) phases have been identified. CaMg₂(MgZn₂ type), Mg₂Ni, and C36 MgNi₂ Laves phases are known complementary phases in the Ca-Ni and Mg-Ni binary systems. The Mg position at A or B sites in the Laves phases results different polytypes formation in the system. From the structural point of view this was widely studied by Komura group in 1980s [29]. Shin-ichi Orimo et al. have studied the nano-structured and amorphous-structured hydriding properties in the binary Mg-based system [30,31]. Akiba and Tsushio have thoroughly investigated the Mg(Ni_1.9M_0.1) (M = Cr, Mn, Fe) with the AB-type Laves phases structure hydriding properties in the ternary system [32,33]. Oesterreicher et al. have reported the (Mg_xR_1−x)Ni₂ (x = 0–1.0, R = Ca, La) hydriding properties [34]. Terashita et al. studied the Mg_1−xCa_xNi₂ hydriding properties [28] and Kohno et al., have studied the Mg-La-Ni hydriding properties [35].

The objective of this study to obtain more data for the hydrogenation of CaNi₂, MgNi₂ and Ca_xMg_1−xNi₂ compounds. The hydrogenation properties of the above-mentioned compounds were analyzed in situ by X-ray diffraction (XRD). In the present article we also synthesized and characterized the binary alloys of CaNi₂, MgNi₂ and ternary alloy of Ca_xMg_1−xNi₂ (x = 0.5, x = 0.67, x = 0.33). Ca_xMg_1−xNi₂ which are pseudo binary alloy based on CaNi₂ (C15 Laves phase) and MgNi₂ (C36 Laves phase).

2. Materials and Methods

2.1. Materials

The Ca_xMg_1−xNi₂ (0 ≤ x ≤ 1) compounds were prepared from the elements Mg powder (99.9%, −325 mesh, Alfa Aesar, Karlsruhe, Germany) Ca granules redistilled (99.9%, metal basis 1 cm (0.4 in) Alfa Aesar, Karlsruhe, Germany) and Ni powder (99.9%, −100 mesh, Aldrich, Seelze, Germany).

2.2. Synthesis

It is rather difficult to get the accurate stoichiometry in the Ca_xMg_1−xNi₂ system because synthesis is hindered due to the low boiling point of calcium 1873 K (1600 °C) which is close to the nickel melting point 1728 K (1455 °C). Therefore, during melting of both elements together leads to losses of calcium by evaporation. Mg addition also leads to the same problems. Different Ca-Mg-Ni samples were prepared by sealing the starting materials into Ta tubes under the argon atmosphere inside the gloves box finally completely sealed by arc melting [36]. Based on stoichiometric amounts of starting materials, an extra 2 wt.% of Ca and 6 wt.% of Mg were therefore added to compensate for the losses of Ca and Mg during reaction. All mixtures were sealed in Ta tubes under argon atmosphere by arc melting and heated from room temperature to T = 920 °C over 100 °C/h, held for 2 h, cooled to T = 800 °C within 1 h and 12 min, then held for 6 h and finally cooled to room temperature under vacuum 10⁻³ mbar.

In an argon-filled gloves box with oxygen and water content below 1 ppm, all alloys were mechanically crushed into powders. Laboratory XRPD data were collected using either a Huber G670 diffractometer with Mo-K1 radiation (flat transmission sample with Lithilen^® grease between two Kapton^® foils) or a Rigaku D/Max 2500VL/PC diffractometer (provided by Rigaku Corporation, Tokyo, Japan) with Cu Ka radiation at 50 kV and 150 mA.

2.3. Hydrogenation

For hydrogenation the powdered intermetallic binary and ternary alloys i.e., Ca_xMg_1−xNi₂ (x = 1, x = 0.67, x = 0.33 and x = 0) (100 mg) each sample was placed in a hydrogen resistant stainless steel (nickel-chromium stainless steel 1.4571) crucible in autoclave made from the same alloy, charged with hydrogen (≥99.9%, Air Liquide, Paris, France) and heated in a vertical tube furnace. The hydrogen pressure was 3.4 MPa and heated to T = 673 K (400 °C), heating rate 100 K/h, hold for 48 h, then cooled to 373 K (100 °C) in 3:30 h finally to room temperature.

