METHOD FOR PRODUCING GRAIN REFINED MAGNESIUM AND MAGNESIUM-ALLOYS

Abstract

The present invention relates to a method for producing grain refined Mg and Mg alloys by solidifying molten Mg or Mg-alloys in the presence of ZrB2 or ZrC particles leading to a grain refined microstructure. The particles have an average particles size less than 20 μm and larger 0.1 μm and are either produced by an in situ formation within the melt or by a high temperature treatment of an intimate mixtures of zirconium oxide and carbon black and in case of ZrB2 additionally a suitable boron source like boron carbide or boron oxide or boron followed by a milling and de-agglomeration procedure. The method according to the invention leads to grain refined Mg or Mg alloys with an average grain size less than 100 μm.

Description

The present invention is directed to a method for producing grain refined Mg and Mg alloys by solidifying molten Mg or molten Mg alloys in presence of a nucleation agent.

BACKGROUND TO THE INVENTION

Mg alloys are used for a wide variety of applications. Mg castings have become more prevalent due to their high specific strength and low weight in transport industries especially automotive [1]. The high demand in the automotive industry for weight savings has resulted in successful applications for Mg products which include seat frames, inner door frames, instrument panels, steering wheels, cylinder head covers etc. The majority of the Mg alloys produced are used for die casting but over the last few years the demand for Mg wrought products has increased [2].

One of the most effective approaches to improve the mechanical properties of Mg alloys is to reduce the grain size of the primary Mg. Due to the hexagonal close packed lattice structure of the Mg and the poor plastic deformation rates at room temperature a fine grain size improves the plastic deformation behavior. The grain size of castings can be influenced by modifying casting parameters or by adding alloying elements which contain nucleants. For Al alloys there are reliable grain refiners commercially available such as TiB₂ and TiC based master alloys. Unfortunately this is not the case for all Mg alloys [3]. Aim of the present invention is to provide a refiner addition working for Mg, for Mg alloys and for Mg—Al alloys in particular.

STATE OF THE ART

In current cast house and foundry practice different approaches are taken to grain refinement depending of the Al content of the Mg alloys. When no Al is present in the Mg alloy Zr addition is a very effective nucleation agent. Zirconium in most cases is added by a master alloy typically containing 33% Zirconium and 67% Magnesium known as Zirmax—being a trade mark from Magnesium electron LtD, MEL. This procedure has been developed by Magnesium Elektron Ltd. in about 1945. Recent literature [US 2005/0161121] proposes alternative procedures to add Zirconium to magnesium e.g. by adding Zirconium as sponge treated by hydrofluoric acid in order to overcome the problem of passivating surfaces. Recent literature [4] explains the benefit of Zirconium metal where Zr acts as a nucleation site as well as a growth restrictor [5,6] permitting grain refinement.

Disadvantages of grain refinement with Zirconium are the relative high costs and necessary high operating temperatures leading to an increased fire risk. In Mg alloys where Al is present (alloy types AZ31, AZ91, AM60) addition of Zr is believed to lead to the formation of an undesired Al3Zr phase which is brittle and therefore detrimental to the mechanical properties and by consuming most of the Zirconium no significant grain refinement is observed [7].

Mg—Al alloys can be grain refined by overheating [IG Farben, GB 359425 (1931) und GB 608941], addition of FeCl [U.S. Pat. No. 2,461,937] and addition of carbon sources [GB 608941, CN 1583327] or by hydrocarbons [GB 608942, GB 653242, U.S. Pat. No. 2,448,993]. While all these methods can achieve grain refinement successfully with small grain sizes depending on cooling rate and alloy concentration [5,6] these methods are often not reproducible. Moreover, overheating leads to increase fire risk and increase Fe concentration from the mild steel crucible and thereby to a decreased resistance against corrosion. Similarly the addition of FeCl will also decrease the corrosion resistance. FeCl and hydrocarbon addition possess a large negative impact on environmental pollution.

Some recent propriety grain refinement agents include compounds such as SiC [8,9] or compounds based on Al (Al4C3, AlN) or Ti (TiC,TiN TiB2) [JP 2001/342528] or mechanically alloyed compounds in specific size ranges [WO 02/46484]. Overall the grain refinement of Mg alloys is poorly controlled [10] and each grain refinement variant has limiting attributes. An overview of the current understanding of grain refinement in Mg alloys is given in the recent scientific literature [4, 8-13].

