Synthetic approach of ternary magnesium niobate (Mg–Nb–O) compounds

Abstract

Magnesium based niobium oxides (Mg–Nb–O) were prepared by solid-state reactions owing to understand the function of transition metal oxides as promoters/catalysts for practical application. Magnesium niobate (Mg₃Nb₆O₁₁) was synthesized for the first time in nearly pure form reported in this context. MgNb₂O₆ and Mg₄Nb₂O₉ were prepared in oxidizing conditions; on the contrary, Mg₃Nb₆O₁₁ preferred reducing environment. Stoichiometric mixtures of the precursor materials MgO, Nb₂O₅ and/or metallic Nb were annealed for the syntheses which revealed the effect of temperature on phase formation, reaction kinetics and heat of reaction. The products were examined by ex-situ, in-situ X-ray diffraction (XRD) and differential scanning calorimetry (DSC). Crystallographic parameters of various binary and ternary compounds (Mg/Nb/O) formed in different calcination conditions, were extracted by Rietveld method. In-situ experiment results in single step reaction for the MgNb₂O₆ synthesis and the heat of formation of the solid-state reaction obtained to be minimum (93 kJ/mol). In contrast, the formation of Mg₄Nb₂O₉ and Mg₃Nb₆O₁₁ compounds towards pure phases rather complicated due to multistep reactions and corresponding heat of formation were estimated to be 140 and 190 kJ/mol. Experimental results have been discussed based on kinetic and thermodynamic constrains.

Introduction

Hydrogen storage represents an important step in the development of hydrogen economy and various storage systems were reported in the literatures. Among these systems, magnesium hydride (MgH₂) is an interesting material for H₂ storage^1,2 owing to high abundance in the lithosphere, cost-effective and less toxic properties^3,4. However, higher temperature is required in H₂ absorption/desorption cycles which are characterized by relatively slow reaction kinetics.

Different strategies have been proposed to overcome these problems such as ball milling and addition of catalysts, particularly, transition metal oxides (TMO), metals, alloys, etc.⁵. Huge interest in TMO was reported in the literatures. It has been revealed from the reviews that niobia (Nb₂O₅) is a competent additive even though its activity is still an unresolved issue as a promoter/additive to alter the reaction kinetics of the system^6,7,8,9.

Milling of MgH₂ with Nb₂O₅ influenced H₂ absorption/desorption kinetics and Mg–Nb–O ternary compounds were formed during H₂ sorption cycles^10,11,12. In particular, a reactive pathway model was proposed by Friedrichs et al.¹³ point out the fact that the reduction of Nb₂O₅ into metallic Nb, followed by successive formation of magnesium based oxides (Mg–Nb–O) which facilitate H₂ transport into the solid structures. Recently, the effect of the presence of Mg–Nb mixed oxide compounds (MgNb₂O₆, Mg₄Nb₂O₉ and Mg₃Nb₆O₁₁) on H₂ absorption properties of MgH₂ was investigated^9,14,15.

A number of possibilities for the formation of binary and ternary compounds with various oxidation states of Nb are available in the literatures^{16,17,18,19,20,21,22,23,24,25}. Abbattista et al.²⁶ obtained an orthorhombic phase of MgNb₂O_3.6 from reduction of MgNb₂O₆. Marinder et al.²⁷ also reported the preparation of Mg₃Nb₆O₁₁ from precursor mixtures of MgO/NbO₂ with different molar ratios of Mg:Nb. The synthesis of some ternary compounds, e.g., MgNb₂O₆, Mg₄Nb₂O₉, Mg₃Nb₆O₁₁ and Mg₅Nb₄O₁₅ were reported elsewhere^24,28. Pagola et al.²⁴ proposed an effective method of solid-state synthetic reactions Mg–Nb oxides basically annealing commercial precursor materials. During preparation of MgNb₂O₆ compound from the starting materials (MgO and Nb₂O₅), MgNb₂O₆ phase of columbite structure is usually obtained including corundum-like Mg₄Nb₂O₉²⁹. Crystallographic, microstructural and morphological features of MgNb₂O₆ and Mg₄Nb₂O₉ compounds were reported elsewhere³⁰. Moreover, they are stable phases at room temperature explored by You et al.³¹. A crystallographic investigation of the orthorhombic columbite-like MgNb₂O₆ phase was done and the fine-structural features of this compound were established from neutron diffraction investigation²⁴. Mg₄Nb₂O₉ shows corundum-type (α-Al₂O₃) structure resulting from order of Mg²⁺ and Nb⁵⁺ ions^32,33. Nonetheless, proper selection of starting materials, calcination conditions to be optimized and reaction mechanism aiming to have pure phases not yet studied in details.

