Dehydriding Process and Hydrogen–Deuterium Exchange of LiBH₄–Mg₂FeD₆ Composites

Abstract

The dehydriding process and hydrogen–deuterium exchange (H–D exchange) of xLiBH₄ + (1 − x)Mg₂FeD₆ (x = 0.25, 0.75) composites has been studied in detail. For the composition with x = 0.25, only one overlapping mass peak of all hydrogen and deuterium related species was observed in mass spectrometry. This implied the simultaneous dehydriding of LiBH₄ and Mg₂FeD₆, despite an almost 190 °C difference in the dehydriding temperatures of the respective discrete complex hydrides. In situ infrared spectroscopy measurements indicated that H–D exchange between [BH₄]⁻ and [FeD₆]⁴⁻ had occurred during ball-milling and was promoted upon heating. The extent of H–D exchange was estimated from the areas of the relevant mass signals: immediately prior to the dehydriding, more than two H atoms in [BH₄]⁻ was replaced by D atoms. For x = 0.75, H–D exchange also occurred and about one to two H atoms in [BH₄]⁻ was replaced by D atoms immediately before the dehydriding. In contrast to the situation for x = 0.25, firstly LiBH₄ and Mg₂FeD₆ dehydrided simultaneously with a special molar ratio = 1:1 at x = 0.75, and then the remaining LiBH₄ reacted with the Mg and Fe derived from the dehydriding of Mg₂FeD₆.

1. Introduction

The complex hydride LiBH₄, consisting of Li⁺ cations and [BH₄]⁻ complex anions, has a high gravimetric hydrogen density of 18.4 mass% and a volumetric hydrogen density of 121 kg H₂/m³ [1]. The main issues to be resolved for developing LiBH₄ as a hydrogen storage material are lowering of its high dehydriding temperature of >420 °C and moderating the harsh rehydriding conditions of 35 MPa H₂ and high temperatures above 600 °C [2]. Many attempts have been made to improve the dehydriding properties of LiBH₄ by incorporating various additives, confining within nanoporous materials, or by preparing reactive composites with metal hydrides, and so on [3,4,5,6,7,8,9,10,11,12].

Recently, we reported that the dehydriding temperature of LiBH₄ is distinctly decreased upon combination with the complex hydride Mg₂FeH₆ composed of Mg²⁺ cations and [FeH₆]⁴⁻ complex anions [13]. For example, the dehydriding temperature of LiBH₄ in xLiBH₄ + (1 − x)Mg₂FeH₆ composite with x = 0.5 is 350 °C, which is 100 °C lower than that of pure LiBH₄.

Besides this decreased dehydriding temperature of LiBH₄, a unique dehydriding process was identified. In thermal gravimetry–mass spectrometry measurements (TG–MS) of xLiBH₄ + (1 − x)Mg₂FeH₆ (0.1 ≤ x ≤ 0.83) composites, when x ≤ 0.5, only one MS peak was observed. When x > 0.5, more than two MS peaks were observed. Moreover, over the entire composition range, the dehydriding temperature of LiBH₄ decreased almost linearly with the proportion of Mg₂FeH₆. These results suggested that when x ≤ 0.5, Mg₂FeH₆ and LiBH₄ dehydrided simultaneously, despite the almost 190 °C difference in the dehydriding temperatures of the respective discrete complex hydrides. Conversely, when x > 0.5, firstly Mg₂FeH₆ dehydrided to Mg and Fe, and then LiBH₄ dehydrided by reacting with the Mg and/or Fe formed.

Several studies on LiBH₄‒rich compositions of xLiBH₄ + (1 − x)Mg₂FeH₆ (x > 0.65) composites have been published. The dehydriding process of 0.8LiBH₄ + 0.2Mg₂FeH₆ was investigated by Langmi et al. [14] and Deng et al. [15]. Both of them reported that the dehydriding temperature of LiBH₄ was decreased by combining with Mg₂FeH₆ by the following two-step dehydriding process. Firstly, Mg₂FeH₆ dehydrided to Mg and Fe, and then LiBH₄ reacted with the Mg and Fe formed. The boron in LiBH₄ was stabilized by Mg and Fe to form MgB₂ and FeB. Ghaani et al. [16] reported the destabilized thermodynamics of 2/3LiBH₄ + 1/3Mg₂FeH₆ composites when compared with that of the respective discrete complex hydrides. In these studies, neither the linear variation in the dehydriding temperature with changing composition nor the evidence of simultaneous dehydriding of LiBH₄ and Mg₂FeH₆ has apparently been noticed. Therefore, clarifying the unique dehydriding process is important to gain a deep knowledge of such composites of complex hydrides.

