This invention relates to metal matrix composite alloys, and more specifically to metal matrix composites comprising an aluminium-containing matrix having a carbon-containing reinforcement material dispersed therein.

It has been previously proposed to incorporate particles of carbon-containing reinforcement material such as titanium carbide into metal matrices to improve the mechanical properties, such as stiffness and strength, of the matrix material.

There are several techniques available for making metal matrix composite materials. One is powder metallurgy, which involves heating and compacting a mixture of the matrix metal and reinforcement material in powder form. However, impurities on the surfaces of the reinforcement and the matrix powders result in a weaker bond between the reinforcement and the matrix, leading to impaired mechanical properties.

U.K. patent specification no. 2288189 A (Brunei University of West London) describes a technique which is designed to eliminate that drawback, specifically as it relates to making metal matrix composites having matrices of aluminium or an alloy of aluminium and, as reinforcement material, a ceramic material such as titanium diboride. The technique comprises mixing the titanium diboride with cryolite or other fluoride flux powder, and melting the mixture together with a melt of the aluminium-based matrix metal, so that the flux can remove impurities from the reinforcement material surfaces.

That technique, involving as it does, mixing of matrix metal with flux material, entails the risk of introducing flux impurities into the metal matrix composites. Extensive work which has been carried out at Nottingham University has shown that it is possible to incorporate ceramic reinforcement materials into aluminium-based matrices, employing fluoride flux-based flux materials to remove impurities from the reinforcement material surfaces, and yet not mix the flux material with the matrix metal. In accordance with the technique developed by that method, reinforcement material is introduced into the matrix by arranging that a body of an appropriate aluminium-containing melt is in contact with a body comprising a mixture of the reinforcement material and molten flux material. When the surfaces of the particles of the reinforcement

material have been fluxed, they are spontaneously transferred into the body of aluminium-containing melt, with negligible transfer of flux into the latter.

We have discovered that when a fluoride flux or other suitable wetting agent is used to treat a titanium carbide reinforcement material prior to its contact with and incorporation into an aluminium-containing melt, by whichever technique, the mechanical properties of the resulting metal matrix composite are often significantly inferior to those which would have been predicted on the basis of the mechanical properties of the matrix and of the reinforcement material. As a result of further work, we have found that those inferior properties can be caused by aluminium carbide associated with the reinforcement particles; aluminium carbide is generally thought to be of the chemical formula A1 ₄ C-, but it is possible that the aluminium carbide may deviate from the stoichiometry implied by that formula. We have further discovered that, where that is the case, the problem can be eliminated or mitigated by arranging that, prior to the reinforcement particles being incorporated into the melt, their surfaces are treated with titanium. Surprisingly small amounts of titanium can be effective for that purpose.

According to the present invention, there is provided a method of introducing a carbon-containing reinforcement material into an aluminium-containing melt, by: treating the reinforcement material with a wetting agent; arranging that the thus-treated reinforcement material contacts the melt under conditions such that it becomes incorporated into it; and subsequently casting the melt; characterised in that, prior to the reinforcement material being incorporated into the melt, its surfaces are treated with titanium, whereby the amount of aluminium carbide associated with the reinforcement material in the cast product is less than otherwise would be the case.

Not only does that reduction in aluminium carbide content lead to an improvement in the metal matrix composite cast product's mechanical properties, but in addition we believe that the composite's resistance to corrosion will also be enhanced, since there will be less scope for damage to it through reaction of aluminium carbide with water to give unwanted aluminium oxide and methane gas.

The titanium used to treat the surfaces of the reinforcement material may be supplied by a titanium-containing salt, e.g. potassium titanofluoride. The latter salt is capable of undergoing

chemical reaction to liberate titanium (for example as titanium metal); an example of such a reaction is that which can occur when potassium titanofluoride is contacted with an aluminium-containing melt. Other ways of supplying the titanium used to treat the surfaces of the reinforcement material will readily occur to those skilled in the art.

Preferably, the titanium is supplied by the wetting agent with which the reinforcement material is treated. This has the advantage of enabling the treatment of the surfaces of the reinforcement material with titanium to be carried out simultaneously with its treatment with the wetting agent. However, the treatment with titanium can instead be carried out before the treatment with the wetting agent. We have found out that a titanium content in the wetting agent of as little as 0.02 % by weight can cause a significant reduction in the amount of aluminium carbide associated with the reinforcement material in the cast product. Indeed, surprisingly, the reduction in the amount of that aluminium carbide is generally not greatly increased by employing contents of titanium in the wetting agent above 0.2 % by weight.

