Improvement in the Electrochemical Lithium Storage Performance of MgH₂

Abstract

Magnesium hydride (MgH₂) exhibits great potential for hydrogen and lithium storage. In this work, MgH₂-based composites with expanded graphite (EG) and TiO₂ were prepared by a plasma-assisted milling process to improve the electrochemical performance of MgH₂. The resulting MgH₂–TiO₂–EG composites showed a remarkable increase in the initial discharge capacity and cycling capacity compared with a pure MgH₂ electrode and MgH₂–EG composite electrodes with different preparation processes. A stable discharge capacity of 305.5 mAh·g⁻¹ could be achieved after 100 cycles for the 20 h-milled MgH₂–TiO₂–EG-20 h composite electrode and the reversibility of the conversion reaction of MgH₂ could be greatly enhanced. This improvement in cyclic performance is attributed mainly to the composite microstructure by the specific plasma-assisted milling process, and the additives TiO₂ and graphite that could effectively ease the volume change during the de-/lithiation process as well as inhibit the particle agglomeration.

1. Introduction

Magnesium hydride (MgH₂) has been intensively investigated as a hydrogen and heat energy storage medium [1,2,3,4] because of its high hydrogen storage capacity (7.6 wt % 110 kg/m³), low cost, and abundance. MgH₂ is an intrinsically ionic compound. The strong Mg–H bond determines the high hydrogen desorption enthalpy (~74 kJ/mol H₂) and the high decomposition temperature (>350 °C) of MgH₂. Additionally, MgH₂ suffers from sluggish hydrogen sorption kinetics due to the slow hydrogen diffusion in the MgH₂ lattice and poor hydrogen dissociation on the surface of Mg [5,6,7,8]. To overcome the thermodynamic and kinetic problems of MgH₂, many effective methods including nanostructuring, alloying, catalyzing, and compositing have been developed by ball milling, film deposition, and chemical synthetic processes [9,10,11,12,13,14].

In 2008, MgH₂ was first reported as a potential anode material for lithium-ion batteries (LIBs) by Y. Oumellal et al. [15] since it exhibited rather high lithium storage capacity (MgH₂ + Li ↔ Mg + LiH, ~2038 mAh·g⁻¹), low work potential (~0.5 V versus Li⁺/Li), and low voltage hysteresis (<0.2 V). The last point is especially superior over other kinds of conversion reaction materials, such as metal oxides, nitrides, fluorides etc. However, the large volume expansion (>85%) during the conversion reaction of magnesium hydride with lithium resulted in fast capacity fading, which is even more severe than other conversion anode materials as the metal hydride has low conductivity and high activity in the liquid electrolyte. Since then, many efforts have been devoted to improving the electrochemical performance of MgH₂ [16,17,18,19,20,21]. As reported by S. Brutti et al. [22], the ball-milled MgH₂ sample with the addition of Super P shows a relatively high discharge capacity (~1600 mAh·g⁻¹) and coulombic efficiency (~60%), which are mainly attributed to the reduction in crystallite size and enhancement of the electronic conductivity of MgH₂. Furthermore, the addition of metal oxides is also helpful in the conversion reaction of MgH₂ with lithium. For example, Kojima et al. [23,24] added the Nb₂O₅/Al₂O₃ into the active material MgH₂, thus enhancing the coulombic efficiency of all solid-state lithium-ion batteries. However, the improved effect in the coulombic efficiency, the reversibility of the conversion reaction, and the cycling performance of a MgH₂ anode is rather limited [25,26,27].

In this work, the expanded graphite (EG) and TiO₂ were milled with Mg/MgH₂ by dielectric barrier discharge plasma-assisted vibratory milling (P-milling), which is especially advantageous for the preparation of composites containing few-layered graphite [28]. The electrochemical performances of the obtained composites, MgH₂–EG and MgH₂–TiO₂–EG, were compared. Additionally, the new electrode preparation method was also developed to improve the performances of hydride anodes. It was demonstrated that the reversibility of the conversion reaction and the cyclic stability of a MgH₂ anode could be greatly enhanced.

