Investigation of Effect of Milling Atmosphere and Starting Composition on Mg 2 FeH 6 Formation

In this study we investigated the synthesis and the hydrogen storage properties of Mg2FeH6. The complex hydride was prepared by ball milling under argon and hydrogen atmosphere from 2Mg + Fe and 2MgH2 + Fe compositions. The samples were characterized by X-ray powder diffraction and scanning electron microcopy. Kinetics of hydrogen absorption and desorption were measured in a Sievert’s apparatus. We found that the milling atmosphere plays a more important role on Mg2FeH6 synthesis than the starting compositions. Ball milling under hydrogen pressure resulted in smaller particles sizes and doubled the yield of Mg2FeH6 formation. Despite the microstructural differences after ball milling, all samples had similar hydrogen absorption and desorption kinetics. Loss of capacity was observed after only five cycles of hydrogen absorption/desorption.


Introduction
The complex hydride Mg 2 FeH 6 has the highest known volumetric density of hydrogen (150 kg m −3 ), which is more than twice higher than hydrogen in liquid state (70.8 kg m −3 ) [1].This high hydrogen density is attractive for solid-state hydrogen storage applications.However, the high operating temperature, owing to the thermodynamic stability of the hydride, limits its use for stationary applications.On the other hand, the important enthalpy of hydride formation makes this material OPEN ACCESS interesting for thermochemical thermal energy storage [2].However, the synthesis of Mg 2 FeH 6 as a pure material is a great challenge, mainly due to the immiscibility of magnesium and iron [3].
In 1984, Didisheim et al. [4] reported the first synthesis of Mg 2 FeH 6 by sintering the metallic elements at high temperature (450-520 °C) under hydrogen pressure (20-120 bar) during several days (2-10 days).Despite the severe sintering conditions, the yield of Mg 2 FeH 6 was only 50%.Later, Selvan and Yvon [5] showed that the most convenient condition to synthesize the Mg 2 FeH 6 was by sintering of 2Mg + Fe at 450 °C under 90 bar of H 2 pressure during 10 days.Huot et al. [6] demonstrated that Mg 2 FeH 6 could be synthesize in mild condition (350 °C, 50 bar of H 2 , 24 h) by ball milling of 2Mg + Fe under argon or hydrogen atmosphere for 20 h before sintering.Furthermore, the synthesis of Mg 2 FeH 6 at room temperature was reported by high energy ball milling of 2MgH 2 + Fe under argon atmosphere [7] and by reactive milling of the metallic elements under hydrogen pressure [8].
The synthesis of Mg 2 FeH 6 by ball milling has been studied by several authors using different milling parameters, such as: type of mill; ball to powder ratio; number and size of balls; milling time; rotational speed; milling atmosphere and hydrogen pressure [9][10][11][12][13][14][15][16][17][18][19][20][21][22][23].Irrespective of the milling conditions, some iron was always detected in the milled powder representing an incomplete Mg 2 FeH 6 formation.Regardless of all these results concerning the Mg 2 FeH 6 synthesis, the effect of milling parameters on the kinetics of hydrogen absorption/desorption has not been thoroughly investigated.
Besides the synthesis, the mechanism of Mg 2 FeH 6 formation has been discussed in several papers.Some authors reported that Mg 2 FeH 6 is formed from the metallic elements [2,4,19,24], while others claimed that MgH 2 is the Mg 2 FeH 6 precursor [9,20,21,25,26].Recently, we studied the transformation of phases during hydrogen absorption of a ball milled 2Mg + Fe mixture [27].We showed that MgH 2 is formed with very fast kinetics, and then, in a much slower reaction, MgH 2 reacts with Fe to form Mg 2 FeH 6 .Danaie et al. [28] observed this transformation of phases by TEM and found that Mg 2 FeH 6 nucleates between MgH 2 and Fe particles, and then, grows with a columnar morphology.Although these results confirm the formation of Mg 2 FeH 6 from MgH 2 , the direct reaction from the metallic elements is also feasible.As Mg 2 FeH 6 is thermodynamically more stable than MgH 2 , the direct reaction can take place under specific conditions of temperature and hydrogen pressure.
In this study, we investigated the effects of milling atmosphere and starting compositions on the microstructure, Mg 2 FeH 6 formation and hydrogen storage properties of ball milled materials.The Mg 2 FeH 6 was synthesized from 2Mg + Fe and 2MgH 2 + Fe compositions by ball milling under argon and hydrogen atmosphere.Short cycles of hydrogen absorption/desorption were performed to evaluate the influence of ball milling on the kinetics and hydrogen storage capacity.The reaction pathway of Mg 2 FeH 6 formation is also discussed.