2.4. In Situ X-ray Powder Diffraction (XRPD)

A Rigaku SmartLab diffractometer with CuKɑ1 radiation was used to collect laboratory XRPD data. Cu-K radiation was used to collect in situ X-ray diffraction data on a SmartLab (Rigaku Corporation, Tokyo, Japan) high-resolution X-ray diffractometer with a HyPix-3000 two-dimensional semiconductor detector. Cu-K has been removed from the Cu-K measurements using a nickel filter. Bragg–Brentano geometry was used to operate the diffractometer. Individual measurements took 10 min to collect data and 6 min to collect data for the second CaNi₂ measurement. Grinding the reactants and placing it in the sample holder of an Anton Paar (Graz, Austria) XRK 900 Reactor Chamber resulted in flat samples. The chamber’s gas supply system was connected, allowing for evacuation and subsequent H₂ gas flooding (H2: Air Liquide, 99.9%). To refine crystal structures, the program FullProf (5.0.0.0) was used [37].

2.5. Characterization

X-ray diffraction (XRD) measurements were performed using a Huber G670 diffractometer with Mo-Kα1 radiation (Provided by Huber Diffraktionstechnik GmbH & Co. KG, Rimsting, Germany) over a 2 θ range of 5–70°. The surface morphology was performed using an Oxford Instruments EDX INCA SYSTEM mounted on a Zeiss LEO 1530 scanning electron microscope with a working distance of 15 mm and an energy dispersive spectrometer (EDS) provided by Oxford company, Scotts Valley, CA, USA. A JEOL 2100 Plus instrument was used for the high-resolution transmission electron microscopy investigations with a point resolution of 0.19 nm.

3. Results

The Rietveld method was used for the evaluation of powder diffraction data, using the FULLPROF software (version: 5.0.0.0). Consecutively, crystal structure refinements were completed. From the Inorganic Crystal structure database (ICSD), the initial’s structure was taken. Pseudo-Voigt were used for all refinement in model reflection profiles. Asymmetry, cell parameter, scale factor and Caglioti parameters were refined but the occupation numbers, thermal displacement and atomic position were fixed for all refinements.

Intermetallic samples were grey powders changing to black and becoming more brittle after hydrogenation. Rietveld refinements based on XRPD data were performed using the structure types of MgNi₂ type and MgCu₂ type as starting models. These structure types could be confirmed by the good fit of the Rietveld refinements (Figure 2a,b). Solid-solutions of Ca_xMg_1−xNi₂ crystallize in the hexagonal Laves phase type (MgNi₂, C36) for x ≤ 0.33 and in the cubic Laves phase type (MgCu₂, C15) for 0.33 < x. The refined lattice parameter i.e., structure type and space group type before hydrogenation are given in (

Table 1. Refined lattice parameters of Ca_xMg_1−xNi₂ compounds based on X-ray powder diffraction at 297(1) K.

X	Structure Type	Space Group Type	a/pm	c/pm
0	MgNi2	P63/mmc	a = 482.513(68)	c = 1582.073(271)
0.33	MgNi2	P63/mmc	a = 482.585(27)	c = 1583.073(106)
0.5	MgCu2	Fd$\overline{3}$m	a = 697.116(32)	A
0.67	MgCu2	Fd$\overline{3}$m	a = 705.112(24)	A
1	MgCu2	Fd$\overline{3}$m	a = 724.063(54)	A

).

After hydrogenation in an autoclave the X-ray diffraction patterns changed completely (Figure 3a,b). Reflections could be assigned to CsCl type cubic structure CaNiH₃ and Ni; Rietveld refinement confirmed this. Most Laves phases retained after incorporation of hydrogen their crystal structure except tetrahedral interstices filled by hydrogen atoms and expanded in volume. A suitable hydrogen position is A₂M₂ in cubic Laves phases (C15 type, Fd$\overline{3}$m), tetrahedral interstices on probable Wyckoff position 96 g. In Ca_xMg_1−xNi₂ (x = 0.67) the peaks present at the Bragg position (2θ°) 16.240, 23.040, 25.800, 28.320, 32.800, 34.860, 36.820, 42.200, 43.880 and corresponds to the reflection from (110), (200), (210), (211), (220), (300), (310), (320), (321) crystal plane assigned to the cubic perovskite type CaNiH₃ and the peaks present at the 2θ° of 20.020, and 40.680 are assigned to Ni. While in the case of Ca_xMg_1−xNi₂ (x = 1) the peaks present at the 2θ° 16.237, 25.855, 28.400, 36.888, 44.054, 50.270, 61.409 and 62.704 are assigned to CaNiH₃ and the peaks present at the 2θ° 20.020, 23.160, 32.980, 38.770, 40.680, 51.880, 53.340, 58.900 are assigned to the Ni. MgNi₂ type structures (C36) such as Ca_xMg_1−xNi₂ when x ≤ 0 absorbed hydrogen at high hydrogen pressure not at low hydrogen pressure confirmed from hydrogenation in autoclave (Figure 3c) and exhibited in hexagonal type crystal structure. The refined lattice parameter i.e., structure type and space group type after hydrogenation are given in (

