METHOD FOR FORMING METAL MATRIX COMPOSITES

The present application relates to a method for forming a reinforced metal matrix, and in particular a metal matrix composite with improved mechanical properties and corrosion resistance.

Metal matrix composites (MMCs) combine two or more materials, typically a

reinforcement phase and a metal matrix in order to enhance the properties of the base matrix alone.

Light metal matrix composites are fundamental not only for the transport sector, i.e. automotive and aerospace industries, but also for others such as the renewable energy industry.

Alternative prior art composites include generic composites with polymer bases and other MMCs that use aluminium as the metallic base but employ alternative fibres such as E-glass (see US20050133123) and carbon (see US4853294).

A basalt fibre and ceramic particle mixed aluminium alloy drill rod material and a method of preparation thereof are disclosed in CN 107043901 A (University Jilin).

WO 2006/134405 (Gogoladze et al.) discloses a method of manufacturing an aluminium- based composite material in which melt aluminium is reinforced with basalt fibres having a diameter of 16-18 micrometres. The method includes forming a 8-10mm layer of the fibres at the bottom of a crucible and then adding the same volume of aluminium. The mixture is then transferred into a furnace at 800-850°C and stirred for 10-15 minutes. It is then removed from the furnace and cast into a mould and cooled. The resulting MMC does not have particularly advantageous properties however.

The present application seeks to provide a method of forming a metal matrix composite with improved mechanical properties.

In accordance with a first aspect of the present invention, there is provided a method for forming a reinforced metal matrix, including the steps of:

(a) at least partially melting a metal or metal alloy,

(b) adding basalt fibres to the melt of step (a) at a proportion from 1 to 10wt%, and (c) stirring the basalt fibres and melt of step (b) at a rate from 400 to 1200rpm for a time from 1 to 60 minutes.

In a preferred embodiment the method additionally includes the step of: (d) sonicating the basalt fibres and melt of step (c) from 1 to 15 minutes at a frequency from 10kHz to 30kHz, a power from 2 to 5kW and an amplitude from 20 to 50μιη.

In a preferred embodiment the method additionally includes the step of:

(e) casting the product of step (d) at a temperature from 200°C to 780°C into a mould which has been preheated at a temperature from 150°C to 500°C for 1 to 2 hours.

In a preferred embodiment, there is provided a method of forming a reinforced metal matrix, including the steps of:

(a) pre-heating basalt fibres at a temperature from 100°C to 500°C for 1 to 3 hours,

(b) roll an aliquot of the fibres in pure aluminium foil,

(c) melting a metal or metal alloy (cast aluminium alloys -at a temperature from 550°C to 750°C; zinc - at a temperature from 400°C to 440°C; magnesium - at a temperature from 600°C to 700°C; tin at a temperature from 200°C to 260°C)

(d) adding from 1 to 5wt% of the basalt fibres of step (a and b) to the melt of step

(e) stirring the basalt fibres and melt of step (d) at 400 to lOOOrpm for a time from 10 to 30 minutes,

(f) sonicating the basalt fibres and melt of step (d) for 5 minutes at a frequency of 17.5kHz, a power of 3.5kW and an amplitude of 40μη"ΐ, and

(g) Casting the product of step (e) at a temperature from 200°C to 720°C into a mould which has been preheated at a temperature from 150°C to 500°C for 1 to 2 hours.

The proposed material preferably involves an aluminium matrix reinforced with basalt fibres, in particular whiskers i.e. short fibres. Basalt fibres are classified as being situated between E-glass (low) and carbon (high) in terms of strength but the fibres are significantly cheaper than carbon fibres. In addition to the excellent mechanical properties basalt fibres show high corrosion and radiation resistance also at

temperatures up to 600°C and they are not combustible, the combination of these properties is unique compared to glass and carbon fibres which makes these fibres a superior alternative for specific applications.

The excellent properties achieved by the Aluminium matrix/basalt fibres composite, produced by the authors, are a breakthrough for a next generation of composites that can be used as an alternative to steel for weight reduction in harsh environments, i.e. marine offshore applications. The material property increase is thought to be based upon a preferential chemical reaction between the matrix and the basalt fibres. We have produced and investigated a manageable reaction between the fibre and matrix to produce strong polyhedral iron-rich phases at the interface. Diffusion of elements occurs from the matrix into the fibre and vice-versa during the melt production and casting of the composite material, producing a stronger (anti-slip) bonding while maintaining a strong fibre core.

The results presented below, indicate a 30% improvement in hardness, tensile strength and corrosion resistance all of which was achieved with only a 5 wt. % fibre addition.

Basalt fibres have been widely used as reinforcement both for thermoset and

thermoplastic polymers, however the applicability of basalt fibres as a reinforcing material for metal-matrix composites has been limited due to the reduced stability of the fibres during processing at elevated temperatures. In particular one of the major problems is the dependence of strength on temperature, usually resulting in a loss of fibre strength during the holding of melts at high temperatures required to enable casting. However the micro-morphology observed was the same prior to the thermal process, a larger volume percentage of the fibres was observed to be associated with porosity. Evaluation of the optimum processing conditions for mixing the fibres in the melt has been conducted by Sabet et al. (Production and optimisation of aluminium basalt composites by hand lay-out technique. In : Proceedings of the World Congress on Engineering, vol III; 4-6 July 2012. London, UK) who consider the strength but did not include detailed observations of the chemical/mineralogical interaction occurring at the interface fibres/metal matrix, hence limiting the findings to the fibres alone without evaluating the overall improvement of the composite.

These previous studies are based on the dispersion of the fibres and their final strength in the composite and are not focused on the fundamental reaction between the fibres and matrix and how this reaction can be favourably utilised in order to improve the chemistry and build on the intermetallic nucleation within the alloy, which as a consequence, impacts and improves the overall characteristics of the composite itself.

Without wishing to be constrained by theory, it is thought that the present invention relies upon the enhancement of the controlled recrystallization process of the fibres as a tool for the nucleation of beneficial polyhedral iron-rich intermetallic phase at the interface fibre/matrix whilst still maintaining an amorphous central core. This

fundamental change to both matrix and reinforcement phase is thought to change the Al/Basalt composite into an Aluminium Matrix Hybrid Composite AMHC. This is thought to maintain the structural integrity of the fibre which hence increases the overall strength of the composite. The microscopic observations of the interface fibres/matrix in the composite produced, points to the important role of the re-crystallisation process on the basalt fibres during the casting process. Due to the high content of iron oxides in the fibres, crystallisation in the basalt fibre begins with oxidation of ferrous cations and the formation of spinel structure phase on the fibre surface. Calcium, Magnesium and Iron diffuse from the interior to the surface where they normally react with the environmental oxygen forming nanocrystalline layers of those elements' oxides. The formation of the oxides is followed by the crystallisation of pyroxene phase nucleating from them.