2. Results and Discussion

The MgH₂–EG composite was prepared by the hydrgogenation treatment of the milled mixture (denoted as Mg–EG) of Mg powder and expandable graphite. Figure 1a shows the XRD patterns of the as-milled and hydrogenated Mg–EG composites. After milling for 10 h, the hexagonal Mg remains the major phase of the composite, and the weak MgO peak is due to the slight oxidation during sample transfer. It is stated that the weak MgF₂ diffractions indicate the reaction of Mg with the electrode bar composed of polytetrafluoroethylene, which was also reported in previous work [20]. After hydrogenation, all the Mg diffraction peaks disappear, indicating complete hydrogenation of the Mg powder. The rutile-type α-MgH₂ phase is indexed by strong and sharp diffraction peaks, implying grain growth of MgH₂ due to the hydrogenation treatment.

Figure 1b shows the galvanostatic discharge/charge curves of the MgH₂–EG electrode at different cycles. In the first discharge profile, the potential drops rapidly from the initial open circuit potential (OCP) to 0.27 V, and then increases to 0.33 V, followed by a well-defined potential plateau assigning to the conversion reaction of MgH₂ with lithium. The slight polarization of the MgH₂–EG composite is owing to the kinetic limitation caused by the poor electronic conductivity of MgH₂ and the weak electronic contact between the active material and the nickel foam [29]. After that, the potential gradually drops further to 0.10 V with another plateau, which is attributed to the alloying of Mg with Li. Upon charging, two potential plateaus ranging from 0.10 V to 0.21 V and from 0.21 V to 0.60 V are assigned to the de-alloying reaction and the reverse conversion reaction of Mg/LiH, respectively. The first discharge capacity, amounting to 717.4 mAh·g⁻¹, is much less than its theoretical capacity (1704.8 mAh·g⁻¹ = 2038 mAh·g⁻¹ × 80 wt % (MgH₂) + 372 mAh·g⁻¹ × 20 wt % (graphite)). This result reflects the loss of active Mg during milling, as well as the kinetic limitation of coarsening the MgH₂ electrode. In addition, in the initial charge process, the MgH₂–EG electrode shows a total charge capacity of 320.6 mAh·g⁻¹, corresponding to an initial coulombic efficiency (ICE) of 44.7%. This low ICE value also indicates the incomplete reversible formation of MgH₂ in the delithiation process.

The cycling performance of MgH₂–EG electrode is shown in Figure 1c. The rapid capacity fading in the initial several cycles may be attributed to the pulverization of active material leading to the loss of electronic contact between the active material and the nickel foam. After 50 cycles, a capacity of only 48.1 mAh·g⁻¹ is maintained in the cell; this value is even less than the capacity (~70 mAh·g⁻¹) contributed by the graphite component. In addition, according to the discharge profile at the 10th cycle (Figure 1b), the plateau corresponding to the conversion reaction of MgH₂ with lithium is invisible. Further, the differential capacity plots (dQ/dV) of different cycles are compared in Figure 1d. The peak centered at 0.61 V, which is assigned to the reverse conversion reaction of Mg with LiH, disappears after 10 cycles. This result further confirms the poor conversion reversibility and cycling performance of the MgH₂–EG electrode.

To improve the electrochemical performances, especially the cycling stability of a MgH₂ electrode, TiO₂ was added to the MgH₂–EG composite to accommodate the large volume variation. In addition, Mg was replaced by MgH₂ as the starting milling material in order to avoid grain growth during the hydrogenation treatment. The XRD patterns of as-milled MgH₂–TiO₂–EG composite with different milling times are shown in Figure 2a. The MgH₂ peaks for the 20 h-milled composite (denoted as MgH₂–TiO₂–EG-20 h) show a relative broadening effect compared to that of the 10 h-milled composite (denoted as MgH₂–TiO₂–EG-10 h), implying a finer grain size of the MgH₂ by longer milling time. The SEM observation shown in Figure 2c,d also displays smaller particle size (~10 μm) for the MgH₂–TiO₂–EG-20 h composite. Actually, the composite particles consist of nanosized primary particles according to the magnified SEM images (Figure 2c,d). With regard to the graphite after P-milling for 20 h, the graphite peak around 26.6° disappears in Figure 2a, implying the formation of a disordered structure of the graphite. It is believed that the graphite could be effectively exfoliated to few-layer graphene (FLG) nanosheets due to the synergic effect of the plasma heating and the impact stress from the milling balls [28,30,31]. Additionally, the hydrogen desorption kinetic curves (Figure 2b) show that both samples could release ~3.4 wt % H₂ within 15 min, corresponding to the actual MgH₂ content of ~44.3 wt % in the MgH₂–TiO₂–EG composite.

