Aluminium-Silicon Lightweight Thermal Management Alloys with Controlled Thermal Expansion

Abstract

With the ever-growing emphasis on global decarbonization and rapid increases in the power densities of electronics equipment in recent years, new methods and lightweight materials have been developed to manage heat load as well as interfacial stresses associated with coefficient of thermal expansion (CTE) mismatches between components. The Al–Si system provides an attractive combination of CTE performance and high thermal conductivity whilst being a very lightweight option. Such materials are of interest to industries where thermal management is a key design criterion, such as the aerospace, automotive, consumer electronics, defense, EV, and space sectors. This paper will describe the development and manufacture of a family of high-performance hypereutectic Al–Si alloys (AyontEX™) by a powder metallurgy method. These alloys are of particular interest for structural heat sink applications that require high reliability under thermal cycling (CTE of 17 μm/(m·°C)), as well as reflective optics and instrument assemblies that require good thermal and mechanical stability (CTE of 13 μm/(m·°C)). Critical performance relationships are presented, coupled with the microstructural, physical, and mechanical properties of these Al–Si alloys.

1. Introduction

With the ever-growing emphasis on global decarbonization and rapid increases in the power densities of electronics equipment in recent years, thermal management has become an escalating engineering challenge. Thus, emphasis on designing systems that allow heat to be transported, stored, or expelled in a manner that enables key system elements to operate with high reliability within a specific temperature range is essential [1]. This raises a need for new methods and lightweight materials to manage heat load as well as interfacial stresses associated with coefficient of thermal expansion (CTE) mismatches between components. Example materials include metal matrix composites (MMCs), such as aluminium–silicon carbide (Al–SiC) and beryllium-beryllium oxide (Be–BeO). However, metallic systems, such as aluminum–beryllium (Al–Be) and aluminum–silicon (Al–Si) alloys, can also be applicable. Each of these material systems presents a unique set of performance properties and manufacturing challenges.

The Al–Si system provides an attractive combination of tailorable CTE and high thermal conductivity, while supporting a lightweight design. Al–Si alloys are of interest for structural heat sink applications that require high reliability under thermal cycling, where a CTE of 17 μm/(m·°C) is desirable to minimize strain due to CTE mismatch between mating copper components. Reflective optics and instrument assemblies are also application areas, in which a CTE close to 13 μm/(m·°C) is necessary to match the nickel plating typically applied to such components whilst providing good thermal stability over broad operating temperature ranges [2,3,4,5].

The production of high Si Al-based alloys via traditional liquid-state processing routes can be problematic due to the challenges posed in achieving good control of the size and morphology of the primary Si phase [6]. Often this can mean that complex alloying additions or unique solidification processing methods are necessary to achieve the properties required for high-performance applications [7,8]. In contrast, manufacturing via a powder metallurgy route enables a wider range of compositions and can provide greater strength and homogeneous-microstructure-associated mechanical integrity [9,10]. Furthermore, powder metallurgy can provide near-net shape capability, which has been proven to lead to reduced waste generation and lower energy consumption per unit mass in comparison to traditional forming methods [11]. By processing in the solid-state, any possible detrimental diffusional or chemical reactions between constituent phases can be minimized. Avoiding molten processing allows the creation of non-equilibrium phase structures. This enables extensive opportunities for novel and interesting material combinations to enhance performance beyond traditional engineering materials.

Mechanical alloying is a solid-state mixing process in which MMC materials and metallic alloys are combined using powder metallurgy methods. As no melting is required, the reinforcement volume fraction and particle size are controlled by raw material selection. Thereby, this process enables precise control of reinforcement particle sizes with higher reinforcement volume fractions, both of which are critical in maximizing and optimizing the property benefits that can be achieved with MMC systems relative to the monolithic material [12]. Mechanical alloying was first developed in the 1960s with the aim of fabricating a nickel-based superalloy with homogeneous distribution of oxide reinforcements for gas turbine applications [13]. Since then, many material combinations have been explored [14,15,16,17,18,19,20,21,22,23,24,25,26]. The basic principle is the cyclical welding and fracture of ductile and brittle powders via repetitive impact, resulting in a distribution of reinforcement within the metallic matrix particles [27]. However, the optimization of key parameters is necessary to achieve a homogeneous and stable process [24,25,26,28,29]. This composite powder then acts as an input into downstream consolidation processes, such as hot isostatic pressing (HIP) or extrusion.

