Synthesis of Highly Activated Magnesium by Niobium and Tantalum Gel Oxide Catalyst

Abstract

As catalysts to improve the kinetics of the reaction between Mg and H₂, Nb₂O₅, Ta₂O₅, and Nb₂O₅–Ta₂O₅ mixture gels with and without heat treatment were synthesized using simple sol-gel methods. Furthermore, MgH₂ was ball milled to superficially disperse 1 mol% of each oxide for 2 h, which was ten times shorter than that of previous works. All the oxides show catalytic effects on using easier and simpler synthesis processes. The catalysis of Nb oxides is better than that of Ta oxides and mixtures. The as-synthesized Nb₂O₅ gel without heat treatment is the best catalyst and improves hydrogenation kinetics at room temperature. The Nb₂O₅ gel with higher catalysis is further reduced compared to the heat-treated one because the gel oxide is more unstable owing to the lesser network between Nb and O atoms due to the existence of –OH groups. By using gel oxides, highly activated Mg can be synthesized under milder conditions compared with previous cases.

Fig. 4 TG results for the hydrogen absorption of Mg with the prepared oxides.

1. Introduction

Energy crisis has been one of the most prevalent issues in the 21st century, and several research and development activities are ongoing to develop new, clean, reliable, flexible, cost efficient, and CO₂-free energy resources, such as solar, wind, and hydro energy. Hydrogen is a promising energy carrier that has been highly regarded because of its advantages, such as high energy density (142 MJ/kg), great variety of potential sources, lightweight, and environmental friendliness.¹⁾ Hydrogen is not a primary energy source; it stores, transports, and delivers the energy.²⁾ An efficient and effective hydrogen storage method must be developed to realize a hydrogen-based energy system. There are various types of hydrogen storage materials, such as conventional metal, complex, and chemical hydrides, as well as liquid organic materials.^3,4) Various types of intermetallic compounds have been developed for stationary applications. However, low hydrogen storage capacity, high material cost, and heavy weight prohibit the mobile application of intermetallic compounds.^5–7) Light s- and p-block elements, such as lithium (Li), sodium (Na), magnesium (Mg), calcium (Ca), and their compounds, are promising candidates for storing massive amounts of hydrogen for stationary and onboard applications.^8,9) Among all the light element-based materials, magnesium hydride (MgH₂) has been considered as an attractive candidate in terms of its gravimetric hydrogen storage capacity (7.7 wt.%) as per the United States Department of Energy, natural abundance, cost-effectiveness, and user safety. Additionally, MgH₂ has the highest energy density (9 MJ/kg Mg) of all reversible hydrides applicable for hydrogen storage.^10–13) The hydrogen absorption and desorption processes of Mg- and Mg-based compounds have been studied extensively. Most of these studies on the Mg–MgH₂ system have targeted the issues of stringent hydrogen absorption and desorption kinetics.^14–17) Remarkable catalytic effects for these phenomena have been found by milling MgH₂ with 3d transition metal elements and oxides.^18–28) Hanada et al. reported that the composite, MgH₂ with 1 mol% Nb₂O₅ milled for 20 h, was able to absorb 4.5 wt.% of hydrogen after full desorption even at room temperature under a pressure lower than 1.0 MPa within 15 s.²⁴⁾ The improvement of the desorption properties was caused by the catalytic effect of reduced niobium oxide.²³⁾ It is noteworthy that Kimura et al. reported the hydrogen absorption of MgH₂ with 1 mol% of Nb₂O₅ even at −50°C.²¹⁾ In addition to the above-mentioned reports as related works, the catalytic effects of characteristic Ta₂O₅ have also been investigated.²⁰⁾ In spite of this much of investigations still the catalytic mechanism of oxides is not completely understood yet in details. The difficulty in understanding the detailed catalysis is caused by the difference between preparation processes of the oxides and their starting states. In addition, ineffective processes, such as extensive milling duration, are sometimes required to obtain highly catalytic active states.¹⁹⁾ Owing to the above-mentioned complicated process, it is difficult to compare the essential catalytic properties of oxides.

In this work, various types of Nb and Ta oxides were synthesized via simple sol-gel methods and dispersed on Mg by ball milling for a shorter time than conventional processes. The catalysis for the hydrogen desorption and absorption of MgH₂ was investigated through thermal and structural analyses. Based on the experimental results, the catalytic properties of Nb and Ta oxides were compared.