For non Al containing Mg-alloys peritectic Zr additions (Zr>0.6 wt. %) have been found to be beneficial in decreasing the grain size effectively. Zr particles act as nucleation sites and free Zr can act as a growth restrictor. The lattice parameters of Zr are very close to those of Mg (a=0.320 nm, c=0.520 nm) and if undissolved Zr particles exist, they may readily act as nucleants. Another problem is the correct particle size because the denser Zr particles settle relatively fast in molten Mg. The driving force is the difference between the densities of Zr and the molten Mg (δ_Zr/δ_Mg≈4). The optimum overall process particle size lies between 2 and 5 μm [12, 13]. For classical nucleation and typical values for the interfacial energy (γ_sl=0.558 J/m²), entropy of fusion (ΔS_f=0.68506*10⁶ J/m³K⁻¹) and undercoolings ΔT=1° C. a critical radius is on order of 1.3 μm. The particle substrate diameter should lie in the same range (Eqn 1) as twice the critical nucleus radius.

$\begin{array}{cc}{r}_{\mathrm{crit}}=\frac{2·{\gamma }_{\mathrm{sl}}}{\Delta S_f·\Delta T_m}& \left(\mathrm{Eqn}.1\right)\end{array}$

r_crit critical nucleation radiusγ_sl surface interfacial energyΔS_f entropy of fusion per unit volumeT temperature

A particle size of ˜3 μm is therefore a good compromise between settling velocity and nucleant potency and efficiency [14].

For effective heterogeneous nucleation the nucleant, the lattice mismatch, and the chemical interaction, such as a peritectic reaction, all play an important role in the effect on the overall interfacial energy (γ_sl) [15; 16].

The role of solute elements on grain size is as important as that of the nucleant. After initial nucleation the growth velocity determines the rate of heat released by recalescence. This decreases the activation undercooling for the available nucleating particles and reduces the number density of growth centers. The solute elements generate a constitutional undercooling in the diffusion layer ahead of the advancing solid/liquid interface, which restricts the grain growth. Therefore in front of the interface, nucleants are more likely to survive and be activated. The growth restriction factor (Q) for equiaxed growth can be estimated from binary phase diagrams:

$\begin{array}{cc}Q=\sum _im_iC_{0,i}\left(k_i-1\right)& \left(\mathrm{Eqn}.2\right)\end{array}$

Where m_i is the slope of the liquidus line, k_i the distribution coefficient and C_0,i the initial concentration of element i [3, 17, 18]. From Table 1 it can be seen that Zr and Fe are very good growth restrictors. Due to the corrosion behavior of Mg, the addition of Fe is not a suitable solution for alloys. However, only solute elements in the melt are freely available for growth restriction. Elements such as Al, Mn, Si, Fe, Ni, Sn and Sb are said to form stable compounds with the Zr [5]. In the case of Al containing alloys, Zr forms Al₃Zr and thereby generates a poisoning effect of the grain refinement mechanism [7].

TABLE 1
Growth Restriction Parameter (m(k − 1)) for various alloying
elements in Mg [5, 6].
Element Q [m(k − 1)]
Fe      52.56
Zr      38.29
Si      9.25
Zn      5.31
Cu      5.28
Al      4.32
Sn      1.47
Mn      0.15

The objective of this invention is provide a method for producing Mg and Mg alloys having a grain refined microstructure, e.g. having an average grain size less than or near to 100 μm.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a schematic of an experimental setup for melting.

FIG. 2 illustrates a TP1 test apparatus.

FIG. 3 illustrates a conical sampling ladle.

FIG. 4 depicts a TP1 sample, including an indication of an area of investigation.

FIG. 5 shows a SEM BSE image, using Z contrast, of a ZrB₂ particle and AlZrTi phases.

FIG. 6 illustrates the EDS spectrum that corresponds to the ZrB₂ particle of FIG. 5.

FIG. 7 is a graph that shows grain size vs. time for ZrB₂ particles formed in situ.

FIG. 8 illustrates an example of the microstructure of non-refined, pure Mg.

FIG. 9 illustrates an example of the Mg alloy microstructure two minutes after adding 0.17 wt. % Zr to the melted Mg of FIG. 8.

FIG. 10 illustrates an example of the Mg alloy microstructure two minutes after adding 0.25 wt. % Zr to melted Mg.

