Hydrogen Sorption Properties of a Novel Refractory Ti-V-Zr-Nb-Mo High Entropy Alloy

Abstract

High entropy alloys belong to a new and promising class of functional materials for solid-state hydrogen storage. In this context, a novel single-phase body centered cubic (bcc) high entropy alloy Ti_0.30V_0.25Zr_0.10Nb_0.25Mo_0.10 was prepared. The physicochemical and hydrogen sorption properties have been determined by both laboratory and large-scale facilities. This alloy can quickly absorb hydrogen up to 2.0 H/M (2.8 wt.%) at room temperature and forms a face centered cubic (fcc) hydride, as proven by synchrotron X-ray diffraction. The Pressure–Composition Isotherm and in situ neutron diffraction during hydrogen/deuterium desorption reaction suggest that the alloy experiences a reversible single step phase transition (bcc↔fcc). PDF analysis from X-ray total scattering data points out that the hydride phase possesses an average fcc structure with random atoms distribution and small lattice distortion. Despite an initial small fading of the capacity, the alloy withstands 20 absorption/desorption cycling without phase decomposition, as demonstrated by kinetic measurements coupled with X-ray diffraction and microstructural study by SEM-EDS. Moreover, the complete hydrogen absorption occurs in less than 30 s at room temperature and the kinetic improves during cycling.

1. Introduction

In recent decades, the reduction in our dependency on fossil fuels has become a necessity due to pollution and environmental issues. To lessen the use of hydrocarbons, which represents one of the main causes of pollution, scientists are searching for a new way to ensure an energy transition from fossil fuels to the use of hydrogen as an energy carrier. The reason for this focus on hydrogen is its abundance and the large amount of chemical energy per unit mass that this element can store. At standard temperature and pressure (STP) conditions pure hydrogen exists only as a low-density gas which hinters its use in practical and efficient applications. One of the most promising solution is to reversibly store hydrogen in solid-state materials forming metal hydrides. Many materials have been tested, such as intermetallic and complex hydrides [1]. Nevertheless, the application of these materials close to STP conditions is still impractical because of their low capacity, unfavorable thermodynamics, sluggish kinetics and loss of reversibility during cycling.

In this context, high entropy alloys (HEA) are a recent class of metallic materials that can be suitable for hydrogen storage. Since their first report by Yeh et al. [2] and Cantor et al. [3], HEAs are defined as disordered solid solutions containing five or more elements in nearly equimolar concentrations (5–35 at.%) [4]. These solid solutions usually adopt simple structures such as body centered cubic (bcc), face centered cubic (fcc) or hexagonal close compact (hcp). In order to predict the formation of solid solutions in this class of materials, two empirical parameters are used. First, the valence electron concentration (VEC) can affect the homogeneity and the structure of a solid solution phase. For example, bcc solid solution of refractory alloys could exist for VEC < 7.6 [5]. Second, the elements in HEAs are randomly distributed on the same crystallographic sites involving important lattice distortion that can be defined as (δ),

$$\delta =\sqrt{{\displaystyle \sum }_{i=1}^n{c}_i{\left(1-\frac{r_i}{\overline{r}}\right)}^2}$$

(1)

where $c_i$, $r_i$ presents the concentration and the radius of the element i and $\overline{r}$ the average atomic radius of the composition. Empirical studies have shown that complete solid solutions are formed for δ < 6.6% [6]. Distortions higher than this critical value may lead to the formation of multiphased alloys with intermetallic precipitations or even amorphous phases for δ > 8% [7].

The mixing of metals forming hydrides into a HEA may result in the creation of a material with interesting hydrogen sorption properties. Moreover, the endless compositional space of HEAs leads to numerous candidates for hydrogen storage.