2. Experimental Procedure

2.1 Sample preparation

Commercial MgH₂ powder (99.8%, FUJIFILM Wako Pure Chemical Corp.) was used as the starting material. All the six oxides were prepared by means of the sol-gel method. Niobium(V) ethoxide (Nb(OC₂H₅)₅) was obtained from FUJIFILM Wako Pure Chemical Corp., while tantalum(V) ethoxide (Ta(OC₂H₅)₅) was purchased from Hokko Chemical Industry Co., Ltd. Ethanol (EtOH; ≥99.5%) was purchased from Nacalai Tesque, Inc. All the reagents were used as received, without further purification. To prepare the Nb₂O₅ gel, Nb(OC₂H₅)₅ (1 mmol) was first dissolved in EtOH (1 mL) by ultrasonication. Thereafter, deionized water (5 mL) was added to the solution for the hydrolysis and condensation of Nb(OC₂H₅)₅. A white precipitate, i.e., Nb₂O₅ gel, was rapidly generated after the addition of water. After mixing with a vortex mixer for 2 min, the obtained product was collected by centrifugation and dried overnight in an oven at 50°C. The Ta₂O₅ gel and Nb₂O₅–Ta₂O₅ mixed gel were prepared in the same manner by substituting Ta(OC₂H₅)₅ or a mixture of Nb(OC₂H₅)₅ and Ta(OC₂H₅)₅ (molar ratio = 1:1) for Nb(OC₂H₅)₅. To remove the –OH groups from these gel samples, heat treatment was conducted at 500°C for 3 h (in air) using a muffle furnace with an alumina crucible. Hereafter, the gel samples without heat treatment are denoted as Nb₂O₅-gel, Ta₂O₅-gel, and Nb₂O₅–Ta₂O₅-gel, while the gel samples subjected to heat treatment are denoted as Nb₂O₅-HT, Ta₂O₅-HT, and Nb₂O₅–Ta₂O₅-HT.

The MgH₂ powder was milled with 1 mol% of each synthesized oxide. Oxide-doped MgH₂ samples were prepared using 300 mg of as-received MgH₂ with 1 mol% of oxide in every batch. The sample was ball milled in a Cr-steel pot (30 cm³) containing 20 pieces of steel balls with diameters of 7 mm. Ball milling was performed in a 1 MPa hydrogen atmosphere at 370 rpm. The sample was ball milled for 2 h with 1 h milling followed by a half-hour pause time pattern. Here, the chosen milling time was one-tenth that of the duration reported in previous works.^21,24) All the MgH₂ sample synthesis processes were performed in a glove box filled with highly purified Ar to avoid oxidation.

2.2 Analyses

Phase identification was performed via powder X-ray diffraction (XRD, RINT-2500v, Rigaku, CuK_α radiation). The samples for the XRD measurements were fixed by grease and covered with a polyimide sheet in the glove box to avoid the influence of moisture and oxygen in the air. The microstructure of the oxide sample was evaluated via scanning electron microscopy (SEM, JSM-6380A, JEOL) equipped with energy dispersive X-ray spectroscopy (EDS). The presence of –OH functional groups in all the oxides was determined by means of thermogravimetry-differential thermal analysis-mass spectroscopy (TG-DTA-MS). The dehydrogenation and hydrogenation properties were investigated through TG-DTA. The TG-DTA-MS apparatus was placed inside an argon (Ar) glove box wherein the oxygen and water contents were maintained below 2 ppm. The gas flow rate was fixed at 300 cm³/min.

3. Results and Discussions

In the XRD measurements of all the oxides, as presented in Fig. 1, no diffraction peaks were observed, suggesting that the catalysts were amorphous in nature. Here, the broad kinks observed around 20–40° and 40–60° originate from the grease and polyimide sheets. Figures 2(a)–(c) and (f)–(h) show the SEM images of all the oxide samples. The mixed states of small and agglomerated particles were observed in all the oxide samples. All the particles were approximately spherical in shape, and there was a random particle size distribution. The gel samples, as well as the heat-treated ones had approximately the same types of images based on particle size, agglomeration, and distribution; there were no large differences among all the oxides. Figures 2(d), (e) and (i), (j) show the SEM-EDS images of Nb₂O₅–Ta₂O₅-gel and Nb₂O₅–Ta₂O₅-HT. The molar ratio obtained by estimation using the EDS results was approximately 1:1, which was consistent with the initial stoichiometric ratio of Nb₂O₅ and Ta₂O₅ (1:1). It was clarified that the Nb and Ta atoms were homogeneously distributed without particle dependence, suggesting that both fine and agglomerated particles were composed of a Nb and Ta mixture. From the above-mentioned structural properties, it can be summarized that the prepared oxides possess similar structures and morphologies.