The present work aims to explore in details the synthetic routes for the preparation of three Mg–Nb–O compounds (e.g., MgNb₂O₆, Mg₄Nb₂O₉ and Mg₃Nb₆O₁₁) in nearly pure phases and the products were characterized by ex-situ and in-situ XRD and DSC experiments.

Methodology

Materials and methods

Commercially available MgO, Nb₂O₅ and Nb (Sigma-Aldrich, Germany) were employed as precursor materials for the solid-state synthesis. MgNb₂O₆ and Mg₄Nb₂O₉ were prepared by annealing MgO/Nb₂O₅ with stoichiometric ratio in oxidizing atmosphere. A mixture of MgO/Nb₂O₅/Nb powders was heated in evacuated quartz ampoules for Mg₃Nb₆O₁₁ preparation. The mixtures were annealed from room temperature (RT) to 473, 673, 873, 1073, 1273 and 1473 K for 24 h and heating rate was set at 10 K/min for all the cases.

Characterization

XRD analysis

Structural analysis of the as-prepared materials was carried out by ex-situ XRD diffractometer (Panalytical) with a radiation source of Cu Kα. A reaction chamber (Anton Paar XRK 900) was employed for in-situ XRD study. XRD patterns were recorded in isothermal conditions following suitable temperature step programs with a step size of 0.017° for 6 min from RT to 1173 K at 10 K/min. The experiment was carried out in a steel-made sample holder under vacuum condition and its thermal expansion was estimated to be shifted nearly 0.1° with a reference of α-quartz. In fact, the shifting of the observed peaks did not alter the lattice constants considerably. Therefore, during in-situ XRD, the surface of the powdered sample moved significantly from the goniometer centre. Continuous vacuum was introduced into the reaction chamber. The diffraction patterns were documented as a function of time and the experiments of the preparation of MgNb₂O₆, Mg₄Nb₂O₉ and Mg₃Nb₆O₁₁ required 10–18 h.

DSC analysis

Differential scanning calorimetry (DSC) thermograms were recorded due to the synthesis of Mg–Nb oxides from RT to 1473 K at 10 K/min using a calorimeter (Setaram) of high temperature with a flow of He and Ar. The reference (αAl₂O₃, ca. 0.2 g) and mixtures of precursor materials were loaded in Pt crucibles.

Crystallographic information of the powder samples built on XRD patterns were evaluated by using MAUD (Material Analysis Using Diffraction). MAUD is oriented to the studies of material science. It is a general analytical program based on diffraction/reflectivity data and mainly supports Rietveld method³⁴.

Results and discussion

Ex-situ experiment

Solid-state synthetic phase evolution is a function of temperature during heating the precursor materials. XRD patterns corresponding to ex-situ measurements of the as-prepared samples were recorded at RT. The XRD patterns monitored along the process of formation of MgNb₂O₆, Mg₄Nb₂O₉ and Mg₃Nb₆O₁₁ are reported in the Figs. 1, 2 and 3, respectively and the phase abundance estimated by Rietveld refinement has been inserted in Tables 1, 2 and 3. XRD patterns of the samples were recorded from 10° to 90°; however, due to simplicity and more clear presentation, only the data from 20° to 50° are reported. A detailed description of phase composition of various binary and ternary oxides grown up with increasing temperatures towards solid state synthesis of pure Mg–Nb–O compounds was evaluated by Rietveld refinement with excellent fittings (R% = 1–4%).

Figure 1

Ex-situ XRD patterns obtained during MgNb₂O₆ preparation: (a) 298 K and (b) 1473 K.

Figure 2

Ex-situ XRD patterns obtained during Mg₄Nb₂O₉ preparation at (a) 298 K, (b) 473 K, (c) 673 K, (d) 873 K, (e) 1073 K, (f) 1273 K and (g) 1473 K.