Elementary and conventional metal hydrides possess a maximum capacity of hydrogen absorption by 2.0 H atoms per metal (H/M) [8]. Nevertheless, Sahlberg et al. reported in 2016 that the TiVZrNbHf single-phase bcc alloy absorbs large amount of hydrogen reaching 2.5 H/M [9]. During the hydrogen absorption the alloy undergoes a phase transition from a bcc lattice to a pseudo fcc structure with a small tetragonal deformation [10]. This very important finding opened the field to the study of refractory HEAs for hydrogen storage with already several interesting reports [11,12,13,14,15,16]. Kunce et al. used laser engineered net shaping (LENS) to synthesize the equimolar TiZrNbMoV [17]. This composition is multiphased with bcc and orthorhombic coexisting phases. The hydrogen absorption capacity of this alloy is 1.79 H/M (2.3 wt.%) and after an additional annealing the capacity drops to 1.45 H/M (1.8 wt.%). After hydrogenation, the material is also multiphased and no clear conclusion could be drawn due to the coexistence of several phases. This suggests that the research efforts should focus on single-phased alloys in order to clarify their interaction with hydrogen and the composition-structure-properties relationship.

Our recent work by Montero et al. focused on the optimization of the Ti-V-Zr-Nb system to obtain single-phased alloy by various synthetic methods [18]. The quaternary Ti_0.325V_0.275Zr_0.125Nb_0.275 alloy is single-phased bcc alloy and forms a dihydride with fcc lattice with a maximum hydrogen absorption capacity of 1.8 H/M (2.7 wt.%). Following this study, we have adopted a new strategy for the investigation of HEAs for hydrogen storage. Our focus is placed on the investigation of the effect of adding 10% of a fifth element to the previous quaternary alloy upon the following formula: Ti_0.30V_0.25Zr_0.10Nb_0.25M_0.10 (M = Al, Mg, Ta). The Ti_0.30V_0.25Zr_0.10Nb_0.25Al_0.10 composition was synthesized by arc melting and can absorb hydrogen forming a bct hydride with the capacity 1.6 H/M (2.6 wt.%) at room temperature [19]. The major improvements of Al addition are related to the desorption and cycling properties of the material. The temperature for hydrogen release was decreased by around 100 °C and the alloy showed superior cycling stability and higher reversible storage capacity than the quaternary counterpart. The Ti_0.30V_0.25Zr_0.10Nb_0.25Mg_0.10 alloy was prepared by high energy mechanosynthesis under inert atmosphere and undergoes a phase transition from a bcc initial lattice to a fcc hydride phase upon hydrogen absorption at room temperature (capacity 1.7 H/M, 2.7 wt.%) [20]. The addition of Mg into the parent quaternary alloy improves the hydrogen absorption/desorption cycling properties. The Ti_0.30V_0.25Zr_0.10Nb_0.25Ta_0.10 alloy can absorb hydrogen up to 2.0 H/M (2.5 wt.%) and the desorption properties were also enhanced, as compared to the parent quaternary alloy [21].

Based on these studies, we are currently exploring the same strategy by using Mo addition into the initial Ti_0.325V_0.275Zr_0.125Nb_0.275 alloy. The physicochemical and the hydrogen sorption properties of a new refractory HEA Ti_0.30V_0.25Zr_0.10Nb_0.25Mo_0.10 are reported here. Our aim is to highlight the effect of 10% addition of Mo on the structure, the microstructure, the chemical homogeneity and the hydrogen sorption performances of the parent composition. Large scale experiments at synchrotron (X-ray powder diffraction and total X-ray scattering for atomic pair distribution function analysis) and neutron sources (in situ neutron powder diffraction) have been employed to clarify the role of Mo and to compare to the analogous compositions previously reported [19,20,21].

2. Materials and Methods

The Ti_0.30V_0.25Zr_0.10Nb_0.25Mo_0.10 alloy was synthesized by high temperature arc melting under Ar atmosphere. The sample was turned and remelted 10 times to enhance the chemical homogeneity.

