Manufacturing Methodology on Casting-Based Aluminium Matrix Composites: Systematic Review

: Ongoing industrial demand for lightweight materials has spiked the research interest in aluminium-based metal matrix composites for its speciﬁc properties. The amount of scientiﬁc publication available on the matter has led to the vast production of knowledge, which highlights the need for a systematic assessment if further progress is expected. In this paper, a systematic review of the published literature is conducted, according to the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses, on the Scopus and Web of Science databases were used in the literature search, which was completed on the 29 August 2020. The data of the research work is structured in the particle pre-processing stage and the melt processing stage. The present review clariﬁes the combined pair-wise effect of particles and the melt treatment performed on their wettability or dispersive or de-agglomerative capability, which allows to achieve their ﬁnal mechanical properties.


Introduction
Current competitiveness in manufacturing industries implies that the overall technological processes and material processing must be optimized, meaning that there is an increasing interest in manufacturing routines based on the least material approaches. Indeed, the most effective route to achieve this goal is the enhancement of the overall mechanical properties, i.e., increasing the load-bearing capacity of materials to reduce their volume in structural applications. This route is aligned with current environmental concerns. For instance, the transportation industry is also searching for innovative materials to reduce the weight of their vehicles and thus decrease fuel consumption and CO 2 emission [1][2][3][4][5].
Light alloys are current prominent solutions when compared to classic steel construction, occupying a significant quota in the manufacturing economy [6][7][8]. Their use generates a significant decrease in density (33% and 77%, respectively for Al and Mg alloys) when compared to steel [9,10]. Additionally, they may present interesting complementary mechanical properties, e.g., Mg alloys are regarded as high vibration damping alloys [11][12][13][14]. Al alloys, even though their density is relatively higher than Mg alloys, their processing may be regarded as simple. Combining both low melting temperatures relatively to steel and additional processing conditions that must be assured in the processing of Mg alloys to prevent severe exothermal reactions (e.g., protective atmosphere with toxic SF6 [15,16] and specific coatings [17] to prevent rapid oxidation in casting processes [18]). In fact, the physical and mechanical properties of aluminium alloys are rather tailorable [19][20][21], considering that the correct processing (e.g., precipitation/work hardening and forming/machining) [22,23] and alloying (e.g., soluble elements and master alloys) [24,25] is   • Ball milling reduced powders aspect ratio and surface roughness • Ball milling induced a close contact between particles along with a uniform distribution • Cold pressing improved wettability, which induced a uniform distribution and better tensile properties. • Cold pressing powders are reported to decrease stirring turbulence   • Ball milling reduced powders aspect ratio and surface roughness • Ball milling induced a close contact between particles along with a uniform distribution • Cold pressing improved wettability, which induced a uniform distribution and better tensile properties. • Cold pressing powders are reported to decrease stirring turbulence  • All method effectively inhibits Al 4 C 3 formation (SnCl 2 > PdCl 2 > NaCl > NH 4 Cl) • Best mechanical properties reported by NaCl (40 g/L at 100 min) followed by PdCl 2 (1 g/L at 120 min), SnCl 2 (40 g/L at 60 min), and NH 4 Cl (15 g/L at 60 min) [  Although shear reduced the variance of the particle size, its material properties often show a more scattered data when compared with the regular impeller stirring

Discussion
This chapter details the discussion regarding composite manufacturing and as such, is subdivided in pre-processing and melt processing.
The pre-processing subchapter addresses any technique or parameters before the introduction of the reinforcement particles in the melt to enhance the wettability between the particle and the melt. These include composite ball milling, wetting agents, method coating, oxidization and heat treatment. Considering that testing the contact angle is not viable in a casting scenario, the assessment of a successful wetting is carried out by qualitative visual inspection of the microstructure [84,85], particle image analysis [86][87][88][89], particle residuals or rejected collection [90,91].
The melt processing sub-chapter addresses any methodology, technique and inherent parameters which aims to distribute and de-agglomerate particles upon introduction in the melt. The presented techniques have an impact on particle wetting and sample microstructure which cannot be evaluated as isolated variables. The assessment of a uniform distribution is carried out by qualitative visual inspection of the microstructure [69,84,85],  [77,92,93], particle image analysis regarding distribution along with agglomeration [88,89,94] and hardness profile along a specific axis [95].