Materion’s mechanical alloying and powder metallurgy technologies have been used for decades to produce commercially available particle reinforced aluminium–SiC SupremEX^® MMCs [30,31,32]. It should be noted that the industrial scale of this technology is highly unique. The total capacity of Materion UK’s mechanical alloying facility is 300 MT per annum. It is understood that light alloy mechanical alloying capability at this scale does not exist anywhere else in the world. A schematic summarizing this powder metallurgy route is given by Figure 1. These MMCs have been shown to display a very fine and homogeneous microstructure, with precise control of particle sizes [33]. The resultant composite materials are known to be isotropic with high modulus, strength, wear, and fatigue performance [30,34], but maintain the machining, fabrication, coating, and processing characteristics of conventional aluminium alloys.

This paper will describe manufacturing of hypereutectic Al–Si alloys via a powder metallurgy route utilizing Materion’s novel mechanical alloying process. Due to the flexibility of the manufacturing route, this lightweight material system has been finely tuned to achieve the specific and desirable CTE values referenced above. A detailed understanding of the relationship between chemistry and CTE has been developed. Critical performance relationships such as this will be presented, coupled with the microstructure and physical and mechanical properties of these Al–Si alloys.

2. Materials and Methods

The powder metallurgy processing described by Figure 1 and [30,31] was leveraged to combine elemental Si with Al alloy powders. The powder mixtures were mechanically alloyed until the Si was distributed within the Al matrix, with an overall refined and homogeneous microstructure. To achieve an optimal balance of mechanical and thermal properties, a 6063-alloy composition was selected, along with high-purity silicon powders. A series of samples with chemistry ranging from 15–55 wt% Si were processed. Initially the mechanically alloyed powder was loaded in a vibratory manner into an aluminium can, before degassing at an elevated temperature under vacuum conditions to eliminate air and any absorbed water vapor. Following this, the HIP process was applied to form fully dense cylindrical billets of material, approximately Ø100 mm × 170 mm in size. This compaction method is summarized by Figure 2.

Samples for CTE evaluation were extracted from these billets and tested per ASTM E228-17 in order to understand the relationship between composition and this critical property. The samples were heat treated to a T6 condition prior to CTE measurement using the following method: solution treatment at 535 °C for 1 h, followed by a cold water (<30 °C) quench (CWQ), and finally, artificial aging at 175 °C for 2 h. All test samples were heat treated with a maximum section thickness less than or equal to 25 mm.

This initial CTE evaluation enabled derivation of the necessary compositions to achieve a CTE close to the target values of 17 μm/(m·°C) and 13 μm/(m·°C) at 28% and 42% Si, respectively. This is depicted graphically by Figure 3. Larger scale cylindrical billets (approx. Ø150 mm × 600 mm) and cuboid billets (approx. 180 mm × 180 mm × 300 mm) were subsequently manufactured at the derived compositions using the same manufacturing conditions to confirm the scalability of the process. These materials were then tested in detail to confirm their microstructure, room temperature tensile properties, density, and electrical conductivity. Tensile testing was carried out using an Instron 3369 loading frame with a 50 kN capacity. Multiple heat treatments were assessed, including the T6 CWQ heat treatment condition defined above, but also the same heat treatment using a 25% polymer-glycol quench (PGQ) medium, as well as the T1 (air-cooled from HIP) condition. Cylindrical specimens were used, with a 5 mm diameter and 25 mm gauge length. Test control was by a constant strain rate of 1.4 × 10⁻⁴ mm/mm/s, until failure. Microstructural analysis was completed using a Zeiss AxioLab5 optical microscope equipped with ZEN core imaging software. CTE was again evaluated per ASTM E228-17 but over a broader temperature range of −100 °C to 200 °C, whilst room temperature thermal diffusivity and specific heat capacity were measured per ASTM E1461-13 and ASTM E1269-11, respectively. This thermal testing was carried out for both the T6 PGQ and T1 heat treatment conditions described above, to assess sensitivity of these properties to heat treatment. Thermal conductivity was calculated as the product of thermal diffusivity, specific heat capacity, and density, per Equation (1) of ASTM E1461-13 [35]. Finally, a series of components representing example applications for each of the alloy components were produced, to assess manufacturability.