Table 2. Refined lattice parameters of Ca_xMg_1−xNi₂ (x = 1, x = 0.67 and x = 0) compounds after hydrogenation based on X-ray powder diffraction at 297(1) K.

X	Structure Type	Space Group Type	a/pm	c/pm
0	MgNi2	P63/mmc	a = 482.637(29)	c = 1583.699(127)
0.67	MgCu2	Fd$\overline{3}$m	a = 705.515(48)	A
1	MgCu2	Fd$\overline{3}$m	a = 705.393(21)	A

).

In situ X-ray powder diffraction data were recorded under hydrogen pressure for three different samples (Figure 4, Figure 5 and Figure 6) i.e., CaNi₂, Ca_0.5Mg_0.5Ni₂ and Ca_0.67Mg_0.33Ni₂. Under 0.3 MPa hydrogen pressure the reaction 2CaNi₂ + 3H₂ → 2CaNiH₃ + 2Ni was followed by in situ X-ray powder diffraction (Figure 4). The reaction temperature was started from 25 to 500 °C completed in 20 °C steps with a 30 °C·min⁻¹ ramp. The hydrogen pressure recorded after each 25 °C absorbed by the sample till the temperature reached 500 °C. The temperature holds for 1 h at 500 °C and then cold to room temperature. The CaNi₂ has a cubic structure and form the cubic CsCl type hydride. There was no intermediate formed during the above reaction. The same reaction conditions were followed for the cubic MgCu₂ type structure Ca_0.5Mg_0.5Ni₂ + H₂ → Ca_0.5Mg_0.5Ni₂H_2.55 (Figure 5) and Ca_0.67Mg_0.33Ni₂ + H₂ → Ca_0.67Mg_0.33Ni₂H_2.70 (Figure 6).

Before the hydrogenation the refined lattice parameters of the three samples Figure 4, Figure 5 and Figure 6 are: a = 724.063(54), a = 697.116(32) and a = 705.112(24), respectively. After hydrogenation of the three samples CaNi₂, Ca_0.5Mg_0.5Ni₂ and Ca_0.67Mg_0.33Ni₂ (4–6), respectively, by in situ X-ray diffraction the lattice parameters of the hydrogenated sample increased, which is confirmed by the Rietveld refinements of its X-ray diffraction profiles. The refined lattice parameters of the (Figure 4, Figure 5 and Figure 6) are a = 724.263(54), a = 697.416(32) and a = 705.512(24), respectively, which confirmed the hydrogen absorption by the synthesized compounds.

Mostly materials are the MgCu₂ (C15) type structure at the composition when x = 0.5, x = 0.67, x = 1 in the Ca_xMg_1−xNi₂ while the materials are the MgNi₂ (C36) type structure at the composition when x = 0, x = 33. The synthesized compounds CaNi₂, Ca_0.5Mg_0.5Ni₂, Ca_0.67Mg_0.33Ni₂, have cubic crystal structure and the Ca_0.33Mg_0.67Ni₂, MgNi₂ have hexagonal crystal structure. It was found from Figure 5 and Figure 6 that x could be increased for Ca_0.5Mg_0.5Ni₂, Ca_0.67Mg_0.33Ni₂, from 2.55 to 2.70. The above materials at 0.1 MPa desorb partially to about x = 1.3. It was concluded that Ca_0.5Mg_0.5Ni₂H_2.6, Ca_0.67Mg_0.33Ni₂H_2.6 roughly corresponds to 1.7% H₂, and that is why these hydrides may be of interest for technological applications.