Materials and Methods
Two compositions, 2MgH 2 + Fe and 2Mg + Fe, were prepared from magnesium hydride (mesh 300, 98%), magnesium (-325 mesh, 99.8%) and iron (-22 mesh, 98%) powders, all provided by Alfa Aesar.The mixtures weighting about 2 grams were loaded together with 20 balls (10 mm in diameter) into a high pressure milling vial of internal volume of 218 cm 3 .Both balls and vial were made of hardened stainless steel.The ball-to-powder weight ratio was 40:1.All handlings were performed inside an argon filled glovebox.
Ball milling was carried out in a Fritsch's vario-planetary mill pulverisette 4. The relative speed ratio between the milling vial and the main disk was set to −1.82, the minus sign meaning that the vial's direction of rotation was opposite to the main disk rotation direction.The samples were milled at 330 rpm (main disk speed) for 12 h.The 2MgH 2 + Fe composition was ball milled under argon atmosphere and 3 MPa of hydrogen pressure.The 2Mg + Fe sample was only ball milled under 3 MPa of H 2 .These samples are respectively named 2MgH 2 + Fe (Ar), 2MgH 2 + Fe (H 2 ) and 2Mg + Fe (H 2 ).
The crystal structure of the ball milled materials was investigated by X-ray powder diffraction (XRPD).For XRPD measurements, a small amount of the ball milled samples was mounted on a sealed flat plate sample holder inside the glovebox.This procedure was used in order to never expose the sample to air.The samples were analyzed on a Bruker's D8 Focus diffractometer with a Bragg-Brentano configuration using CuKα radiation.The percentage of phases as well as the average crystallite size were evaluated from the XRPD patterns by Rietveld method using GSAS [29] and EXPGUI [30] softwares.Morphology of ball milled powders was characterized by scanning electron microscopy (SEM) using a Tescan Vega3-SB microscope (ICMCB, Bordeaux, France).
The concentration of Fe in the ball milled powders was measured by atomic absorption spectrometry (AAS) to evaluate a possible Fe contamination from the balls and vial by milling.Samples of approximately 100 mg were dissolved in a 100 mL mixture of HCl (10% in volume) and distilled water, and analyzed in a Varian's SpectrAA 50/55 spectrometer.
Kinetic curves of hydrogen absorption/desorption were measured using a homemade Sieverts' apparatus.Samples of around 100 mg were loaded and sealed into the sample holder inside the glovebox.The first hydrogen desorption was studied under 100 kPa of H 2 during heating at 10 °C min −1 from room temperature up to 400 °C.After completed hydrogen desorption, 5 cycles of hydrogen absorption/desorption were measured at 400 °C.The kinetics of hydrogen absorption was measured for 1 hour under initial H 2 pressure of 2.5 and 1.7 MPa.In the case of hydrogen desorption, the H 2 pressure of the system was set to 100 kPa.
To investigate the transformation of phases upon hydrogenation, samples were "quenched" during the kinetic experiments under H 2 pressure."Quenching" was achieved by closing the sample valve and cooling the sample holder down to room temperature in a few seconds using a water bath.As MgH 2 and Mg 2 FeH 6 are both highly stable hydrides, the changes of phase composition and hydrogen capacity are minimal after quenching.Thus, the XRPD patterns of quenched samples truly reflect the phases present during the reaction of hydrogen absorption.