By using a method in accordance with the invention, it is generally possible, by control of the conditions employed, to ensure that the reinforcement material in the cast product is substantially free of associated aluminium carbide. .Although, as shown above, surprisingly little titanium is required to produce a significant reduction in the proportion of aluminium carbide associated with the reinforcement material, increasing the amount of titanium used to treat the reinforcement material generally leads to a reduction in the proportion of aluminium carbide associated with the reinforcement material. However, as also indicated above, one soon reaches a titanium level where the law of diminishing returns would make it not worthwhile further to increase that content merely to further reduce the proportion of aluminium carbide associated with the reinforcement material.

Our work on the invention has related to the case where the reinforcement material is titanium carbide, but it is equally applicable to other carbon-containing reinforcement materials containing titanium, such as titanium carbonitride, for example. We believe the invention to be applicable to other carbon-containing reinforcement materials, such as silicon carbide or boron carbide, for example.

The titanium carbide we have used in our work on the invention is produced by a vacuum carburisation process. That process normally involves the vacuum carburisation of titanium dioxide, which may be effected by heating suitable mixtures of carbon and titanium dioxide at high temperature under a high degree of vacuum. Such a process normally yields a product having a certain amount of free carbon, typically in the range of 0.5 to 1.5 % by weight, but it is known for titanium carbide so produced to have a free carbon content of as high as 3 % by weight. Even with such a high free carbon content it is possible, using the process of the present invention, to achieve a dramatic reduction in the level of aluminium carbide associated with the titanium carbide. While not wishing to be bound by any theory as to how the invention works, we believe that the treatment of the surfaces of the reinforcement material with titanium may achieve the reduction of aluminium carbide associated with reinforcement material by reacting with free carbon associated with the titanium carbide, which otherwise would have been able to react with aluminium of the matrix metal to form aluminium carbide, the reaction of the titanium with that carbon perhaps forming titanium carbide.

The reinforcement material may be of paniculate form, preferably having a mean particle size in the range from 2 to 20 micrometers. However, other forms of reinforcement material, such as fibres, for example, may be employed.

Any wetting agent which is capable of treating the reinforcement material to enable it to wet an aluminium-containing melt may be employed. Generally, it should be one which will absorb surface oxidic contaminants from the reinforcement and matrix materials.

We prefer that the wetting agent should have a melting point below 700 °C.

A preferred type of wetting agent is one which comprises a fluoride-based slag. Such a wetting agent may comprise potassium fluorotitanate. In such a case titanium is supplied by the wetting agent, and it is not necessary to add any additional titanium to reduce the amount of aluminium carbide associated with the reinforcement material in the cast product. A slag comprising potassium fluorotitanate may include potassium borofluoride. A suitable such slag comprises predominantly potassium fluorotitanate and potassium borofluoride in the molar ratio of about 1:2.0.

However, wetting agents comprising potassium fluorotitanate are relatively expensive. We have discovered that almost as good results can be achieved in accordance with the invention using a K-Al-F slag as the wetting agent, with an addition of titanium. Surprisingly, the proportion of titanium required to be effective significantly to reduce the amount of aluminium carbide associated with the reinforcement material (where there otherwise would be such an association) is small. We have found that a particularly suitable K-Al-F slag for this purpose is one containing KA1F ₄ and K ₃ A1F ₆ in a molar ratio in the range from 5:1 to 10:1. The reader is referred to the phase diagram published in International patent publication no. 94/29072 (Lexor Technologies Limited), as Figure 3 of the drawings. With hindsight, that diagram provides a partial explanation of the usefulness of K-Al-F slags of composition within that range, on the basis of the relatively low melting points of such slags.