The galvanostatic charge/discharge curves of the MgH₂–TiO₂–EG composite with different milling times are compared in Figure 3a,b. In the first discharge profile, the MgH₂–TiO₂–EG-10 h composite electrode delivers a total discharge capacity of ~1224.6 mAh·g⁻¹, which is very close to its theoretical capacity (1193.9 mAh·g⁻¹ = 2038 mAh·g⁻¹ × 50 wt % (MgH₂) + 335 mAh·g⁻¹ × 30 wt % (TiO₂) + 372 mAh·g⁻¹ × 20 wt % (graphite)), and this result is also much higher than that of the MgH₂–EG electrode mentioned above. The ICE for the MgH₂–TiO₂–EG-10 h composite electrode is 46.4%, which is a little higher than that for the MgH₂–EG electrode. As seen in Figure 3c, the MgH₂–TiO₂–EG-10 h composite electrode exhibits a discharge capacity of 179.1 mAh·g⁻¹ at the 100th cycle, with a capacity retention of 33%. Compared with the MgH₂–TiO₂–EG-10 h electrode, the MgH₂–TiO₂–EG-20 h electrode shows a similar initial discharge capacity (~1218.6 mAh·g⁻¹) and ICE (48.1%) but possessing a much higher cycling capacity of 305.4 mAh·g⁻¹ after 100 cycles and a capacity retention of ~31%. As also shown in Figure 3c, while the MgH₂–TiO₂–EG-10 h composite electrode experiences rapid capacity fading within the first several cycles, the MgH₂–TiO₂–EG-20 h electrode delivers more stable capacity within 10 cycles, and it also shows higher coulombic efficiency throughout the cycling. Additionally, as shown in Figure 3d, the distinct anodic peak in the differential capacity plots (dQ/dV) at the 100th cycle clearly demonstrates the reversible formation of MgH₂, which indicates that the enhanced cyclic stability of the MgH₂–TiO₂–EG-20 h composite electrode is due to the enhanced conversion reaction reversibility of MgH₂.

XRD analysis was performed to the change of phase structure of the MgH₂–TiO₂–EG-20 h electrode after cycling, but the result (not shown here) shows no diffractions and implies poor crystallinity of the active materials. SEM observation was also carried out to investigate the microstructural evolution of the composite electrodes, and the results are shown in Figure 4. Before cycling, the electrode surface of both the MgH₂–TiO₂–EG-10 h and the MgH₂–TiO₂–EG-20 h electrodes are composed of irregular particles with sizes less than 20 μm (Figure 4a,c), and there is no remarkable morphological difference between them. After 100 cycles, the surface morphology of both electrodes experiences obvious particle coarsening, which is due to the lithiation and delithiation of the active material, which causes repeated powder pulverization and agglomeration. This result also explains the capacity loss of the electrodes during cycling. Further, it is also shown that the particle coarsening effect for the MgH₂–TiO₂–EG-10 h electrode is more serious than for the MgH₂–TiO₂–EG-20 h electrode and the large voids between coarse particles are clearly observed. This microstructural difference indicates that large-volume changes are better accommodated by the TiO₂ and graphite additives with finer microstructure and by longer P-milling time, which help to maintain the structural integrity of the electrode.