Ball Milling
The XRPD patterns of the ball milled samples are shown in Figure 1.The presence of a high background at low angles in all XRPD patterns is due to the protective dome of the sealed sample holder.The use of this protective dome was required to prevent air exposure and formation of MgO phase [31].The XRPD patterns show that Mg 2 FeH 6 was synthesized by ball milling regardless of starting compositions and milling atmospheres.However, diffraction peaks of α-Fe phase were identified in all samples.This means an incomplete formation of Mg 2 FeH 6 despite the long milling time.In the case of the 2MgH 2 + Fe composition processed under argon atmosphere, the presence of iron after ball milling was expected due to stoichiometry of the reactants and milling atmosphere.The XRPD pattern of this sample also presented small diffraction peaks corresponding to β-MgH 2 and Mg phases.For the 2MgH 2 + Fe and 2Mg + Fe samples ball milled under H 2 pressure, only Mg 2 FeH 6 and α-Fe phases were clearly identified.As some free Fe was still present after milling, it means that one would expect to identify Mg and/or β-MgH 2 phases in these two samples.However, it should be taken into account that the detection of these phases can be quite difficult due to their smaller number of electrons and volume fraction in comparison to Mg 2 FeH 6 and α-Fe.Another explanation could be an iron contamination from the balls and vial by milling.To assess this possible contamination, we measured the Fe concentration on the ball milled samples by atomic absorption spectrometry (AAS).Table 1 presents the results of AAS of the ball milled samples.For comparison, the nominal compositions of Fe before ball milling are also indicated.The results showed that the Fe concentrations on the processed samples were similar to nominal values.Therefore, ball milling did not modify the iron composition of the samples.Rietveld refinement was performed on the XRPD patterns to get a quantitative analysis and evaluate the effects of starting composition and milling atmosphere on the Mg 2 FeH 6 formation.Table 2 summarizes the phase abundance in wt.% as well as the average crystallite size calculated from the Rietveld refinement.As expected, ball milling 2MgH 2 + Fe under argon atmosphere produced the lowest yield of Mg 2 FeH 6 (about 41 wt.%).Ball milling under H 2 pressure doubled the amount of Mg 2 FeH 6 phase.The average crystallite size estimated for all phases were of the same order of magnitude (10 to 20 nm).

Microstructural Characterization
The microstructure of the ball milled materials was investigated by scanning electron microscopy (SEM).Figure 2 shows the SEM micrographs using a secondary electron (SE) and back-scattered electron (BSE) detectors.From the SE micrographs, the difference in particle size between the samples processed under H 2 pressure and Ar atmosphere is evident.The particle sizes of the 2MgH 2 + Fe (Ar) sample were in the range of 3 to 10 µm.For the samples prepared under H 2 pressure, the particle sizes were smaller than 3 µm.It shows that milling atmosphere plays a more important role on the final particle size than the nature of the starting composition.Furthermore, the BSE micrographs of the sample milled in Ar presented brighter spots of 30-40 nm distributed over the larger particles.These spots correspond to Fe-rich regions.This feature was barely seen for the others samples.Thus, ball milling under H 2 also results in better homogeneity and microstructure refinement.
Our results are different from the study of Castro and Gennari [11].These authors investigated the synthesis of Mg 2 FeH 6 by ball milling under H 2 pressure.They showed that the synthesis time of Mg 2 FeH 6 from a 2MgH 2 + Fe composition was almost twice longer and yielded practically half of Mg 2 FeH 6 than when a 2Mg + Fe composition was milled at the same experimental condition.The authors attributed the results to the unlike mechanical properties and microstructures of the starting compositions.It should be pointed out that they synthesized less than 30 wt. % of Mg 2 FeH 6 by ball milling 2Mg + Fe for 60 h.In the case of the 2MgH 2 + Fe composition, the maximum yield of Mg 2 FeH 6 was only 15.6 wt.% after 100 h of ball milling.In our case, the microstructure and the synthesis of Mg 2 FeH 6 were mostly affected by the milling atmosphere than the nature of starting composition.A possible explanation of the discrepancies between their results and ours could be the milling parameters.They used a less energetic milling device (Uni-ball-Mill II) and lower hydrogen pressure (0.5 MPa), which could have hindered the Mg 2 FeH 6 formation.