Coating
Given the low wettability of ceramic particles [96,97], metallic coatings or oxidation of the interface is often performed before the reinforcement is introduced in the melt to improve reinforcement-matrix bonding. Among the available options, nickel and copper are the most common choices for electroless deposition coating [72,[98][99][100][101].
Pourhosseini et al. have studied the effects of Cu, Ni and Co coatings on the wettability of Al 2 O 3np through electroless deposition in Al/Al 2 O 3 stir casting [72]. When compared to uncoated samples, Ni, Cu and Co report enhancement with regards to tensile properties and hardness, respectively. Ni coating performance was justified given its distribution and lesser agglomeration as well as the formation of the Al 3 Ni phase in the Al 2 O 3 interface. Mousavian et al. have studied the effects of Cu, Ni and Co coating of SiC mp through electroless deposition in Al/SiC Stir casting [102]. The resultant mechanical properties show that Cu, Co, Ni have effectively reinforced the Al-based composite. The authors hypothesised that higher incorporation percentage is promoted by the endothermic reactions with molten aluminium. Yang et al. have studied the effect of Ti coatings on the wettability and distribution of particles in Al2014/SiC p stir-casting [73]. It was reported that the formation of a Ti 5 Si 3 in the SiC/Ti interface would decrease its contact angle in a shorter period for 800 • C, leading to higher interfacial energy when compared with the uncoated samples.

Oxidation
While oxygen solubility is extremely low in liquid aluminium, the Al element presents a very high affinity for oxygen. It is suggested that Al can alter the interfacial chemistry by forming an oxide layer with a lower O/metal ratio than in the bulk oxide that may improve wetting [60]. Zhang et al. have studied the bonding effects of oxidised, Ni and Cu coated SiC p of an Al/SiC through semi-solid stir casting [74]. Results report that the SiC p oxidization has the most significant impact on yield strength, ultimate tensile strength and fracture strain of all tested samples. The authors show that the developed SiO 2 layer improves the interfacial wettability by reactions with Al and Mg elements in the melt, thus producing its oxidised phases Al 2 O 3 , MgO and MgAl 2 O 4 . The wettability effect of SiO 2 is in agreement with other works [88]. Similarly, Mahesh et al. have investigated the effects of heat treating B 4 C p on the microstructure and mechanical properties of Al/B 4 C by stir casting [70]. Even though a uniform distribution of the particles was achieved with heat treatment at 250 • C. The authors state that above 300 • C particles form a glassy boric oxide phase (B 2 O 3 ), which causes sintering and will eventually induce agglomeration in the melt.

Pre-Heating
When a mismatch of thermal expansion coefficients exists between the reinforcement and the main matrix, frequent for most ceramic-reinforced MMC, higher temperatures can induce porosity and dislocations in the vicinity of the particle [75]. Additionally, the increase in reinforcement content, higher casting temperature increases the probability of gas solution in the melt, which will later result in solidification defects [103]. In fact, reinforcement particles tend to agglomerate in lower melt temperatures due to higher viscosity [75] (p. 4479).
Azadi et al. compared the effect of ball milling and heat treatment of SiO 2np on mechanical and tribological properties. The presence of SiO 2p enhanced hardness, elastic modulus, compression yield strength but decreased tensile strength, elongation and toughness. It was reported that the pre-heating of 1 wt.% reinforcement particles at 400 • C improved its wetting and material properties while promoting porosity defects [104]. Ghandvar  process. The authors compared the effects of different particle pretreatments regarding surface treatment, preheating and the use of a wetting agent [105]. It is reported that the pretreatment enhanced the particle/matrix wettability as well as its distribution, which was then further improved with the addition of 1 wt.% Mg. The increase in hardness, UTS and elongation was associated with the removal of impurities during cleaning and the formation of SiO 2 during the heating.

Cleaning
Adeosun et al. have studied the effects of cleaning time and reagents in a 6011/SiC p composite particle wettability and mechanical properties, including the use of NaCl, SnCl 2 , NH 4 Cl and PdCl 2 [69]. Wetting reagents were proven effective as they improved interfacial bonding and inhibited the formation of Al 4 C 3 , although residual particle clustering could be observed. Optimal ductility and hardness are reported by SnCl 2 and PdCl 2 with clearing period of 100-120 min followed by a 60 min cleaning with NaCl or NH 4 Cl. Similar conclusions were drawn out by other authors [105,106].

