Field of the Invention

This invention relates to hydrogen storage materials and particularly relates to a cast alloy, or other rapidly solidified alloy, which can be used as a hydrogen storage material. Background of the Invention

As the world's population expands and economic activity increases, there are ever increasing signs that increasing atmospheric concentrations of carbon dioxide are warming the earth causing climate change. While the eventual depletion of the world's oil and fossil fuel energy sources will inevitably require other economic energy sources to be found, the more noticeable signs of global warming have increased pressures for global energy systems to move away from carbon rich fuels whose combustion produces carbon monoxide and carbon dioxide gases.

Hydrogen energy is attracting a great deal of interest and is expected to eventually be a replacement for petroleum based fuels. However, there are still several technical issues and barriers that must be overcome before hydrogen can be adopted as a practical fuel, the main obstacle being the development of a viable hydrogen storage system. While hydrogen can be stored as a compressed gas or a liquid, the former requires heavy steel cylinders which hold relatively low hydrogen mass and the latter is energy intensive to produce, reducing any environmental benefits. In addition, both gaseous and liquid hydrogen are potentially dangerous should the pressure storage vessel rupture.

A safer, more compact method of hydrogen storage is to store it within solid materials. When infiltrated with hydrogen at relatively low pressures, metals and inter-metallic compounds can absorb large quantities of hydrogen in a safe, solid form. The stored hydrogen can be released when required by simply heating the alloy. Storage of hydrogen as a solid hydride can provide a greater weight percentage storage than compressed gas. However, a desirable hydrogen storage material must have a high storage capacity relative to the weight of the material, a suitable desorption temperature, good kinetics, good reversibility and be of a relatively low cost. The exact level of each of these properties will depend on the application and other constraints. For instance, in some applications cost may be an important consideration. In other applications, slower desorption kinetics might be desirable.

Pure Mg theoretically has sufficient hydrogen carrying capacity of 7.6 mass%, a value not normally improved by making alloying additions. However, pure Mg does not find any useful application as a hydrogen storage material because the reactions rates are exceedingly slow, requiring the addition of catalytic elements in order to rectify this deficiency.

As an aside, much research has focussed on nanoscale powder metallurgy techniques which offer limited control over the crystallographic structure of the phases (i.e. interfaces, twins etc), are dangerous as the powder may be highly explosive, and would be prohibitively expensive for large-scale mass production of commercial hydrogen storage components.

The optimal properties of any hydrogen storage material will depend on the application to which they are applied. Thus, it would be desirable to provide a means by which the properties of a range of materials could be tailored to provide a suite of materials suitable for various applications. Of particular relevance is the notion of cost- performance trade-offs. Some applications, for instance military or automotive applications, may have an emphasis on high performance in critical applications, where the gains for a high-performance material outweigh relatively high costs. Other applications, for instance conventional industrial storage applications, may place a moderate emphasis on both cost and performance, leading to the need to balance these by close economic analysis of the application and how it is addressed. Other applications again, for instance large scale bulk or buffer storage, may place almost overwhelming emphasis on cost and require only a minimum focus on performance. Furthermore, the cost structures of both hydrogen storage systems and the applications they target can change gradually but unpredictably over time. The price of raw metals for the creation of the hydrogen storage material may fluctuate, and the relative cost components of the application (labour costs, energy costs, cost of capital and so forth) may also change. For these reasons, the preferred material for any given application is liable to change over time.

Reference to any prior art in the specification is not, and should not be taken as, an acknowledgment or any form of suggestion that this prior art forms part of the common general knowledge in Australia or any other jurisdiction or that this prior art could reasonably be expected to be ascertained, understood and regarded as relevant by a person skilled in the art.

Summary of the Invention The inventors have discovered that the addition of an alloying element or elements, X, to Mg to form a Mg-X alloy, where X is at least one element that forms a eutectic with Mg selected from the group consisting of Al, Ag, Au, Ba, Bi, Ca, Ce, Cu, Eu, Ga, Gd, Ge, Hg, La, Ni, Pb, Sb, Si, Sm, Sn, Sr, Th, Tl, Y, Yb and Zn, and which results in or is predicted to result in an enhancement of the hydrogen storage properties of the Mg-X alloy. If X includes Ni, then at least one other element X should also be present in the alloy.

Since some of these elements suffer from one or more of the following drawbacks (e.g. scarcity, high cost, high density, toxicity, radioactivity, amount required) they cease to be useful from a practical point of view, despite their ability to potentially provide desirable benefits. From the potential list we therefore derive a shorter list of practically useful elements for Mg-X. More preferably this revised listing includes Al, Bi, Ca, Ce, Cu, La, Ni, Sb, Si, Sn, Sr and Zn, with a most preferred short list of Al, Ca, Cu, Ni, Si and Zn. As above, if X includes Ni, then at least one other element X should also be present in the alloy.

In most cases, it is preferable to add the alloying element or elements, X, at hypo- eutectic amounts, that is, at concentrations less than the eutectic composition for that particular element. In that way, the microstructure will consist of a primary phase of magnesium with the eutectic being the secondary phase. However, in the case of silicon (Si), since the eutectic composition is at a very low Si level (1 .34 mass%), beneficial effects can also be obtained by adding it at certain hypereutectic amounts, that is, at concentrations greater than 1 .34 mass% to achieve a microstructure consisting of primary Mg ₂ Si with the eutectic being the secondary phase. This is discussed further in the section titled "Detailed descriptions of the embodiments."

Binary Mg-X Alloys

In a preferred aspect of the invention there is provided a Mg-based alloy consisting essentially of: a Mg-X alloy normally having a primary crystallized Mg phase, except in the case of Si at hypereutectic amounts which has a primary crystallised Mg ₂ Si phase; an amount of X such that the Mg-X alloy is hypoeutectic, where X is an element that forms a eutectic with Mg selected from the group consisting of Al, Ag, Au, Ba, Bi, Ca, Ce, Cu, Eu, Ga, Gd, Ge, Hg, La, Pb, Sb, Sm, Sn, Sr, Th, Tl, Y, Yb and Zn; more preferably from the group of Al, Bi, Ca, Ce, Cu, La, Sb, Sn, Sr and Zn; most preferably from the group of Al, Ca, Cu and Zn;

or, an amount of X such that the Mg-X alloy is either hypoeutectic, eutectic or hypereutectic, where X is Si.

greater than zero and up to 2 mass% Na as a performance enhancing element; and

the balance Mg and incidental impurities.

In a second preferred aspect of the invention there is provided a hydrogen storage material consisting essentially of the as-cast Mg-based alloy of the above aspect.