Fig. 1

XRD patterns of the synthesized oxides.

Fig. 2

SEM images of (a) Nb₂O₅-gel, (b) Ta₂O₅-gel, (c) Nb₂O₅–Ta₂O₅-gel, EDS mapping of (d) Nb and (e) Ta for Nb₂O₅–Ta₂O₅-gel, SEM images of (f) Nb₂O₅-HT, (g) Ta₂O₅-HT, (h) Nb₂O₅–Ta₂O₅-HT, EDS mapping of (i) Nb and (j) Ta for Nb₂O₅–Ta₂O₅-HT.

The hydrogen desorption kinetics of the catalyzed MgH₂ have been studied using the TG-DTA measurements at a constant heating rate of 5°C/min, and the results are presented in Fig. 3. The XRD patterns of all the samples after ball milling show no diffraction peaks corresponding to the byproducts except for MgH₂ (see Fig. A1). Hydrogen desorption took place according to the following reaction:   

\begin{equation} \text{MgH$_{2}$} \to \text{Mg} + \text{H$_{2}$}. \end{equation} (1)

Here, it was confirmed through XRD measurements that the MgH₂ phases completely changed to Mg after the TG measurements (see Fig. A2). The hydrogen desorption peak temperatures for MgH₂ + Nb₂O₅-gel and -HT the were least and approximately 225°C, while those for MgH₂ + Ta₂O₅-gel and -HT were the highest and approximately 255°C. For the hydrogen desorption reaction, the catalytic effects of the gel and heat-treated oxides are rather the same. Here, although the hydrogen desorption temperature of MgH₂ + Nb₂O₅-HT is slightly lower than that of the MgH₂ + Nb₂O₅-gel, the difference between peak temperatures is within 5°C. It is essentially difficult to discuss the small difference for the ball-milled samples under thermally activated conditions. The hydrogen desorption profiles of MgH₂ + Nb₂O₅ are broader than those of Ta oxides, suggesting that the ball-milling conditions are not optimized to obtain the best catalysis.¹⁹⁾ For the mixtures, the catalytic effects are rather intermediate between those of Nb and Ta oxides. As described in the experimental section, the molar amount of Nb₂O₅ in MgH₂ + Nb₂O₅–Ta₂O₅ is half the value of MgH₂ + Nb₂O₅ because the total amount is fixed at 1 mol% for MgH₂. Thus, the catalytic effects of the mixture might be explained by the amount of Nb in the samples. However, the single peak profile and temperature of hydrogen desorption cannot be explained by only the amount. As a reasonable possibility, it is considered that the solid solution or ternary oxide (Nb–Ta–O) composed of both oxides is formed during the synthesis process by the sol-gel method and a new catalytic state different for each oxide is generated. If each oxide in the mixture independently produces the catalytic active sites, it is expected that the DTA profile shows two hydrogen desorption peaks, which are observed around 220°C and 250°C due to the Nb and Ta oxide effects, respectively. Notably, two peak hydrogen desorption profiles were reported for MgH₂ with different active sites.¹⁹⁾ However, a single hydrogen desorption peak was observed for MgH₂ with Nb₂O₅–Ta₂O₅ samples in this work. Thus, although synergetic effects to enhance the catalysis of these oxides were not found unfortunately, the intermediate effect was observed for the synthesized oxide containing both Nb and Ta, suggesting that a characteristic mixed state such as a solid solution or ternary compound is formed by the sol-gel methods. From the above results, it is expected that homogeneous oxides containing two or more metal elements can be synthesized by the sol-gel methods to find more active catalysts. Among these oxides, Nb₂O₅–Ta₂O₅-HT showed a multipeak profile, indicating that the catalytic states were non-homogeneous. The difference in the weight loss observed in the TG results for Nb₂O₅, Ta₂O₅, and Nb₂O₅–Ta₂O₅ could be explained by the weight difference between Nb and Ta. Interestingly, the catalytic effects of the Nb₂O₅ gels with and without heat treatment are comparable to those of previously reported MgH₂ with excellent performances prepared by ball milling for 20 h.^19,21,24) For the conventional Nb₂O₅, a long milling time is required to form a homogeneous and highly catalytic active state, which leads to single peak hydrogen desorption at low temperatures. When the milling time is only 2 h, the hydrogen desorption shows two peaks because some parts of Nb₂O₅ are not changed to highly active states.¹⁹⁾ On the other hand, the Nb₂O₅ synthesized by the sol-gel method can be converted to highly active states even by ball milling for only 2 h, and the hydrogen is desorbed by one peak at low temperature. This is an advantage of the sample used in this work for practical applications.