Because the hydrogen released from LiBH₄ and Mg₂FeH₆ cannot be differentiated in MS measurements, the evidence for simultaneous dehydriding when x ≤ 0.5 was not conclusive and the assignment of the multiple MS peaks observed to the dehydriding of LiBH₄ and Mg₂FeH₆ when x > 0.5 was not unequivocal. In this study, instead of Mg₂FeH₆ we have used Mg₂FeD₆ to prepare xLiBH₄ + (1 − x)Mg₂FeD₆ composites with two compositions, x = 0.25 and 0.75, in the expectation of distinguishing the dehydriding processes of the respective components by the MS signals of H₂ and D₂. During the actual measurement, H–D exchange between LiBH₄ and Mg₂FeD₆ was observed, and the relationship between this H–D exchange and the simultaneous dehydriding processes is discussed herein.

2. Results and Discussion

2.1. Dehydriding Property of 0.25LiBH₄ + 0.75Mg₂FeD₆

The TG–MS profile of 0.25LiBH₄ + 0.75Mg₂FeD₆ is shown in Figure 1. The dehydriding profile was almost the same as that of 0.25LiBH₄ + 0.75Mg₂FeH₆, in which only one MS peak was observed, implying that both LiBH₄ and Mg₂FeH₆ were dehydrided, as explained in the Introduction. Taking the purity of Mg₂FeD₆ into consideration, the experimental weight loss of 9.5 wt% was in reasonable agreement with the theoretical weight loss (10.5 wt%) for full dehydriding of LiBH₄ and Mg₂FeD₆ and the general reaction equation may be as follows:

0.25LiBH₄ + 0.75Mg₂FeD₆ → 1.5Mg + 0.75Fe + 0.25Li(H,D) + 0.25B + 2.625(H,D)₂

(1)

Actually, XRD analysis of the sample collected after TG measurement (400 °C), as illustrated in Figure S1 in the Supplementary Material, showed that the dehydrided products contained Mg, Fe, Fe₂B, and LiH(D).

Figure 1. TG–MS profile of 0.25LiBH₄ + 0.75Mg₂FeD₆ (solid line) and 0.25LiBH₄ + 0.75Mg₂FeH₆ (dash line). m/e = 2, 3, and 4 are signals of H₂, HD, and D₂, respectively.

Figure 1. TG–MS profile of 0.25LiBH₄ + 0.75Mg₂FeD₆ (solid line) and 0.25LiBH₄ + 0.75Mg₂FeH₆ (dash line). m/e = 2, 3, and 4 are signals of H₂, HD, and D₂, respectively.

In the MS measurement, peaks due to H₂ and D₂ were observed at the same temperature of 325 °C. This proved our expectation that LiBH₄ and Mg₂FeD₆ dehydrided simultaneously. Besides the H₂ and D₂ peaks, a peak due to HD was also observed, indicative of H–D exchange between LiBH₄ and Mg₂FeD₆.

To prove this H–D exchange, in situ IR spectra were recorded, and the results are shown in Figure 2. As shown in the spectrum at 25 °C, H–D exchange has already occurred during ball-milling process. Referring to theoretical data, for isotopically pure LiBH₄, a BH stretching peak at ν ≈ 2350 cm⁻¹ and bending peaks at ν ≈ 1300 and 1100 cm⁻¹ should be observed [17]; for isotopically pure Mg₂FeD₆, an FeD stretching peak should appear at ν ≈ 1260 cm⁻¹ [18]. The experimentally observed peak at ν = 2340 cm⁻¹ at 25 °C was assigned to the BH stretching mode, and that at ν = 1310 cm⁻¹ was assigned to the FeD stretching mode. The missing BH bending peak and the broadened peak of the BH stretching mode indicate that the symmetry of [BH₄]⁻ had been broken owing to the part replacement of H atoms by D atoms [19]. The peak at ν ≈ 1840 cm⁻¹ was assigned as the FeH stretching mode in Mg₂FeD₅H [18]. A very broad peak in the region ν = 1600–1900 cm⁻¹ was considered to be due to merged FeH and BD stretching peaks [18,20]. It was difficult to distinguish the FeH stretching peak at ν = 1794 cm⁻¹ and the BD stretching peak at ν = 1775 cm⁻¹ within the spectral resolution.

Figure 2. In situ IR spectra of ball-milled 0.25LiBH₄ + 0.75Mg₂FeD₆. The heating rate was 5 °C/min. The atmosphere in the sample holder was 0.1 MPa Ar without gas flow.

Figure 2. In situ IR spectra of ball-milled 0.25LiBH₄ + 0.75Mg₂FeD₆. The heating rate was 5 °C/min. The atmosphere in the sample holder was 0.1 MPa Ar without gas flow.