Structural characterizations were carried out by X-ray powder diffraction using both laboratory and synchrotron sources. The former investigation was carried out by the help of a D8 advance Bruker diffractometer (Cu Kα radiation, Bragg–Brentano geometry). Synchrotron X-ray powder diffraction (SR-XRD) was performed at the Cristal beamline in SOLEIL facility (λ = 0.7289 Å). A capillary tube of 0.2 mm of diameter was filled with finely grounded hydride powder. The acquisition time was approximately 10 min for a scanning range from 1° to 85° (2θ). High-energy synchrotron XRD measurements were carried out ID15A beamline at ESRF, Grenoble (France) [22]. Incident flux was normalized using a diode placed in front of the sample. Diffraction patterns were collected in transmission geometry by using a Pilatus3 X CdTe 2M hybrid photon counting detector. A number of diffraction patterns were collected at slightly different detector positions for each sample, in order to minimize inhomogeneity effects and cover dead zones. The detector was off-centered with respect to the 99 keV incident beam and positioned close to the sample to access q values up to ~ 30 Å⁻¹. The images were corrected for detector geometry, response, distortion, transparency and for the background contributions and X-ray polarization. Data were azimuthally integrated using a modified version [23] of the PyFAI library [24] to give one dimensional scattering patterns. The G(r) were calculated from these patterns using routines from the Diffpy-CMI library [25] with local modifications for outlier rejection and treatment of background. The structural analysis was carried out using the PDFgui free software [26].

In situ and ex situ powder neutron diffraction (nD) experiments on a deuterated sample were performed on the D1B beamline at the Institute Laue-Langevin (λ = 1.2858 Å) with a scanning range from 1 to 128° (2θ°) DOI: 10.5291/ILL.DATA.CRG-2768. The ex situ measurements were performed using a vanadium container while the in situ measurements were performed on a silica tube containing the deuterated powder sample. The silica tube was connected to a secondary vacuum and the desorption step was carried out with a constant heating rate with 1 °C/min while the evolved gas pressure was recorded by the help of a vacuum gauge.

For the sake of comparison, X-ray powder diffraction, total X-ray scattering at synchrotron and neutron powder diffraction experiments have been also recorded for the quaternary alloy Ti_0.325V_0.275Zr_0.125Nb_0.275 under the same conditions.

The X-ray and neutron powder diffraction data were refined using the Rietveld method with Fullprof software [27]. The peak shape was described by Thompson-cox Hastings Pseudo Voigt function.

The microstructure of the alloy before and after cycling was analyzed by Scanning Electron Microscopy (SEM) acquiring back-scattered electron signal (BSE) using a Zeiss Merlin microscope. The chemical mapping of the sample was performed by energy dispersive X-ray spectroscopy (EDS). Prior to the microstructure measurement, the powder sample was embedded in a resin and coated with 1.9 nm of Pd.

Hydrogen sorption properties were determined by measuring the Pressure–Composition Isotherms (PCI) at several temperatures, the kinetics of absorption at room temperature and the desorption profile by thermal desorption spectroscopy (TDS).

The PCI curves were recorded using a homemade volumetric device thermostated at 25 °C. The sample was loaded in a stainless-steel container and activated by a heat treatment at 350 °C under dynamic vacuum for 3 h. For the calculation of the hydrogen uptake, the real equation of state for hydrogen was acquired from the program GASPAK V3.32 (Cryodata. Inc., USA). Sequential absorption/desorption cycling measurements were performed using the same volumetric device. The absorption was performed at 50 bar H₂ pressure at 25 °C, while the desorption was carried out by heating at 450 °C under dynamic vacuum for 12 h. The kinetic curves were recorded only during the absorption step.

Thermal desorption spectroscopy (TDS) was used to determine the hydrogen desorption properties. Prior to the measurement, the sample was loaded in a quartz tube and evacuated under secondary vacuum. During desorption with a temperature ramp of 5 °C/min up to 445 °C, a mass spectrometer was used to record the evolved gases partial pressures and to follow the H₂ desorption profile [28].

3. Results and Discussions

The composition Ti_0.30V_0.25Zr_0.10Nb_0.25Mo_0.10 possesses the following empirical parameters: VEC = 4.7 and δ = 5.17%. The data for the calculation of δ were taken from the reference [29].