However, this is different when the fibres are within the anoxic environment of the molten metal. In this case, as will be explained later, the fibres still reach the re- crystallisation temperature, hence the mineral phases start to recombine, but no oxidation can occur, what happens instead is that the cations mentioned above in addition to silicon react with the intermetallics forming in the matrix, that help with the nucleation of beneficial polyhedral iron-rich phases around the recrystallized outer core of the fibre (Figures 4, 5 and 6). The lack of oxygen limits the full recrystallization of the fibre maintaining an integral internal core which enhances the strength of the fibre and in parallel iron/silicon phases migrate from the matrix towards the interface with the fibres where they can combine with the cations exiting from the fibres themselves. This process starves the intermetallics around the fibres of iron limiting their growth. This is particularly beneficial for iron-rich aluminium alloys such as those derived from recycled aluminium, making the proposed method for the addition of basalt fibres, the perfect tool to achieve high spec composites using recycled aluminium.

The basalt fibres used have been sourced from ZaoMineral. The production method of these fibres is disclosed in : UA89145, UA90360, UA90361, UA90362.

A number of preferred embodiments of the invention will now be described with reference to the drawings, in which :

Figure 1 is a schematic diagram of distributive mixing equipment for carrying out a method in accordance with the invention;

Figure 2 is a schematic illustration of ultrasonic melt processing apparatus for carrying out a method in accordance with the invention;

Figure 3 is an optical microscopy image of basalt fibres diffused in an aluminium matrix prepared in accordance with the invention; Figure 4 is a SEM BSE image of a fibre in a metal matrix prepared in accordance with the invention and a corresponding EDS spectrum;

Figure 5 is a SEM BSE image of the cross-section of a fibre prepared in accordance with the invention and corresponding spectra of the same;

Figure 6 is a SEM BSE image of the cross-section of a further fibre prepared in accordance with the invention and corresponding spectra of the same;

Figure 7 is a stress/strain graph of an LM25 reference alloy;

Figure 8 is a stress/strain graph of an LM25-5% basalt fibre MMC;

Figure 9 shows mechanical test data, namely Ultimate Tensile Strength (9A), 0.2% Offset Yield Strength (9B), and Vickers hardness (9C);

Figure 10 shows burnt fibres skimmed off the melt before casting;

Figure 11 shows an LM 25 prepared at 800°C SEM image (A), EDS spectra of the fibre rims at the interface with the matrix (B);

Figure 12 shows an LM 25 prepared at 720°C;

Figure 13 shows an LM 20 prepared at 800°C;

Figure 14 shows a fibre wetted within LM 25 at 720°C (A-D), EDS area map of elemental composition (E);

Figure 15 shows an LM 20 prepared at720°C;

Figure 16 shows an LM 20 prepared at 710°C;

Figure 17 shows an LM 25 prepared at 710°C;

Figure 18 shows an LM 20 prepared at 690°C;

Figure 19 show san Mn-enriched fibres wetted in LM 20 prepared at 720°C;

Figure 20 shows micrographs of Al- lOFe prepared at 780°C (A), 750°C (B), 720°C (C);

Figure 21 is an optical micrograph of LM 25 stirred at 500 rpm for 10 minutes (A and B), LM 20 stirred at 1000 rpm for 10 minutes (C);

Figure 22 is an optical micrograph showing the agglomeration effect of non-application of ultrasonication;

Figure 23 is an optical micrograph of composite after 10 minutes stirring at 1000 rpm (A), 20 minutes (B) Chart showing % of fibres in micrographs A and B (C); Figure 24 is an EDS area analysis of sample after 5 minutes exposure (A), 10 minutes (B), 30 minutes (C);

Figure 25 is an optical micrographs of cast LM 25/ Basalt after 3 minutes melt exposure (A), 10 minutes (B), 30 minute (C) and 80 minutes (D);

Figure 26 shows Basalt fibre reactions in LM 20, namely A) EDS area scan of Silicon with a basalt fibre dipped in LM 20 for 5 minutes, B) EDS area scan of Silicon with a basalt fibre dipped in LM 20 for 10 minutes, C) EDS area scan of Silicon with a basalt fibre dipped in LM 20 for 30 minutes;

Figure 27 is a picture of a fibre inside an intermetallic phase (A) and SEM spectrum of the intermetallic (B), the centre part of the fibre (C) and the recrystallized part (D);

Figure 28 shows pictures of the fibres inside the intermetallic phases (A,B), SEM spectrum of the intermetallic phase (C) and the fibre (D), Intermetallics nucleating from a fibre (E,F);

Figure 29 shows pictures of the fibres inside the AI65- Ti35 alloy (A) EDS spectrum of spot 1 ;

Figure 30 shows fibres wetted into LM 0 matrix;

Figure 31 shows an EDS analysis of the fibres inside the LM 20 alloy: 5 minutes exposure (A), 10 minutes (B), 30 minutes (C);

Figure 32 is a micrograph of LM 20 with basalt reinforcement;

Figure 33 is an SEM of the fibres inside the LM 25 alloy (A) Fibre and intermetallic sliced in transverse direction (B) Fibre and intermetallic sliced at an angle in the longitudinal direction (C) Intermetallic formation on exposed fibre;

Figure 34 is an EDS area mapping and EBSD of LM 25/Basalt: Aluminium map (A), Silicon map (B), EBSD map (C);

Figure 35 is an SEM image of the fibre in LM 12 and resulting IMC (A) EDS spectrum of spot 1 at the fibre rim showing Al, Si, Fe and Cu (B);

Figure 36 are pictures of the fibres inside the AZ91 alloy (A) Higher mag. Image showing intermetallic formation (B) Intermetallic formation on exposed fibre (C);

Figure 37 is a BSE image of intermetallics at fibre interfaces. Corresponding EDX line scan analysis indicating that the intermetallics are Al-Fe rich; Figure 38 is an optical micrograph showing a good distribution of fibres within the matrix. AZ91 with 5wt% fibres processed at 667 C (A) EDS spectrum of bright intermetallics (B);

Figure 39 shows intermetallic formation around fibres in AZ91 (A), EDS spectra taken from the IMCs (B-D);

Figure 40 is an optical micrograph of intermetallic formation within a mixed alloy after introduction of basalt fibres;

Figure 41 is an LM 25/Composite Stress/Strain curve;

Figure 42 is an LM 20/Composite Density data;

Figure 43 is an LM 20/Composite Wear test curves;

Figure 44 are corrosion resistance samples LM 25/composite/316L; and

Figure 45 is a corrosion resistance curve.