3. Results

3.1. Microstructure

An ultra-fine and homogeneous distribution of Si particles within the Al alloy was achieved for all Si contents tested, as Figure 4 shows. It can also be observed that the mechanically alloyed powder was generally blocky and irregular in shape across all compositions. This was consistent with past Al–SiC materials mechanically alloyed by the same process [30,31]. In the consolidated (post-HIP) form, the microstructure was seen to be consistent in all tested directions, indicating isotropic properties as expected. Typical optical micrographs of the 6063 + 42% Si alloy are presented Figure 5. Here, the ultrafine and homogeneous microstructure is again evident. The primary silicon size was observed to be an average of 3–4 μm in diameter, with almost all silicon particles within the 1–7 μm range.

3.2. Coefficient of Thermal Expansion

The achieved CTE matches to commercially pure copper and nickel are summarized by Figure 6 via comparison to reference data from the literature [36,37,38,39]. This shows an exact mean CTE match to copper for the 20–50 °C temperature range. The 28% Si alloy was determined to have a marginally lower CTE for temperatures below room temperature and a marginally higher CTE at temperatures greater than 50 °C. The mean CTE value at all temperatures, as well as the rate of change of CTE with respect to temperature for the Al–Si alloy, was significantly lower than that of unreinforced aluminium alloys (~23 μm/(m·°C)), thus providing a relatively excellent CTE match to copper. The prior heat-treatment condition of this Al–Si alloy was seen to have negligible influence over the temperature range tested.

A close CTE match was observed to nickel, with the 42% Al–Si alloy displaying a marginally higher CTE for the temperatures tested. The rate of change of CTE with respect to temperature for this Al–Si alloy and nickel was similar over the temperature range tested. Again, no influence of the prior heat treatment condition was observed.

3.3. Additional Physical and Mechanical Properties

Beyond CTE, the additional key material properties of the tested Al–Si alloys are presented in

Table 1. Summary of achieved typical properties. Ranges indicate dependence on heat treatment.

Property		Unit	AyontEX 17	AyontEX 13
Composition		-	6063 + 28% Si	6063 + 42% Si
Density		g/cm3	2.60	2.54
Elastic Modulus		GPa	87	103
Specific Stiffness		Gpa/g/cm3	33	41
Mean CTE	(−100–20 °C)	μm/(m·°C)	15.6	12.1
	(20–50 °C)		16.8	13.2
	(20–100 °C)		17.2	13.7
	(20–200 °C)		18.4	14.6
Thermal Conductivity		W/mK	160–170	134
Specific Heat Capacity		J/g/K	0.88	0.85
0.2% Proof Strength		Mpa	170–300	300–340
Ultimate Tensile Strength		Mpa	240–355	325–345
Specific Strength		Mpa/g/cm3	92–137	128–136

. A good balance of strength and conductivity for thermal management applications was displayed by these alloys. Given the large ranges in strength for AyontEX 17 quoted in Table 1, typical engineering stress–strain curves for this alloy are provided ii Figure 7 and true stress–strain equivalents can be seen in Figure 8.

4. Discussion

The simple rule of mixtures calculation model provided a close approximation to the measured relationship between CTE and Si content, as shown by Figure 3. The measured values were consistently slightly lower than predicted and trended away from the model at lower and higher Si content values. A possible explanation for this deviation between the model and the measured values is that the rule of mixtures calculation does not account for phases other than the constituent elements (e.g., Mg₂Si precipitates), or the different elastic properties of the matrix, reinforcement, and associated thermal strains [40]. Particularly at the more extreme values of Si reinforcement (≥50%), the so-called percolation threshold may have been passed, meaning it would be inappropriate to consider the Si phase and the elastic region surrounding it to be individual particles, but more likely a continuous path of reinforcement [41]. This could explain the increased deviation from the model observed at the highest Si contents tested. The negligible effect of heat treatment condition on the CTE of these alloys can be explained by the high Si content necessary to achieve such low CTEs. This thereby limits the ability to influence properties via heat treatment.