Fig. 3

TG-DTA results for hydrogen desorption of MgH₂ with the prepared oxides.

Figure 4 shows the isothermal hydrogen absorption profile performed at 40°C under a pressure of 0.1 MPa for all the catalyzed dehydrogenated samples. The starting point of the experimental data was determined as the switching time of the carrier gas from Ar to H₂. The hydrogen absorption reaction proceeded for all the samples, although the pristine Mg required thermal activation.²¹⁾ Among all the samples, the MgH₂ with Nb₂O₅-gel showed the highest hydrogen absorption rate, and the catalysis was better in the order of Nb₂O₅ > Nb₂O₅–Ta₂O₅ mixtures > Ta₂O₅. The obtained tendency is the same as that of the hydrogen desorption experiments. Here, the formation of the MgH₂ phase and the remaining Mg phase with a reasonable amount corresponding to the TG results were confirmed via XRD measurements (see Fig. A3). The difference in catalysis for the hydrogen absorption process between the samples is significant compared with those observed in the desorption process because the hydrogen desorption reactions of MgH₂ are affected by the combination effects of the catalysts and thermal activation from the thermodynamic requirement. The gel oxides for all the series revealed had higher catalysis than the heat-treated oxides, indicating that a special reason exists for the gel oxide to be converted into further activated state.

Fig. 4

TG results for the hydrogen absorption of Mg with the prepared oxides.

To understand the catalytic active states, XPS analysis was carried out on the Nb₂O₅ series as a representative system, which showed higher catalytic effects among the evaluated oxides in this work. The XPS spectra of the samples are shown in Fig. 5. The Nb 3d peaks were examined to determine the variation in the chemical state by the ball milling process with MgH₂. The Nb₂O₅-gel and Nb₂O₅-HT oxides were completely assigned to Nb⁵⁺ in its pristine form and then reduced to lower valence states after ball milling. The XPS spectra were expressed by fitting curves corresponding to the Nb⁴⁺ and Nb⁰ states, where the reduction of Nb during ball milling with MgH₂ exhibited a phenomenon similar to that presented in previous reports.^18,23) The peak area ratio of Nb⁰/Nb⁴⁺ for MgH₂ + Nb₂O₅-gel and MgH₂ + Nb₂O₅-HT was estimated to be approximately 0.95 and 0.80, respectively, thereby suggesting that the Nb₂O₅-gel without heat treatment is easily reduced compared to that of the other. In a previous work, it was reported that a more reduced Nb state is more active as the catalyst and the catalytic activity of Nb⁴⁺ is not high.²⁹⁾ In addition, Shirasu et al. reported that the multi-valence state of transition metals was important for the catalytic effects of the reaction with H₂ in the case of the vanadium membrane.³⁰⁾ Based on the analogy of this report, it is possible that the balance of the existing Nb valence states is effective for obtaining a highly active state as the catalyst. It is evident from the above-stated results that reduction states are important for obtaining high catalysis on Nb oxide for the hydrogen absorption and desorption reaction of Mg.

Fig. 5

XPS spectra of Nb 3d for Nb₂O₅-gel, Nb₂O₅-HT, MgH₂ + Nb₂O₅-gel and MgH₂ + Nb₂O₅-HT.