EXPERIMENTAL PROCEDURES

Materials

Commercially available LM25, an aluminium casting alloy (USA designation A356) was used as a matrix material and was supplied by Norton Aluminum Ltd., Staffordshire, UK in the form of ingots (actual composition (wt. %) : AI-7.50Si-0.30Mg-0.003Cu-0.05Fe- 0.005Mn-0.003Zn-0.11Ti). The reinforcement used was in the form of Basalt short fibres i.e. whiskers (weight % composition as shown in Table 1).

Compound _si02 AI203 Fe203 CaO MgO Na20 K20 Ti02 P205 MnO O203

Wt% 52.8 17.5 10.3 8.59 4.63 3.34 1.46 1.38 0.28 0.16 0.06 Mixing processes

The process for synthesizing the low cost Al-metal matrix composites primarily consists of two steps. The first part of the fabrication process, entails distributive mixing of the reinforcement in the matrix whereas the second part is dispersive mixing using ultrasonic cavitation prior to shaping.

Distributive mixing

Distributive mixing employs conventional mechanical stirring to pre-mix the Al alloy with reinforcement basalt whiskers. A distributive mixing setup is schematically shown in Figure 1. The mixing equipment consisted of a driving motor to create the torque on the impeller, a lifting mechanism for the rotation drive unit and stirrer assembly and a transfer tube for introducing the reinforcement whiskers into the melt. An ingot was melted in a cylindrical crucible inside a top loaded resistance furnace at 750 °C.

Accurately measured reinforcement (5 wt% of basalt fibres) was pre-heated at 250 °C for 1 hour in the furnace. A four bladed titanium impeller was coated with boron nitride to prevent a reaction with molten aluminium. A controlled argon atmosphere was maintained inside the furnace throughout the whole experiment to prevent melt oxidation. With the help of the transfer tube (in the form of whiskers wrapped in aluminium foil), the preheated reinforcement whiskers were transferred slowly and continuously into the melt at the side of the vortex created by stirring the melt at 400 - 600 rpm for 10 mins.

The distribution of the reinforcing phase is strongly influenced by many aspects of the processing route. According to Bernoulli's theorem, the negative pressure differential existing at the vortex helps to draw externally added whiskers into the liquid metal. The force provided by stirring the melt with a mechanical stirrer helps to overcome the surface energy barriers due to the poor wettability of basalt whiskers by Al alloys. Once the whiskers are transferred into the liquid, the distribution is strongly affected by certain flow transitions. The application of a shear force to the added whiskers contributes to distribute them relatively uniformly. In order to produce preferable melt circulation and homogeneous distribution of the reinforcement throughout the melt, the impeller design plays a crucial role. To avoid stagnant zones, necessary axial and radial flows are preferably provided by an impeller having an impeller diameter / crucible diameter ratio of 0.4 and a width of impeller / diameter of impeller ratio of 0.35. Stirring of the mixture is carried out in the liquid regime to achieve spatial homogeneity of particulates (similar volume fraction in any part of the mixture).

The distributive mixing stage is important as a means to incorporate the reinforcement whiskers and to coarsely distribute them in the melt. The degree of mixing is governed by the momentum transfer from the position of the stirrer to the clusters located away from the stirrer position. In this mixing step, due to the presence of velocity gradients within the liquid media, the shear force applied to the liquid that is in contact with the impeller results in a low shear force when averaged out over the whole volume of the liquid media. A lack of sufficient shear force in distributive mixing results in

agglomerates in relatively stagnant zones (e.g. near crucible walls). To break-up these agglomerates into individual whiskers in the liquid metal, the applied shear stress (τ) on reinforcement clusters should overcome the average cohesive force or the tensile strength of the cluster. Hence dispersive mixing needs ideally to be performed under ultrasonic cavitation.

Dispersive mixing under ultrasonic cavitation

In order to achieve enhanced mechanical properties in MMCs especially those prepared via casting, it is necessary that the Basalt whiskers are homogeneously and uniformly dispersed within the molten Aluminium. The agglomeration of fine-sized whiskers prevents high quality castings being achieved. While cohesive forces hold the

reinforcement agglomerates together, high hydrodynamic forces are needed to break them up. The process of mixing under ultrasonication utilises ultrasonic cavitation forces to overcome the cohesive force or the tensile strength of the cluster in order to disperse the fibres. Therefore, the mechanically stirred Al-Basalt composite slurry was transferred to the apparatus shown in Figure 2 and ultrasonicated for 5 mins (17.5 kHz, 3.5 kW, 40 m amplitude, Niobium sonotrode) to ensure the dispersion of the basalt whiskers.

Introducing ultrasonic waves within the melt results in acoustic cavitation and streaming effects. Acoustic cavitation occurs through the formation, growth and collapse of cavitation bubbles under alternate acoustic pressure wave cycles. Acoustic streaming is a liquid flow due to an acoustic pressure gradient, leading to highly effective stirring. Molten Al under ambient always contains dissolved gases. Above the cavitation threshold, entrapped gas generates numerous tiny cavities. Upon pulsation, these cavitation bubbles grow by rectified diffusion of gases from the melt. Cavitation bubbles of a particular size implode during positive pressure cycle generating a temperature of 5000 °C and pressure of 1000 MPa forming a 100 m/s liquid jet in the vicinity of the cavitation implosion region. Simultaneously, acoustic streaming of 0.1 m/s formed from the pulsation of the cavitation region ensures a continuous stirring effect throughout the melt. During ultrasonication, cavitation and acoustic streaming are believed to improve the wettability of the reinforcement. Also, cavitation pressure and the liquid jet created by the implosion of the bubbles disintegrate the agglomerates. The composite slurry is then cast at 720 °C into a tensile specimen mould which was preheated to 400 °C to ensure preferred melt fluidity and porosity within the casting.