We have found that a particularly desirable K-Al-F slag is one containing KA1F ₄ and K ₃ A1F ₆ within the above-mentioned 5:1 to 10:1 molar ratio range which is a residual slag which has been formed as a result of reacting a salt comprising potassium fluorotitanate, optionally also comprising potassium borofluoride, with an aluminium-containing melt. We have found that such a slag will have a residual titanium content which will enable the surfaces of the reinforcement material, when being treated with the slag wetting agent, also to be treated to secure that the amount of aluminium carbide associated with the reinforcement material is less than otherwise would be the case. Thus, this is another embodiment of the invention in which titanium is supplied by the wetting agent. Preferably, that content of titanium in the residual slag is in the range from 0.02 to 0.2 % by weight. The K-Al-F slag of the kind described in this paragraph is preferably boron-free, as we have found the presence of boron in the slag to be detrimental.

In order to minimise the risk of introducing significant quantities of slag or other wetting agent into the metal matrix, we very much prefer that the technique employed to introduce the reinforcement material into the aluminium-containing melt should not involve agitation of the melt. In our preferred technique for achieving that end, reinforcement material is introduced into the aluminium-containing melt by arranging that a body of the aluminium-containing melt is in contact with a body comprising the titanium and the reinforcement material, the reinforcement material being in admixture with the wetting agent, whereby the surfaces of the particles of the reinforcement material, having been fluxed and treated with titanium, are spontaneously

transferred into the body of aluminium-containing melt. This is the technique developed at

Nottingham University which is referred to above, but enhanced by the teachings of the present invention.

Once the reinforcement material has been introduced into the melt, it can be evenly dispersed into it by agitating the melt. That agitation should, of course, be carried out in such a way as not to involve incorporation of wetting agent into the melt. Normally, that would be ensured by removing the wetting agent before performing the agitation.

We have successfully tested the method of the invention with aluminium-containing melts of the following kinds:

(a) Commercially pure aluminium.

(b) An Aluminium Association 2000 series alloy. (The Aluminium Association is at 900 19th Street N.W., Washington D.C. 20006, U.S.A.). Such alloys are defined as aluminium alloys having a minor proportion of copper. An example is a binary aluminium-copper alloy containing up to 6 % (e.g. 4 %) by weight of copper.

(c) .An aluminium-silicon alloy. A preferred type is a hypoeutectic aluminium-silicon alloy (which may contain up to 12 % by weight of silicon), e.g. one containing up to 9 % by weight of silicon. A specific example is Aluminium Association alloy type A356, which comprises 7 % by weight silicon and 0.3 % by weight magnesium.

We have found where the aluminium-containing metal of the matrix of the cast product is to contain magnesium, it is generally desirable that the magnesium content is not supplied until after the wetting agent has been removed. Where magnesium is present in the aluminium melt at the stage at which the treatment with the mixture of titanium carbide and flux is applied, the presence of magnesium can seriously reduce the efficiency of transfer of the reinforcement material. Also, with all of the types of wetting agent mentioned above, the wetting agent would be deleterious to the magnesium content.

After the introduction of the reinforcement material the melt may undergo holding prior to casting, for example where it has been alloyed with magnesium as described above; a holding

period in the range from 10 to 30 minutes is generally suitable. Holding generally should be at temperatures within the range from 720 to 800 °C.

With careful practising of the invention as described above using titanium carbide as the reinforcement material, it is possible to achieve that the majority of the particles of reinforcement material are within the grains of the cast alloy, rather than at the grain boundaries. It is also possible to achieve a high volume fraction of the reinforcement material in the cast alloy; volume fractions of over 0.2, and indeed up to 0.3, have been achieved.

As a result of careful scanning electron microscope studies, we have discovered that, where the reinforcement material is not treated with titanium in accordance with the method of the invention, then the reinforcement material in the cast product is subject to a considerable amount of cracking, to the extent that there can be a substantial amount of disintegration of the reinforcement material; and that, by contrast, where the reinforcement material is treated with titanium in accordance with the method of the invention, the amount of cracking in the reinforcement material is very much less. While not wishing to be bound by this hypothesis, we suspect that the difference may be due to better wetting of the reinforcement material in the product made by the method of the invention.

In order that the invention may be more fully understood, some embodiments in accordance with it will now be described with reference to the accompanying drawings, in which:

Figs, la, 2a, 3a, 4a and 5 are optical photomicrographs, showing the metal matrix composites produced in Examples 1, 2, 3, 4 and 5, respectively, the magnification being at 150, with the exception of Fig. 5, which is at 400; and

Figs, lb, 2b, 3b and 4b are scanning electron micrographs, corresponding to the respective optical photomicrographs of Figs, la to 4a, the magnifications all being at 3,500.