Figure 3

Ex-situ XRD patterns obtained during Mg₃Nb₆O₁₁ preparation at (a) 298 K, (b) 473 K, (c) 673 K, (d) 873 K, (e) 1073 K, (f) 1273 K and (g) 1473 K.

Table 1 Abundance of MgNb₂O₆ compound obtained by Rietveld refinement of XRD patterns at 298 and 1473 K (MC = monoclinic, OR = Orthorombic, O₆ = MgNb₂O₆).

Table 2 Abundance of Mg₄Nb₂O₉ compound obtained by Rietveld refinement of XRD patterns at different temperatures (MC = monoclinic, OR = Orthorombic, O₆ = MgNb₂O₆, O₉ = Mg₄Nb₂O₉).

Table 3 Abundance of Mg₃Nb₆O₁₁ compounds obtained by Rietveld refinement of XRD patterns at different temperatures (MC = monoclinic, OR = Orthorombic, O₆ = MgNb₂O₆, O₉ = Mg₄Nb₂O₉, O₁₁ = Mg₃Nb₆O₁₁).

During the synthesis of MgNb₂O₆ (Fig. 1 and Table 1), the binary precursor materials (MgO and Nb₂O₅) reacted according to the following reaction:

$${\text{MgO}}\, + \,{\text{Nb}}_{2} {\text{O}}_{5} \to {\text{MgNb}}_{2} {\text{O}}_{6}$$

(1)

Detailed synthetic method of MgNb₂O₆ was discussed in the previous study³⁵. In the current issue XRD patterns at RT and maximum reaction temperature 1473 K are shown (Fig. 1a,b). In fact, no solid state reactions were carried out from RT to less than 673 K followed by calcination at 673–1073 K; the evolution of columbite-like MgNb₂O₆ phase was observed. A nearly pure MgNb₂O₆ phase was obtained at 1273 K and the solid-state reaction of the desired compound was completed at 1473 K (97 wt%) excepting traces of unreacted more stable monoclinic Nb₂O₅ (Table 1).

In the case of Mg₄Nb₂O₉ preparation (Fig. 2 and Table 2) the precursor oxides are basically unaffected by the treatment in the range RT-873 K (Fig. 2a,b,c,d). The MgNb₂O₆ phase appeared as intermediate after annealed at 1073 K (Fig. 2e). With increasing temperature up to 1273 K, diffraction patterns showed simultaneous presence of various binary and ternary compounds (Mg/Nb/O) (Fig. 2f). In this step, the fraction of MgNb₂O₆ decreased and a considerable amount of the corundum-type Mg₄Nb₂O₉ was formed. At higher test temperature (1473 K), almost pure Mg₄Nb₂O₉ phase was observed (Fig. 2g). The whole process can be described by the following stoichiometry:

$$4{\text{MgO}}\, + \,{\text{Nb}}_{2} {\text{O}}_{5} \to {\text{Mg}}_{4} {\text{Nb}}_{2} {\text{O}}_{9}$$

(2)

However, the final formation of the mixed phase which started at 1473 K occurs on the basis of a reaction between MgO and MgNb₂O₆ according to the following reaction:

$$3{\text{MgO}}\, + \,{\text{MgNb}}_{2} {\text{O}}_{6} \to {\text{Mg}}_{4} {\text{Nb}}_{2} {\text{O}}_{9}$$

(3)

The phase evolution leading to Mg₃Nb₆O₁₁ is more complicated than in the previous two cases owing to the presence of more reactive metallic Nb in the precursor mixture. The chemical composition remains nearly constant up to 673 K (Fig. 3a,b). Upon annealing at 873 K, niobium oxides (NbO₂ and NbO) of lower oxidation states were appeared.