To understand these phenomena, thermal analyses of the oxides were performed. Figure 6 shows the TG-MS profiles under Ar flow of the as-synthesized oxides. All the gel oxides emitted water (m/z = 18) in a wide temperature range from 50–300°C, and the water emission was negligible after the heat treatment. The results suggest that only the drying process at 50°C is insufficient to eliminate the –OH groups included in the sample during the synthesis process by the sol-gel method. The gel oxides contain the –OH groups as a bridge-like role in the network of Nb and O atoms considering the synthesis process, namely, it would be a precursor of stable Nb₂O₅ phases.^31,32) During heat treatment, the –OH groups disappear by releasing H₂O to form stable Nb and O bonds. Therefore, the Nb₂O₅-gel as an intermediate is a metastable state and easier to form in the reduced state, which is catalytically active for the reactions of Mg. By using the metastable Nb₂O₅-gel prepared by the simple sol-gel method without any heat treatments, the highly active Mg as hydrogen storage materials can be synthesized under milder conditions, such as the only 2 h ball milling relative to those reported in previous works.

Fig. 6

TG-MS results of the synthesized oxides.

4. Conclusion

In this work, various types of amorphous Nb and Ta oxides were synthesized via simple sol-gel methods. These oxides are dispersed on the MgH₂ surface by ball milling for 2 h, which is ten times shorter than that of the previous synthesis process using ball milling. All the oxides evidently reveal catalysis for the hydrogen desorption and absorption reactions of Mg. In particular, hydrogen absorption can proceed around 40°C, suggesting that high Mg activation can be achieved. From the experimental results of Nb and Ta oxides, it is apparent that the catalysis of Nb₂O₅ is higher than that of Ta₂O₅. The synergetic effects of the synthesized Nb₂O₅–Ta₂O₅ mixed oxide are not found for catalysis; however, it is indicated that the mixed oxide state such as solid solution or ternary oxide (Nb–Ta–O) synthesized by the sol-gel methods shows characteristic catalytic effects differently from that of each oxide. Namely, the sol-gel method would be useful to synthesize the homogeneous oxides containing two or more metal elements and find more active catalysts. The catalytic effects of the gel oxides are higher than those of the oxides heat-treated at 500°C. It is expected that the difference in the catalysis is caused by the stability of the synthesized oxides. Therefore, the gel oxides are recognized as suitable precursors to produce highly activated Mg under simple and mild conditions.

Acknowledgments

We would like to thank the Center for Functional Nano Oxide at Hiroshima University and the JSPS Core-to-Core Program for their financial support. The authors are thankful to Prof. Masahiro Sadakane for supporting this research.