The XRD of the as-cast material and the related Rietveld analysis indicated that the alloy crystallizes in a simple bcc type structure (space group Im$\overline{3}$m) with the unit cell $a_{bcc}$ = 3.240 (3) Å (Figure 1). For comparison, the lattice parameter of the single phase bcc quaternary alloy Ti_0.325V_0.275Zr_0.125Nb_0.275 prepared by the same method is higher than the present quinary composition, $a_{bcc}$ = 3.261 (1) Å [18]. Thus, the addition of 10 at.% Mo to Ti_0.325V_0.275Zr_0.125Nb_0.275 reduces the bcc lattice parameter. This can be explained by the smaller atomic radius of Mo as compared to the other elements, except for V: r_V = 1.31 Å < r_Mo = 1.36 Å < r_Nb = 1.43 Å < r_Ti = 1.45 Å < r_Zr = 1.59 Å. The average of radius of the quaternary composition is $\overline{r}$= 1.445 Å and decreases after the addition of 10 at.% Mo to $\overline{r}=$ 1.428 Å, in line with the observed lattice parameter reduction. Moreover, the δ parameter decreases from 6% to 5.17% upon addition of 10 at.% of Mo into the quaternary alloy Ti_0.325V_0.275Zr_0.125Nb_0.275.

The chemical homogeneity of the as-cast alloy was investigated by SEM-EDS (Figure 2).

The SEM-EDS analysis suggests a dendritic microstructure with small Nb and Mo enrichment, whereas the Zr concentration is slightly higher in the interdendritic regions (

Table 1. Chemical composition of the as-cast alloy in the overall, dendritic and interdendritic regions.

Region	Ti (at.%)	V (at.%)	Zr (at.%)	Nb (at.%)	Mo (at.%)
Dendritic	28.2 (0.5)	22.9 (0.2)	6.2 (0.4)	29.8 (0.7)	12.9 (0.4)
Interdendritic	30.4 (0.4)	25.5 (0.7)	23.6 (1.7)	15.9 (2.4)	4.9 (1.2)
Overall	30.1 (0.7)	24.8 (0.5)	10.3 (1.4)	25.1 (1.4)	9.7 (0.8)
Nominal	30	25	10	25	10

). The Ti and V are evenly distributed over both zones. Despite this dendritic microstructure occurring during the solidification of HEA, the overall composition is close to the nominal (Table 1).

Both XRD and SEM-EDS confirmed that the as-cast sample crystalizes in a single-phase bcc lattice with an almost homogeneous chemical distribution.

The hydrogen sorption properties were studied using Sievert’s method. The absorption PCI curves of the Ti_0.30V_0.25Zr_0.10Nb_0.25Mo_0.10 alloy were recorded at different temperatures (in the range 25 °C to 300 °C) to explore the effect of the temperature on the sorption properties (Figure 3).

All PCI curves show a single plateau at very low pressure, below the accuracy of the pressure transducer (<10⁻² bar), irrespective of the temperature. No obvious signature of a two-step transition is noticed, as commonly encountered in the conventional bcc alloys, [30] or less often in some refractory HEAs [14]. The maximum hydrogen uptake at around 50 bar decreases with the increasing temperature from 2.0 H/M (2.8 wt.%) to around 1.0 H/M (1.4 wt.%) with increasing the temperature from 25 °C to 300 °C. The absence of measurable plateau pressure with increasing the temperature hints to a very stable hydride phase, as also confirmed by our recent thermodynamic modelling for bcc multicomponent alloys (see calculations details in Part SI-1) [31].