RESULTS

Under the melting conditions described above the fibres were dispersed in the matrix (Figure 3) and a chemical interaction between the matrix and fibre occurred whilst maintaining the physical integrity of the fibre. Detailed evaluation of the intermetallic phases was performed and the analysis indicates that polyhedral phases of Fe bearing intermetallics (AIFeMnSi) form at the fibre-matrix interface (Figure 4).

Furthermore, the migration of silicon from the core of fibre and aluminium from the matrix to the core of the fibre are visible indicating the fibre-matrix reaction (Figure 5). In particular, the SEM BSE image of the cross-section of a fibre shows the newly formed polyhedral phases on the surface of the fibre and their corresponding composition (Spectrum 5). This reveals the changes occurring within the core of the fibre which loses silicon and becomes aluminium whilst the silicon moves to the surface of the fibres forming the new intermetallic phases.

The same phenomenon is visible in Figure 6 where the decomposition of the fibre and its recrystallization can be seen in conjunction with the formation of the intermetallic phases. In particular the core of the fibres maintains the chemical composition of the original fibre, albeit showing a partial enrichment in aluminium (Spectrum 1) while the recrystallized area (Spectrum 3) shows a composition limited to aluminium, magnesium and oxygen, with the silicon that was initially present, migrating from the recrystallized area to form the intermetallic phases (Spectrum 4).

The tensile strength of the basalt reinforced aluminium composite showed a 30% improvement (see Figure 8) in comparison with the unreinforced alloy (see Figure 7).

Further experimental testing was carried out as summarised below. The conclusions are given without wishing to be constrained by theory and are non-limiting. EXPERIMENTAL PROGRAM

1.0. Investigation on the effects of different processing conditions

stirred

1.1. Investigation on the effects of different alloys To investigate the potential reactions incurring with different alloys, a series of experiments were completed, focusing each time on one specific alloying addition. The major elements investigated were: Iron (Fe), Nickel (Ni), Titanium (Ti), Silicon (Si), and Copper (Cu)

Investigation on the mechanical property improvements relative to the reference

1.3. Sample preparation

Once a melt has been created and allowed to cool the material is sectioned and prepared using standard metallography protocols - [Mounting in resin, planar grinding, polishing (SiC), Polishing (Diamond), final polishing (Colloidal silica)].

Each sample is analysed using a Zeiss SUPRA FEG-SEM equipped with EDAX Octane Superior EDS detector. The samples displayed hereafter have been imaged using the SEM and a spectra of the elemental composition is taken using the EDS system.

2.0. INVESTIGATION ON THE EFFECTS OF DIFFERENT PROCESSING

CONDITIONS

2.1. Experimental program

A series of experiments have been completed to investigate key attributes of the formation and manufacture of the composite material. These attributes are the weight percent of fibre addition, temperature of the processing range, exposure of the fibre to the melt and stirring.

2.1.1. Fibre wt%

To determine the nominal weight fraction of basalt reinforcement to add to aluminium alloy. Tensile tests of the composites, reinforced with 1 wt%, 2.5 wt% and 5 wt% basalt, were performed with 4 specimens for each mode. Cast tensile samples of 10 mm gauge diameter, 50 mm gauge length and 14 mm 0 grips were produced and tensile testing was completed on an Instron®5569 test machine. A strain rate of 0.400 strain/minute was used throughout for both composite samples and reference samples with all tests completed at ambient temperature. Results provided offset yield and ultimate tensile strength along with calculated Young's Modulus.

2.1.2. Temperature range

The effect of temperature range on the formation of intermetallics within the composite was investigated. The processing temperature was varied between 680 and 800°C and metallography samples taken. The samples are then analysed using EDS and SEM observation. 2.1.3. Exposure

Two sets of experimental data are presented to evaluate the effects of fibre/matrix exposure and melt stirring. The fibre/matrix exposure time experiments were completed by dipping the fibres into molten aluminium LM 20 to investigate the interfacial and fibre reactions. A series of small interval casts were made using LM 25 and 5 wt% of basalt fibre to investigate the intermetallic formation over time.

2.1.4. Stirring

The stirring experiments involve using a powered blade immersed in the melt to stir the fibres into the aluminium, the blade speed and stirring time is varied. In all cases the resulting samples were prepared for microstructural observation and EDS analysis.

2.2. Wt. % of Fibres

The reference sample represents the baseline for assessing the fibre addition weight percentage. The ultimate tensile strength is the highest load the material can experience before the onset of necking and failure. Adding 2.5 wt% of fibres improves the UTS by a small margin (~5%) as shown in Figure 9A, the addition of both 1 wt% and 5 wt% reduces the UTS below the reference sample baseline and so are seen as potentially detrimental. The fibre alignment is important with respect to improving the tensile strength, with those fibres aligned perpendicular to the direction of the applied load reducing it. Data for the 0.2% offset yield point were calculated and presented in Figure 9B. This data shows that the 0.2% offset yield strength is improved by all additions of fibres, however, it is the 2.5 wt% sample that improved by the largest margin. The 5 wt% sample did achieve a significant increase in offset yield strength and so both the 5 and the 2.5 wt% samples were subjected to further testing.

Hardness testing using a Vickers Micro hardness indenter was performed with the results shown to be in agreement with the tensile data. Both composite samples achieved an increase in hardness over the reference sample as shown in Figure 9C

Adding significantly more basalt to the composite is extremely difficult. 5 wt% is most achieved in these experiments. Higher fibre additions have been attempted but led to the formation of a layer of burnt fibres within the usual casting dross. This is shown in Figure 10. before casting any scale is skimmed from the surface of the melt, removing any burnt fibres. Conclusions

The larger the volume fraction of fibres the longer it takes to add the fibres. The difficulty in adding large wt. % of fibres is the time it takes to stir in large volumes of fibres. 5 wt. % was the largest achievable

Larger fibre additions resulted in consumed fibres floating and forming a soot layer

2.5 wt. % was the optimum in this range of experiments. An increase in both offset yield and UTS presents a general increase in mechanical properties. 2.3. Temperature range experiments

High temperatures: A series of small casting experiments were completed at an unusually high temperature in order to investigate the upper limits of the casting process when adding the fibres and also to see what intermetallics form within the composite samples. Aluminium alloys were prepared in small ceramic crucibles and entered the furnace to be heated to 800°C. 5 wt% of basalt fibres were added, stirred and then left to react in the furnace for 30 minutes before being poured, sectioned and analysed. Figure 11A shows the difficulties in ensuring suitable fibre wetting. The LM 25 samples produced at 800°C showed agglomeration of fibres and reduced reactions due to the manual mixing and the low volume of melt. However, the fibres still re-crystallise and have an initial reaction with the matrix showing an aluminium oxide layer at the fibre's rim which is the trigger of intermetallics formation. Figure 11B shows the EDS spectra for the investigated IMC.