When we take a look at the IR spectra during the heating process, the area of the BD/FeH stretching peak did not change discernibly when compared to that of the steadily diminishing BH stretching peak, although the whole peak intensities weakened upon heating due to deterioration of the optical focusing by the thermally expanding sample. In addition, the FeD stretching peak at ν = 1310 cm⁻¹ was shifted to higher wavenumber and broadened significantly during the heating process. Referring to previous H–D exchange studies of LiBH₄ and Mg(BH₄)₂ [19,20,21], these results suggest that H–D exchange was promoted during the heating process. At 325 °C, all of the peaks faded as a result of the dehydriding reaction, consistent with the TG–MS measurements shown in Figure 1. The peaks disappeared and the spectrum did not change further at 350 °C, suggesting completion of the dehydriding reaction.

To further assess the extent of H–D exchange, we attempted a quantitative analysis based on the areas of MS signals that directly related to the amount of gas released. If it is supposed that H and D atoms are firstly released from H–D exchanged Mg₂FeH_y/3D_6–y/3 and LiBH_4–yD_y, and then mix and combine freely to form H₂, HD, and D₂ gas molecules, then statistically the area ratio of the MS signals of H₂, HD, and D₂ should be 25:12:1. In fact, the experimentally measured area ratio of these MS signals was 9:5:1, quite different from the statistical distribution. Thus, the processes of H–D exchange and dehydriding need to be interpreted differently. Here, we suppose that:

(a)

H–D exchange occurred during ball-milling and the heating process but stopped as soon as the dehydriding started;

(b)

H₂, HD, or D₂ molecules were directly released from either Mg₂FeH_y/3D_{6 − y/3} or LiBH_{4 − y}D_y; H or D atoms derived from the two different complex hydrides cannot combine to form gas molecules.

Based on this assumption, the extent of H–D exchange was estimated from the area ratio of MS signals. The result shows that the extent of H–D exchange, y, was around 2.5 immediately prior to the onset of dehydriding:

0.25LiBH₄ + 0.75Mg₂FeD₆ → 0.25LiBH_{4 − y}D_y + 0.75Mg₂FeH_y/3D_{6 − y/3} (y ≈ 2.5)

(2)

According to the estimation, less than one D atom in [FeD₆]⁴⁻ was replaced by H atom before the dehydriding and more than two H atoms in [BH₄]⁻ were replaced by D atoms. This result is in good agreement with the IR spectra: even though a shift to higher wavenumber was observed, the area of the FeD stretching peak did not decrease markedly because the replacement in [FeD₆]⁴⁻ was slight. Conversely, the BH stretching peak at ν ≈ 2320 cm⁻¹ was weakened and the BD stretching peak at ν ≈ 1740 cm⁻¹ was significantly intensified since the symmetry of [BH₄]⁻ was severely disrupted.

2.2. Dehydriding Property of 0.75LiBH₄ + 0.25Mg₂FeD₆

In contrast to the situation for 0.25LiBH₄ + 0.75Mg₂FeD₆, multiple MS peaks were observed for the dehydriding process of 0.75LiBH₄ + 0.25Mg₂FeD₆. The TG–MS profile is shown in Figure 3, together with that of 0.75LiBH₄ + 0.25Mg₂FeH₆ as a reference. The total weight loss was 9.0 wt%, which suggests full dehydriding of both LiBH₄ and Mg₂FeD₆. The dehydrided products were confirmed as Mg, FeB₂, and LiH(D) by XRD analysis, as shown in Figure S2 in the Supplementary Material.

Figure 3. TG–MS profile of 0.75LiBH₄ + 0.25Mg₂FeD₆ (solid line) and 0.75LiBH₄ + 0.25Mg₂FeH₆ (dash line).

Figure 3. TG–MS profile of 0.75LiBH₄ + 0.25Mg₂FeD₆ (solid line) and 0.75LiBH₄ + 0.25Mg₂FeH₆ (dash line).