The quaternary Ti_0.325V_0.275Zr_0.125Nb_0.275 has a maximum hydrogen capacity of 1.75 H/M (2.47 wt.%) [18] while the quinary composition Ti_0.30V_0.25Zr_0.10Nb_0.25Mo_0.10 reaches a maximum hydrogen absorption capacity of 2.0 H/M (2.8 wt.%). Therefore, the addition of 10% of Mo to the quaternary Ti_0.325V_0.275Zr_0.125Nb_0.275 increases the maximum amount of the hydrogen uptake by 0.25 H/M reaching full capacity of the dihydride phase. Our previous studies on the effect of the addition of 10 at.% of Mg [20], Al [19] and Ta [21] in the same quaternary alloy demonstrated that only Ta shows the same beneficial effect as Mo on the capacity, whereas Mg and Al additions decrease the maximum uptake from 1.75 H/M to 1.72 H/M and 1.6 H/M for Mg and Al, respectively. A comparison with the values of maximum capacity, as expressed in wt.%, from previous results is given in

Table 2. Comparison of hydrogen storage properties of several refractory bcc HEAs.

Composition	Alloy	Hydride	Capacity (wt.%)	Reference
Ti0.30V0.25Zr0.10Nb0.25Mo0.10	bcc	fcc	2.8	present study
Ti0.325V0.275Zr0.125Nb0.275	bcc	fcc	2.47	[18]
Ti0.30V0.25Zr0.10Nb0.25Al0.10	bcc	bct	2.6	[19]
Ti0.30V0.25Zr0.10Nb0.25Mg0.10	bcc	fcc	2.7	[20]
Ti0.30V0.25Zr0.10Nb0.25Ta0.10	bcc	fcc	2.5	[21]
TiVZrNbHf	bcc	bct	2.7	[9]
TiVNb	bcc	fcc	3	[12]
TiVNbTa	bcc	fcc	2	[12]
TiVNbCr	bcc	fcc	3.1	[12]
TiZrNbHfTa	bcc	fcc	1.7	[14]

.

In terms of gravimetric capacity, the studied quinary composition has a higher capacity of absorption than other bcc alloys as TiV₂ (2.6 wt.%) [32], classical intermetallics such as, LaNi₅ (1.4 wt.%) [32] or other HEAs like TiVNbTa (2 wt.%) [12] and TiZrNbHfTa (1.7 wt.%) [14].

The crystalline structure of the dihydride phase Ti_0.30V_0.25Zr_0.10Nb_0.25Mo_0.10H₂ was studied by synchrotron X-ray powder diffraction (SR-XRD) (Figure 4a). The Rietveld analysis on SR-XRD patterns of the dihydride shows a single-phase hydride with a fcc structure (CaF₂-type with Fm$\overline{3}$m) and lattice parameter $a_{fcc}$= 4.459 (1) Å (

Table 3. The values of the fcc lattice constant for the Ti_0.325V_0.275Zr_0.125Nb_0.275H_1.7 and Ti_0.30V_0.25Zr_0.10Nb_0.25Mo_0.10H₂ hydrides from SR-XRD (CRISTAL at SOLEIL synchrotron), nD (D1B at ILL) and PDF analysis (ID15A at ESRF synchrotron). The mean squared displacement of elements, U_iso, is obtained from PDF modelling.

Sample	FCC Lattice Parameter, a (Å)			Uiso(Ų)
	From SR-XRD	From nD	From PDF	From PDF
Ti0.325V0.275Zr0.125Nb0.275H1.7	4.478 (1)	-	4.482 (2)	0.009 (1)
Ti0.325V0.275Zr0.125Nb0.275D1.7	-	4.475 (1)	-	-
Ti0.30V0.25Zr0.10Nb0.25Mo0.10H2	4.459 (1)	-	4.466 (2)	0.014 (5)
Ti0.30V0.25Zr0.10Nb0.25Mo0.10D2	-	4.450 (1)	-	-

). For complementary structural investigation of the hydride ex situ neutron powder diffraction was performed (Figure 4b). The deuterium was preferred to hydrogen in this experiment because of the high incoherent scattering of hydrogen atoms causing significant background. The fcc lattice parameter by nD is $a_{fcc}$= 4.450 (1) Å (Table 3), in good agreement with SR-XRD results. This technique of characterization enables us to locate the position of deuterium atoms inside the fcc lattice of the deuteride. Using the Rietveld refinement, the deuterium atoms were located at the tetrahedral sites. This result is in agreement with previous studies on similar HEAs [10]. The SR-XRD and nD patterns together with the corresponding Rietveld refinements for Ti_0.325V_0.275Zr_0.125Nb_0.275H_1.7 are shown in Figures SI-2 and SI-3, while the obtained values for the fcc lattice parameter are listed in Table 3.