Hydrogen Desorption under Heating
After ball milling, the first hydrogen desorption was studied by heating the samples under 100 kPa of H 2 from room temperature up to 400 °C at 10 °C min −1 .Figure 3 presents the simultaneous curves of hydrogen capacity and temperature ramp of the ball milled powders.The samples showed a similar desorption behavior but with different amounts of hydrogen released.The 2MgH 2 + Fe composition ball milled under argon atmosphere had the lowest hydrogen gravimetric capacity (3.59 wt.%).This value is very close to the theoretical hydrogen capacity of the starting composition (3.62 wt.%).However, this material absorbed 0.32 wt.% of hydrogen from 130 up to 245 °C.As shown in Figure 1, Mg diffraction peaks were identified on the XRPD pattern of this sample.Thus, considering the hydrogen capacity of MgH 2 (7.6 wt.%), one can calculate that a hydrogen absorption of 0.32 wt.% requires 3.89 wt.% of magnesium.This percentage of Mg is close to the relative quantity estimated by Reitveld refinement in Table 2 (4.4 wt.%).Therefore, this hydrogen absorption occurring at around 150 °C could be attributed to the formation of β-MgH 2 from magnesium.As expected, ball milling the 2MgH 2 + Fe and 2Mg + Fe compositions under hydrogen atmosphere led to higher capacities of 4.94 and 4.32 wt.%, respectively.Assuming that Mg 2 FeH 6 is the only hydride phase present on these samples, as suggested by the XRPD results, one can determine the percentage of Mg 2 FeH 6 from the hydrogen capacity.Dividing the capacity of hydrogen desorption by the theoretical hydrogen capacity of Mg 2 FeH 6 (5.4 wt.%), we find that the percentage of Mg 2 FeH 6 phase on the 2MgH 2 + Fe (H 2 ) and 2Mg + Fe (H 2 ) samples are respectively 91.5 wt.% and 80 wt.%.These values are in agreement with the results of Rietveld refinement.The difference in hydrogen capacity between these two samples could be explained by the reaction pathway of Mg 2 FeH 6 formation.During ball milling of the metallic elements (2Mg + Fe) under H 2 pressure, the β-MgH 2 phase is firstly formed and then, it reacts with Fe to form the complex hydride Mg 2 FeH 6 [22].Using instead MgH 2 as starting material is a shortcut to Mg 2 FeH 6 formation and result in higher hydrogen capacity.

Kinetics of Hydrogen Absorption and Desorption
After complete hydrogen desorption, five cycles of hydrogen absorption and desorption were performed at 400 °C.The initial H 2 pressure was set to 2.5 MPa for absorption and 100 kPa for desorption.The kinetic curves of hydrogen absorption and desorption are presented in Figure 4.The "Reacted fraction" y-axis represents the ratio of measured hydrogen capacity to the maximum theoretical capacity of the samples (5.4 wt.%).Despite having different particle sizes and phase compositions after ball milling, all samples presented similar kinetics of hydrogen absorption and desorption.Hydrogen absorption was practically completed in less than five minutes but the samples did not absorb more than 75% of the theoretical hydrogen capacity.Moreover, from cycle 1 to 3, we observed for all samples a decrease in hydrogen capacity.From third to fifth cycle, the hydrogen capacities were practically constant.In the last cycle, the hydrogen capacity of the 2MgH 2 + Fe (Ar) sample was around 60% of the theoretical capacity.For the 2Mg + Fe (H 2 ) and 2MgH 2 + Fe (H 2 ) samples, the hydrogen absorption respectively stabilized around 65 and 68% of the maximum capacity.In the case of hydrogen desorption, similar kinetics was observed for all samples.According to the study of Bogdanović et al. [2], the equilibrium pressure at 400 °C of Mg 2 FeH 6 and MgH 2 are 1.12 and 1.97 MPa, respectively.Therefore, the formation of both β-MgH 2 and Mg 2 FeH 6 phases are thermodynamically favorable at 400 °C under 2.5 MPa of H 2 .We recently showed that under conditions where the formation of both hydrides are thermodynamically possible, Mg 2 FeH 6 is preferentially formed from a two-step reaction where β-MgH 2 plays the role of Mg 2 FeH 6 precursor [27,28].Despite being thermodynamically less stable than Mg 2 FeH 6 , β-MgH 2 is firstly formed due to its faster kinetics of formation [27].Afterward, Mg 2 FeH 6 nucleates between β-MgH 2 and α-Fe phases, and grows with a columnar morphology in a slow diffusional process [28].Consequently, the kinetics of this diffusional reaction restrains the formation of Mg 2 FeH 6 and result in lower hydrogen storage capacity.However, this is not the only reaction pathway for Mg 2 FeH 6 formation.The complex hydride can also be formed directly from the metallic elements.This reaction is possible because MgH 2 has a higher equilibrium pressure than Mg 2 FeH 6 .Thus, if the H 2 pressure is higher than the equilibrium pressure of Mg 2 FeH 6 but lower than MgH 2 , the formation of MgH 2 would be thermodynamically restricted and Mg 2 FeH 6 would form directly from magnesium and iron.To confirm these two pathways of Mg 2 FeH 6 formation, we investigated the transformation of phases upon hydrogenation by quenching two samples under H 2 pressure.For this investigation we used the 2MgH 2 + Fe (H 2 ) sample after complete hydrogen desorption.Quenching was carried out after 10 minutes of hydrogen absorption at 400 °C under two hydrogen pressures: 1.7 MPa where only the direct reaction from metallic elements is possible and 2.5 MPa where the two-steps reaction is possible.The kinetic curves of hydrogen absorption before quenching are presented in Figure 5.The hydrogen absorption under 2.5 MPa of H 2 was faster and the hydrogen capacity was 35% higher than when the absorption was performed under 1.7 MPa.This result is attributed to the fast formation of MgH 2 , as confirmed by the XRPD patterns shown in Figure 6.In the case of the sample quenched under 1.7 MPa, the XRPD pattern shows the presence of Mg diffraction peaks but, as expected, none from β-MgH 2 phase.This confirms that under this hydrogenation condition, Mg 2 FeH 6 is formed directly from the metallic elements.Results of Rietveld refinement shown in Table 3 indicate that a higher proportion of Mg 2 FeH 6 is formed in the direct reaction than in the two-steps reaction.However, hydrogenation under higher H 2 pressure resulted in faster kinetics and higher hydrogen capacity due to the formation of β-MgH 2 .The effect of cycling under direct reaction condition was studied.Five consecutive cycles of absorption/desorption at 400 °C under respectively 1.7 MPa and 100 kPa were measured and are shown in Figure 7.Despite different pathways of Mg 2 FeH 6 formation, the kinetic curves of the direct (Figure 7) and two-step (Figure 3) reactions presented similar loss of hydrogen storage capacity.