A second experiment was completed at a lower temperature of 720°C. This can be seen in Figure 12. Wettability of the fibres is improved due to higher volume of melt and stirring (lOOOrpm). Figure 12A shows an improved wettability of the fibres and 12B shows the formation of an intermetallic compound at the fibre/matrix interface. Analysis of the fibre's rim shows composition similar to the one presented in Figure 11B indicating the layer of oxide as linked with the formation of the intermetallics. The diffusion of silicon outside the rim and the nucleation of the oxide layers triggers the nucleation of dendritic intermetallics (Figure 11B - initial shape prior to full crystallisation).

The 800°C experiment was then repeated for LM 20 but showed the same wettability issues due to the hand stirring procedure. However, re-crystallisation of the fibres rim is visible and reaction products can be observed. Figure 13. Shows the fibres within the composite showing an unaltered core (white in figure 13B) an amorphous layer enriched in aluminium (light grey figure 13B) and a fully recrystallised aluminium oxide layer (dark grey figure 13B).

The influence of alloying elements tests completed in part for section 3 also contain experimental work conducted at the high end of the temperature range (800°C). Please see section 3 for more detail.

Both the LM 20 and the LM 25 were successfully cast at 720°C, showing excellent wettability and extensive fibre reaction. Figure 14. shows the fibres wetted in the melt (A-D) with IMCs forming at the interface. Figure 14E shows an EDS area scan of the fibre displayed in D. This highlights the formation of various IMCs.

Figure 15. shows fibres within LM 20 prepared at 720°C. Excellent wetting is achieved and significant reaction has occurred between the fibre and the matrix alloy.

Experiments performed at lower temperature (710 °C) with LM20 and LM25 show fibres recrystallisation and intermetallics formation at the interface fibre/matrix (Figs 16 and 17). The observed phenomena seem to be congruent with the observations at 720 °C described above.

The LM20 and LM25 alloy (which are the most common casting alloys, hence were used more frequently in the experimental procedure) were also tested at 690 °C. At this temperature fibres re-crystallisation rim is already visible, however the intermetallics (forming perpendicular to the fibres rims) tend to remain with a Chinese-script morphology instead of acquiring a more polyhedral one as demonstrated in Figure 18.

In fact the effect of temperature seems to be more related to the habit of the formed intermetallics:

At 690 Chinese-script

- At 710/720 polyhedral

- Above 720 block-like

Nevertheless, the interaction between the fibres and the matrix remain the same and the link between the fibre/matrix reaction and the formation of intermetallics seems to remain unchanged.

To evaluate the importance of fibres composition; linked to their potential wettability and chemical reaction with the matrix, fibres doped with manganese (Mn) were subjected to the same testing. Added to LM 20 at 720°C the fibres, wet and react just the same as the previous experiment. Figure 19. shows the Mn enriched fibres, wetted in LM 20 prepared at 720°C.

In a separate experiment, Al- lOFe was prepared with fibre addition of 5% at 780, 750 and 720°C. Figure 20 shows micrographs of the composite prepared at those

temperatures. It can be observed that the intermetallics tend to form surrounding the fibres at all temperatures. However, due to the high iron content in this case the intermetallics tend to have a block-like morphology.

Conclusions:

The range of temperatures showed that in the presence of molten, or partially molten metal two reactions take place:

• The fibres start recrystalising forming a reactive oxide rim that triggers the

formation of intermetallics in the matrix;

• A chemical reaction also happens at the interface, silicon and iron diffuse out of the fibres fostering the formation of intermetallics while aluminium diffuses into the fibres fostering the formation of the oxide layer which helps the formation of the intermetallics.

• Different habits of IMC formation, namely, dendritic, polyhedral and block-like form at different processing temperatures.

2.4. Fibre distribution - Stirring rpm and stirring time

The distribution and overall wetting of fibres within the matrix alloy is determined by the stirring speed, stirring time and the application of ultrasonication. The time of contact is also a factor as too long in the melt will cause full recrystallisation of the fibres. Different stirring speeds were trialled with the optimum lying between 500 and 1000 rpm. The higher rotation speeds do provide a more homogeneous distribution but can fracture the fibres and in extreme conditions can form an unwettable powder. Figure 21A and 21B show an LM 25 based sample after 10 minutes mixing at 500 rpm, Figure 21C shows an LM 20 alloy after 10 minutes stirring at 1000 rpm. A longer stirring time is useful but cannot exceed the exposure time as defined in section 2.5. The effect of application of ultrasonication is evident from Figure 22. This shows a sample produced using just 500 rpm/10 minutes of stirring but without any ultrasonication. It is clear to see that the fibres are still agglomerated which significantly reduces mechanical properties. Hence stirring is a necessary step (as was observed in section 2.3), while ultrasonication is a preferred step. Figure 23 shows the effect of stirring time on the distribution of fibres. Figure 23A is the LM 20 alloy after 10 minutes stirring, Figure 23B shows the same alloy after 20 minutes stirring. Figure 23C shows data on the % or fibres present in the optical micrograph. This confirms the improvement in fibre entering the melt and homogenously distributing in the matrix when stirred for 20 minutes.

Conclusions

• Stirring the basalt fibres in the melt is an essential step.

• Stirring between 10 and 20 minutes is preferred

• Stirring for a long period time is preferred to present good distribution of fibres.

• Ultrasonic cavitation is a refinement step that is useful in producing defect free castings and helps separate fibre clumps.