FIG. 11 illustrates an example of synthetic ZrB₂ particles.

FIG. 12 is a graph that shows grain size vs. time for 0.1 wt. % ZrB₂ particles and 0.5 Wt. % ZrB₂ particles.

FIG. 13 illustrates an example of the Mg alloy microstructure two minutes after adding 0.1 wt. % ZrB₂ particles to melted Mg.

FIG. 14 illustrates an example of the Mg alloy microstructure two minutes after adding 0.5 wt. % ZrB₂ particles to melted Mg.

FIG. 15 illustrates an example of synthetic ZrC particles.

FIG. 16 is a graph that shows grain size vs. time for Mg+2 wt. % Al+ZrC, with a significant reduction of grain size.

FIG. 17 illustrates an example of the Mg alloy microstructure two minutes after adding 0.1 wt. % ZrC particles to melted Mg.

FIG. 18 illustrates an example of the Mg alloy microstructure two minutes after adding 0.5 wt. % ZrC particles to melted Mg.

FIG. 19 is a graph that shows grain size vs. time for AZ31 with 0.1 wt. % ZrB₂, with a minimal grain size of below 100 μm after two minutes.

FIG. 20 illustrates an example of the alloy microstructure after adding 0.1 wt. % ZrB₂ to AZ31, with a typical grain size about 130 μm.

FIG. 21 illustrates an example of the alloy microstructure after adding 0.1 wt. % ZrB₂ to AZ91, with a typical grain size about 80 μm.

FIG. 22 illustrates an example of the alloy microstructure after adding 0.1 wt. % ZrC to AZ31, with a typical grain size about 80 μm.

FIG. 23 is a graph that shows grain size vs. time for AZ31 with 0.1 wt. % ZrC.

FIG. 24 illustrates an example of the alloy microstructure five minutes after adding 0.1 wt. % ZrC to AZ31, with a typical grain size about 140 μm.

SUMMARY OF THE INVENTION

The method according to the present invention for producing grain refined Mg and Mg alloys by solidifying molten Mg or molten Mg alloys in presence of a nucleation agent is characterized in that as said nucleation agent ZrB₂ and/or ZrC is used.

The nucleation agent has a particle size preferably within the range of 0, 1 μm and 20 μm, more preferred within the range of 1 μm and 4 μm.

The nucleation agent may be added to the molten Mg or Mg alloy or may be produced within the molten Mg or Mg alloy.

It has been shown that best results are achieved when ZrB₂ and/or ZrC is present in the molten Mg or Mg alloy in a concentration within the range of 0.01 wt. % and 1 wt. %.

Good grain refinement of Mg and Mg alloys, in particular Mg—Al alloys, can be obtained by producing in situ ZrB₂ particles by a reaction of conventional Al grain refiner master alloys (Al—Ti—B) with Al—Zr master alloys. In situ ZrB₂ particles are formed in the master alloy and result in the grain refinement of Mg alloys containing Al. Ultimate grain sizes of 100 μm can be obtained. Fading of the grain refining performance is observed after stopping to stir after an additional approximately 60 minutes. This appears to be caused by sedimentation of the nucleant particles and could be partly overcome by renewed stirring. Some effect of growth restriction can be observed, the higher Q-value results in a finer grain size. A metallographic examination showed some Al₃Zr phases in the samples which were refined using the produced master alloy containing Zr.

Even better grain refinement performance compared to the addition method of the pre-produced master alloys, is the addition of synthetic ZrB₂ and ZrC particles which offer further advantages, because a finer grain size and even less fading is achieved. An ultimate grain size of 60 μm and 70 μm after 2 minutes could be obtained for ZrB₂ and ZrC respectively despite only moderate growth restriction by free Al in the alloy. No Al₃Zr phases were present in the microstructure. Grain refinement was mainly obtained by the provision of nucleation sites with a higher number density of heterogeneous particles resulting in a finer grain size. Surprisingly, fading of the grain refining mechanism was less than that for the in situ grain refining particles.

By this invention the best solidification procedure for Mg or Mg-alloy in terms of grain refining effect is achieved by introducing an amount of 0.01-1 wt. % of ZrC and or ZrB₂ particles preferably between 0.10-0.20 wt. % with an average grain size of 0.1-10 μm preferably 1-4 μm into the Mg or Mg-alloy melt.