In a third preferred aspect of the invention there is provided a method of modifying the hydrogen absorption and/or desorption characteristics of a Mg-based alloy, the method including the steps of:

forming a Mg-X melt, wherein X is an element that forms a eutectic with Mg selected from the group consisting of Al, Ag, Au, Ba, Bi, Ca, Ce, Cu, Eu, Ga, Gd, Ge, Hg, La, Pb, Sb, Si, Sm, Sn, Sr, Th, Tl, Y, Yb and Zn; more preferably from the group of Al, Bi, Ca, Ce, Cu, La, Sb, Si, Sn, Sr and Zn; most preferably from the group of Al, Ca, Cu, Si and Zn;

adding the alloying element (any except Si) at an amount that achieves a hypoeutectic alloy composition; or, adding the alloying element (where Si) at an amount that achieves a hypoeutectic, eutectic or hypereutectic alloy composition; adding Na up to 2 mass% in the melt as a refining performance enhancing element to the Mg-X melt; and

casting the melt to produce the Mg-based alloy.

In a fourth preferred aspect, there is provided a method of producing a Mg-based alloy for use as a hydrogen storage material including the steps of: forming a Mg-X melt, wherein X is an element that forms a eutectic with Mg selected from the group consisting of Al, Ag, Au, Ba, Bi, Ca, Ce, Cu, Eu, Ga, Gd, Ge, Hg, La, Pb, Sb, Si, Sm, Sn, Sr, Th, Tl, Y, Yb and Zn; more preferably from the group of Al, Bi, Ca, Ce, Cu, La, Sb, Si, Sn, Sr and Zn; most preferably from the group of Al, Ca, Cu, Si and Zn;

adding the alloying element (any except Si) at an amount that achieves a hypoeutectic alloy composition; or, adding the alloying element (where Si) at an amount that achieves a hypoeutectic, eutectic or hypereutectic alloy composition; adding Na up to 2 mass% in the melt as a performance enhancing element to the Mg-X melt; and

casting the melt to produce the Mg-based alloy.

In each of the above aspects, it is preferable that the Mg-based alloy include, as X, an element selected from the group consisting of Al, Bi, Ca, Ce, Cu, La, Sb, Si, Sn, Sr, Y and Zn; more preferably, X is selected from the group consisting of Al, Ca, Cu, Si and Zn. The preferred amounts of the X addition are as follows in the list below and also tabulated in Table 1 . The rationale and method for determining the amount is provided in the Section entitled "Detailed description of the embodiments":

When X is Al, it is preferable to include from about 3 to about 9 rmass% in the Mg- based alloy; more preferably, Al is included in a range of from about 5 to about 7 mass%.

When X is Ca, it is preferable to include from about 1 to about 7 mass% in the Mg-based alloy; more preferably, Ca is included in a range of from about 3 to about 6 rmass%.

When X is Cu, it is preferable to include from about 1 to about 9 mass% in the Mg-based alloy; more preferably, Cu is included in a range of from about 4 to about 7 mass%.

When X is Si, it is preferable to include from about 0.8 to about 6 mass% in the Mg-based alloy; more preferably, Si is included in a range of from about 2 to about 4 rmass%. When X is Zn, it is preferable to include from about 2 to about 9 mass% in the Mg-based alloy; more preferably, Zn is included in a range of from about 5 to about 8 rmass%.

When X is one of the remaining elements Bi, Ce, La, Sb, Sn, or Sr, it is preferable to include mass% amounts in the range shown in Column 2 of Table 1 for element X in the Mg-based alloy; more preferably, X is included in the range shown in Column 3 of Table 1 for element X in the Mg-based alloy.

Table 1 Preferred and more preferred addition ranges (approximate values) for alloying elements X in Mg-X alloys

Ternary and Higher Order Mg-X Alloys

In a fifth aspect of the invention there is provided a Mg-based alloy consisting essentially of:

a Mg-X alloy normally having a primary crystallized Mg phase, except in the case where one element is Si at hypereutectic amounts which will have a primary crystallised Mg ₂ Si phase;

amounts of X such that the Mg-X alloy is hypoeutectic with respect to those elements, wherein X is two or more elements that each form a eutectic with Mg selected from the group consisting of Al, Ag, Au, Ba, Bi, Ca, Ce, Cu, Eu, Ga, Gd, Ge, Hg, La, Ni, Pb, Sb, Sm, Sn, Sr, Th, Tl, Y, Yb and Zn; more preferably from the group of Al, Bi, Ca, Ce, Cu, La, Ni, Sb, Sn, Sr and Zn; most preferably from the group of Al, Ca, Cu, Ni and Zn;

or, an amount of Si, when added, such that the Mg-X alloy is either hypoeutectic, eutectic or hypereutectic with respect to the element Si, together with an amount of another X such that the Mg-X alloy is hypoeutectic with respect to that other element, wherein the other X is an element that forms a eutectic with Mg selected from the group consisting of Al, Ag, Au, Ba, Bi, Ca, Ce, Cu, Eu, Ga, Gd, Ge, Hg, La, Ni, Pb, Sb, Sm, Sn, Sr, Th, Tl, Y, Yb and Zn; more preferably from the group of Al, Bi, Ca, Ce, Cu, La, Ni, Sb, Sn, Sr and Zn; most preferably from the group of Al, Ca, Cu, Ni and Zn;

greater than zero and up to 2 mass% Na as a performance enhancing element; and

the balance Mg and incidental impurities.

In a sixth aspect of the invention there is provided a hydrogen storage material consisting essentially of the as-cast Mg-based alloy of the above aspect.

In a seventh aspect of the invention there is provided a method of modifying the hydrogen absorption and/or desorption characteristics of a Mg-based alloy, the method including the steps of:

forming a Mg-X melt, wherein X is two or more elements that each form a eutectic with Mg selected from the group consisting of Al, Ag, Au, Ba, Bi, Ca, Ce, Cu, Eu, Ga, Gd, Ge, Hg, La, Ni, Pb, Sb, Si, Sm, Sn, Sr, Th, Tl, Y, Yb and Zn; more preferably from the group of Al, Bi, Ca, Ce, Cu, La, Ni, Sb, Si, Sn, Sr and Zn; most preferably from the group of Al, Ca, Cu, Ni, Si and Zn;

adding the alloying element(s), except Si, at an amount that achieves a hypoeutectic alloy composition; or, adding the alloying element, where Si, at an amount that achieves either a hypoeutectic, eutectic or hypereutectic alloy composition;

adding greater than zero and up to 2 mass% Na in the melt as a performance enhancing element to the Mg-X melt; and

casting the melt to produce the Mg-based alloy. In an eighth aspect of the invention there is provided a method of producing a Mg-based alloy for use as a hydrogen storage material including the steps of:

forming a Mg-X melt, wherein X is two or more elements that each form a eutectic with Mg selected from the group consisting of Al, Ag, Au, Ba, Bi, Ca, Ce, Cu, Eu, Ga, Gd, Ge, Hg, La, Ni, Pb, Sb, Si, Sm, Sn, Sr, Th, Tl, Y, Yb and Zn; more preferably from the group of Al, Bi, Ca, Ce, Cu, La, Ni, Sb, Si, Sn, Sr and Zn; most preferably from the group of Al, Ca, Cu, Ni, Si and Zn;

adding the alloying elements, except Si, at an amount that achieves a hypoeutectic alloy composition; or, adding the alloying element, where Si is present, at an amount that achieves either a hypoeutectic, eutectic or hypereutectic alloy composition;

adding greater than zero and up to 2 mass% Na in the melt as a performance enhancing element to the Mg-X melt; and

casting the melt to produce the Mg-based alloy.

