Vanadium Hydride as Conversion Type Negative Electrode for All-Solid-State Lithium-Ion-Battery

Yasuhiro Matsumura; Keiji Takagishi; Hiroki Miyaoka; Takayuki Ichikawa

doi:10.2320/matertrans.MT-M2019105

Abstract

Possibility as an electrode by using vanadium hydride, which is one of typical metal hydrides as hydrogen storage alloy, was examined for an all solid lithium ion battery. The results obtained show vanadium hydride as a negative electrode material by conversion reaction were clarified for the first time in this work. By analyzing the charging/discharging properties and XRD profiles for each process, the reaction mechanism of the conversion reaction of vanadium system and lithiation reaction of hard carbon was clarified.

Fig. 1 Galvanostatic charge-discharge curves for V-LiH electrode in voltage range of 0.005 V–1.0 V at 125°C.

1. Introduction

Electrode materials such as silicon (Si) and metal hydrides (MH) with higher capacity than graphite have been actively studied.¹^–¹⁰⁾ In 2008, Oumellal et al. reported that magnesium hydride (MgH₂) can be used as a negative electrode material for LIBs using organic liquid electrolyte.¹¹⁾ The theoretical capacity of MgH₂ is 2034 mAh/g, which is significantly higher than that of graphite intercalation mentioned above. The reaction of metal hydride with lithium ions proceeds through the following reaction and termed as conversion reaction,

\begin{equation} \text{M} + \text{LiH} \Leftrightarrow \text{MH$_{x}$} + \text{$x$Li$^{+}$} + \text{$x$e$^{-}$}. \end{equation}

(1)

For the practical development of metal hydride electrode, the understanding of reaction mechanism is necessary in order to improve the properties such as cyclic performance, overpotential, and kinetics. In addition to the electrode materials, investigation on the electrolyte itself is also important to safely utilize LIBs. In the case of conventional LIBs using the organic liquid electrolyte, there is a risk of fire accidents due to the leakage of the liquid electrolyte. In recent years, solid state inorganic electrolytes have attracted the attention of scientific community. The sulfur-based solid electrolyte shows excellent lithium ion conductivity at room temperature.¹²^–¹⁶⁾ Kanno et al. have synthesized thio-LISICON which also shows the high ion conductivity at room temperature.¹⁷^–²¹⁾ Besides, Matsuo et al. recently reported that the complex hydride LiBH₄, which is well-known as a promising hydrogen storage material reveals super ionic conductor above a phase transition at 115°C.²²^–²⁸⁾ Thus, all solid state lithium ion batteries using metal hydride electrodes and LiBH₄ electrolyte are expected as a next generation electric storage device because of the high capacity and safety.²⁹⁾ Liang et al. reported the excellent electrochemical performance and the mechanism of conversion reaction in all solid state LIB composed of MgH₂ as electrode and LiBH₄ as electrolyte.³⁰^,³¹⁾ Moreover, Kawahito et al. investigated the electrochemical properties of TiH₂ using the LiBH₄ electrolyte.³²⁾ Although the metal hydride is recognized as a potential and interesting electrode material to establish high performance batteries as seen above, there are only few reports on the other hydrides. In this paper, vanadium hydride is focused as the promising negative electrode material for the all solid state lithium ion batteries.

Vanadium (V), having a bcc structure, is one of the attractive hydrogen storage materials. It can absorb and release 3.8 mass% of hydrogen³³^–⁴²⁾ via following absorption and desorption reactions,

\begin{equation} \text{2V($\alpha$)} + \text{1/2H$_{2}$} \Leftrightarrow \text{V$_{2}$H($\beta$1)} \end{equation}

(2)

\begin{equation} \text{V$_{2}$H($\beta$1)} + \text{1/2H$_{2}$} \Leftrightarrow \text{2VH($\beta$2)} \end{equation}