Higher additions than 1 wt. % particles may still lead grain refinement but not by the action of a trace element but as an alloying element leading ultimately to a metal matrix composite. Coarser particles can also be effective in grain refining if they comprise of agglomerates of finer primary particles. Solid particles larger 10 μm shall be excluded in order to avoid harmful influence on the mechanical properties or the Mg or Mg-alloy. However, larger solid particles are very likely to settle in the melt. It is necessary to introduce the particles into the melt and avoid unsuccessful procedures which keep the particles on the surface of the melt. Successful procedures can be blowing the powder into the melt pool by inert gas flow, introducing as a softy pressed tablet, wrapped into a foil or as a cored wired or any equivalent procedure. Furthermore it is recommended to distribute the particles evenly in the melt by mechanical stirring.

The particles can preferably be obtained from a carbothermal reduction of an intimate mixture of zirconium oxide with a carbon source and in case of ZrB₂ an additional boron source. The carbon source is not necessarily but preferably carbon black e.g. type N-991. The boron source can be boron oxide or boron carbide powder. Elemental Boron can also be used but is not recommended due to high risk of ignition. The ratio of the compounds of the intimate powder mix needs to be balanced in a way that all moles of oxygen are balanced by moles of carbon in order to form carbon monoxide at the subsequent heat treatment and enough excess carbon or boron is still present to form the ZrC or ZrB₂ respectively.

A simultaneous production of a mixture of ZrC+ZrB₂ is also possible. The heat treatment needs to be carried out under vacuum or inert gas flow at temperatures and times which are sufficient to remove the oxygen from the oxides and to form the desired phases. These conditions depend on the quality of the mixture. Typical conditions are heating a mixture of ZrO₂+C+B₄C to 1900° C. in a graphite crucible and to keep the temperature for 2 hours or at least as long as no carbon monoxide formation can be detected. After subsequent cooling to room temperature the material needs to be crushed in standard crushers e.g. a jaw crusher and milled e.g. in a standard ball mill preferably with cemented carbide balls. Subsequent sieving and, or classifying is recommended to exclude large solid particles and to adjust the desired average particle size.

Other procedures leading to primary particles in the desired particle size range are also possible but not recommended due high manufacturing costs and difficult to control process conditions. These processes could be the synthesis from Zirconium—and/or Boron—halides with hydrocarbon gases or arc fusion processes with subsequent heavy milling or processes related to the described principles. However, it needs to be pointed out that the proposed synthesis has the advantage that it results in a high yield of particles already in the desired particle size range.

By way of the following Examples preferred embodiments of the invention are further described. Example 1 refers to the in situ formation of ZrB₂ in the melt, Examples 2 and 3 refer to the addition of ZrB₂-particles and ZrC-particles respectively into the melt and Example 4 refers to the addition of both ZrB₂-particles and ZrC-particles into the melt of commercial Mg alloys.

EXAMPLE 1

For the in situ formation of ZrB₂, two master alloys were used. The first was an Al—Ti—B master alloy containing 94 wt % Al, 5 wt. % Ti and wt. % B. This was a typical master alloy used in refining commercial Al alloys. The second was an Al—Zr alloy, which contained 10 wt. % Zr.

For the production of the new master alloy an induction furnace and a clay bonded crucible were used. 500 g of each alloy were added. The alloys were heated up to 800° C. under an Ar atmosphere to protect the melt surface from excessive oxidation. After 30 minutes of continuous stirring the alloy was poured into a graphite coated steel ingot. After pouring, the alloy was examined metallographically, chemically by induced coupled plasma spectroscopy (ICP-AES method) and by a scanning electron microscope (SEM) equipped with EDS. Table 2 shows the chemical composition of the produced master alloy.

TABLE 2
Chemical composition of the produced grain refiner alloy.
	Zr	Ti	Al
	[%]	[%]	[%]
	Chemical	5.1	1.3	balance
	composition

Two grain refining tests with different added Zr contents of 0.17 wt. % and 0.25 wt. % were performed. Pure Mg (Q-value of 0.50) was used and melted in an electrical resistance furnace using a graphite coated mild steel crucible. A schematic of the experimental setup is shown in FIG. 1. The overall Al content of the first melt was 2.7 wt. % (0.17 wt. % Zr) and for the second melt 3.7 wt. % (0.25 wt. % Zr) which resulted in Q-values of 18.17 and 25.55 respectively. The melt temperature was 730° C. The TP1 test apparatus (FIG. 2) was used for cooling the Mg samples. According to the TP1 test procedure described in the “Standard Test Procedure for Al Alloy Grain Refiners 1990, TP-1” [13] the samples are taken before and after adding the previously produced grain refiner additions. The graphite coated conical sampling ladle (fabricated of mild steel) was preheated to a temperature of 320° C. (FIG. 3). To protect the melt surface form the excessive oxidation a gas mixture of Nitrogen and 1.25 vol. % SF₆ was used.