In each of the above aspects, it is preferable that the Mg-based alloy include, as X, two or more elements selected from the group consisting of Al, Bi, Ca, Ce, Cu, La, Ni, Sb, Si, Sn, Sr, Y and Zn; more preferably, at least one X is selected from the group consisting of Al, Ca, Cu, Ni, Si and Zn.

Where two X elements are present in the Mg-based alloy, the preferred amounts of the two X additions are indicated in the Figures 1 -15. Fifteen graphs are shown and provide all of the possible two element combinations (the order of elements is not important) from the most preferred element listing, that is, from the list Al, Ca, Cu, Ni, Si and Zn. Element addition amounts bounded within the dashed lines are claimed as the more preferred addition ranges; while those shown bounded by the solid lines are claimed as the most preferred addition ranges.

The rationale and method for determining the amounts in these graphs is provided in the Section entitled "Detailed description of the embodiments". The same methodology can be used to determine the amounts of X required to be added to the Mg-based alloys, where more than two X elements are employed, and/or where one or more X is selected from the broader element listings. These other possible combinations have not been tabulated but the ranges so calculated by the described methodology as being preferred and most preferred are thus claimed for each of the possible element combinations.

Applicable to All Alloys

In each of the above aspects, it is preferable that the Mg-based alloy include Na as a performance enhancing element.

When the performance enhancing element is Na, it is preferable to include it from about 20 ppm to about 20,000 ppm in the Mg-based alloy; more preferably, Na is included in a range of from about 200 to about 4000 ppm; most preferable Na is included in the range of about 800 to about 2000 ppm.

In each of the above aspects, it is preferable that the performance enhancing element be added to the Mg-X melt under a non-oxidising atmosphere.

In practice, we have demonstrated that the addition of Na to the Mg-X eutectic class of materials can perform one or more of several functions:

Improve the initial activation rate of alloys for hydrogen storage, e.g. Mg-Si, Mg-Cu, Mg-AI;

Activate alloys that are otherwise unable to absorb hydrogen, e.g. Mg-Sn, Mg-Y;

Improve the absorption rate of hydrogen under practical cycling conditions, e.g. Mg- Si;

Improve the early stage cycling activation rate for alloys, e.g. Mg-Cu;

The addition of Na may achieve the above effects by one or more of several means:

• Altering the nature of the surface oxide which acts as an interface through which hydrogen must pass to enter or exit the alloy;

• Acting as a grain boundary segregating element that provides an easy medium for hydrogen transport paths into and out of the alloy; · Acting as a grain boundary segregating element that provides a brittle interface between grains that favours easier fracture during chip formation and hydrogen cycling;

• Providing a catalytic effect over and above that achieved by the presence of the major alloying element, if any; • Refining the microstructure thus providing shorter hydrogen transport and diffusion distances for the hydrogen atoms.

It should be noted that some or all of these mechanisms, along with others not yet identified, may be operative in various combinations in different alloys. In other words, this patent rests on the discovery that the beneficial effects of Na in Mg-based alloys (that contain eutectic phases) for hydrogen storage applications extend broadly to quite different alloy systems by acting in numerous distinct ways.

In Examples 1 and 2, Mg-Si alloys are demonstrated to attain their peak material performance after the first few cycles of hydrogen charging and discharging that is accompanied by the breakdown of the alloy material. This process is facilitated by the presence of particular brittle eutectic intermetallics, for example Mg ₂ Si in Mg-Si alloys. This process occurs by fracturing of the material due to the stresses induced by the hydrogen reaction. Fracturing is a process that is enhanced by increased brittleness such as conferred by particular intermetallics and by Na.

Further examples would be the Mg-Sn and Mg-Y systems (demonstrated later in Examples 3 and 4, respectively). Both of these systems show essentially no ability to store hydrogen in their non-enhanced form. This is most likely due to the inability for Sn or Y additions to act as a catalyst to the hydrogen uptake reaction. However, when Na is added to each of these systems they each surprisingly become capable of hydrogen storage.

Another example is the Mg-AI system. The refinement mechanism for Na previously suggested, related to a particular class of eutectic systems - the "faceted-nonfaceted" systems. However the Mg-AI system is not a "faceted-nonfaceted" system, and instead shows nonfaceted-nonfaceted behaviour. Nevertheless, the addition of Na still confers a surprising benefit to the hydrogen performance of the alloy. This demonstrates that the benefit of Na in eutectic systems is not confined to structural refinement of a particular type of eutectic. This is demonstrated later in Example 5.

Several other examples (6-10) are also presented later that discuss the effects of adding more than one eutectic-forming element to Mg in the absence and the presence of Na as a performance enhancing element. These examples are for various alloys of the Mg-AI-Ca, Mg-Cu-Ni, Mg-Cu-Si and Mg-Ni-Si systems. Brief Description of the Figures

Figure 1 illustrates the preferred and most preferred element addition ranges for Mg-AI- Ca alloys

Figure 2 illustrates the preferred and most preferred element addition ranges for Mg-AI- Cu alloys

Figure 3 illustrates the preferred and most preferred element addition ranges for Mg-AI- Ni alloys

Figure 4 illustrates the preferred and most preferred element addition ranges for Mg-AI- Si alloys

Figure 5 illustrates the preferred and most preferred element addition ranges for Mg-AI- Zn alloys

Figure 6 illustrates the preferred and most preferred element addition ranges for Mg-Ca- Cu alloys

Figure 7 illustrates the preferred and most preferred element addition ranges for Mg-Ca- Ni alloys

Figure 8 illustrates the preferred and most preferred element addition ranges for Mg-Ca- Si alloys

Figure 9 illustrates the preferred and most preferred element addition ranges for Mg-Ca- Zn alloys

Figure 10 illustrates the preferred and most preferred element addition ranges for Mg- Cu-Ni alloys

Figure 1 1 illustrates the preferred and most preferred element addition ranges for Mg- Cu-Si alloys

Figure 12 illustrates the preferred and most preferred element addition ranges for Mg- Cu-Zn alloys

Figure 13 illustrates the preferred and most preferred element addition ranges for Mg- Ni-Si alloys

Figure 14 illustrates the preferred and most preferred element addition ranges for Mg- Ni-Zn alloys Figure 15 illustrates the preferred and most preferred element addition ranges for Mg- Si-Zn alloys

Figure 16 is a Mg-Si binary equilibrium phase diagram

Figure 17 displays microstructures (SEM-BSE) of nominal Mg-1 %Si alloy with and without Na addition

Figure 18 are hydrogen activation curves for nominal Mg-1 %Si alloy with and without Na addition

Figure 19 is a Mg-Sn binary equilibrium phase diagram

Figure 20 are microstructures (SEM-BSE) of Mg-24%Sn alloy with and without Na addition

Figure 21 are hydrogen activation curves for Mg-24%Sn alloy with and without Na addition. The target composition of the alloy was 25%Sn; the measured composition was 24%Sn.