Examplβ 1 (Invention)

lOOg of CP aluminium (commercial purity aluminium; in this case 99.7 % by weight) was placed in a Salamander crucible and melted in an electric induction furnace and held at 750 to 800 °C for 15 minutes prior to processing. (We have found that a radiative heating furnace can be used instead of the electric induction furnace.)

17.6g of titanium carbide powder and 8.8g of flux powder were added together and blended until a reasonably homogeneous mixture, as assessed by eye, was produced.

The titanium carbide was from the commercial production of London & Scandinavian Metallurgical Co Limited. It was slightly porous, and had the following properties:

titanium content: 80.56 % by weight total carbon content: 19.74 % by weight free carbon content: 0.69 % by weight

FSSS (Fisher sub-sieve sizer) particle size: 5.2 micrometers.

The flux was a material commercially available from London & Scandinavian Metallurgical Co Limited under the designation potassium aluminium fluoride (PAF). It is a residual slag, of the K-Al-F type, which is formed as a result of reacting a salt comprising potassium fluorotitanate, optionally also comprising potassium borofluoride, with an aluminium-containing melt. Chemically, it comprises a mixture of KA1F ₄ and K ₃ A1F ₆ , within the preferred molar ratio of 5: 1 to 10: 1 referred to above, and also contains titanium (originating mainly from the potassium titanofluoride; we believe that if the titanium had been added to the flux as titanium metal it would have been almost as effective). The flux used fell within the following specification:

potassium: 25 to 30 % by weight aluminium: 18 to 20 % by weight fluorine: 45 to 50 % by weight titanium: 0.1 % by weight maximum (0.07 % by weight in the material used in the Examples). melting range: 560 to 580 °C density: approximately 2.7 g/cm ³ .

At the end of the 15 minutes holding period, the oxide skin which had formed was skimmed from the surface of the aluminium melt, and the mixture of titanium carbide powder and flux powder was poured onto the surface of the liquid aluminium. The mixture was allowed to remain on that surface for 2 to 5 minutes. During that period, when there was no stirring of the aluminium or of the mixture of flux and titanium carbide, the flux melted to produce a liquid K-Al-F flux, and titanium carbide particles spontaneously transferred into the aluminium melt. At the end of that period, when the maximum degree of titanium carbide particle transfer had been achieved, the liquid slag was removed from the melt surface using a ladle. The titanium carbide-containing aluminium melt was stirred briefly (and for the first time since before the mixture of flux and titanium carbide had been added), using a graphite rod (we have found a Syalon rod to be equally suitable), in order to homogenise the dispersion of titanium carbide particles in the melt.

The composite was then cast into a cast iron, wedge-shaped mould. The type of mould used was such that the cooling rate in materials cast into it varies from about 60 °Ks'' near the tip to about 0.5 °Ks ^" ' 85 mm away towards the remote end.

A section from the mid-plane of the casting was mounted in a resin material, ground on silicon carbide paper, polished using 6 micrometer and then 1 micrometer diamond pastes. It was then viewed in a polarising microscope. Fig. la is a photograph of that view. The aluminium matrix can be seen at 11, and typical titanium carbide particles can be seen at 12. The dark areas such as at 13 are porosity.

Visual examination of several micrographs from the sample showed that, overall, a medium concentration of titanium carbide particles in the matrix had been achieved, and that the particle distribution was good, and the porosity was low.

A second sample from the mid-plane of the casting was prepared in the same manner as with Fig. la, and was then examined with a scanning electron microscope, the image being obtained in backscattered electron mode. Fig. lb shows the resulting scanning electron micrograph. As in Fig. la, the aluminium matrix can be seen at 11, appearing as a dark grey background, and typical titanium carbide particles can be seen at 12; they appear mid-grey. There is some porosity in this view; it appears black as shown at 13. There is a relatively small number of aluminium carbide particles, which appear as white particles, examples being shown at 14. There is a small amount of cracking in some of the titanium carbide particles, as can be seen at 15, for example.