The formation of niobium monoxide (NbO) and dioxide (NbO₂) can be explained considering that the synthesis was carried out in static vacuum in order to avoid oxidation of metallic Nb. However, in these conditions, for a semiconducting oxide like Nb₂O₅, oxygen depletion easily occurs and metallic Nb is thus partially oxidized by the released oxygen (Reactions 4 and 5). Similar effects were also reported for Mg–Ng–O compounds¹⁶.

$${\text{Nb}}_{2} {\text{O}}_{5} \to {\text{Nb}}_{2} {\text{O}}_{{5{-}x}} + {\text{x}}/2{\text{O}}_{2}$$

(4)

$${\text{xO}}_{2} + {\text{Nb}} \to 2{\text{NbO}}_{{\text{x}}}$$

(5)

Only at higher temperature (1073 K) MgNb₂O₆ appeared as a predominant phase (Fig. 2e). In addition, at this temperature various peaks related to Mg₃Nb₆O₁₁ were observed and its phase quantity gradually increased with the temperature. With raising the synthetic temperature to 1273 K, the abundance of both MgNb₂O₆ and binary oxides decreased and some amount of Mg₄Nb₂O₉ was observed together with the desired Mg₃Nb₆O₁₁ phase (Fig. 3f). Then at higher temperature (1473 K), approximately a single phase of the ternary compound (92 wt%) was obtained with some contamination only by MgNb₂O₆ (8 wt%) (Fig. 3g).

In-situ experiment

In-situ experiments, aimed to study kinetics of the preparation of Mg–Nb–O compounds, were carried out from RT to 1173 K in isothermal conditions (Figs. 4, 5, 6) and corresponding amount of phases obtained by Rietveld analysis are reported in Tables 4, 5 and 6.

Figure 4

In-situ XRD patterns obtained during MgNb₂O₆ preparation at (a) 298 K and (b) 1173 K.

Figure 5

In-situ XRD patterns obtained during Mg₄Nb₂O₉ preparation at (a) 298 K, (b) 1073 K, (c) 1098 K, (d) 1123 K, (e) 1148 K and (f) 1173 K.

Figure 6

In-situ XRD patterns obtained during Mg₃Nb₆O₁₁ preparation at (a) 298 K, (b) 1073 K, (c) 1098 K, (d) 1123 K, (e) 1148 K and (f) 1173 K.

Table 4 Abundance of MgNb₂O₆ phase compound obtained by Rietveld refinement of in–situ XRD patterns (MC = monoclinic, OR = Orthorombic, O₆ = MgNb₂O₆).

Table 5 Abundance of Mg₄Nb₂O₉ compound obtained by Rietveld refinement of in–situ XRD patterns (MC = monoclinic, OR = Orthorombic, O₆ = MgNb₂O₆).

Table 6 Abundance of Mg₃Nb₆O₁₁ compound obtained by Rietveld refinement of in–situ XRD patterns (MC = monoclinic, OR = Orthorombic, O₆ = MgNb₂O₆).

Considering that no appreciable reactions take place among the parent materials at lower temperature during in-situ experiments. After an initial pattern acquisition at RT, the solid-state synthetic temperature was increased directly to 973 K during MgNb₂O₆ preparation and up to 1073 K in the other two cases (Figs. 4, 5, 6 and Tables 4, 5, 6).

In-situ XRD data for MgNb₂O₆ preparation are shown in Fig. 4 and corresponding phase compositions are inserted in Table 4. In this case the MgNb₂O₆ phase appeared at 1123 K following Reaction 1 and the abundance of this phase increased with the temperature as proved by ex-situ measurements. Phase evolution with temperature (298–1173 K) are described in our previous work³⁵.

In the case of in-situ Mg₄Nb₂O₉ preparation (Fig. 5) and resultant phase compositions are introduced in Table 5, where only the MgNb₂O₆ phase was detected for the reasons reported earlier. The presence of this phase, however, confirms the compound representing a key step in the Mg₄Nb₂O₉ phase formation (Reactions 1 and 3).

Similarly, MgNb₂O₆ phase was observed as an early stage during in-situ study of Mg₃Nb₆O₁₁ (Fig. 6) and respective phase contents are placed in Table 6. In fact, this phase evolution is much more complicated than the other two phases as also detected in ex-situ experiment, might be owing to the addition of more reactive metallic Nb. The evolution of the various phases indicates that, even in vacuum condition, metallic Nb is easily oxidized at lower temperatures hindering the formation of Mg₃Nb₆O₁₁.

Through this approach of the ternary compound synthesis, in contrast to ex-situ method, no pure phases were appeared due to temperature and kinetic restrictions. In fact, the highest temperature that can be reached by the in-situ equipment is lower than that of ex-situ preparation because of instrumental limitation. Moreover, incomplete phase transformation is determined by the reaction time inferior to the ex-situ experiments; however, the in-situ test is very important to comprehend the kinetics of the solid-state synthesis. Besides, low-valence niobium oxides (NbO and NbO₂) were evidenced for all the cases during in-situ experiment focus the fact of vacuum conditions at which the reactions were carried out.