Figure 22 is a Mg-Y binary equilibrium phase diagram Figure 23 are microstructures (SEM-BSE) of Mg-12%Y alloy with and without Na addition

Figure 24 are hydrogen activation curves for Mg-12%Y alloy with and without Na addition. The target composition of the alloy was 17.5%Y; the measured composition was 12%Y.

Figure 25 is a Mg-AI binary equilibrium phase diagram

Figure 26 are hydrogen activation and absorption/desorption curves for a nominal Mg- 27%AI alloy with and without Na addition

Figure 27 is a Mg-Cu binary equilibrium phase diagram; and

Figure 28 is a Mg-Ni binary equilibrium phase diagram Detailed Description of the Embodiments

A general elemental alloy addition (an alloying addition, designated as X) is different to an elemental addition made to act as a performance enhancer. An alloying addition is the addition of an element to combine that element with the base material or another main element to form specific phases of a desired composition, for example, a solution or mixture of Fe and Cr of a specific composition, or the formation of an intermetallic like Mg ₂ Cu as a chemical compound.

Performance enhancing elements (e.g. refining elements) have a much more specific role and are intended to alter the physical and/or chemical processes that occur during the formation of materials (for instance during solidification), such that the size, morphology or distribution of the resultant species are favorably altered, or the hydrogen uptake kinetics are altered by one or other means, for example, a catalytic effect. That is, the addition of performance enhancing elements is not primarily intended to alter the chemical composition of the materials, that is, the combination with the other alloying elements is not particularly sought or intended.

A hypoeutectic alloy is different to a hypereutectic alloy. A hypoeutectic alloy is one where the composition of the alloy is richer in the element on the left hand side of the phase diagram (in this case magnesium) than that of the added eutectic-forming element. In other words, when the composition of a hypoeutectic Mg-X alloy is plotted on a Mg-X phase diagram it appears to the left of the eutectic composition. A hypoeutectic Mg-X alloy forms a primary Mg phase before the corresponding eutectic forms. The composition of a hypereutectic Mg-X alloy, by contrast, will appear to the right of the eutectic point and will therefore be richer in element X than is required to form the eutectic. A hypereutectic Mg-X alloy forms a primary intermetallic phase before the corresponding eutectic forms. The same intermetallic phase is also present as a constituent of the eutectic.

What defines the lines and regions visible in any given phase diagram is the sequence of phase formation that occurs during the cooling or heating of material of various specific compositions. Although both hypoeutectic and hypereutectic compositions will form material of the eutectic composition (the "eutectic phase") at the final stages of cooling from the melt, the ranges of each are differentiated by which phase forms prior to the eutectic reaction. The range of hypereutectic compositions, being characterized by an amount of the right-hand element greater than the eutectic composition, will first form the nearest stable (or meta-stable) phase to the right of the eutectic composition on the phase diagram. Conversely, hypoeutectic alloys, characterized by an amount of the left-hand element less than the eutectic composition, will first form the nearest stable or meta-stable phase to the left of the eutectic composition. In the case of the magnesium-based system, this is either near-pure magnesium or a magnesium solid solution containing some limited amount of element X (see Table 2).

Table 2 - Elements that form eutectics with Mg in Mg-X alloys and their main eutectic properties

Yb 8.0 45.5 Mg ₂ Yb 77.1

Zn 6.2 51 .3 Mg ₇ Zn ₃ 53.6

The resultant microstructures in each case display a "primary" phase (that is, the first to solidify) and the "eutectic" phase. It is the composition of the primary phase that determines whether a material is hypoeutectic or hypereutectic. To restate, in a magnesium-based system, a hypoeutectic material will display a primary phase of essentially pure magnesium, whereas a hypereutectic material will display a primary phase of Mg-X intermetallic compound.

There are various ways the alloy composition, and the amount of performance enhancing element added, can be defined. The most common method to define the alloy composition is by weight-based or mass-based percentage (wt% or mass%) used. In this method, the value refers to the mass of the alloying element compared to the total mass of the alloy (for example, 7 mass% Si in a Mg-Si alloy means there are 7 grams of Si for every 100 g of Mg and Si combined). To define the amount of performance enhancing element added, either a mass% notation as above can be used, or a ppm notation can be used. When ppm is used, the term ppm refers to the parts of element present per million parts of the total alloy. For example, 4000 ppm Na in a Mg-Si-Na alloy means there are 4000 parts (by mass) of Na for every 1 million parts (by mass) of Mg, Si and Na combined.

The methods according to the invention include the step of forming a Mg-X melt, the element(s) X being selected from the group consisting of Al, Ag, Au, Ba, Bi, Ca, Ce, Cu, Eu, Ga, Gd, Ge, Hg, La, Ni, Pb, Sb, Si, Sm, Sn, Sr, Th, Tl, Y, Yb and Zn, and being added in hypoeutectic amounts for all elements except Si which can be in hypoeutectic, eutectic or hypereutectic amounts.

A skilled person would understand how to determine the eutectic point for any element X (that is, the eutectic composition C _eu t and temperature T _eut ), and therefore what levels of X are hypoeutectic or hypereutectic amounts for each of the alloying elements. In any case, these eutectic points and other important data, for instance, the equilibrium solid solubility (C _ss ) of element X in Mg, are provided in Table 2. As an example, the eutectic point of a Mg-Sn system can be seen from Table 2 to be 36.9 mass% Sn, with the remainder Mg; anything less than this amount of Sn would be considered a hypoeutectic amount. The equilibrium solid solubility (C _ss ) of X in Mg is the maximum amount of X which can be present (in the dissolved state) within the Mg phase before the eutectic phase will start to form. This condition occurs at the eutectic temperature. This value is only achieved during equilibrium solidification conditions (that is, extremely slow cooling). There will be no eutectic phase present in the alloy structure if the amount of X added is less than the solid solubility of X at the eutectic temperature if slow equilibrium cooling prevails (for example, 14.5 mass% for Sn, Table 2).

In practical casting operations however, non-equilibrium solidification conditions prevail (that is, much faster cooling than above) and therefore the effective solid solubility of X at the eutectic point is reduced and eutectic phase is formed at lower compositions than otherwise expected as given by the equilibrium data. This non-equilibrium solidification is often called "Scheil" solidification. Without being bound by theory, the Applicant's propose that the non-equilibrium solid solubility of element X in Mg can conveniently be expressed as a percentage of C _ss . For example, 20% of C _ss , or 0.2 times C _ss , is a practically useful value for the purposes of calculation described below. In the case of the example Sn, eutectic phase may therefore be expected to form down to a Sn content of 2.9 mass%, rather than only above 14.5 mass%.