The short and medium-range orders of the full hydride phase were investigated by X-ray total scattering and related PDF analysis. PDF is described with the G(R) formalism, which indicates the probability of finding a pair of atoms separated by a distance R. The PDF obtained from Fourier transformation of the total scattering data shows the local structure of the hydride in terms of real-space distribution of interatomic distances. The modelling of the hydride phase was carried out using a fcc lattice with random distribution of all atoms on the crystallographic sites. For the sake of comparison, PDF data and related structural fit are compared to the quaternary hydride, see Figure 5.

The PDF data of Ti_0.325V_0.275Zr_0.125Nb_0.275H_1.7 and Ti_0.30V_0.25Zr_0.10Nb_0.25Mo_0.10H₂ hydrides have almost the same profile proving similar local order despite the addition of 10 at.% of Mo into the initial quaternary alloy (Figure 5a). The difference curves between experimental and calculated PDF (Figure 5b) suggest that the undistorted random fcc model can well describe both the medium structure in the high-R region and the local structure in the low-R range, irrespective of the composition. Therefore, the average random fcc structure is able to explain all observed features in the PDF data without any evidence of local ordering or strong lattice distortion at short-range. Previous reports revealed that deviations from average structure are often considerable for R < 4 Å especially, in the presence of large elements, such as Zr and Hf [33]. However, this is not the case presently, and we hypothesize that the small Zr content (12.5 at.% and 10 at.% for the quaternary and quinary alloys, respectively) might be responsible for the formation of an average structure with random atoms distribution and small lattice distortion. Moreover, previous literature clearly demonstrated by combined X-ray and neutrons PDF analysis that the amount of lattice distortion is relieved when hydrides with CaF₂-type structures are formed from the bcc refractory HEAs [34].

The refined fcc lattice parameters (a) from PDF modelling and the structural refinements of SR-XRD and nD data are listed in Table 3 together with the mean squared displacement of elements, U_iso, as obtained from PDF analysis. The values of the fcc lattice parameter from PDF are slightly larger than the ones from the diffraction assessments, irrespective of composition. Interestingly, U_iso seems to increase by the addition of a fifth element, as expected, although this is within the error bar.

Further investigations were made to study the hydrogen desorption properties from the Ti_0.30V_0.25Zr_0.10Nb_0.25Mo_0.10H₂.

The thermo-desorption spectroscopy was performed on the deuteride Ti_0.30V_0.25Zr_0.10Nb_0.25Mo_0.10D₂ (Figure 6a). The onset temperature is around 150 °C marking the beginning of a plateau-like desorption followed by a main peak at 256 °C indicating the maximum desorption rate and finally a second plateau-like desorption up to 400 °C. The TDS profile of Ti_0.325V_0.275Zr_0.125Nb_0.275 is relatively different from the quaternary alloy which is composed of one main peak with an onset temperature at 250 °C and a maximum desorption rate at 287 °C [18,21].

The addition of 10 at.% Mo to the quaternary composition decreases the temperature of the maximum desorption rate from 287 °C to 256 °C and the onset temperature from 250 °C to 150 °C. Our previous studies proved that the addition of 10 at.% of other elements to the quaternary Ti_0.325V_0.275Zr_0.125Nb_0.275 results in a modification of TDS profile. For example, the desorption profile of Mg_0.10Ti_0.30V_0.25Zr_0.10Nb_0.25 is similar to the quaternary alloy with one main desorption peak at relatively high temperature [20], whereas for Al_0.10Ti_0.30V_0.25Zr_0.10Nb_0.25 the TDS profile drastically changes with a main desorption peak at low temperature (~130 °C) followed by a small and broad peak at high temperature [19]. The desorption profile is also modified for Ta_0.10Ti_0.30V_0.25Zr_0.10Nb_0.25 with two main desorption events at low and high temperatures (~160 °C and 260 °C, respectively) [21]. Thus, the desorption properties change drastically depending on the nature of the fifth element into the initial quaternary composition.