Example 2 (Invention)

The experiment of Example 1 was repeated exactly, except that the flux employed (again in an amount of 8.8g) was a mixture of potassium borofluoride, KBF ₄ , and potassium titanofluoride, J TiFg, in a molar ratio of 2:1. The weight percentages of titanium and boron in the flux, as determined by chemical analysis, were 9.95 % and 4.35 %, respectively.

Fig. 2a is a photomicrograph of a sample from the mid-plane of the cast product, produced in the same manner as Fig. la. By analogy, numerals 21, 22 and 23 show respectively aluminium matrix, typical titanium carbide particles, and porosity. Visual examination of several micrographs from the sample showed that, overall, a high concentration of titanium carbide particles in the matrix had been achieved, and that the particle distribution was good, and the porosity was very low.

Fig. 2b is a scanning electron micrograph of a second sample from the mid-plane of the cast product, produced in the same manner as Fig. lb. As in Fig. 2a, the aluminium matrix can be seen at 21, appearing as a dark grey background, and typical titanium carbide particles can be seen at 22; they appear mid-grey. There is some porosity in this view; it appears black, as shown at 23. There is a very small number of aluminium carbide particles, an example of which is

shown at 24; again they appear as white particles. There is a small amount of cracking in some of the titanium carbide particles, as can be seen at 25, for example.

Example 3 (Comparative)

The experiment of Example 1 was repeated exactly, except that the flux employed (again in an amount of 8.8g) was a K-Al-F flux commercially available under the trade mark Nocoloc from Solvay Fluor und Derivate GmbH. It contained no titanium or zirconium. From published literature we understand that it has the following chemical analysis: potassium: 28 to 30 % by weight aluminium: 16 to 18 % by weight fluorine: 49 to 51 % by weight.

Our own analysis indicates that it substantially consists of KA1F ₄ .

Fig. 3a is a photomicrograph of a sample from the mid-plane of the cast product, produced in the same manner as Fig. la. By analogy, numerals 31, 32 and 33 show respectively aluminium matrix, typical titanium carbide particles, and porosity. Visual examination of several micrographs from the sample showed that, overall, a low concentration of titanium carbide particles in the matrix had been achieved, and the particle distribution was uneven, and the porosity was medium.

Fig. 3b is a scanning electron micrograph of a second sample from the mid-plane of the cast product, produced in the same manner as Fig. lb. As in Fig. 3a, the aluminium matrix can be seen at 31, and typical titanium carbide particles can be seen at 32. There is a significant amount of porosity in this view; it appears black as shown at 33. Aluminium carbide particles can be seen at 34, for example; again they appear as white particles, but in this case there is much more aluminium carbide present than in Figs, lb and 2b, and the particles are generally larger. A large amount of cracking of titanium carbide particles can be seen, as at 35, for example. Indeed, the cracking of titanium carbide particles is so extensive that in many cases the titanium carbide particles have completely disintegrated.

Example 4 (Comparative)

The experiment of Example 1 was repeated exactly, except that the flux employed (again in an amount of 8.8g) was a K-Al-F flux consisting of a mixture of KA1F ₄ and K _j AIF^ in a molar ratio of 10: 1. It contained no titanium or zirconium.

Fig. 4a is a photomicrograph of a sample from the mid-plane of the cast product, produced in the same manner as Fig. la. By analogy, numerals 41, 42 and 43 show respectively aluminium matrix, typical titanium carbide particles, and porosity. Visual examination of several micrographs from the sample showed that, overall, a medium concentration of titanium carbide particles in the matrix had been achieved, and the particle distribution showed light clustering, and the porosity was low.

Fig. 4b is a scanning electron micrograph of a second sample from the mid-plane of the cast product, produced in the same manner as Fig. lb. As in Fig. 4a, the aluminium matrix can be seen at 41, appearing as a dark grey background, and typical titanium carbide particles can be seen at 42; they appear mid-grey. There is a significant amount of porosity in this view; it appears black as shown at 43. Compared with Figs, lb and 2b, there is a much larger amount of aluminium carbide (again visible as white particles). Examples of the aluminium carbide particles can be seen at 44; they are generally larger than those to be seen in Figs, lb and 2b. A large amount of cracking of titanium carbide particles can be seen, as at 45, for example. Indeed, the cracking of titanium carbide particles is so extensive that in many cases the titanium carbide particles have completely disintegrated.