For each alloying element X, the corresponding eutectic phase is a mixture of Mg solid solution and an intermetallic phase consisting of Mg and X atoms in fixed atomic proportions illustrated by the chemical formulae of the phases shown in Table 2. This can be also represented as mass% of X in the intermetallic phase, designated C _in t.

The respective amounts of primary Mg solid solution phase (or primary Mg ₂ Si intermetallic for hypereutectic amounts of Si in Mg) and binary eutectic (which is co- formed material consisting of Mg and the associated Mg _a X _b intermetallic) can be calculated for any given starting alloy composition (C ₀ ) by the application of certain principles well-known to those skilled in the art.

From experience we propose that for optimum hydrogen adsorption/desorption characteristics in an Mg-X alloy it is not a certain fixed mass% range of element X that governs performance, but rather the total mass% amount of the eutectic intermetallic phase(s) that are formed in the microstructure for that composition range. Therefore since the values of C _ss , C _eu t and C _in t vary for each X, the amount of X that needs to be added in the alloy (C ₀ ) to obtain the desired total mass% of intermetallic phase(s) will also differ. As a simplifying practice, when two or more X elements are added, we treat each X as if it operates as a unique binary eutectic (using the values from Table 2) in order to calculate and then simply summate the amounts of eutectic intermetallic phases that form at any given composition for each X. We deliberately avoid complications presented by the formation of ternary eutectic intermetallic phases of the form Mg _a X _b Y _c or similar. The data for these phases are not always well established, phase stoichiormetry is sometimes variable, and the precise sequence of formation of the possible intermetallic phases can vary depending on the exact alloy composition and the cooling conditions. By this simplified method we are able to determine optimum alloy composition ranges for hydrogen performance using readily available binary alloy data.

The preferred and most preferred ranges of alloy compositions shown in Figures 1 -1 5 were prepared using the above methodology and then applying the following selection criteria based on the total amount (mass%) of eutectic intermetallic phases formed.

All examples (given in Figures 1 -15) were calculated using a 0.2 factor for solid solubility to simulate non-equilibrium cooling conditions during casting, with selection criteria for alloy content combinations based on the following values of mass% of intermetallic phases:

1 . Lower Preferred Range between 0.2 and 0.5 times 1 2.8 mass% of total intermetallic phases (the relevance of this number is described below), that is, between about 2.6 to 6.4 mass% of total intermetallic phases;

2. Middle Most Preferred Range between 0.5 and 1 .0 times 12.8 mass%, that is, between 6.4 and 1 2.8 mass% of total intermetallic phases;

3. Upper Preferred range between 1 .0 and 1 .25 times 12.8 mass%, that is, between 1 2.8 and 16.0 rmass% of total intermetallic phases.

In practice, these preferred and most preferred composition ranges of element(s) X can be further limited based on the final cost and density of the alloy relative to a suitable reference point, and/or by imposing a maximum allowable combined mass% of additions.

It should be noted that the value of 1 2.8 mass% of intermetallic phase, used in the calculations above, is defined for use in the specification and claims as a reference value based on the total amount of intermetallic phase known to be present in a well- established model Mg-based alloy that is used for hydrogen storage. This amount of intermetallic phase represents a good balance between catalytic behaviour, hydrogen storage capacity and alloy cost. This balance of properties derives from having a microstructure with sufficient intermetallic phase to provide useful catalytic effects and good hydrogen transport pathways, while retaining a large enough amount of Mg solid solution in the structure to provide a sufficiently large hydrogen occluding sink. By replicating this total amount of intermetallic phase, or particular ranges of total amounts of intermetallic phase derived from this value, in other eutectic alloy systems, it is possible to design and generate alternative alloy compositions that also have acceptable levels of catalytic behaviour, hydrogen storage capacity and alloy cost. This is demonstrated in the Examples presented later in this application. The described predictive approach to determine the total amounts of intermetallic phase in the microstructures is supportable by the results of XRD studies of the constituent phases of the alloys.

The method of forming the Mg-X melt may be known by those skilled in the art. Preferably, solid element (X) is added to molten Mg, however it may also be possible to add solid Mg to a molten metal (X). When there is more than one X, each X may be added simultaneously or sequentially, before, during or after the melting of the Mg. The temperatures necessary to prepare the melts are determinable by the skilled person. The melt may then be mixed to provide a homogenised mix of the Mg and element(s) X. The temperatures necessary to cast and solidify the melts are also determinable by the skilled person.

The methods according to the invention also include the step of adding Na up to 2 mass% as a performance enhancing element. The addition of Na may occur prior to, simultaneous with, or following the addition of the element(s) X to the molten Mg. Preferably, the Na is added afterwards and the molten mixture is again mixed to form a homogeneous mixture of Mg, element(s) X and Na. Preferably, the molten steps are conducted under a protective, non-oxidising atmosphere to prevent the Mg-based alloy from combusting. Typical atmospheres include SF ₆ and/or HFC-1 34a mixed in varying proportions with a carrier gas such as air, C0 ₂ or Ar.

The methods according to the invention further include the step of solidifying the melt to produce the Mg-X alloy. This may be done in any manner known in the art. For instance, the metal may be cast by a suitable casting procedure (for example, pouring) into suitable moulds. Suitable casting methods include, but are not limited to, sand casting, gravity die casting, and squeeze casting. The Mg-X alloy is preferably able to be used as a hydrogen storage material at this stage, that is, without further microstructure-altering processing such as heat treatment. Note, that this does not preclude post-cast processing of the cast alloy such as finely dividing the cast alloy, for example, by chipping, rasping, machining, filing, or crushing into chips or particles. The material so divided can be loaded into a vessel intended to support the storage of hydrogen in magnesium by way of the relevant temperatures and pressures.

The mechanical properties of the resultant Mg-X alloys will vary and be dependent upon the composition and method of manufacture, as would be understood by the skilled person.

The hydrogen activation, absorption and desorption characteristics will be dependent upon the alloy composition and method of manufacture. These characteristics are not predictable using first principles; however it is clear from our studies that the addition of the performance enhancing element Na leads to an improvement (as compared to the same material without the presence of Na). In practice, we have demonstrated that the addition of sodium to the Mg-X eutectic class of materials can perform one or more of several functions:

• Improve the initial activation rate of alloys for hydrogen storage, e.g. Mg-Si, Mg- Cu, Mg-AI; · Activate alloys that are otherwise unable to absorb hydrogen, e.g. Mg-Sn, Mg-Y;

• Improve the absorption rate of hydrogen under practical cycling conditions, e.g. Mg-Si;

• Improve the early stage cycling activation rate for alloys, e.g. Mg-Cu;

The addition of sodium may achieve the above effects by one or more of several means:

• Altering the nature of the surface oxide which acts as an interface through which hydrogen must pass to enter or exit the alloy;

• Acting as a grain boundary segregating element that provides an easy medium for hydrogen transport paths into and out of the alloy; • Acting as a grain boundary segregating element that provides a brittle interface between grains that favours easier fracture during chip formation and hydrogen cycling;

• Providing a catalytic effect over and above that achieved by the presence of the major alloying element or elements, if any;

• Refining the microstructure so providing shorter hydrogen transport and diffusion distances for the hydrogen atoms.