2.5. Exposure time

A series of dipping experiments were completed to investigate the effect of molten aluminium on the basalt fibres. To complete these tests a melt of LM 20 was prepared and held at 700°C. Basalt fibres were prepared and aligned before being dipped into the molten metal and held in place for a determined length of time (5, 10 and 30 minutes). The fibres were then removed from the melt and allowed to cool before sample preparation and analysis via SEM. EDS area scans of the fibres show that there is significant diffusion from the fibre into the surrounding metal, and also from the surrounding melt, into the fibre. Figure 24 shows the area analysis for both Aluminium and Silicon. The image shows an abundance of Silicon and a lack of Aluminium within the fibre after just 5 minutes (Fig. 24A). After 10 minutes (Fig. 24B) the fibre re-crystallises at the fibre/matrix interface and the migration of elements begins. The fibre slowly gains aluminium and loses silicon. After 30 minutes the fibre is almost completely transformed (Fig. 24C). These experiments show that the exposure time of the Basalt fibres is an important factor to obtaining the preferred chemistry within the final produced

composite. The exposure time is ideally at least 5-10 minutes to ensure chemical reactions begin and ideally no more than 30 minutes to prevent complete conversion of the fibres. In order to understand the reaction of the fibres in LM 25 a small series of interval castings were completed at 5, 10, 30 and 80 minutes of exposure. Figure 25 shows the optical micrographs of the LM 25 samples after various exposure times. It is visible that reactions occur after as little as 3 minutes (Fig. 25A) and the fibre has almost completely crystallised after 30 minutes (Fig. 25C and 25D). From these experiments it is clear that the reactions occur much faster but are consistent in terms of how quickly the fibre is fully transformed. The exposure time for LM 25 should be more than 3 but less than 30 minutes. Conclusions

• From the experiments above it is possible to define the exposure time.

Defining Exposure Time: The exposure time for basalt fibre addition should ideally be longer than 5 minutes but less than 30 minutes to ensure reactions occur but the amorphous core of the fibre remains intact within a sheath of crystallised material.

3.0. ALLOYING ELEMENT REACTIONS (SI, FE, TI, CU, NI, MG)

3.1. Investigation on the reaction between aluminium alloys and basalt fibres : The previously mentioned experiments on the exposure time and fibre wt.% direct the element reaction experiments of this chapter. A suitable amount of fibres needs to be added but not too much to hamper stirring the low volume melt. The addition and stirring should ideally also be completed in a relatively short time, no more than 30 minutes total exposure. This is illustrated again below in Figure 26 for the reaction with silicon. The fibre displays no visible reaction to the hot aluminium after just 5 minutes. However, after 10 minutes (Fig 26B) the fibre begins to crystallise, primarily at the matrix/fibre interface, with the core of the fibre remaining amorphous. After the longest exposure time (30 minutes) the fibre has undergone extensive crystallisation and only a small fraction of the amorphous core remains, as shown in Fig 26C Conclusions

• The fibre addition sequence and timing is important to the success of the

composite. • Total exposure should ideally not exceed 30 minutes

• Fibres react very quickly, observed in as little as 3 minutes

3.2. Investigation on the effects of different alloy elements:

The following experiments have been completed to observe any reactions within the composite between the major alloying elements and added basalt fibre reinforcement. In this series of experiments a standard method was employed. In each experiment the following steps were taken

The alloy is selected for the experiment based on the concentration of the primary alloying element. Different alloys with a high proportion of specific elements (Fe,Ni,Mg,Ti,Cu)

40g of the selected alloy is heated in a furnace for lhr at 800°C

Approximately 5wt.% of basalt fibres are added. Addition of a large volume fibres is necessary to ensure a suitable percentage of wetted fibres to observe the interface reaction.

Hand stirring of the melt and fibres is completed for 10 minutes, (no ultra- sonication is performed).

After the addition of fibres and stirring, the melt is left within the furnace for Vi hour at 800°C.

The crucible is removed from the furnace and left to cool under ambient conditions.

3.2.1. Iron Fe: AI90- Fe8- Sil

An iron rich aluminium alloy (AI90-Fe8-Sil) was selected to investigate any reaction occurring between the fibres and the matrix alloy. Figure 27 below shows the basalt fibres encapsulated by a large block shaped intermetallic. A series of spectra have been collected to evaluate the composition of the various regions within the micrograph. Figure 27B shows the composition of the large intermetallic (light region), this is mostly aluminium and iron. Figure 27C and D show the fibre core (C) and fibre sheath (D). Conclusions:

• Figure 27 shows the formation of a block-like intermetallic around a pair of basalt fibres with the resulting spectra from the EDS showing the elemental

composition.

· The basalt fibres appear to react with the iron-rich intermetallics, which cause them to change shape (from a needle shape to a block-like shape - related to the temperature of the intermetallics crystallisation) and the fibres to recrystallize.

•

3.2.2. Nickel Ni: AI8- NI2

A nickel-rich aluminium alloy was selected to investigate any reaction occurring between the fibres and the matrix alloy. High nickel content alloys present themselves with large block like intermetallics. Figure 28A shows a low magnification image of the alloy after processing at 800°C. Figure 28B shows a higher mag. image of a basalt fibre

encapsulated within a large block-like intermetallic. Figure 28C and D show the related spectra for the EDS analysis.

When reacting with nickel-rich alloys, basalt fibres cause the formation of much larger intermetallic phases than in other aluminium alloys. Figure 28E and F shows a detailed image and EDS analysis of basalt fibre with intermetallics nucleating from the

crystallized sheath, with the core of the fibre presenting as amorphous.

Conclusions:

• Nickel intermetallics present as large and block-like.

• The intermetallics nucleate and grow from the crystallised fibre sheath.

• The interaction fibres/matrix triggers the nucleation of the intermetallics.

· When reacting with nickel-rich alloys, basalt fibres causes the formation of large intermetallic phases compared to those found in other aluminium alloys.

3.2.3. Titanium Ti: AI65- Ti35

The investigation of the reaction fibres/matrix in the titanium-rich aluminium alloys indicates that also in this case the fibres recrystallise and form an aluminium oxide rim (Fig. 29). Close to the surface of the fibre a small intermetallic growth is visible (spot 4). However, as in this case of silicon the fibre does not appear to be directly responsible for the formation of aluminides intermetallics.

Conclusions:

• The fibres in this alloy system do not appear to react in order to nucleate

intermetallics.

• Aluminium oxide (alumina) forms at the rim of the fibre as in the other alloys.

3.2.4. Commercial Purity Aluminium : LM 0

No elemental analysis was carried out as the fibres only seemed to recrystallize and didn't form any intermetallics Figure 30 shows fibres successfully wetted within the LM 0 matrix material. At this stage there does not seem to be any reactions causing nucleation and growth of intermetallic compounds surrounding the fibre.