The evolution of the various Mg–Nb oxide phases during ex-situ experiments are discussed earlier and compared them to the in-situ outcomes. Since it was not possible to obtain the Mg₄Nb₂O₉ and Mg₃Nb₆O₁₁ phases with the in-situ approach; a comparison between the two approaches can be done only for MgNb₂O₆. Nevertheless, the following considerations can be done:

Nearly 24 wt% of the MgNb₂O₆ phase was obtained at 873 K and 1073–1173 K by ex-situ and in-situ tests, respectively point out that in-situ measurement required higher temperature (Fig. 7, Tables 1 and 4)³⁵. This indicates that MgNb₂O₆ formation is promoted by an oxidizing ambient. Moreover, the reasons behind Mg₄Nb₂O₉ that was not appeared during in-situ experiments can be explained as the growing of the compound started at higher temperature. This evidence also can be interpreted that the formation of Mg₄Nb₂O₉ is composed, at least, by two steps. Both the experiments indicate that the MgNb₂O₆ phase represents the first step during Mg₄Nb₂O₉ preparation (Reactions 2 and 3). This phase then evolved to the final product at higher temperature. On the contrary, Mg₃Nb₆O₁₁ formation is not parallel to the others and its mechanism is still an open question. Ex-situ experiment shows that the Mg₄Nb₂O₉ and Mg₃Nb₆O₁₁ phases were simultaneously appeared at about 1073 K and at this temperature, Mg₄Nb₂O₉ formation is favoured (15%) rather than Mg₃Nb₆O₁₁ (10%). In-situ measurement clearly shows that the grown-up of MgNb₂O₆ and Mg₄Nb₂O₉ compounds were hampered in vacuum conditions. Moreover, they possibly react as follows (Reaction 6) at higher temperature to form Mg₃Nb₆O₁₁ as evidenced by the quantity of the reactants obtained from Rietveld analysis decreased and that of the product increased (1273–1473, Table 3):

$$2{\text{MgNb}}_{2} {\text{O}}_{6} + {\text{Mg}}_{4} {\text{Nb}}_{2} {\text{O}}_{9} + {\text{NbO}}_{2} \to 2{\text{Mg}}_{3} {\text{Nb}}_{6} {\text{O}}_{11} + {\text{NbO}}$$

(6)

Figure 7

Comparison of the evolution of various ternary Mg-Nb–O phases (O₆ = MgNb₂O₆, O₉ = Mg₄Nb₂O₉ and O₁₁ = Mg₃Nb₆O₁₁) with temperature during ex-situ and in-situ experiments.

In addition, during the formation of Mg₃Nb₆O₁₁ compound, the MgNb₂O₆ and Mg₄Nb₂O₉ represent the intermediate steps which evolve to the final product at higher temperature according to the following reactions (Reaction 7 and 8):

$${\text{MgNb}}_{2} {\text{O}}_{6} + 2{\text{MgO}}\, + \,{\text{Nb}}_{2} {\text{O}}_{5} + 4{\text{Nb}} \to {\text{Mg}}_{3} {\text{Nb}}_{6} {\text{O}}_{11} + 2{\text{NbO}}$$

(7)

$${\text{Mg}}_{4} {\text{Nb}}_{2} {\text{O}}_{9} + 2{\text{MgO}} + 3{\text{Nb}}_{2} {\text{O}}_{5} + 8{\text{Nb}} \to 2{\text{Mg}}_{3} {\text{Nb}}_{6} {\text{O}}_{11} + 4{\text{NbO}}$$

(8)

Taking into account Reactions 4 and 5, the previous mechanism also explains why oxides with low oxidation states (NbO and NbO₂) are always observed during the preparation of Mg₃Nb₆O₁₁.

DSC studies

The preparation of the Mg–Nb–O compounds was studied by DSC analysis (Fig. 8) monitoring the final products by XRD with the aim of measuring the heat of reaction of the synthesis during the experiment. In the case of MgNb₂O₆ and Mg₄Nb₂O₉, pure phases were observed at the highest test temperatures; whereas, a yield of 70% was obtained for Mg₃Nb₆O₁₁ (Figure not attached due to simplicity).