As an example, one of the unusual properties of sodium to improve hydrogen alloy performance are discussed in a forthcoming technical paper by Tran et al entitled "Effect of trace Na addition levels on the hydriding kinetics of hypoeutectic Mg-Ni alloys", in which the authors conclude that "the improvement is found more strongly dependent on the chemical effect of Na compared to other microstructural factors".

Examples

Manufacture and Testing for Examples 1 to 10 Cast alloys of nominal composition Mg-1 %Si, Mg-2%Si, Mg-25%Sn, Mg-1 7.5%Y, Mg- 27%AI, Mg-5%AI-3%Ca, Mg-3.5%Cu-3.5%Ni, Mg-4%Cu-1 %Ni, Mg-4%Cu-1 %Si and Mg-1 %Ni-1 .6%Si, each with and without addition of Na as a refining element, were prepared using a furnace to melt the ingredients. The contents of the elements in the alloys are also shown in Table 3. Note: all alloy compositions in the following examples are expressed in mass %. In more detail, the Mg-1 %Si, Mg-2%Si, Mg-25%Sn, Mg- 17.5%Y, Mg-27%AI, Mg-5%AI-3%Ca, Mg-3.5%Cu-3.5%Ni, Mg-4%Cu-1 %Ni, Mg-4%Cu- 1 %Si and Mg-1 %Ni-1 .6%Si alloys were produced by first melting industrial grade Mg under a protective atmosphere in an electric resistance furnace at 750 °C. Si, Sn, Y, Al, Ca, Cu and Ni, as required, were then added as particles and stirred into the melt. Each melt was held for 150 min to ensure full dissolution. Next, Na metal was added to each melt as the performance enhancing element. An amount of 4,000ppm Na was added to each melt, however a large percentage of the Na is lost through combustion and evaporation, the alloy typically retaining nominally 2000 ppm Na in the cast product. After homogenisation, the melt was cast into steel moulds (preheated to 250 °C) of cylindrical dimension (15mm internal diameter, 200mm height and 30 mm wall thickness). After solidification, samples were sectioned perpendicular to the longitudinal axis of the cylinder, at a location 1 0 mm from the bottom, and polished for metallographic analysis. Micro-structural characterisation was carried out by scanning electron microscopy (SEM) for bulk 'as-cast' samples.

For the Mg-0.5%Si, Mg-25%Sn, Mg-17.5%Y and Mg-27%AI alloys, the hydrogen absorption/desorption testing was conducted using a gravimetric-type automated apparatus (Technosystem Ltd. PCTM-5000A). The alloy material in each case was tested in the form of bulk swarf/chips produced from the as-cast alloy materials. Testing and measurement using the Technosystem device was as follows: Activation was performed at a constant temperature of 350°C and pressure of 2 MPa for 20 hours. Following activation, rate-limited absorption/desorption cycles were performed at conditions of constant temperature of 350°C using a pressure of 0.2 MPa for 24 minutes for desorption, and a 2-stage absorption also carried out at 350°C (1 MPa for 24 minute, followed by 1 .5 MPa for a further 24 minutes). Table 3 - Test Alloy Compositions, Amount of Intermetallics, and Hydrogen

Performance Data for Examples 1 - 10

For the Mg-2.1 %Si, Mg-5%Ai-3%Ca, Mg-3.5%Cu-3.5%Ni, Mg-4%Cu-1 %Ni, Mg-4%Cu- 1 %Si and Mg-1 %Ni-1 .6%Si alloys, the hydrogen absorption/desorption testing was conducted using an in-house designed and constructed gravimetric-type automated apparatus (designated X2). Testing in the X2 device was as follows: Activation was performed using a rapidly ramped temperature up to 340°C at a pressure of 1 MPa for 20 hours. Following activation, rate-limited absorption/desorption cycles were performed at the following conditions: desorption for 1 hour at 360°C and 0.1 MPa, followed by absorption for 3 hours at 340°C and 1 MPa.

Both of these testing protocols are capable of determining the key hydrogen performance values of the test material. These include: (1 ) Activation time (also known as "incubation time"), which is the time delay observed in the first test cycle before the material starts to physically absorb and occlude the hydrogen; (2) Saturation hydrogen level, which is the maximum amount of hydrogen (in mass%) absorbed at the end of the first cycle of 20 hours; and, (3) Cyclic effective peak hydrogen (EPH) value, which is the amount of hydrogen (in mass%) absorbed at the end of the final absorption period of a fixed number of absorption/desorption cycles (that is, the 15th cycle for Technosystem testing and the 4th cycle for X2 testing). Example 1: Mg-1%Si alloy

According to the equilibrium Mg-Si phase diagram shown in Figure 16, the alloy system has a eutectic composition of 1 .34 mass% Si. The expected mass % of intermetallic Mg ₂ Si phase in a Mg-1 %Si alloy is 2.7% according to the Lever rule. This represents a point close to the lower bound of the preferred intermetallic mass% range, that is, about 2.6, or 0.2 multiplied by 1 2.8.

Figure 17 shows SEM-BSE images of Mg-1 %Si and Mg-1 %Si-Na alloys. The addition of Na appears to break-up the brittle Mg ₂ Si phase in the eutectic which otherwise appears as a dense grain boundary phase in the absence of Na. The presence of Na also provides a brittle grain boundary layer that further facilitates the breakdown of alloy particles during repeated hydrogen absorption and desorption cycles which in turn leads to improved hydrogen kinetics.

The hydrogen activation (absorption) measurements in Figure 18 show that the Mg- 1 %Si alloy demonstrates a gradual increase in absorption capacity during the activation under conditions of 350 °C and 2.0 MPa. The activation curve of the alloy with Na added demonstrates the superior hydrogen absorption kinetics of this alloy (3 hour activation, saturation hydrogen level of 6.0%) over the Na-free alloy (4.5 hours activation, saturation hydrogen level of 5.9%). Following 15 absorption/desorption cycles the effective peak hydrogen values (EPH) of the Na-containing and the Na-free alloy were 5.5% and 4.0%, respectively.

Si is a lightweight and cheap alloy addition compared to more traditional and expensive elemental additions and therefore may be suitable on a cost-for-cost basis or a weight- for-weight basis even though hydrogen storage performance levels are modest.

Example 2: Mg-2%Si alloy According to the equilibrium Mg-Si phase diagram shown in Figure 16, the Mg-2%Si alloy has a hypereutectic composition. The expected mass% of intermetallic Mg ₂ Si phase in a Mg-2%Si alloy is 5.5% according to the Lever rule, or about 0.4 multiplied by 12.8. This represents a point closer to the lower bound of the more preferred intermetallic mass% range than Example 1 .