3. Experimental

3.1. Materials Preparation

To synthesize the MgH₂–EG composite, 1.48 g Mg powder (99.9% purity, ~50 μm) and 0.4 g expandable graphite (EG) with a mass ratio of 80:20 were handled in a steel vial. The expandable graphite (99.9% purity, 100 mesh) was preheated at 1000 °C and held for 90 s under air atmosphere to obtain the worm-like expandable graphite. The handling process was operated in an argon-filled glovebox with an O₂ and H₂O content of less than 1 ppm to minimize the contamination. The milling was carried out on a dielectric barrier discharge plasma-assisted vibratory miller with the ball to powder weight ratio of 50:1; the details of plasma-assisted milling (P-milling) have been described in previous work [30,31]. After ball milling for 10 h, the as-prepared Mg–EG sample was hydrogenated at 450 °C under 6 MPa H₂ for 6 h.

To synthesize the MgH₂–TiO₂–EG composite, 2 g mixture of MgH₂ powder (hydrogen-storage grade), TiO₂ powder (99.0% purity, ≥325 mesh), and the worm-like EG with a weight ratio of 5:3:2 were handled in a steel vial and milled with the same parameters for 10 h and 20 h and denoted as MgH₂–TiO₂–EG-10 h and MgH₂–TiO₂–EG-20 h, respectively.

3.2. Material Characterization

X-ray diffraction (XRD, Empyrean diffractometer, PANAlytical Inc., Almelo, The Netherland) with Cu Kα radiation was used to characterize the phase structure of the samples. The microstructure was observed by using a scanning electron microscope (SEM, Carl Zeiss Supra 40, Oberkochen, Germany). To determine the hydrogen content of the as-prepared MgH₂–TiO₂–EG sample, the desorption kinetics were measured at 400 °C using a Sievert-type automatic apparatus.

3.3. Electrochemical Measurement

The electrochemical properties of the active materials were measured using coin-type half-cells (CR2016) assembled in an Ar-filled glovebox. For preparation of the MgH₂–EG electrode, the active material was cold pressed directly on the nickel foam with the pressure of 20 MPa. For preparation of the MgH₂–TiO₂–EG electrode, the active material was first mixed with the conductive agent (Super-P) and the binder (polyvinylidene fluoride (PVdF)) in a mass ratio of 8:1:1 and then dissolved in solvent (N-methyl-2-pyrrolidinone (NMP)) to make a slurry with the appropriate viscosity. The slurry was then manually spread onto a Cu foil in the glovebox filled with Ar and dried in a vacuum oven at 80 °C for 12 h. The loading of the active material was ~1.0 mg·cm⁻². The cell used Li foil as the counter and reference electrode and a Celgrad 2400 membrane as the separator. The electrolyte was 1 M LiPF₆ in ethylene carbonate and diethyl carbonate (1:1 by volume) with 10 wt % fluoroethylene carbonate (FEC).

The galvanostatic charge/discharge tests were performed in a voltage range of 0.01 V to 2.0 V (vs. Li/Li⁺) at the current density of 100 mA·g⁻¹ using a Land test system (Wuhan, China) at a constant temperature (30 °C).

4. Conclusions

In summary, the electrochemical lithium storage properties of MgH₂ were greatly improved by compositing with graphite and TiO₂ via the discharge plasma milling process. The resulting MgH₂–TiO₂–EG composites show a remarkable increase in the initial discharge capacity and cycling capacity compared to pure MgH₂ and MgH₂–EG composite electrodes with different preparation processes. The 20 h-milled MgH₂–TiO₂–EG-20 h composite delivered a stable discharge capacity of 305.5 mAh·g⁻¹ even after 100 cycles, and the reversible conversion reaction of MgH₂ has been greatly enhanced. This work demonstrates the potential of the MgH₂–TiO₂ graphite composite by plasma-milled milling for electrochemical applications. The next goal is to obtain a higher cyclic capacity by suppressing the fast capacity fading within the initial discharge/charge cycle, and further elevate the reversible conversion reaction of MgH₂. It is also stated that the possible hydrogen release from hydride materials during discharge/charging should be avoided and given more attention.