To better understand the reversibility hydrogen absorption and desorption for this composition, we investigated the crystalline structure of the desorbed sample after the TDS measurement. The XRD pattern and the corresponding Rietveld refinement of the desorbed sample (Figure 6b) shows a full recovery of the initial bcc structure with a lattice parameter $a_{bcc}$= 3.235(2) Å close to the initial value. No phase segregation nor decomposition could be noticed by desorption proving complete reversibility under these conditions.

To further investigate the structural phase transition accompanying the hydrogen desorption, we carried out in situ neutron diffraction (Figure 7).

The deuterium desorption from the Ti_0.30V_0.25Zr_0.10Nb_0.25Mo_0.10D₂ occurs within three steps, as shown from the vacuum pressure reading during in situ neutron diffraction (Figure 7 right), in agreement with the TDS profile. First event is a plateau-like desorption followed by a major peak at 242 °C and finally a broad desorption is noticed at high temperature.

The initial phase corresponding to the fcc deuteride (marked with #) is stable with rising the temperature up to around 200 °C where the peaks intensity decreases. Surprisingly, the first desorption event starting at around 150 °C is not accompanied by any noticeable phase transition or, at least, loss in the peak’s intensity. Further increase in temperature triggers a sharp phase transition at 242 °C (in agreement with the TDS experiment) from fcc deuteride to bcc phase (marked with *). The diffraction signals of the desorbed bcc phase are not intense due to the low thermal neutron cross section of the elements in the composition, in agreement with our previous studies [19,20,21]. Furthermore, this composition might be a good candidate for an in-core, light water reactor environment [35].

These experiments confirmed that the hydrogen absorption and desorption reactions are fully reversible in Ti_0.30V_0.25Zr_0.10Nb_0.25Mo_0.10 and occur within a single-step transformation from bcc ↔ fcc, as also proposed for the quaternary alloy [21]. Therefore, it will be interesting to study the hydrogen absorption/desorption stability properties for 20 cycles (Figure 8a). For each absorption step at room temperature the kinetics curves were measured (Figure 8b). The desorption was realized by heating at high temperature under dynamic vacuum.

The quaternary alloy synthesized by arc melting loses capacity from 1.8 H/M (2.7 wt.%) to 1.5 H/M (2.3 wt.%) after the first four cycles followed by stabilization [18]. A similar drop of the capacity is also observed for the quinary alloy from 2.0 H/M (2.8 wt.%) to 1.7 H/M (2.47 wt.%) followed by stabilization (Figure 8a). Fading of the capacity during cycling is currently observed in many HEAs, although this issue is still not clearly understood [36,37]. Compositions with the refractory elements were also studied previously for their cycling properties, for example, Nygård et al. stated that TiVNbTa alloy has a capacity of 0.85 H/M after the first cycle but the reversible capacity is reduced to zero after only 5 cycles [12]. Montero et.al. studied the cycling properties of Al_0.10Ti_0.30V_0.25Zr_0.10Nb_0.25 alloy which proves an increased reversible gravimetric capacity as compared to the quaternary alloy from 2.25 wt.% to 2.45 wt.%, respectively, i.e., cycling stability boosts from 85% to 94% by adding Al [19].

In conclusion, the effect of 10% addition of Mo into the quaternary alloy increases the capacity of the first cycle and the reversible uptake relative to the initial composition.

The capacity loss could be explained by both extrinsic and intrinsic factors. The former ones include vacuum or gas contaminants that can affect the surface of the sample either by oxidation or adsorption of hydrocarbons on the surface thus reducing the volume of the alloy able to react with hydrogen. The intrinsic factors comprise the phenomenon of phase disproportionation accompanied by the formation of stable monohydride phases such as TiH₂, VH₂, the inhomogeneities in the chemical composition, the kinetic issues that may lead to incomplete absorption or desorption and the decrease in the degree of crystallinity.

