Composition of Mg and AZ31, AZ61, AZ91 alloys.
Magnesium (Mg) is hydrogenated as core-shell-type hydride. Therefore, increase of absorption capacity to the theoretical hydrogen capacity is still one of the most important issues for the hydrogen storage materials. In this study, the procedure of the core-shell structure as well as effect of Al concentration in Mg on the growth MgH2 in Mg were investigated. MgH2 was formed on the surface as well as inside of unreacted Mg core. The inside MgH2 was formed in a granular form on Mg grain boundary and its size increased by applying plastic deformation. Thickness of the surface MgH2 and size of the internal MgH2 increased with an increase in hydrogenation time until the hydride surface was completely covered with MgH2. However, the growth of the surface and internal MgH2 came to a halt after the surface was covered with MgH2. From these results, supplying H from metal side was dominantly contributed for growth of the surface and internal MgH2 because diffusion rate of H in Mg was much higher than that in MgH2. In addition, the growth of internal MgH2 as well as control of surface MgH2 can contribute to the promotion of the complete hydrogenation of Mg-based hydrogen storage materials.
- grain boundary
Magnesium (Mg) can store 7.6 mass% of hydrogen after formation of magnesium hydride (MgH2), which has attractive features for hydrogen storage material such as low cost, abundant resource and light weight . However, dehydrogenation temperature is very high (560 K at 0.1 MPaH2) because MgH2 is thermodynamically stable (ΔrH = −72.8 ± 4.2 kJ mol−1, ΔrS = −142 ± 3 J K−1 mol−1) . In addition, their hydrogenation/dehydrogenation kinetics is also lower, then the conditions of hydrogenation and dehydrogenation are severe and core-shell-type hydride is formed. In order to obtain MgH2 completely from Mg and effectiveness of hydrogenation/dehydrogenation process, it is necessary to finely pulverize, severe plastic deformation, heat treatment for a long time, and addition of catalyst [3, 4, 5, 6, 7, 8, 9, 10].
Mg is a metal and when it reacts with H2, MgH2 forms an ionic bond and a weak covalent bond between Mg-H and the valence number of the ion is indicated as Mg1.91+ and H0.26− . The diffusion coefficient of H in MgH2 is several order of magnitude lower when compared to that in Mg: DHMg = 7 × 10−11 m2 s−1(300 K) in Mg and DHMgH2 = 1.1 × 10−20 m2 s−1 (305 K) in MgH2 [12, 13]. Based on these characteristics, powder Mg forms core-shell-type structure hydride, MgH2 as a shell and unreacted Mg remains in the core [14, 15] making progress of completely hydrogenation difficult. On the other hand, the hydrogen partial pressure has a great influence on the progress of the hydrogenation. When the hydrogen partial pressure is high, since MgH2 quickly covers the Mg powder surface, hydrogenation halts and the amount of hydride concentration decreases markedly, whereas when the hydrogen partial pressure is low, the time until MgH2 covers the Mg surface extends, then the hydride concentration increases . Therefore, to accomplish the efficient hydrogenation, the process of hydrogenation should be revealed.
As a result of investigation aiming at efficient hydrogenation, some curious microstructural characteristics were obtained, that is shell of MgH2 and core Mg in addition to MgH2 in the core Mg. In the following, MgH2 on the surface layer named as MgH2(sur) and that in Mg core named as MgH2(int). MgH2(int) formed in the Mg core is distinguished from MgH2(sur). The particle size of MgH2(int) in Mg−6 mass%, Al-1 mass%, Zn alloy was larger than that in pure Mg. This result shows the grain size of MgH2(int) would be in correlation with Al concentration. Therefore, in this study, the influence of Al concentration in Mg on formation of MgH2(int) is clarified. To reveal the mechanism of MgH2(int) formation, coarse-type Mg-based hydrogen storage materials will be developed. Bulky hydrogen storage materials are attractive for handling, safety, and applying large module.
2. Materials and methods
Pure Mg manufactured by Rare Metallic Co., Ltd. and Mg- (3, 6, 9) mass%, Al-1 mass%, Zn alloy (hereinafter referred to as AZ31, AZ61, AZ91) manufactured by Sankyo Aluminum Co., Ltd. were used as samples. Table 1 shows the composition of the samples.