(3)

\begin{equation} \text{VH($\beta$2)} + \text{1/2H$_{2}$} \Leftrightarrow \text{VH$_{2}$($\gamma$)} \end{equation}

(4)

where the β1 phase formed at first reaction step is thermodynamically stable even in the air atmosphere at room temperature. However, the γ phase as the final hydrogenated state is unstable phase and releases hydrogen easily without hydrogen partial pressure at room temperature.³³^,⁴³^,⁴⁴⁾

On the other hand, the conversion reaction of vanadium (V) and lithium hydride (LiH) can be assumed as follows:

\begin{equation} \text{V} + \text{$x$LiH} \Leftrightarrow \text{VH$_{x}$} + \text{$x$Li} \end{equation}

(5)

From the Nernst equation, theoretical reaction potential corresponding to each hydride phase can be obtained as follows.

\begin{align} \text{E}& = \frac{1}{xF}(\varDelta H^{\circ}(\textit{VH$_{x}$}) - T(\varDelta S(\textit{VH$_{x}$}) - \varDelta S(V))\\ &\quad + xT(\varDelta S(\textit{LiH}) - \varDelta S(\textit{Li})) - x\varDelta H^{\circ}(\textit{LiH})) \end{align}

(6)

where the enthalpy and entropy values corresponding to LiH are obtained from a databook and the values corresponding to VH_x should be obtained from PC isothermal profiles of VH_x. According to thermodynamic considerations, the total conversion reaction between V and LiH can be written as follows:

\begin{equation} \text{V} + \text{2LiH} \leftrightarrow \text{VH$_{2}$} + \text{2Li$^{+}$} + \text{2e$^{-}$} \end{equation}

(7)

where the theoretical reaction voltage and the capacity are obtained as 0.72 V and 1013 mAh/g. In this work, the electrochemical properties of V + LiH as the electrode material for all solid state LIB were investigated. A possible reaction mechanism is proposed.

2. Experiments

Since VH₂ is an unstable material under an ambient conditions, a mixture of V and LiH was used as starting materials. First of all, V powder (99.9%, Rare metallic) and LiH powder (99.4%, Alfa Aesar) were mixed by mechanical ball-milling method with the following procedure. V and LiH were weighted in a molar ratio of 1:3 to be 300 mg in total. This mixture was mechanically milled for 2 h with 370 rpm by a planetary ball-milling apparatus (Fritsch, P7). The purchased LiBH₄ (≧ 95% in purity, Sigma Aldrich) was hand milled in a mortar for 15 min in order to increase the interface with the electrode material. Then, LiBH₄ and acetylene black were heated at 200°C for 12 h under a dynamic vacuum in order to remove moisture adsorbed on the surface before further mixing. Finally, the V-LiH mixture was mixed with LiBH₄ as a Li-ion conducting material and acetylene black as an electron conducting material in a weight ratio of 40:30:30 by the ball-milling for 2 h with 370 rpm under 0.1 MPa of Ar.

Li foil as a counter electrode and LiBH₄ as a solid electrolyte were placed on a circular stainless-steel plate and transferred into a pellet maker and pressed at 10 MPa for 5 min. Then, the synthesized V-LiH electrode sample (about 10 mg) was spread on LiBH₄ and pressed at 40 MPa for 5 min. The obtained pellet was sealed in coin type cell.

Charge and discharge properties were investigated by electrochemical test apparatus (HJ1001SD8, Hokuto Denko Co.). For the electrochemical measurements, voltage window was fixed to 0.005–1.0 V, and current density was kept as 10 mA/g (0.05 C). The battery tests were carried out at 125°C in oil bath because LiBH₄ as the solid electrolyte behaves a superior ionic conductivity at this temperature.²²⁾ And the capacity unit would be calculated due to the gravimetric amount of VH₂. To investigate structural change of the electrode material, X-ray diffraction (XRD) measurements were performed by using Cu-Kα radiation at room temperature (RINT-2500V, Rigaku Co.). All the processes of the above sample synthesis, assembly/disassembly of the coin cell, and the preparation of samples for the XRD measurements were carried out in a glove box filled with purified Ar (99.9%, 1 MPa).