For the 0.25LiBH₄ + 0.75Mg₂FeD₆ composition, In situ IR confirmed that H-D exchange between LiBH₄ and Mg₂FeD₆ occurred during ball-milling and was promoted during the heating process, as shown in Figure S3 in the Supplementary Material. With the same premises as in the case of the 0.25LiBH₄ + 0.75Mg₂FeD₆ composition, the extent of H-D exchange immediately prior to dehydriding was estimated as y ≈ 4.5 from the area ratio of the MS signals (sum of the two peaks) of H₂, HD, and D₂. This result shows that more than one H atom in [BH₄]⁻ was replaced by D:

0.75LiBH₄ + 0.25Mg₂FeD₆ → 0.75LiBH_{4 − y/3}D_y/3 + 0.25Mg₂FeH_yD_{6 − y} (y ≈ 4.5)

(3)

The slope of the TG profile changed and the MS curves separated at around 370 °C. The weight loss was 5.4 wt% before the inflection temperature followed by 3.6 wt% until completion of the dehydriding. Considering the estimated extent of HD exchange, the weight loss indicated that 0.25LiBH_{4 − y/3}D_y/3 + 0.25Mg₂FeH_yD_{6 − y} dehydrided before the inflection temperature and then the residue 0.5LiBH_{4 – y/3}D_y/3 dehydrided. The area ratio of the first MS peak was 14:7:1, and this changed to 8:6:1 for the second MS peak. This result supports the interpretation that the first MS peak corresponds to the simultaneous dehydrogenation of isotopically exchanged LiBH₄ and Mg₂FeD₆ and the second MS peak corresponds to the dehydriding of LiBH_{4 − y/3}D_y/3. Therefore, it is evident that even though multiple MS peaks were observed, LiBH₄ and Mg₂FeD₆ were still dehydrided simultaneously. Following on from our previous report on the dehydriding properties of other compositions of xLiBH₄ + (1 − x)Mg₂FeH₆ composites, when x ≤ 0.5 (molar ratio of LiBH₄:Mg₂FeD₆ = 1:1), only one MS peak is observed [13]. It can be surmised that the molar ratio 1:1 is a special composition: LiBH₄ can dehydride simultaneously with Mg₂FeH₆/Mg₂FeD₆ up to this molar ratio; if there is more LiBH₄ in the composite, the residual LiBH₄ will subsequently dehydride by reacting with Mg and Fe derived from the dehydriding of Mg₂FeH₆/Mg₂FeD₆.

3. Experimental Section

Mg₂FeD₆ was synthesized by pressing a 2Mg + Fe mixture into pellets and subjecting it to heat treatment at 400 °C for 20 h under 3 MPa D₂. The product yield was 91% according to TG measurement and the isotopic purity was almost 100% according to the MS measurement. Mg₂FeD₆ was then mixed with LiBH₄ (95%, Aldrich, St. Louis, MO, USA) and xLiBH₄ + (1 − x)Mg₂FeD₆ composites with compositions x = 0.25 and 0.75 were prepared by planetary ball-milling (Fritsch P-5, Fritsch, Idar-Oberstein, Germany) for 5 h under argon.

The dehydriding properties were examined by TG–MS measurements (TG8120, Rigaku, Tokyo, Japan, Ar flow of 150 mL/min, heating rate of 5 °C/min). Powder X-ray diffraction (XRD) measurements were conducted on an X’Pert-Pro diffractometer (Cu-K_α radiation, PANalytical, Almelo, The Netherlands). In situ infrared spectroscopy measurements were performed on a iZ10 infrared spectrometer (diffuse-reflectance mode, heating rate 5 °C/min, resolution 4 cm⁻¹, Thermo Nicolet, Thermo Fisher Scientific, Waltham, MA, USA). The samples were always handled in a glove box filled with purified argon.

4. Conclusions

We have investigated the dehydriding processes of xLiBH₄ + (1 − x)Mg₂FeD₆ (x = 0.25, 0.75) composites in detail. For both of these compositions, H–D exchange between LiBH₄ and Mg₂FeD₆ occurred during ball-milling and was promoted during the heating process, as confirmed by in situ infrared spectroscopy and mass spectrometry measurements. The extent of H–D exchange immediately prior to the dehydriding reaction was estimated from the area ratio of MS signals. For the composition with x = 0.25, more than two H atoms in [BH₄]⁻ were replaced by D atoms and for that with x = 0.75, one to two H atoms in [BH₄]⁻ were replaced by D atoms.

For the composition with x = 0.25, only one MS peak was observed, which resulted from the simultaneous dehydriding of isotopically exchanged LiBH₄ and Mg₂FeD₆. For the composition with x = 0.75, two MS peaks were observed, which resulted from partial simultaneous dehydriding of isotopically exchanged LiBH₄ and Mg₂FeD₆, and subsequent dehydriding of the residue isotopically exchanged LiBH₄. A special molar ratio of 1:1 has been identified as the limit for simultaneous dehydriding of LiBH₄ with Mg₂FeH₆/Mg₂FeD₆. Experiments aimed at delineating the detailed thermodynamics of the dehydriding process based on pressure–composition isotherm analysis is underway and the kinetics of the H–D exchange is being investigated by in situ Raman spectroscopy.