Conclusions
We investigated the effect of milling atmosphere and starting compositions on Mg 2 FeH 6 formation.Samples of 2MgH 2 + Fe and 2Mg + Fe compositions were processed by ball milling under argon and hydrogen atmosphere.The milling atmosphere played a more important role on the Mg 2 FeH 6 synthesis than the nature of the starting compositions.Ball milling under hydrogen resulted in smaller particle sizes, better homogeneity and microstructure refinement, and moreover, doubled the yield of Mg 2 FeH 6 formation.Despite having different particle sizes and phase compositions after ball milling, all samples presented similar kinetics of hydrogen desorption.The transformation of phases upon hydrogenation was investigated by quenching samples under H 2 pressure.Depending on the hydrogenation conditions (temperature and H 2 pressure), the complex hydride can be formed from two reaction pathways: (1) directly from the metallic elements; or (2) in a two-step reaction where MgH 2 plays the role of Mg 2 FeH 6 precursor.Regardless of reaction pathway, loss of hydrogen capacity was measured after only five cycles of hydrogen absorption/desorption.

Figure 1 .
Figure 1.XRPD patterns of ball milled samples.The position of X-ray diffraction peaks for the identified phases are indicated below the patterns.

Figure 2 .
Figure 2. SEM micrographs of ball milled samples.The left and right columns show respectively the secondary (SE) and back-scattered electron (BSE) micrographs.

Figure 3 .
Figure 3. Curves of hydrogen desorption under 100 kPa of H 2 during heating at 10 °C min −1 .

Figure 5 .
Figure 5. Kinetic curves of hydrogen absorption at 400 °C before quenching under 1.7 and 2.5 MPa of H 2 .

Figure 7 .
Figure 7. Kinetic curves of hydrogen (a) absorption and (b) desorption at 400 °C of 2MgH 2 + Fe (H 2 ) sample.The hydrogen absorption and desorption were respectively measured under 1.7 MPa and 100 kPa of H 2 .

Table 1 .
Iron concentration (wt.%) of ball milled samples measured by atomic absorption spectrometry.The nominal composition before ball milling is also presented for comparison. 2MgH

Table 2 .
Relative quantities and crystallite size of phases identified in the ball milled samples as calculated by Rietveld refinement.In parenthesis are uncertainties on the last significant digit.

Table 3 .
Relative quantities and crystallite size of identified phases in the "quenched" samples during hydrogen absorption tests under 1.7 and 2.5 MPa of H 2 as calculated by Rietveld refinement.In parenthesis are uncertainties on the last significant digit.