Figure 8

DSC curves for Mg–Nb–O ternary phases: (a) MgNb₂O₆, (b) Mg₄Nb₂O₉ and (c) Mg₃Nb₆O₁₁.

Heating the starting materials for the synthesis up to 1400 K, three curves are obtained as shown in Fig. 8. All curves are characterized by exothermic signal spreads on a very large range of temperature (~ 200 K). Two peaks for each compound can be observed. In the case of MgNb₂O₆, the presence of two peaks nearly at 1090 and 1180 K (Fig. 8a) suggests that the overall process for the compound formation is complicated than the simple solid-state reaction reported earlier (Reaction 1). The presence of the two steps in the synthetic reaction can be interpreted through the following arguments:

The heating method connected to the DSC measurement is quite similar to that of the in-situ experiment where it was carried out in the reducing environment. In-situ results in fact also point out that the reaction is kinetically low and the intermediate products were formed as NbO and NbO₂ in the range of 1023–1173 K.

The low temperature peaks in the DSC thermograms can be explained on the basis of the formation of such intermediates which, in this particular case, further reacted at higher temperature since they were not detected by XRD of the final product. A similar behaviour was revealed from the DSC curves for Mg₄Nb₂O₉ and Mg₃Nb₆O₁₁ (Fig. 8b,c). In these cases two peaks were also observed but they were shifted at higher temperature (1090 K for the both, and 1220 K and 1205 K for the Mg₄Nb₂O₉ Mg₃Nb₆O₁₁, respectively) (Fig. 8) indicating a multi-step pathway in the ternary oxide formation.

The heat of formation (HF), in the case of MgNb₂O₆, was obtained to be minimum (93 kJ/mol) calculated from the DSC results. In addition, the HF of Mg₄Nb₂O₉ and Mg₃Nb₆O₁₁ compounds towards pure phases were found 140 and 190 kJ/mol, respectively. Considering that the formation reaction is fully accomplished for MgNb₂O₆ and Mg₄Nb₂O₉. Moreover, it is possible to obtain the HF for the two oxides from the DSC curves (Fig. 9) which results in − 286.17 kJ/g atom and − 294.95 kJ/g atom for MgNb₂O₆ and the Mg₄Nb₂O₉, respectively. The driving force (DF) of the compounds can be calculated from available database when the synthesis is performed starting from MgO and Nb₂O₅ (Fig. 9).

Figure 9

Heat of formation (H.F.) for MgNb₂O₆ and Mg₄Nb₂O₉ (D.F. = Driving Force).

This DF values are very close to the cases being 10.41 kJ/g atom and 9.40 kJ/g atom for the MgNb₂O₆ and Mg₄Nb₂O₉, respectively thus MgNb₂O₆ representing the more stable phase in agreement with XRD results which clearly showed that during Mg₄Nb₂O₉ preparation, the MgNb₂O₆ was always obtained as intermediate. Similar calculation for MgNb₂O₆ was reported earlier by our group³⁵.

Conclusions

In this work three Mg–Nb–O compounds of MgNb₂O₆, Mg₄Nb₂O₉ and Mg₃Nb₆O₁₁ were successfully prepared by solid-state reactions and the formation mechanism was characterized by means of XRD performed both in-situ and ex-situ. XRD analysis shows that the solid-state reactions leading to the ternary compounds are kinetically slow; therefore, high temperature with long time was needed for calcination of the precursor materials to obtain nearly pure phases. Moreover, the formation of MgNb₂O₆ and Mg₄Nb₂O₉ was hampered when the synthesis was carried out in absence of oxygen. XRD also shows that MgNb₂O₆ represents an intermediate in the Mg₄Nb₂O₉ and Mg₃Nb₆O₁₁ formation. For the first time the heat of formation for the MgNb₂O₆ and Mg₄Nb₂O₉ phases were estimated by means of DSC analysis which shows that the MgNb₂O₆ is thermodynamically more stable phase. The as-prepared three pure Mg–Nb–O phases, which are utilized as catalysts/additives to improve H₂ sorption kinetics of MgH₂ towards practical application, will be discussed in the forthcoming issues.