As pretreatment, these samples were cut into prismatic shapes (8 mm in width, 5 mm in length, 3 mm in thickness) and then annealed, evacuated to 2 × 10−4 Pa in a stainless steel container, and replacing it with Ar gas (nominal purity 99.9999 vol%) was repeated several times and then heated under an Ar atmosphere of 3 kPa at a temperature of 723 K and a treatment time for 32.4 ks. For AZ61, annealed material (annealed) and 10% cold-rolled sample (cold-rolled) were subjected to hydrogenation (129.6 ks), respectively. An observation of the cross-sectional structure was carried out using an optical microscope (OM). Mg alloys are known to produce MgH2 at low temperature by applying severe plastic deformation . Table 2 shows the crystal grain sizes of annealed Mg and AZ31, AZ61, and AZ91 alloys.
|Material||Grain size, d/μm (standard deviation)|
For hydrogenation of Mg alloys, these samples after annealing was dry-polished with SiC abrasive paper (# 320) in Ar gas circulation-type glove box (dew point −55 to −70°C, oxygen concentration 1 ppm or less), dry polished and immediately sealed in a stainless steel reaction vessel with a gasket made of stainless steel with Ag plating. In this reaction vessel, a sheath-type thermocouple was installed inside the container and the temperature in the vicinity of the samples was set. In order to generate MgH2, hydrogenation was carried out at a hydrogen pressure of 3.61 MPa and a temperature of 673 K using a Siebert’s type apparatus, and the nominal purity of the hydrogen gas used was 99.99999 vol%. The time was maintained in the range of 0–259.2 ks, and it is shown in parentheses here after.
The X-ray diffraction (XRD) instrument was used for the identification of MgH2(sur). X-ray source of CuKα with the tube voltage of 40 kV and the tube current of 15 mA was used. Microstructure of MgH2 was judged with presence or absence of charge up by using of scanning electron microscope (SEM). Crystal structure of MgH2(int) was investigated using an electron beam backscatter diffraction (EBSD) analyzer.
Samples for microstructural observation and analysis were cut at a position of 4 mm which is half in the width direction and roughly polished using SiC polishing paper (# 320 → # 800 → # 1000) followed by an Al2O3 suspension (0.05 μm) or SiO2 suspension (0.04 μm). Dehydrated methanol was used as the extension liquid. The sample after polishing was promptly washed with an ultrasonic washing machine in dehydrated methanol. After washing, the sample was dried using pure nitrogen gas (99.99%).
Three-dimensional analysis of MgH2(sur) and MgH2(int) was carried out based on data constructed by a continuous serial sectional method, and an optical microscope (OM) was used for microstructural observation. Information at the position and observation depth was obtained by driving a pyramidal indentation into the sample using a Vickers hardness instrument. ImageJ was used for geometrical analysis and the construction of the three-dimension image [18, 19].
Figure 1 described quantitative analysis method for obtaining information on the thickness of the MgH2(sur) layer. In the analysis, five or more images are selected and the MgH2 was observed as a dark area. The mean thickness (dav) of MgH2(sur) was calculated by dividing the area of each observed MgH2(sur) by the length of MgH2(sur) on the surface. Next, the shape of the surface is extracted. The surface line was translated in the direction of the inside. The maximum thickness of MgH2(sur) was evaluated as the translated length at point of the surface line parted from MgH2(sur) particle. The minimum thickness (dmin) was evaluated at the first point of intersection between the shape line and Mg. The average particle size of MgH2(int) is calculated by calculating the area of MgH2(int), and the value of the diameter calculated when assuming MgH2(int) as a spherical calculation on average.
Figure 2 shows the XRD pattern of hydrogenated Mg, AZ31, AZ61, AZ91 alloys for 1.8 ks. In all the samples, formation of MgH2 was observed. These samples were not polished, therefore the information from XRD was mainly from MgH2(sur). The peaks from Mg were shifted to higher angle side with Al concentration due to substitutional solid solution formation. On the other hand, in the peak pattern of MgH2, obvious peak shifts were not observed with increase of Al concentration.
Figure 3 shows cross-sectional OM images of annealed and cold rolled AZ61 alloy. The annealed and cold rolled AZ61 exhibited both MgH2(sur) (white arrow) and MgH2(int) (black arrow). There were no difference on MgH2(sur) between annealed and cold rolled AZ61 alloy as dav was 9 μm (dmax: 22 μm, dmin: 2 μm) in annealed and dav was 8 μm (dmax: 22 μm, dmin: 2 μm) in cold rolled condition. However, particle size of MgH2(int) was lager in cold rolled condition than in annealed condition as mean diameter was 15 μm (maximum: 38 μm, minimum: 3 μm) in annealed condition and mean diameter was 53 μm (maximum: 125 μm, minimum: 14 μm) in cold rolled condition.