One sample was taken before the addition of the master alloy to the melt and subsequent samples were taken at intervals of 1, 2, 5, 10, 20, 40, 60, 80, 100, 120, 180 and 185 minutes. After 120 minutes the continuous mechanical stirring was stopped. At 180 minutes a sample was taken and the stirrer restarted. After 5 minutes of stirring a sample was taken at 185 minutes to investigate if grain refiner performance is restored.

The produced grain refiner alloy composed of Al—Zr and Al—Ti—B was examined with a scanning electron microscope, in order to locate ZrB₂ particles present in the samples. FIG. 5 shows a scanning electron micrograph, in which two small particles are visible as bright rectangular particles in the vicinity of a larger grey Al₃(Zr,Ti) phase (FIG. 5). EDS spectrums from these small bright particles revealed a large Zr peak and a smaller B and Ti peak (FIG. 6). This suggests that the particles are mainly ZrB₂. Analysis from the surrounding matrix indicates the presence of Al and little some remaining Ti. Because of the excellent crystallographic match with the Mg these in situ formed ZrB₂ particles should readily act as heterogeneous nucleants. Furthermore, the in situ formation of the particles has the advantage, that there is no oxide on the surface of the particles.

Thereby the crystallographic match of the particle with Mg will not be influenced by oxides on the surface of the nucleant. As it can be seen in FIG. 7 the grain refinement effect is small with a Zr content of 0.17 wt. %. A smaller grain size of ≈100 μm can be achieved when the Zr content is increased up to 0.25 wt. %. It should be noted that both Zr concentrations are below 0.6 wt. % Zr i.e. the additions are hypoperitectic that is off the peritetic horizontal. The smaller grain size with the higher Zr content suggests a growth restriction effect. After 80 minutes the grain sizes increases slightly. At 120 minutes stirring was stopped for 60 minutes. A further sample was taken after 180 minutes. The grain size increased dramatically and subsequently declined strongly after 5 minutes stirring to approximately the size observed at 120 minutes.

This implies that the borides settle to the bottom of the steel crucible i.e. behave similar to boride particles in Al melts [15]. Likewise, with stirring it was possible to reactivate the settled particles but not to obtain the ultimate grain size measured after 5 minutes.

FIG. 8 shows the non refined state of the Mg showing a columnar structure. The microstructure altered significantly from columnar to equi axed with grain refinement. FIGS. 9 and 10 show typical examples of uniformly equiaxed grains in the microstructure 2 minutes after adding the master alloy to the Mg melt.

The transition of a columnar to uniformly equiaxed microstructure can be seen as the minimum requirement for successful grain refinement. It should be noted that some Al₃Zr phases are present in the grain refined microstructure which may be detrimental to further mechanical processing.

EXAMPLE 2

A batch of ZrB₂ particles was produced by mixing 37.67 kg ZrO₂ (commercially available 99.5% ceramic grade), 4.78 kg C (commercially available carbon black type N991), 8.55 kg B4C (commercially available 1-5 μm) in a high energy shear mixer type Lödige for 4 hours. This powder blend was filled in graphite boats and pushed under Argon atmosphere through a resistance heated graphite tube furnace. The temperature was set at 1800° C. and the material was kept for 6 hours in the heating zone. After cooling to room temperature the material was crushed in a jaw crusher and milled in a ball mill with ZrO2 ceramic balls for 1 hour.

Then the material was control sieved with a 100 mesh sieve. The analysis was C: 1.05%, N: <0.01%, O: 0.54% and an average grain size D50: 3.72 μm measured by a Cilas 1064 laser diffraction instrument in wet conditions.