According to the Orimo at al., amorphous MgNi and its hydrides have local structure related to CsCl type structures [37,38]. It is known that Mg is an alkaline-earth element as well as Ca, so it is possible that Mg and Ca have some crystallographic structural similarities. The present result obtained from the Rietveld refinement supports their results for the MgNi hydride. However, there are some differences between the CaNi and MgNi hydrogenation properties. The MgNi has hydrogen content 1 H/M and CaNi has hydrogen content 1.5 H/M which is more than MgNi [39,40]. The MgNi shows a reversible hydrogenation and dehydrogenation while CaNi does not show the reversibility with hydrogen gas for the reaction.

The SEM analysis was carried out to for morphological examination and homogenized nature of the prepared derivation of Ca_xMg_1−xNi₂ (0 ≤ x ≤ 1) as shown Figure 7a–d. The micrographs reveal the formation of larger particles due to the aggregation of small particles. The large size aggregates are unevenly distributed and, due to the irregular shape and size, it is difficult to identify the exact morphology and particles size from the SEM micrographs.

The elemental composition and percentage purity of the synthesized derivatives of Ca_xMg_1−xNi₂ (0 ≤ x ≤ 1) were evaluated through EDX analysis. EDX elemental mapping shows the uniform distribution of Ca, Ni and Mg in their respective samples as shown in Figure 8a–d. The EDS spectrum for CaNi₂, (Figure 8b) exhibits the sharp signals at 0.35, 3.7, 4.0 keV for Ca and 0.7, 7.5 and 8.25 keV for Ni, where the weight percent of 24.96 and 75.05 were reported for Ca and Ni, respectively. The composition of Ca_0.5Mg_1–0.5Ni₂ is evident from the EDS spectrum given in Figure 8b, where Ca, Mg and Ni are present in the sample with weight percent of 15.51, 10.69 and 73.80, respectively. The composition of Ca_0.33Mg_0.67Ni₂ has been confirmed from the EDS spectrum shown in Figure 8c. The weight percent of the Ca, Mg and Ni are 9.48, 3.23 and 87.29 determined during elemental analysis. The Figure 8d reveals the composition of MgNi₂, the peaks for the Ni were observed at 0.85, 7.5 and 8.35 keV and present Mg in the sample was confirmed by the peak at 1.25 keV. The weight percent for Ni and Mg was found to be 64, 83 and 35.17, respectively.

To confirm the above XRD results HRTEM were performed for the CaNi₂ and Ca_xMg_1−xNi₂ (x = 0.5). Figure 9a–d show the TEM images of CaNi₂ and Ca_xMg_1−xNi₂ (x = 0.5). From the TEM images it was confirmed that the synthesized compounds are crystalline. Crystallographic information and d-spacing of the crystal planes were determined by selected area electron diffraction. Figure 10a showed the SAED of CaNi₂ and Figure 10b, c of Ca_xMg_1−xNi₂ (x = 0.5). Selected area electron diffraction patterns along the different directions of CaNi₂ and Ca_xMg_1−xNi₂ (x = 0.5) were taken. Figure 10 showed that for these images of single crystal, it is simply dots, and that these dots are located at the same position with regard one another. Thus, there is a complete order. Figure 10 showed that the diffraction gives discrete parts because in a single crystal all atoms are in the same order to all the edges of materials. The surface morphology of the synthesized materials was rough and porous, which is suitable for hydrogen storage.

4. Conclusions

For technological applications the hydriding behavior of these materials make them interesting as potential hydrogen storage with tailored properties to exploit the adverse kinetic effects upon Mg incorporation for hindering decomposition. After hydrogenation, there is in autoclave growth of crystallites of CaNi hydride and Ni. Elemental hydride absorbed hydrogen (CaH₂ + NiH_0.7 = CaNiH_2.7) so that the hydrogen absorbed by the synthesized CaNiH₃ in autoclave is approximately equal. Laves phases based on MgCu₂ (C15) and MgNi₂ (C36) retained their crystal structures upon hydrogenation. The MgNi₂ type structures (C36) such as Ca_xMg_1−xNi₂ absorbed hydrogen at high hydrogen pressure when x ≤ 0, not at low hydrogen pressure, which was concluded from hydrogenation in autoclave and exhibited in the hexagonal type crystal structure. The MgCu₂ type structures (C15) such as Ca_xMg_1−xNi₂ when x ≤ 1 absorbed hydrogen at a low pressure, confirmed by in situ X-ray diffraction and in the autoclave hydrogenation study, and as exhibited in cubic crystal structure.