The Mg-2%Si alloy shows similar activation characteristics to the Mg-1 %Si alloy, achieving start of activation after 6 and 3 hrs for the Na-free and Na-containing alloy variants, respectively. Although the cyclic EPH value of the Na-free Mg-2%Si alloy is lower than the Na-free Mg-1 %Si alloy, that is 2.6% compared to 4.0%, the cycled EPH value in the Na-containing alloy is higher than the Na-containing Mg-1 %Si alloy, or 5.5% compared to 5.2%, respectively. This suggests that as Si is increased, the activation and hydrogen storage benefits of Na addition become greater, in part as a result of the Na providing a brittle layer in the grain boundaries that increases the ease of particle fragmentation.

Example 3: Mg-24%Sn alloy

According to the equilibrium Mg-Sn phase diagram shown in Figure 19, the system has a hypoeutectic composition between about 14.5 and 37 mass% Sn. The expected mass% of intermetallic Mg ₂ Sn phase in a Mg-24%Sn alloy is 33.9% according to the Lever rule, or about 2.6 multiplied by 12.8. This level is well above the 16% upper bound of the preferred range.

Figure 20 shows SEM-BSE of images of the Na-free and Na-containing Mg-24%Sn alloy. Although the Na-free alloy shows a fine lamellar eutectic structure, upon the addition of Na, the eutectic region appears to have become substantially more refined.

The hydrogen activation (absorption) measurements shown in Figure 21 demonstrate that the Mg-24%Sn alloy, even though it has a fine lamellar structure, has only a very limited hydrogen absorption capacity (0.3%) at the end of the 20 hour activation cycle under conditions of 350 °C and 2.0 MPa, that is, after a very long activation period (1 1 hours) possibly coinciding with the breakdown of oxides formed during handling. The activation curve of the alloy with Na added demonstrates far superior hydrogen absorption kinetics (with a much shorter activation time of less than 1 hour, much faster uptake and a much higher saturation hydrogen level of 3.3%) compared to the Na-free alloy. Following 15 absorption/desorption cycles the effective peak hydrogen values (EPH) of the Na-containing and the Na-free alloy were 2.8% and 0%, respectively.

In this example it is evident that there is too much intermetallic present in the microstructure and this limits the amount of available Mg for hydrogen uptake and occlusion. The Mg ₂ Sn intermetallic is also an ineffective catalyst which can however be substantially improved with the addition of Na as a performance enhancing element. Example 4: Mg-12%Y alloy

According to the equilibrium Mg-Y phase diagram shown in Figure 22, the Mg-12%Y alloy has a hypoeutectic composition. The expected mass% of intermetallic Mg ₂₄ Y5 phase in a Mg-12%Y alloy is 26% according to the Lever rule, or about 2 multiplied by 12.8. This level is well above the 1 6% upper bound of the preferred range.

Figure 23 shows SEM-BSE images of the Na-free and Na-containing Mg-12%Y alloy. The Na-free alloy shows a relatively coarse irregular eutectic, which remains largely unchanged, and possibly is coarsened, when Na is added to the alloy.

The hydrogen activation (absorption) measurements in Figure 24 show that the Mg- 12%Y alloy did not show any hydrogen absorption capacity during the 20 hour activation period under conditions of 350 °C and 2.0 MPa. This may be due to highly stable oxide resisting breakdown and allowing hydrogen uptake. As above, the presence of Y and its corresponding Mg ₂₄ Ys intermetallic does not instil any catalytic effect in the alloy. The activation curve of the alloy with Na added demonstrates far superior hydrogen absorption kinetics (with shorter activation time of 3.5 hour, much faster uptake and a much higher saturation hydrogen level of 2.0 mass%) compared to the Na-free alloy. Following 15 absorption/desorption cycles the effective peak hydrogen values (EPH) of the Na-containing and the Na-free alloy were 3.1 % and 0%, respectively. In this case, Na is demonstrated to exert a positive influence that does not derive from any refinement of the eutectic phase, perhaps instead exerting its own catalytic effect. Again we note the modest EPH value achieved in the Na-containing alloy and attribute this to the large proportion of ineffective intermetallic in the structure and the concomitant reduced amount of hydrogen-occluding Mg.

Example 5: Mg-27%AI alloy According to the equilibrium Mg-AI phase diagram shown in Figure 25, the Mg-27%AI alloy has a hypoeutectic composition. The expected mass% of intermetallic Mg ₁₇ AI ₁₂ phase in a Mg-27%AI alloy is 64.5% according to the Lever rule, or about 5 multiplied by 12.8. This level is well above the 1 6% upper bound of the preferred range.

The hydrogen activation (absorption) and cyclic absorption/desorption measurements in Figure 26 show that the Na-free Mg-27%AI alloy started absorbing hydrogen after only a very short activation time, approximately 0 hour, and the saturation hydrogen level reached 3.2% after 20 hours at 350°C and 2.0 MPa. This is quite different to the rather unusual response in the Na-containing alloy where activation was delayed until after an activation time of 7.5 hours but where hydrogen absorption was then rapidly accelerated to achieve 4.7% in 20 hours. In this case, Na appears to retard nucleation of hydride cells but either produces more efficient nucleant particles in the long run or else provides easier hydrogen transport pathways, compared to the alloy without Na, where nucleation is easy but either the nucleants so formed are ineffective and/or the hydrogen transport is limited.

Following 1 5 absorption/desorption cycles the effective peak hydrogen values (EPH) of the Na-containing and the Na-free alloy were 3.6% and 2.9%, respectively. In this case, Na is demonstrated to exert a clear positive influence that presumably derives from an influence of hydride nucleation and possibly accelerated hydrogen diffusion. Again we note the modest EPH values achieved in the Na-free and Na-containing alloy and attribute this to the large proportion of ineffective intermetallic in the structure and the corresponding reduced amount of hydrogen-occluding Mg. Example 6: Mg-5%AI-3%Ca alloy

In order to demonstrate the concept that X might be more than one element from the list of eutectic forming elements (namely Al, Ag, Au, Ba, Bi, Ca, Ce, Cu, Eu, Ga, Gd, Ge, Hg, La, Ni, Pb, Sb, Si, Sm, Sn, Sr, Th, Tl, Y, Yb and Zn), we chose to test a Mg-5%AI- 3%Ca ternary alloy without Na and with a nominal level of 2000 ppm of Na. This alloy composition is hypoeutectic with respect of both Al and Ca, and contains a comparable total mass% of eutectic-forming elements as a known commercial Mg-based alloy used for hydrogen storage. Based on the eutectic compositions, the alloy should also form a similar mass% amount of intermetallic compounds as the commercial alloy, that is, 12.7% compared to 12.8%. The composition lies at the upper bound of our most preferred range as illustrated in Figure 1 .