3.2.5. Aluminium Casting Alloy LM 20: Al-Si l2

The casting alloy LM 20 is similar to LM 25 (to follow in section 3.2.6) with the exception of a slightly higher Silicon content. A dipping experiment conducted at 700-730°C presents a reaction between the silicon and aluminium within the matrix and fibres. The fibre starts aluminium deficient and gains aluminium; however, the fibre also starts with an abundance of silicon and loses a small amount to the matrix. Compared to the LM25 in this case the matrix contains more silicon, hence the diffusion of silicon on the fibre is limited because the alloy contains almost its maximum of silicon. Figure 31 shows an EDS map of the dipping experiment. As shown, the matrix already begins with a large percentage of silicon dispersed in the alloy (indicated with a white A). In Figure 31C there are more prominent regions of silicon abundance extending from the fibre/matrix interface. This phenomenon is the trigger of the intermetallics formation an optical micrograph of the LM 20 alloy with reinforcements added can be viewed in Figure 32. The fibre has reacted with the alloy and produced block like IMCs surrounding the fibre. Conclusions:

• The high silicon contents in the alloy reduces the loss of silicon from the fibres, however this diffusion phenomenon still acts as trigger for the nucleation and IMCs form around the fibre sheath.

3.2.6. Aluminium Casting Alloy LM 25: Al-Si7.5

Samples created using LM 25 as the matrix alloy were tested for intermetallic formation. With a high fraction of silicon present (however, less than LM 20 (Al-Sil2 the same migration of Aluminium and Silicon was observed. Figure 33 shows optical micrographs of basalt fibres within the matrix. Fig 33A and B shows the formation of silicon rich intermetallic compounds around the fibre sheath. Fig. 33C shows the intermetallic anchored to the fibre revealed on a fracture surface after tensile testing. Figure 34 shows EDS mapping and EBSD mapping to support the localised concentration of silicon in and around the fibre. Fig 34B shows the EDS silicon map, in the top right-hand corner the formation of a silicon/aluminium intermetallic surrounding the fibre is visible

(arrowed). Fig 34C shows the EBSB map of this region. The fibres having been immersed in the melt result in recrystallized zones (regions of small crystals). The smaller concentration of silicon in the matrix in this case allows for further silicon diffusion at the interface fostering the generation of intermetallics.

Conclusions:

• The presence of fibres in the matrix and their re-crystallisation process fosters diffusion of silicon at the interface fibre/matrix triggering the generation of polyhedral intermetallics that work as an anchor between the fibres and the matrix. 3.2.7. Copper Cu : LM 12

The fibres show full re-crystallisation and copper aluminium silicide intermetallic (Fig. 35B) crystallise perpendicularly to the fibre (Fig. 35A)

Conclusions:

• As in the other alloys the fibre here re-crystallises and the silicon diffuses in the matrix triggering the crystallisation of the fibre.

3.3. Magnesium Mg : AZ91 (2.5% Wt.)

Basalt fibres were successfully added to a magnesium alloy the alloy is designated as AZ91 and thus the major alloying elements present are Aluminium (9%) and Zinc (1%). AZ91 was placed in a steel crucible and melted at 730°C for 3 hours. Sulphur hexafluoride (SF ₆ ) was used as cover gas. The mixing temperature was 678°C. 5 wt% of basalt fibres was added and mixed with a rotation speed of 350 rpm for 5 minutes. Ultrasonic cavitation was completed on the mixture for 5 minutes. Once completed the mixture was heated to the pouring temperature of 667°C and gravity cast into a book mould producing metallurgical samples.

Magnesium and aluminium make up 99% of the material in the alloy and both elements are present within the fibre so a reaction was expected. Formation of intermetallics are observed as shown in Figure 36A-36C and Figure 37. EDX line scan analyses have revealed the IMCs to be Al-Fe rich (Fig. 37). Good wettability was achieved with a fairly homogeneous distribution as shown in Figure 38. Some initial mechanical properties for the magnesium composite are presented in table 3.3.1.

UTS (MPa) YS (MPa) E (GPa) Hv5 HBVV2.5 - 187.5

AZ91 Ref 100 87 45 54.4 64.0

AZ91 5Wt% 45 45 26 68.1 75.5 fibres

Table 3.3.1. Mechanical properties of AZ91 and composite sample

Despite a good distribution of fibres, the mechanical properties were found to decrease with the exception of hardness which increased from 54.45 to 68.1, and 64 to 75.5. As the hardness increases the decrease observed in the mechanical properties may be attributed to casting defects introduced in the manufacturing process.

Figure 39. shows an SEM image of IMCs forming around the fibres just like in the previous Aluminium matrix experiments. The EDS analysis shows these IMCs to be rich in magnesium and aluminium oxides, but also contain small amount of iron and manganese.

Conclusions

• In a similar way to the alumina formation in section 3.2.4. (LM 0) the fibres

contain magnesium oxide and thus form magnesium based intermetallics

· One of the major alloying elements is Aluminium and so it is expected to see some alumina/spinel formation.

3.4. Alloy mixtures

Basalt fibres have been shown to react with alloying elements that are present within the basalt fibres, these have been tested in isolation with only a single element considered at a time. The addition of more than one of these elements provides similar results, multiple intermetallics form around the fibre/matrix interface which are rich in specific material present in both the fibre and the matrix. For example, adding AI65- Ti35 to AI8 -Ni2. Figure 40 shows a micrograph of the intermetallic formation around fibres within this alloy mixture. Large block intermetallic are present as expected along with the oxide layer at the rim of the fibres. The formation of this phase is deemed responsible for the trigger of the intermetallic nucleation through the release of silicon from the fibres. When silicon is very low in the matrix all of the silicon diffuses and the intermetallics are found to be aluminides aluminates, when the silicon is present in the matrix in a higher concentration then the addition of the silicon diffusing from the fibres generates silicide intermetallics. In all cases the intermetallics formed are polyhedral with temperature below 750 degrees and block-like when the temperature of the melt is above 750°C. In all cases the percentage of acicular intermetallics is extremely limited.

Conclusions:

• Composites created from mixed alloy sources present IMC reactions and

formation after fibre additions.

4.0. INVESTIGATION INTO MECHANICAL PROPERTIES

4.1. Tensile testing

A series of mechanical tests have been completed to assess the performance of the composite. The first set of tests are designed to evaluate the ultimate tensile strength as well as the offset yield strength to define the onset of plastic deformation within the sample. The hardness of the resulting composite is then assessed.