3. Results and Discussion

Figure 1 shows charge/discharge profiles of the V-LiH electrode for the 5 cycles. Because VH₂ is unstable without hydrogen pressure, the initial state of 1st cycle is V + 3LiH and the charging properties are examined at first. As is shown in Fig. 1, the charging process are only started from 0.53 V. However, the other cycles for charging are started from 0 V. For the XRD profile corresponding to 0.53 V (start), not only V but also VH_0.81 are confirmed although the starting materials are V and LiH. This phenomenon indicates that a reaction of V + 0.81LiH → VH_0.81 + 0.81Li partially proceeds as a solid-solid reaction during mechanical ball-milling process before the electrochemical test. And, the plateau voltage around 0.65 V for the 1st charging should be corresponding to VH_0.81 + 1.19LiH → VH₂ + 1.19Li. However, the amount of 1.19Li has only ca. 600 mAh/g capacity. The capacity of 600 mAh/g obtained for the 1st charging is corresponding to the length of plateau, which is much smaller than that of total capacity of 1069 mAh/g. The extra 469 mAh/g is due to de-lithiation from lithiated carbon which might be generated during the mechanical ball-milling process from the reaction between acetylene black and Li mentioned above. And then, electrochemical reaction for the 1st charging process shows conversion process to generate VH₂ for the plateau region and de-lithiation process from CLi_x for the slope region. However, as shown in Fig. 2 corresponding to 1.0 V, the generated product is VH_0.81 although the expected product should be VH₂, indicating that the generated product does not change even after charging process with 600 mAh/g capacity. However, as mentioned above, VH₂ is unstable without H₂ pressure. The generated VH₂ must be decomposed to incomplete hydride phase VH_0.81 with slightly high stability during the XRD measurement.

Fig. 1

Galvanostatic charge-discharge curves for V-LiH electrode in voltage range of 0.005 V–1.0 V at 125°C.

Fig. 2

Ex-situ XRD patterns of V-LiH electrode at the various states of charging-discharging process. The profiles of 0.6 (start) and 1.0 V are corresponding to initial and final states for the 1st charging, respectively. The profiles of 0.5, 0.3 and 0.005 V are corresponding to each voltage for the 1st discharging.

With respect to the 1st discharge process, about 1230 mAh/g capacity in total was shown with a prompt slope (1–0.6 V; ca. 30 mAh/g), a first plateau (0.6–0.45 V, ca. 720 mAh/g), and the other slope (0.45–0 V, ca. 480 mAh/g), indicating that the total capacity is larger than 1013 mAh/g corresponding to reaction (7). Because the expected coulombic efficiency corresponding to VH₂ conversion reaction cannot be reached to 100%, this total capacity of 1230 should contain not only conversion process but also lithiation process to acetylene black (hard carbon). Of course, V single phase was confirmed in the 0 and 0.3 V XRD profiles of Fig. 2. Here, focusing on the 2nd to 5th discharging cycles, the profiles corresponding to more than 0.6 V and less than 0.3 V are quite similar, and we assume that the contribution of lithiation to the hard carbon is 470 mAh/g, which is essentially equivalent to the graphitized carbon black so far reported,⁴⁵⁾ and this value doesn’t drastically change with the increasing cycles. The remaining part should be corresponding to conversion reaction.

Figure 3 shows the discharging profiles of each cycle, which are subtracted by the expected lithiation profile corresponding to acetylene black shown in the inset of Fig. 3. As shown in Fig. 3, three plateaus are clearly recognized in the early stages of cycles. And the lengths of the low plateaus around 0.35–0.4 V shown by horizontal arrows did not change with the increasing cycles, indicating about 150 mAh/g. However, the lengths corresponding to the high voltage region shown by vertical arrows were decreased drastically from 600 to 50 mAh/g for the 1st cycle and 5th cycle, respectively. It is noteworthy that the high and low plateau voltages are decreased with increasing cycles but middle plateau voltage was kept with even increasing cycles.