The alloy without sodium addition performed poorly with an activation time of 15 hours, saturation hydrogen level of 0.6% after 20 hours and a cyclic EPH value of 1 .1 % after the 4th cycle. This was significantly improved with the nominal level of 2000 ppm Na yielding a shortened activation time of 8 hours, saturation hydrogen level of 2.1 %, and an increased cyclic EPH value of 3.7%. This clearly demonstrates a significant positive benefit for a Na addition to an otherwise moderately-performing alloy. It should be noted that although the EPH value is moderate, it is equivalent to or better than that the values achieved in the alloys in Examples 3-5 above. These values were achieved using much lower elemental addition levels than those used in Examples 3-5. In addition, they demonstrate the benefit of limiting the total mass% amounts of intermetallic phases in the microstructure to within the preferred composition ranges predicted using the calculation method described earlier.

Example 7: Mg-3.5%Cu-3.5%Ni alloy

According to Figures 27 and 28, the Mg-3.5%Cu-3.5%Ni alloy composition is hypoeutectic with respect to both Cu and Ni. As for Example 6, it contains a comparable total mass% of eutectic-forming elements as a known commercial Mg-based alloy used for hydrogen storage. Based on the eutectic compositions, the alloy should also form a similar rmass% amount of intermetallic compounds as the commercial alloy, that is, 12.6% compared to 1 2.8%. The composition lies at the upper bound of our most preferred range as illustrated in Figure 10.

In the Na-free variant of the alloy, hydrogen absorption started after an activation time of 5 hours, and achieved a good cyclic EPH value of 5.1 % following the 4 ^th absorption cycle. This is significantly better than the Na-free alloy in Example 6 and demonstrates that the nature of the intermetallic phase types is an important determining factor, in addition to the total amount of interrmetallics present.

When Na is added to the Mg-3.5%Cu-3.5%Ni alloy as a performance enhancing element the activation time was reduced to 2 hours and the cyclic EPH value was increased to a high level of 6.0%. In this variant, the combined benefits of intermetallic types, total amount of intermetallic and the presence of the performance-enhancing element Na is clearly demonstrated. Example 8: Mg-3.5%Cu-3.5%Ni alloy

According to Figures 27 and 28, the Mg-4%Cu-1 %Ni alloy composition is hypoeutectic with respect to both Cu and Ni. Compared to Example 7, it contains a reduced total mass% of eutectic-forming elements. This is lower than the amount found in a known commercial Mg-based alloy used for hydrogen storage. Based on the eutectic compositions, the alloy form a significantly smaller mass% amount of intermetallic compounds as the commercial alloy, that is, 8.9% compared to 1 2.8%. The composition lies centrally between the upper and lower bounds of the most preferred range as illustrated in Figure 10.

In the Na-free variant of the alloy, hydrogen absorption started after an activation time of 4 hours, and achieved a good cyclic EPH value of 5.1 % following the 4 ^th absorption cycle. This is similar to the Na-free Mg-3.5%Cu-3.5%Ni alloy in Example 7 and demonstrates that the nature of the intermetallic phase types is an important determining factor, and in addition that it can be beneficial to reduce the total amount of intermetallics present when they are highly effective catalysts, as similar performance can be achieved at lower cost and weight penalties.

When Na is added to the Mg-4%Cu-1 %Ni alloy as a performance enhancing element the activation time was reduced to 0 hours and the cyclic EPH value was increased to a very high level of 6.5%. In this variant, the combined benefits of intermetallic types, reduced amount of total intermetallics and the presence of the performance-enhancing element Na is clearly demonstrated. Example 9: Mg-4%Cu-1%Si alloy

According to Figures 16 and 27, the Mg-4%Cu-1 %Si alloy composition is hypoeutectic with respect to both Cu and Si. Compared to Example 8, it contains a similar total mass% of eutectic-forming elements. This is lower than the amount found in a known commercial Mg-based alloy used for hydrogen storage. Based on the eutectic compositions, the alloy form a smaller rmass% amount of intermetallic compounds than the commercial alloy, that is, 9.8% compared to 12.8%. The composition lies centrally between the upper and lower bounds of the most preferred range as illustrated in Figure 1 1 .

In the Na-free variant of the alloy, hydrogen absorption started after an activation time of 1 1 hours, and achieved a modest cyclic EPH value of 2.6% following the 4 ^th absorption cycle. This is lower than the Na-free Mg-4%Cu-1 %Ni alloy in Example 8 and demonstrates that the nature of the intermetallic phase types is an important determining factor in addition to the total amount of intermetallics present.

As noted in Examples 1 and 2, the Mg ₂ Si intermetallic phase is not a particularly effective catalyst unless Na is also present as a performance enhancing element. When Na is added to the Mg-4%Cu-1 %Si alloy the activation time was reduced to 3 hours and the cyclic EPH value was increased to a very high level of 6.2%, both values not as good as for Na-containing Mg-4%Cu-1 %Ni but nevertheless highly acceptable hydrogen storage values, particularly when the significant cost benefit of replacing a 1 %Ni addition with a 1 % Si addition is taken into account.

Example 10: Mg-1%Ni-1.6%Si alloy According to Figures 16 and 28, the Mg-1 %Ni-1 .6%Si alloy composition is hypoeutectic with respect to both Ni content and hypereutectic with respect to Si content. Compared to the previous ternary alloys (Examples 6-9), it contains a significantly reduced total mass% of eutectic-forming elements. This is also lower than the amount found in a known commercial Mg-based alloy used for hydrogen storage. Based on the eutectic compositions, the alloy forms a smaller mass% amount of intermetallic compounds than the commercial alloy, that is, 6.2% compared to 12.8%. This Mg-1 %Ni-1 .6%Si alloy composition lies close to the lower bound of the most preferred range as illustrated in Figure 13.

In the Na-free variant of the alloy, hydrogen absorption started after an activation time of 6 hours, and achieved only a low cyclic EPH value of 1 .6% following the 4 ^th absorption cycle. This is lower than all of the Na-free ternary alloys except for Mg-5%AI-3%Ca in Example 8. However, on a pro-rata amount based on the total amount of intermetallic the EPH value may be considered much more favourably than expressed by its base value.

However, it is extremely noteworthy that when Na is also present as a performance enhancing element, the hydrogen storage characteristics are dramatically improved. The Na-containing Mg-1 %Ni-1 .6%Si alloy shows a reduced activation time of 3 hours and a high cyclic EPH value of 5.8%. As for Example 9, these are highly acceptable hydrogen storage values, particularly when the significant cost benefit of the low overall elemental addition level is considered.

This example combined with other examples above, clearly demonstrate that commercially useful hydrogen storage alloys can be designed based on significantly lower amounts of intermetallic phase(s) present in the microstructure than is normally considered viable. The methodology used to predict and design these alloy compositions is supported by the data of these examples. However, it has also highlighted that although the total amount of intermetallics is a good indicator of hydrogen storage performance it is not the sole determinant of good or exceptional performance; the contributions of the different intermetallic phases vary as does their response to the performance-enhancing element Na.

Further Examples

Extensive testing has been undertaken on various preferred alloy compositions. The full list of these compositions and results are presented in Table 4 below. In all cases it is clear that the addition of Na confers a benefit to the EPH, and in many cases the incubation and saturation performance are also improved.