FIG. 11 shows a scanning electron micrograph of the ZrB₂ particles used. The ZrB₂ particles displayed a hexagonal platelet morphology. The good lattice match on the ZrB₂ particles (table 1) occurs only on the basal planes. The average ZrB₂ particle size lies in the range of 0.5 to 4 μm and corresponds well with the stated average diameter given by the supplier. However, it is noticeable that clustering can occur and cluster sizes can be in the order of 10 μm. The particle sizes are therefore of the same magnitude as a critical homogeneous nuclei (diameter: dent 3 μm) with an undercooling of 1° C.

Two series with different particle contents were cast. Approximately 2 wt. % Al (99.9%), which results in a Q factor of 9.12, was added to each Mg melt. The metal was melted in a resistive furnace using a graphite coated steel crucible. The melt temperature was constant at 730° C. in all tests. A mixture of 1.25 vol. % SF₆ in N₂ was used for the melt protection.

To test the grain refiner performance in commercial alloys, addition of 0.1 wt % ZrB₂ where made to the alloys AZ91 and AZ31 under TP1 standard test conditions.

Approximately 6 kg of Mg alloy was used per test. 0.1 wt. % and 0.5 wt. % of the ZrB₂ powder were added in each test under continuous mechanical stirring (axial flow stirrer) to the melt. The stirrer was made of mild steel and constructed to distribute the particles throughout by mechanical stirring. Samples were taken at 1, 2, 5, 10, 20, 40, 60, 80, 100 and 120 minutes after the ZrB₂ or ZrC addition to the melt, in accordance with the TP1 test procedure. All conical samples were sectioned at the same height of 38 mm and only the 12×12 mm mid section was used for metallographic examination (FIG. 4). After the preparation the samples were etched with picric acid and the grain size measured (horizontally and vertically) with the intersect method as described in the ASTM standard E 112-88, with more than 100 intercepts counted for each grain size determination.

In FIG. 12 two test series with different contents of ZrB₂ particles resulted in significant grain refinement despite only a moderate growth restriction effect. In the first series with 0.1 wt. % ZrB₂, an ultimate grain size of 100 μm was achieved with a contact time of 5 minutes. After this the grain size increased slightly to 130 μm, within experimental error. FIG. 13 presents a typical microstructure after 2 minutes showing a uniform grain size with equiaxed morphology. Some twinning was visible within the microstructure.

The second series was performed with a ZrB₂ content of 0.5 wt. %. An ultimate grain size of 60 μm was obtained after 5 minutes. Further holding resulted in an increase of grain size up to 120 μm, within experimental error.

FIG. 14 shows the effect of increasing the addition level to 0.5 wt. % ZrB₂ which, in comparison to FIG. 13, revealed a much finer grain size. Again an equiaxed uniform microstructure was visible with some twinning. Generally the crystal structure changed in both series from a columnar to an equiaxed structure.

EXAMPLE 3

A batch of ZrC particles was produced by mixing 38.69 kg ZrO2 (commercially available 99.5% ceramic grade), 11.31 kg C (commercially available carbon black type N991), 8.55 kg B4C (commercially available 1-5 μm) in a high energy shear mixer type Lödige for 4 hours. This powder blend was filled in graphite boats and pushed under Argon atmosphere through a resistance heated graphite tube furnace. The temperature was set at 1900° C. and the material was kept for 9 hours in the heating zone. After cooling to room temperature the material was crushed in a jaw crusher and milled in a ball mill with cemented carbide balls for 1 hour under Argon atmosphere due danger of self ignition. Then the material was control sieved with a 100 mesh sieve. The analysis was C: 11.05%, N: <0.01%, O: 1.34% and an average grain size D50: 3.22 μm measured by a Cilas 1064 laser diffraction instrument in wet conditions FIG. 15 shows a scanning electron micrograph of the synthetic ZrC particles used. The ZrC particles displayed a faceted cubic morphology. The good lattice match on the ZrC particles (table 1) occurs on the close packed facets. The average ZrC particle size lies in the range of 1 to 5 μm and corresponds well with the stated average diameter given by the supplier. However, it is noticeable that clustering can occur and cluster sizes can be in the order of 10 μm. The particle sizes are therefore of the same magnitude as a critical homogeneous nuclei (diameter: d_int˜3 μm) with an undercooling of 1° C. similar to the case of ZrB₂ particles.

The Mg casting tests were carried out as described in example 2 only replacing ZrB₂ by ZrC. In FIG. 16 two test series with different contents of ZrC particles resulted in significant grain refinement despite only a moderate growth restriction effect. In the first series with 0.1 wt. % ZrC, an ultimate grain size of 70 μm was achieved with a contact time of 5 minutes. After this the grain size increased slightly to ˜110 μm, within experimental error. FIG. 17 presents a typical microstructure after 2 minutes showing a uniform grain size with an equiaxed morphology. Some twinning was visible.