Example 5 (Comparative)

In view of the chemical similarity between zirconium and titanium, we had expected that practising the method of the invention, but treating the surfaces of the reinforcement material with zirconium instead of titanium, would also be capable of reducing the amount of aluminium carbide associated with the reinforced material in the cast product.

In order to test that hypothesis, the experiment of Example 2 was repeated, with the following exceptions:

(a) The amounts of titanium carbide powder and flux powder were respectively 20.29g and 10.70g.

(b) The flux consisted of a mixture of 5.04g of potassium borofluoride, KBF^ and 5.66g of potassium zirconium fluoride, K-ZrFg; thus, the molar ratio of KBF ₄ to K-ZrF ₆ was substantially 2 to 1.

(c) No scanning electron micrograph of the product was produced.

Fig. 5 shows the photomicrograph of the product. The aluminium matrix can be seen at 51, and typical titanium carbide particles can be seen at 52. The dark areas such as can be seen at 53 are porosity. Contrary to our expectation, the product contained considerable amounts of aluminium carbide, to the extent that it is even visible in the photomicrograph, as at 54. Thus, the attempt to use zirconium to reduce the amount of aluminium carbide in the product was entirely unsuccessful.

Example 6 (Mechanical Properties - Comparative and Invention)

The method of the invention was practised substantially as described in Example 1 above, but with a range of titanium carbide contents in the mixture of titanium carbide and flux applied to the aluminium melts in successive experiments. Cast samples were selected having nominal titanium carbide contents of 5, 10 and 15 % by volume, respectively. Further samples were cast from untreated melts of CP aluminium (99.7 % by weight). In addition, an aluminium-based titanium diboride metal matrix composite, containing 7 % by volume of titanium diboride, was made by a method similar to that described in Example 1 of International patent publication no. 93/05189 (London & Scandinavian Metallurgical Co Limited). Samples were taken from the mid-plane of each of the castings, and were all tested to determine the Young's modulus (YM), ultimate tensile strength (UTS) and percentage elongation at failure (Elongation) of the materials. The results are shown in the Table below.

Tahle

YM UTS Elongation

GPa MPa %

Comparative

Unreinforced 63 50 37

7 % titanium diboride 76 110 7

Invention

5 % titanium carbide 69 75 35

10 % titanium carbide 80 110 19

20 % titanium carbide 95 135 15

It can be seen that, compared with unreinforced aluminium, titanium carbide composites produced by a method in accordance with the invention give enhanced stiffness and strength. The improvements in stiffness are particularly significant, and this is important industrially. The results show an average of 2.5 % improvement in stiffness (YM) per volume percent added. This rate of improvement is similar to that achieved by titanium diboride based composites produced according to the teachings of International patent publication no. 93/05189, and significantly higher than those achieved by other known composites based on aluminium oxide and silicon carbide, which range from 1.2 to 1.7 % increase per percent addition.

The above YM values for the titanium carbide composites produced by the method of the invention are in each case greater than 97 % of the respective value predicted on the basis of the Tsai-Halpin model. This indicates that the bonding of the reinforcement particles to the matrix is extremely good.

Generally, molten composite alloys are less fluid than unreinforced alloys, but we have found that composites produced according to the invention can show less deterioration than other composites, enabling higher levels of titanium carbide to be added (up to 30 % by volume). This good castability is important in commercial applications and contributes to the good mechanical properties.

The Table also shows that the titanium carbide composites retain a higher level of ductility (as shown by the elongation at failure) than the titanium diboride composite; typical silicon carbide and aluminium oxide reinforced composites have even lower ductility. The combination of high stiffness and high ductility is particularly advantageous.

The titanium carbide composites produced by the method of the invention can also have an unusually fine grain size depending on the titanium carbide addition level and the cooling rate; substantially lower than the 100 to 200 micrometers grain size in normal grain-refined 99.7 % aluminium. This should enhance toughness and improve workability. Calculations have shown that the grain sizes in those composites are of the order of those predicted on a one particle per grain analysis modelled by Humphreys (see Humphreys F. J. and Hatheriy M., Recrystallisation and Related Annealing Phenomena, Pergamon Press, 1995, First Edition) at all cooling rates and volume fractions.