Fig. 3

Discharging curves subtracted by the assumed profile of hard carbon shown in the inset. Vertical arrows showed the end points corresponding to higher plateaus regions and horizontal arrows showed remaining capacities of each profile.

As the discharging processes included not only the conversion reaction of V-LiH to VH₂ but also the lithiation reaction of acetylene black, the charging processes after 2nd cycle should include the de-lithiation reaction with 470 mAh/g capacity. Again, focusing on the 2nd cycle charging profile shown in Fig. 1, the slope regions below 0.45 V (130 mAh/g) and above 0.75 V (90 mAh/g) show about 220 mAh/g capacity, indicating that remaining capacity of 250 mAh/g should be hidden in the plateau regions between 0.45 and 0.65 V. Therefore, the remaining capacities of ca. 250 mAh/g corresponding to 0.45 V plateau and of ca. 580 mAh/g corresponding to 0.6 V plateau can be explained by the conversion reactions of V + 0.81LiH → VH_0.81 + 0.81Li and VH_0.81 + 1.19LiH → VH₂ + 1.19Li, respectively. And, with increasing cycles, the charging processes are reproduced well to have slimilar profiles in the conversion and lithiation reactions. Of course, the capacities corresponding to more than 0.5 V for the charging show the slightly decreased phenomena, indicating a good agreement with the decreasing trend of higher plateau region capacity of the 2nd to 5th discharging processes.

Finally, the capacity corresponding to the de/lithiation reaction of acetylene black would be discussed. The capacity of 470 mAh/g should be calculated due to the gravimetric amount of VH₂. Therefore, the capacity calculated from the amount of carbon should be different. Actually, the amount acetylene black is 30/40 ratio compared with the amount of V + 3LiH. Because the total molecular weight corresponding to V + 3LiH is 75, the corresponding VH₂ and carbon amounts should be 53 and 56, respectively. Therefore, 470 mAh/g capacity due to the VH₂ weight should be converted to 444 mAh/g due to the carbon weight, indicating this value is quite reasonable as is assumed by lithiation of hard carbon.

4. Conclusion

In this study, we focused on the mechanism of V and LiH conversion reaction by using the electrochemical properties for all solid state LIBs. As a result of the electrochemical performance, the initial state of V and LiH should be partially converted to VH_0.81 during initial mixing by mechanical ball-milling. Of course, the by-product Li should be reacted with acetylene black, generating the lithiated hard carbon. Therefore, the 1st charging process shows slightly low capacity even though the capacity is decreased with increasing cycles. The 1st charging plateau at 0.65 V should be explained by the following conversion reaction, VH_0.81 + 1.19LiH → VH₂ + 1.19Li.

In the over all charging/discharging processes, the stable and extra capacity of 470 mAh/g is included as de-lithiation/lithiation processes, respectively. For the discharging processes after subtracted by hard carbon contribution, there are three plateau are obtained. Although the corresponding reaction to each conversion reaction are not clear, high plateau region shown between 0.5 and 0.6 V should be explained by the opposite reaction mentioned above, in which the capacity is drastically decreased with increasing cycles. On the other hand, the capacity corresponding to middle and low plateau regions shown around 0.4 V is not changed with even increasing cycles. Finally, in the charging processes, the plateaus of 0.5 and 0.6 V are, respectively, corresponding to the conversion reactions to generate VH_0.81 and VH₂, and as extra capacity corresponding to the slope regions of charging process, the de-lithiation reaction proceeded.

Acknowledgements

The authors acknowledge the financial support of the JSPS SAKURA Program.

REFERENCES

Corresponding author

Register with J-STAGE for free!