The compositions of the two now-commercially-available alloys, AyontEX 17 (with a CTE of 17 μm/(m·°C)) at 28% Si and AyontEX 13 (with a CTE of 13 μm/(m·°C)) at 42% Si, were derived from the measured CTE curve in Figure 3 as described in Section 3. The fine and homogeneous microstructures achieved (Figure 4 and Figure 5) are important in enabling the application of high-throughput and available downstream machining and finishing processes to these alloys. Ease of machinability relative to equivalent liquid-state processed Al–Si alloys or other lightweight, low-CTE materials (e.g., MMCs) can provide a significant benefit in overall part cost by increasing throughput and decreasing tooling costs. Such benefits can be clearly realized in the manufacture of complex, high-performance components, where fine surface finish and high tolerance is required. Several common machining processes, including milling, turning, drilling, tapping, and electrical discharge machining (EDM), have been successfully performed on this material, many of which are highlighted by Figure 9. Here, the thin wall sections at the top of the part are 0.8 mm thick, and the EDM-processed slot at bottom is 0.5 mm thick and 32 mm deep. These features were achieved with standard, high-throughput methods using no special parameters.

Example components representing typical applications for each of the alloys are presented by Figure 10. The miniature structural heatsink in Figure 10a is 58 mm × 40 mm × 5 mm in size, highlighting the ability to carry out precise milling, drilling, and tapping operations with the AyontEX 17 material. The AyontEX 13 mirror substrate presented by Figure 10b was approx. 150 mm in diameter. This was produced via rough machining using standard carbide tooling and completed with poly-crystalline diamond tooling. Given the low-density (2.54 g/cm³) and complex lightweighting geometry, this mirror had a mass of 0.2 kg. Such lightweight mirror design is highly beneficial for airborne applications. More detailed demonstration of lightweight mirror manufacture in this material and associated application-specific testing has been covered elsewhere [42,43].

An emerging topic in lightweight optical mirror design is additive manufacturing to enable complex structures for material and load optimization [44,45,46]. It is then an interesting topic for future exploration to develop and apply additive manufacturing techniques to the AyontEX 13 material, given this alloy already exists in a powder form, as per the powder metallurgy route presented here. Such development would combine the material property benefits discussed here with the design freedom of additive manufacturing for high-performance applications.

Close CTE matches with mating materials are important in the applications discussed above. The results presented in Figure 6 confirm that precise CTE matches to the reference data for copper and nickel materials were achieved with the manufactured alloys over the full −100–200 °C test range. Particularly, for AyontEX 13, the rate of change of CTE with respect to temperature was very consistent with that of commercially pure nickel. This is important for reflective optical and instrument systems, where precise CTE matches to nickel plating layers are critical for both dimensional and thermal stability over broad operating conditions [2,3,4,5,42,43]. It should be noted that in the case of electroless plating, the nickel material will contain some level of phosphorus. The CTE curve for NiP [3,4,47], is very similar to the commercially pure nickel CTE data used for comparison in Figure 6. It can therefore be said the AyontEX 13 material provides an excellent CTE match to both electrolytic and electroless Ni plating for application in high-precision optical mirrors. Of course, the ability to Ni plate the material is critical for such applications. An example of successful application of electrolytic nickel plating to AyontEX 13 is given by Figure 11. Electroless NiP plating solutions are also compatible [43], should this be preferred for the application.

The prior heat treatment condition of the material (T1 v. T6) was seen to have negligible impact on CTE for all compositions tested. It can therefore be understood that controlling the thermal expansion of these hypereutectic Al–Si alloys is best achieved through adjusting the Si content, as opposed to via heat treatment optimization.

As expected, increasing the Si content resulted in a linear decrease in density and a linear increase in elastic modulus. Increasing Si content was seen to result in a decrease in thermal conductivity, but also a decrease in the dependence of heat treatment condition on the thermal conductivity. For AyontEX 17 (6063 + 28% Si), the mean thermal conductivity was determined to be 160 W/mK in the T6 condition, but 170 W/mK in the T1 (air-cooled from HIP) condition, a difference of ~6%. In contrast, for AyontEX 13 (6063 + 42% Si), the heat treatment condition was found to have a negligible impact on the thermal conductivity.