Tests to assess the ability of the MMC to resist wear were carried out by applying a dry sliding abrasion test using a pin-on-disc setup. Wear loss is calculated form the mass of material lost at 250, 500 and 750m.

In all cases the results from the tests are compared to reference samples that have been prepared and processed in the sample method as the MMC.

Sample Strain at 0.2° o Offset Ultimate Tensile Young's

Failure Yield Strength Modulus

U.C. : Ultrasound % MPa MPa GPa

Cavitation

LM 25 Reference 4.7 76.43 166.9 45

5wt With U.C 4.4 97.99 189.0 60

5wt% Without 4.0 81.56 163.4 51 Table 4.1.1 LM 25/Composite Tensile test data

a e . . ompos te ens e test ata

The tensile strength of the composite material is assessed based upon a reference sample with no addition of basalt fibres but using the same preparation process as the Basaltium samples. Table 4.1.1 shows the data for the LM 25 alloy, and also specifies the need to apply ultrasonic cavitation to achieve the superior property set. Table 4.1.2 shows the data from the LM 20 experiments, from these results the optimum weight % of fibre addition has been deemed as 2.5 wt%. This sample produces the highest UTS and a significant increase in the offset yield. Figure 41 shows the stress/strain curves for the LM 25 samples. It is clear that the 2.5 wt% sample has the best performance in terms of ultimate tensile strength and offset yield but this comes at a price of a reduction in the achievable strain to failure.

The ultimate tensile strength of the material has been tested, Figure 9A shows the data for the LM 20 samples with varying fibre additions. It is clear from the data that both 1 and 5 wt% additions cause a reduction in UTS compared to the reference samples. The only set of samples to see a rise in the UTS and therefore superior properties is the set containing 2.5 wt% of fibres.

The offset yield gives an indication as to the load required to initiate plastic deformation of the sample. The data shown in Figure 9B gives a clear indication that the higher wt% fibre additions are more favourable. Both the 5 and 2.5% samples show an increased yield strength compared to the reference sample 16.92 and 22.16% respectively but with the 2.5 wt% sample once again proving to contain the most favourable properties.

Conclusions:

Offset yield is increased by all fibre additions

2.5 wt. % presents the largest increase in offset yield and also and increase in UTS

UTS is only increased with 2.5 wt. % and is reduced in 1 and 5 wt. % samples.

4.2. Hardness

The tensile data is also confirmed by the expected results derived from the hardness testing. Both the LM 20 and LM 25 tensile results show certain samples displaying increased load to failure. This increased UTS demonstrates an increase in hardness and loss in overall ductility of the material. The hardness results displayed in Figure 9C show an increase in hardness for both the 5 and 2.5 wt% samples, when compared to the reference sample.

The mechanical testing shows an improvement in hardness and tensile strength with the addition of fibres. Further improvement is visible with the additional step of

ultrasonication. It is believed that ultrasonication helps the final distribution of the fibres and minimizes the presence of porosity, overall improving the quality of the final cast.

Conclusions:

• Hardness is increased compared to the reference sample.

· 2.5 wt. % fibre addition provides the highest material hardness.

4.3. Wear Resistance

Wear resistance testing was carried out on LM 20 composite samples and compared to a reference sample with no addition of basalt. Using the pin on disk wear test and subsequent values of material density (Figure 42) a wear rate can be calculated. Figure 43 shows the wear test curves for the reference, 5 and 2.5 wt% samples. It is proven that the wear results are linked to the hardness of the material, a higher hardness value would suggest a corresponding rise in abrasive resistance. From the wear curves in Figure 43 and the results Table 4.3.1 the best performing sample is once again the 2.5 wt% fibre addition sample. This sample suffers 50% less abrasive wear than the reference sample.

a e . . ompos e ear es a a

A clear improvement in wear resistance is visible with the addition of fibres.

Investigation by SEM showed that no fibres pull out is visible hence demonstrating that the intermetallics formation works as anchoring system for the fibres.

Conclusions:

• 2.5 wt.% fibre addition presents the best resistance to wear.

• Basal tium composite exhibits improved wear resistance compared to the LM 20 reference sample.

4.4. Corrosion Resistance

A series of tests were completed to measure the corrosion resistance of the composite as shown in table 4.4.1. The solutions used simulated the potential harsh environments that the composite may interact with. As the primary focus of the material is maritime, salt water simulations are the most suitable. 3.5 and 10% solutions were used in addition to a 1 mol solution. However, to test under particularly aggressive environments a solution containing Hydrochloric acid (HCL) was used. Tests completed with salt water showed such a low corrosion rate for the composite samples that the more aggressive HCI solution was preferred. The results from the accelerated corrosion tests using HCI are presented in table 4.4.2.

Table 4.4.1. Corrosion tests, Sample list and Parameters.

Table 4.4.2. Corrosion test data from samples tested in 1 Mol HCL

General corrosion : These tests are completed with the test piece under zero load conditions. This is achieved by immersing the samples in a heated tank of the solution (60°C) for 48 hours to observe any corrosion. Stress corrosion : These tests are completed with the samples held under tension in a specially designed stress-corrosion tank. Samples were machined smooth and subjected to the testing in a heated bath of the corrosive solution (60°C Sea water) for 48 hours. This accelerates the corrosion test, simulating weeks' worth of loading in a shorter time frame. Each sample was loaded to 75% of the material yield stress

Figure 44 shows examples of the samples subjected to the corrosion testing. In addition to the reference and composite samples, a series of tests was also completed on 316L stainless steel to act as benchmark for corrosion resistant materials currently used in marine applications. The reference samples (no fibre addition) produces the most corrosion reaction product. This surface corrosion is visible on both the general and stress corrosion samples.

Figure 45 shows a corrosion resistance curve for the LM 25 composite and reference samples. The corrosion rate is proportional to Icorr (kinetic value corrosion current density), so this graph shows that the composite sample has a reduced corrosion rate.

Conclusions

• Composite samples showed decreased levels of surface corrosion

• Composite sample presented a reduced corrosion rate compared to the reference samples.

All optional and preferred features and modifications of the described embodiments and dependent claims are usable in all aspects of the invention taught herein. Furthermore, the individual features of the dependent claims, as well as all optional and preferred features and modifications of the described embodiments are combinable and

interchangeable with one another.

The disclosures in UK patent application number 1714401.5, from which this application claims priority, and in the abstract accompanying this application are incorporated herein by reference.