The second series was performed with a ZrC content of 0.5 wt. %. An ultimate grain size of 70 μm was obtained after 5 minutes. Further holding resulted in an increase of grain size up to 120 μm, within experimental error.

FIG. 17 shows the effect of increasing the addition level to 0.5 wt. % ZrC which, in comparison to FIG. 18, revealed a similar grain size. Again an equiaxed uniform microstructure was visible with some twinning. Generally the crystal structure changed in both series from a columnar to an equiaxed structure.

EXAMPLE 4

The addition of grain refining particles in laboratory alloy can be flawed as undetected nucleating particles from undefined impurity trace elements may affect the grain refiner performance. Therefore 0.1 wt. % addition of ZrB₂ or ZrC were added to the commercial alloys AZ91 and AZ31 following the same experimental conditions as described in example 2 and 3. The low addition rate of only 0.1 wt. % of particles permit to stay within the tight alloy specification of the wrought alloy AZ31 and still permit grain refinement in achieve equiaxed grains and smaller grain sizes. Grain refinement by other additions into the AZ31 alloy by current or state of art particles (e.g. SiC) is only successful with much larger amounts but remaining outside the specification of the AZ31 alloy.

For the alloys AZ31 and AZ91 significant grain refinement can be seen after 2 to 5 min contact time in FIGS. 19 to 21 for ZrB₂ in and FIGS. 22 to 24 for ZrC addition. Smaller grain size can be obtained for the solute richer AZ91 alloy while the AZ31 alloy is somewhat more difficult to refine. It is noteworthy that to achieve contrast for the grain size measurement annealing in AZ31, it was necessary to highlight grain boundaries by annealing. Therefore the grain sizes measured here will have grown during annealing and the grain size in the as cast state can be expected even smaller than observed for the AZ31 alloy.

Claims

1. A method for producing grain refined Mg and Mg alloys by solidifying molten Mg or molten Mg alloys in presence of a nucleation agent, wherein a nucleation agent ZrB2 and/or ZrC is used.

2. A method according to claim 1, wherein said nucleation agent has a particle size within the range of about 0.1 μm and about 20 μm.

3. A method according to claim 2, wherein said nucleation agent is added to the molten Mg or Mg alloy.

4. A method according to claim 2, wherein said nucleation agent is produced within the molten Mg or Mg alloy.

5. A method according to claim 2, wherein at least one of ZrB2 and ZrC is present in the molten Mg or Mg alloy in a concentration within the range of about 0.01 wt. % and about 1 wt.

6. A method according to claim 2, wherein said nucleation agent has a particle size within the range of about 1 μm and about 4 μm.

7. A method according to claim 6, wherein said nucleation agent is added to the molten Mg or Mg alloy.

8. A method according to claim 6, wherein said nucleation agent is produced within the molten Mg or Mg alloy.

8. A method according to claim 6, wherein at least one of ZrB2 and ZrC is present in the molten Mg or Mg alloy in a concentration within the range of about 0.01 wt. % and about 1 wt.

9. A method according to claim 1, wherein said nucleation agent is added to the molten Mg or Mg alloy.

10. A method according to claim 9, wherein at least one of ZrB2 and ZrC is present in the molten Mg or Mg alloy in a concentration within the range of about 0.01 wt. % and about 1 wt.

11. A method according to claim 1, wherein said nucleation agent is produced within the molten Mg or Mg alloy.

12. A method according to claim 11, wherein at least one of ZrB2 and ZrC is present in the molten Mg or Mg alloy in a concentration within the range of about 0.01 wt. % and about 1 wt.

13. A method according to claim 1, wherein said nucleation agent is added to the molten Mg or Mg alloy.

14. A method according to claim 13, wherein at least one of ZrB2 and ZrC is present in the molten Mg or Mg alloy in a concentration within the range of about 0.01 wt. % and about 1 wt.

15. A method according to claim 1, wherein at least one of ZrB2 and ZrC is present in the molten Mg or Mg alloy in a concentration within the range of about 0.01 wt. % and about 1 wt.

16. A method according to claim 1, wherein said molten Mg alloy and a Mg—Al alloy is used.