This high thermal conductivity of AyontEX 17 in the T1 condition, coupled with the precise CTE match to Cu over the 0–100 °C temperature range (see Figure 6), highlights the materials’ applicability for use in lightweight structural heatsink assemblies. As a direct replacement for 6061 Al, for example, AyontEX 17 provides an increased modulus and decreased CTE and density, whilst maintaining thermal conductivity. This is critical in enabling high power density devices, where heat load must be efficiently managed, as well as interfacial stresses driven by CTE mismatches to mating Cu components.

Similar trends with respect to Si loading were observed regarding the mechanical properties, and in particular the yield strength. The 0.2% proof strength of AyontEX 17 was seen to almost double from 170 MPa in T1 to 300 MPa in T6, with significant differences in ultimate tensile strength and strain to failure. These differences are depicted by the stress–strain curves given by Figure 7 and Figure 8. This data was generated on billet material directly following the HIP process. The application of secondary forming processes, such as forging and extrusion, will enhance the ductility of the finished product. This remains an interesting topic for further exploration to allow close to shape forming operations. By comparison, the mechanical properties in the higher Si vol% materials had a much lower dependence on the heat treatment condition, as outlined by Figure 12 and Figure 13. These effects can be explained by simply considering that as the Si vol% was increased, the vol% of the 6063-alloy decreased. This in turn meant that less Mg was available within the alloy for Mg₂Si precipitation strengthening effects through heat treatment. Additionally, as the quantity of the primary Si phase particles was increased, the relative distance between adjacent Si particles was decreased, and thus strengthening because of the primary Si phase became a more dominant mechanism relative to the Mg₂Si precipitation phases.

The increasing influence of the Si phase as the Si vol% was increased to more extreme values (≥50%) led the ensuing material to become more brittle. This resulted in an associated decrease in both 0.2% proof strength and ultimate tensile strength, as captured by Figure 12 and Figure 13. For such compositions, the proportion of the primary Si phase is so significant that fracture can occur at the alloy–Si particle interface, due to Si particle clustering or the percolation effects described above. The severity of such effects within particulate MMC materials is well understood [48]. In contrast, at a more modest Si vol%, such as those used for the AyontEX 17 and AyontEX 13 materials, fracture occurs predominantly through the ductile 6063 alloy matrix phase.

Deeper exploration of the microstructural effects and phases at play in these hypereutectic Al–Si alloys provides an interesting avenue for future study. Investigating the impacts of using recycled Al alloy as an input raw material, and an associated high Fe content, could prove interesting. Previous studies have identified effective Si to have a significant impact on the performance of 6000-series aluminium alloys with high Fe content [49]. Exploration of how this effect translates to the more extreme Si levels in these hypereutectic alloys would be an interesting topic, particularly so with a view towards possible incorporation of recycled or secondary aluminium sources in the AyontEX alloys or SupremEX MMCs to increase the circularity of these high-performance materials.

5. Conclusions

A family of now commercially available high-performance hypereutectic Al–Si alloys (AyontEX™) have been developed with precise CTE matches to copper and nickel. Core material properties for the intended applications have been characterized, along with the key manufacturability considerations. The fine and homogeneous microstructure provides enhanced performance and manufacturability relative to equivalent liquid-state processed alloys. Further development work for these Al–Si alloys should be directed towards the use of mechanical alloying technology for lower CTE values, as well as testing the CTE of existing alloys over a broader temperature range. The development of additional fabricated forms via forging and extrusion for increased ductility and product form flexibility is also under investigation. Exploration of the incorporation of secondary or recycled aluminium sources as a raw material in this manufacturing process, and subsequent understanding of the effects of associated impurity elements, is an area for possible future study also.

Another potential application for Materion’s mechanical alloying technology and these alloys moving forward is additive manufacturing. Particularly for lightweight, high-precision optical mirror manufacture, the AyontEX 13 composition is interesting. More broadly, however, the mechanical alloying process employed in this study allows for compositions to be tailored to suit the requirements for additive methods in terms of alloy and reinforcement composition and microstructures. The process also has sufficient scalability to provide an industrial solution. Due to this ability to mix a wide range of materials to produce light metal composite powders, it may be possible to produce new and novel powders as an input for additive processes. This could lead to additive manufacturing material and property options that do not exist today.

These research areas provide exciting opportunities for further exploitations of mechanical alloying technology, and powder metallurgy in general, towards developing new products to meet the needs of customers in global markets such as within the aerospace, high-performance automotive, space, and defense sectors.