The absorption kinetics (Figure 8b) indicate that around 90% of the full capacity is reached within 30 s on the first cycle and 20 s for further cycling. This behavior can be explained by decrepitation of the material during cycling: large and clean surfaces are created to react and dissociate hydrogen molecules together with faster hydrogen atoms diffusion into smaller grains of the alloy.

To further investigate the effect of cycling on the microstructure and the chemical homogeneity we have performed SEM-EDS after cycling (Figure 9).

The BSE image shows the effect of hydrogen embrittlement on the material with the formation of numerous cracks. The EDS chemical mapping shows the same homogenous distribution of the elements as the initial sample. The same microstructure as the as-cast material is observed despite cycling: Nb- and Mo-rich dendritic zones coexist with Zr-rich interdendritic regions.

The hydrogen cycling might affect the crystalline structure of the alloy. To assess this effect XRD and Rietveld analysis have been carried out for the cycled sample (Figure 10).

These results show that the cycled sample is single phased with a bcc lattice, the same as the initial material. Consequently, no phase segregation is noticed after cycling that might explain the initial fading of the capacity during cycling. The lattice parameter of the cycled alloy is $a_{bcc}$ = 3.219 (4) Å, which is slightly smaller than the as-cast sample ($a_{bcc}$ = 3.240 (3) Å). However, after 20 cycles, a peak broadening occurs as also demonstrated by the increase of the full width at half maximum (FWHM) listed in

Table 4. Lattice parameter, FWHM of (110), (200) and (211) diffraction peaks for the as-cast sample and the desorbed sample after 20 cycles.

	As-Cast		After 20-Cycles
a (Å)	3.240 (3)		3.219 (4)
Miller Indices	2θ (°)	FWHM (°)	2θ (°)	FWHM (°)
(110)	39.54	0.642	39.57	1.286
(200)	57.21	1.017	57.39	1.620
(211)	71.49	0.993	71.65	2.314

for the three main diffraction peaks. This peak broadening might be responsible for the discrepancy in the lattice parameter via enhanced uncertainties in the lattice determination. Moreover, peak broadening might suggest a slight loss of the degree of crystallinity that may account for the initial loss of capacity during cycling, although this latter issue is still under debate.

4. Conclusions

A single-phased bcc alloy with composition Ti_0.30V_0.25Zr_0.10Nb_0.25Mo_0.10 was prepared by arc melting with excellent chemical homogeneity, as confirmed by XRD and SEM-EDS. This alloy rapidly absorbs hydrogen at room temperature forming a hydride phase with maximum 2.0 H/M capacity. The phase transition from metal to hydride seems to occur within a single step, irrespective of the temperature. The hydride phase was thoroughly characterized by synchrotron X-ray scattering complemented by neutron diffraction. Synchrotron X-ray and neutron diffraction reveal the formation of a single-phased fcc hydride whereas, at more local level, PDF analysis points out that the hydride phase possesses an average fcc structure with random atoms distribution and small lattice distortion. Hydrogen atoms occupy the tetrahedral interstitial sites of the fcc hydride phase. The desorption reaction takes place at high temperature. The reaction with hydrogen is fully reversible since the initial bcc phase is recovered after absorption/desorption cycling. The in situ neutron diffraction experiment during deuterium desorption seems to indicate a single step phase transition from fcc hydride to initial bcc lattice. Despite an initial fading in the first cycles, the absorption capacity stabilizes to around 1.7 H/M (2.47 wt.%) for further cycling. Cycling is accompanied by a small loss of crystallinity that might explain the initial capacity decrease. Moreover, the cycling preserves the alloy and hydride structures without introducing phase segregation or chemical inhomogeneities and increases the absorption reaction kinetics at room temperature. The findings here demonstrate that the addition of 10 at.% of Mo into the quaternary Ti_0.325V_0.275Zr_0.125Nb_0.275 alloy improves the hydrogen storage performances. Owing to the vast possibilities of compositions these insights can help the progress of the research for the design of novel and performant HEAs as solid-state stores.