Figure 4 shows the SEM image and results of EBSD analysis of hydrogenated AZ31. Figure 4 (a) shows the backscattered electron image (BSE), (b) shows the phase map in the same area with (a), (c) is image quality (IQ) map, and (d) is an inverse pole figure (IPF). In phase map (b), Mg was shown in red and MgH2 was shown in green color. Comparing (a) and (b), dark area in BSE (a) corresponds to MgH2 area in phase map (b). As focused on IQ map (c), image quality of MgH2 area was lower than that of Mg area. The decrease of quality at MgH2 was due to high volume change between Mg to MgH2 about 32% as the result of residual strain and/or nonconducting material; MgH2 tends to charge up.
The MgH2(int) observed in Figures 3 and 4 was surrounded with metallic Mg, and the MgH2(int) has not interface with hydrogen gas at glance. Therefore, three-dimensional analysis was conducted at the area shown in Figure 5. Figure 5 taken from cold rolled and hydrogenated AZ61 sample show (a) part of OM micrograph, (b) MgH2(sur) and MgH2(int) extracted as ROI area contrast inverted, and construction of three-dimensional image rotated from 0 to 60° ((c) 0°, (d) 20°, (e) 40°, and (f) 60°). From Figure 5c–f, the MgH2(int) was not in contact with hydrogen gas and MgH2(sur).
Focusing on MgH2(sur), homogeneous hydride layer does not growth from surface and structure with variations in thickness was spread out in a planar manner with respect to the surface. The white arrow point of MgH2(sur) was grown abnormally like as heteroepitaxially, however size of the part of MgH2(sur) was same size with surrounding MgH2(int), since the part of MgH2(sur) would be formed by collision between MgH2(sur) and MgH2(int) at near surface.
Figure 6 shows an OM micrograph of hydrogenated Mg and a schematic image. Focusing on MgH2(int), the shape of MgH2(int) was granular and precipitated on the Mg grain boundary. In addition, two types of MgH2(int) were observed as growth to two adjacent Mg grains (at pointed two grains) and on only one side of Mg grain. They were indicated in the inset image.
Figure 7 shows the SEM-BSE images of the MgH2(sur) of hydrogenated Mg, AZ31, AZ61, and AZ91 alloys for 1.8–259.2 ks. Focusing on images of hydrogenated in relatively short time, for example at 1.8 ks, MgH2(sur) formed with granular and dotted and that was not formed uniformly like a layer at the time when the entire surface was not covered with hydride. With increase of hydrogenation time, the granular hydride, MgH2(sur), grew in the direction parallel to the surface and formed a rough layer. For images of samples exposed to relatively long hydrogenation time, the thickness of MgH2(sur) decreased with an increase in Al concentration. Figure 8 shows the thickness of MgH2(sur) of Mg, AZ31, AZ61, and AZ91 alloys with hydrogenation time. The white rhombi and the reverse triangles are shown as dmax and dmin, respectively. The black circles are shown as dav and the vertical lines are expressed as standard deviation. Focusing on dmin, dmin = 0 means that the surface was not completely covered with MgH2(sur). Therefore, paying attention to the value of dmin, it is possible to estimate the time when entire surface is covered with MgH2(sur), and the time describes as time of halt (τh). τh was 32.4 ks for Mg and AZ31, 64.8 ks for AZ61, 129.6 ks for AZ91, and τh increased with increasing Al concentration. Next, looking at dav, it was found that dav grew greatly with increasing hydrogenation time before τh, whereas the growth rate drastically decreased after the time of τh. These results indicate that the growth of MgH2(sur) halted when sample surface was covered with MgH2(sur).
To reveal the influence of Al concentration on thickness of MgH2(sur), the relationship between average of dav after τh and Al concentration is shown in Figure 9. As shown in Figure 9, it was found that the thickness of MgH2(sur) decreased with the increase of Al concentration, and the decrease ratio is remarkable from 0 to 3 mass% Al than from 6 to 9 mass% Al.
Changes in the growth rate of MgH2(int) are described as follows. Figure 10 shows the average particle size of MgH2(int) of Mg, AZ31, AZ61, and AZ91 alloys with hydrogenation time. As with the changes of MgH2(sur), the growth rate of MgH2(int) was high over short time and low on longer hydrogenation time side. The time of infection point was almost correlated with τh as 3.6 ks for Mg and AZ31, 64.8 ks for AZ61, 129.6 ks for AZ91. Figure 10 shows the relationship between the average particle size of MgH2(int) after τh and Al concentration. The average particle size of MgH2(int) increased with increase of Al concentration, contrary to the change of thickness of MgH2(sur).
4.1 Formation of MgH2 in Mg core
After hydrogenating Mg and AZ alloys, two types of MgH2 were formed: on the surface, MgH2(sur) and in the unreacted Mg core, MgH2(int). MgH2(int) formed only along Mg grain boundary, and did not form in Mg crystal grain. It is known that the diffusion of H atoms in Mg made the grain boundary a preferential route . Therefore, the reason why MgH2(int) preferentially formed on grain boundary would be one factor of the fast diffusion of H atoms at Mg grain boundary. In addition, MgH2 preferentially formed with existence of defect as nanocrystallize Mg showed fast hydrogenation and dehydrogenation . This observation also supports the reason why MgH2(int) formed on Mg grain boundary.