REFERENCES

• 1)   I.P.  Jain,  P.  Jain and  A.  Jain: J. Alloy. Compd. 503 (2010) 303–339.
• 2)   L.  Schlapbach and  A.  Züttel: Nature 414 (2001) 353–358.
• 3)  N.K. Zhevago: Other methods for the physical storage of hydrogen, (Compendium of Hydrogen Energy, Elsevier, 2016) pp. 189–218.
• 4)   A.  Jain,  S.  Agarwal and  T.  Ichikawa: Catalysts 8 (2018) 651.
• 5)   V.  Charbonnier,  J.  Zhang,  J.  Monnier,  L.  Goubault,  P.  Bernard,  C.  Magén,  V.  Serin and  M.  Latroche: J. Phys. Chem. C 119 (2015) 12218–12225.
• 6)   S.  Kumar,  A.  Jain,  T.  Ichikawa,  Y.  Kojima and  G.K.  Dey: Renew. Sustain. Energy Rev. 72 (2017) 791–800.
• 7)   M.  Spodaryk,  N.  Gasilova and  A.  Züttel: J. Alloy. Compd. 775 (2019) 175–180.
• 8)   A.  Züttel: Mater. Today 6 (2003) 24–33.
• 9)   H.  Wang,  H.J.  Lin,  W.T.  Cai,  L.Z.  Ouyang and  M.  Zhu: J. Alloy. Compd. 658 (2016) 280–300.
• 10)   F.  Leardini,  J.R.  Ares,  J.F.  Fernández,  J.  Bodega and  C.  Sánchez: Int. J. Hydrogen Energy 36 (2011) 8351–8357.
• 11)   A.S.  Pedersen,  J.  Kjøller,  B.  Larsen and  B.  Vigeholm: Int. J. Hydrogen Energy 8 (1983) 205–211.
• 12)   I.P.  Jain,  C.  Lal and  A.  Jain: Int. J. Hydrogen Energy 35 (2010) 5133–5144.
• 13)  U. S. Department of Energy: http://www.energy.gov/.
• 14)   M.  Dornheim,  N.  Eigen,  G.  Barkhordarian,  T.  Klassen and  R.  Bormann: Adv. Eng. Mater. 8 (2006) 377–385.
• 15)   C.  Pistidda et al.: J. Power Sources 270 (2014) 554–563.
• 16)   J.C.  Crivello,  B.  Dam,  R.V.  Denys,  M.  Dornheim,  D.M.  Grant,  J.  Huot,  T.R.  Jensen,  P.  de Jongh,  M.  Latroche,  C.  Milanese,  D.  Milčius,  G.S.  Walker,  C.J.  Webb,  C.  Zlotea and  V.A.  Yartys: Appl. Phys., A Mater. Sci. Process. 122 (2016) 97.
• 17)   V.A.  Yartys et al.: Int. J. Hydrogen Energy 44 (2019) 7809–7859.
• 18)   O.  Friedrichs,  F.  Aguey-Zinsou,  J.R.A.  Fernández,  J.C.  Sánchez-López,  A.  Justo,  T.  Klassen,  R.  Bormann and  A.  Fernández: Acta Mater. 54 (2006) 105–110.
• 19)   N.  Hanada,  T.  Ichikawa and  H.  Fujii: J. Alloy. Compd. 404–406 (2005) 716–719.
• 20)   S.  Kumar and  G.P.  Tiwari: J. Mater. Sci. 52 (2017) 6962–6968.
• 21)   T.  Kimura,  H.  Miyaoka,  T.  Ichikawa and  Y.  Kojima: Int. J. Hydrogen Energy 38 (2013) 13728–13733.
• 22)   T.  Kimura,  H.  Miyaoka,  N.  Hanada and  T.  Ichikawa: J. Japan Inst. Met. Mater. 79 (2015) 107–111.
• 23)   N.  Hanada,  T.  Ichikawa,  S.  Isobe,  T.  Nakagawa,  K.  Tokoyoda,  T.  Honma,  H.  Fujii and  Y.  Kojima: J. Phys. Chem. C 113 (2009) 13450–13455.
• 24)   N.  Hanada,  T.  Ichikawa,  S.  Hino and  H.  Fujii: J. Alloy. Compd. 420 (2006) 46–49.
• 25)   O.  Friedrichs,  D.  Martínez-Martínez,  G.  Guilera,  J.C.S.  López and  A.  Fernández: J. Phys. Chem. C 111 (2007) 10700–10706.
• 26)   G.  Barkhordarian,  T.  Klassen and  R.U.  Bormann: J. Alloy. Compd. 364 (2004) 242–246.
• 27)   G.  Barkhordarian,  T.  Klassen and  R.  Bormann: Scr. Mater. 49 (2003) 213–217.
• 28)   G.  Barkhordarian,  T.  Klassen and  R.  Bormann: J. Phys. Chem. B 110 (2006) 11020–11024.
• 29)   S.  Isobe,  K.  Kudoh,  S.  Hino,  K.  Hara,  N.  Hashimoto and  S.  Ohnuki: Appl. Phys. Lett. 107 (2015) 081602.
• 30)   Y.  Shirasu,  T.  Nambu,  K.  Omata,  H.  Yukawa and  Y.  Matsumoto: Mater. Trans. 60 (2019) 2174–2178.
• 31)   M.  Ristić,  S.  Popović and  S.  Musić: Mater. Lett. 58 (2004) 2658–2663.
• 32)   K.  Nakajima,  Y.  Baba,  R.  Noma,  M.  Kitano,  J.N.  Kondo,  S.  Hayashi and  M.  Hara: J. Am. Chem. Soc. 133 (2011) 4224–4227.

Appendix

Fig. A1

XRD patterns of the MgH₂ samples with oxide additives.

Fig. A2

XRD patterns of the MgH₂ samples with oxide additives after the dehydrogenation.

Fig. A3

XRD patterns of the MgH₂ samples with oxide additives after the dehydrogenation.