As shown in Figure 6, MgH2(int) grew for adjacent Mg crystal grain or only on one crystal grain. In addition, some Mg grain boundaries did not form MgH2(int). Some orientation relationship between Mg and MgH2 was reported as (0001)Mg//(001)MgH2, [−1–120]Mg//MgH2  and (−2110)Mg//(12–1)MgH2, Mg//MgH2 . Because interface energy between new phase and mother phase was low in orientation relationship, solid-solid transformation easily proceeded in orientation relationship with a small lattice mismatch. On the other hand, hydrogenation rate of Mg increased with formation of (0002) crystal texture [23, 24]. Form above results, nucleation and growth of MgH2(int) would also depend on adjacent Mg crystal orientation. However, influence of Mg crystal texture for MgH2 formation had other formation of MgO factor , and particle shape of MgH2(int) was ellipsoidal. Therefore, consideration with complex effects for formation of MgH2(int) should be needed such as orientation relationship, precipitates on Mg grain boundary, interface energy with adjacent Mg grains, and flux of H atom and so on.
4.2 Relationships between particle size of MgH2(int) and Al concentration
Before τh, MgH2(sur) formed in granular form and spread and the growth rate of average thickness was larger than that after τh. After τh, the growth of MgH2(sur) apparently halted. These growth rate change was same in MgH2(int). These results be attributed to extremely low diffusion rate of H in MgH2 when compared to that in Mg [12, 13]. When the whole surface was covered with MgH2(sur), the supply rate of H to unreacted internal metallic Mg or AZ significantly decreased, halting the growth of MgH2(sur) and MgH2(int).
Some studies have reported that the amount of MgH2 formed depends on Gibbs free energy (ΔrG) of MgH2 from Mg and H2 gas [16, 25, 26]. The microstructure of MgH2(sur) at initial state was formed as granular and scattered. Concerning this result, the nucleation rate of MgH2(sur) was low and the nucleation rate would depend on absolute value of ΔrG. Applying low value of ΔrG, growth of MgH2(int) proceeded because nucleation rate of MgH2(sur) was low and τh is shift to longer time. However, when high value of ΔrG was applied, the nucleation rate of MgH2(sur) increased and immediately the surface covered with thin MgH2(sur) rapidly and the particle size of MgH2(int) was explained to be small size.
The average particle size of MgH2(int) increased with increasing Al concentration. As shown in Figure 10, the growth rate of MgH2(int) decreased when the surface was covered with MgH2(sur). The thickness of MgH2(sur) decreased with increasing Al concentration. From these results, the reasons why the particle size of MgH2(int) decreased with increasing Al concentration would be shifting of long time τh and small diffusion distance of H in MgH2 at high Al content. Therefore, the amount of supplying H increased and large size of MgH2(int) formed with increasing Al concentration.
In this study, focusing on the formation of MgH2 in the Mg core, the effects of Al concentration in Mg for microstructure of hydrogenated Mg and Mg-Al-Zn alloys were investigated. MgH2(int) formed at Mg grain boundary and the growth rate of MgH2(int) was investigated including plastic deformed condition. From three-dimensional analysis, it was found that the MgH2(int) was surrounded by metallic Mg or Mg-Al-Zn alloys and they had not interfaced with H2 gas and MgH2 on the surface area(sur). The time when the surface covered with MgH2(sur) was described as time of halt, τh. Comparing growth rate of MgH2(sur) and MgH2(int) before and after τh, the growth rate of both MgH2 were higher before τh than after τh. The growth of MgH2(sur) and MgH2(int) were observed to stop after τh because H supply route change from in Mg to MgH2. After τh, the thickness of MgH2(sur) decreased and particle size of MgH2(int) increased with increasing Al concentration. This result could be explained by increase of supplied H amount to MgH2(int) due to the shifting τh to longer time and small diffusion distance of MgH2(sur) which had low value of diffusion coefficient.
Findings from this research point out in following:
Hydrogenated Mg plate formed MgH2 on surface(sur) and in internal area(int).
The thickness of MgH2(sur) decreased with increase of Al concentrations in Mg.
The particle size of MgH2(int) increased with increase of Al concentrations in Mg.
This is a product of research which was financially supported by JSPS KAKENHI Grant Number 19K15278, and the Environment Research and Technology Development Fund (2RF-1801) of the Environmental Restoration and Conservation Agency of Japan.