Melt Processing
Shared by any MMC processing method, the reinforcement processing can potentially limit the efficiency for particle-matrix bonding. In the MMC casting route, high contact angle and poor chemical bonding between the reinforcement particle-matrix are often caused by poor wettability [107]. Contact angle can be understood as the angle formed by the interface between a non-reactive liquid and flat, smooth and chemically homogeneous solid surface in a state of equilibrium at a given temperature [60]. As depicted in Figure 3, the impact of temperature on the wetting behaviour of the metal/ceramic pair is evidently positive. Even so, there is a sparse distribution of the data points that can be attributed to the interaction with the available atmospheric content, as thoroughly reported in [96,[108][109][110][111][112][113][114], being dependent on sample purity, pre-treatment, crystallization, oxidization with available oxygen molecules, the experimental procedure itself, among others [115]. through a semisolid process. The authors compared the effects of different particle pretreatments regarding surface treatment, preheating and the use of a wetting agent [105]. It is reported that the pretreatment enhanced the particle/matrix wettability as well as its distribution, which was then further improved with the addition of 1 wt.% Mg. The increase in hardness, UTS and elongation was associated with the removal of impurities during cleaning and the formation of SiO2 during the heating.

Cleaning
Adeosun et al. have studied the effects of cleaning time and reagents in a 6011/SiCp composite particle wettability and mechanical properties, including the use of NaCl, SnCl2, NH4Cl and PdCl2 [69]. Wetting reagents were proven effective as they improved interfacial bonding and inhibited the formation of Al4C3, although residual particle clustering could be observed. Optimal ductility and hardness are reported by SnCl2 and PdCl2 with clearing period of 100-120 min followed by a 60 min cleaning with NaCl or NH4Cl. Similar conclusions were drawn out by other authors [105,106].

Melt Processing
Shared by any MMC processing method, the reinforcement processing can potentially limit the efficiency for particle-matrix bonding. In the MMC casting route, high contact angle and poor chemical bonding between the reinforcement particle-matrix are often caused by poor wettability [107]. Contact angle can be understood as the angle formed by the interface between a non-reactive liquid and flat, smooth and chemically homogeneous solid surface in a state of equilibrium at a given temperature [60]. As depicted in Figure 3, the impact of temperature on the wetting behaviour of the metal/ceramic pair is evidently positive. Even so, there is a sparse distribution of the data points that can be attributed to the interaction with the available atmospheric content, as thoroughly reported in [96,[108][109][110][111][112][113][114], being dependent on sample purity, pre-treatment, crystallization, oxidization with available oxygen molecules, the experimental procedure itself, among others [115].

Melt Temperature
Abdizadeh and Baghchesara have studied the optimal melt temperature (750 °C-950 °C) for 5-15 vol.% of B4C to produce an A356 alloy-based MMC [75]. It was observed that

Melt Temperature
Abdizadeh and Baghchesara have studied the optimal melt temperature (750 • C-950 • C) for 5-15 vol.% of B 4 C to produce an A356 alloy-based MMC [75]. It was observed that porosity increased alongside melt temperature and reinforcement content. The authors show that the inferior mechanical properties at lower temperatures were caused by the higher agglomeration tendency as well as poor wettability. Optimised toughness, elongation, tensile and yield strength were obtained at 850 • C reinforced with 10 vol.% B 4 C. The best hardness values were shown at highest particle volumetric content and melt temperature from the experiment. Akbari et al. have studied the effects of different particle sizes (50 µm, 10 µm and 20 nm) of Al 2 O 3 particles at different melt temperature (750, 800, 950 • C) and volumetric content (0.5 to 7 vol.%) in an A356 alloy [116]. Particles were milled together with Al powders in a 1:1 weight ratio and introduced through gas injection. Stirring was carried out for 10 min at 450 rpm before pouring into a steel die, after which samples were heat treated (T6). Porosity is shown to increase with higher volume fractions and smaller sized particles, to which the authors justify with the increase of contact surface area and particle agglomerate (i.e., entrapped gas). For a 1.5 vol.% content, porosity aggravates at higher melt temperatures, especially for nanoparticles. Results report an optimum volume fraction of 1.5 vol.% for tensile properties and 2.5 vol.% for hardness for nanoparticles. For microparticles, the optimised volume fraction is in the range between 3 and 5 vol.%. Gurusamy et al. have studied the effect of melt and die temperature on the reinforcement of 10 vol.% SiC p on an A356 alloy by squeeze casting [117]. The processing procedure consists of melting the A356 alloy at 750 • C in a graphite crucible, followed by the introduction of the preheated reinforcement particles and a stirring period of 10 min before pouring into a preheated die. In this work, the authors have employed a melt temperature of 750-900 • C for single die temperature (400 • C) and produced samples with die temperatures of 250-400 • C for a single melt temperature (800 • C). Results show that these parameters had a significant effect on the microstructure and mechanical properties, while the effect on the MMC distribution was minimal. Optimal UTS, impact strength and hardness are achieved at 850 • C and 350 • C melt and die temperature, respectively. Abdizadeh et al. has compared the fabrication A356/MgO np at different particle content (1.5, 2.5, 5 vol.%) by stir casting and powder metallurgy at different processing temperatures [118]. In the stir casting method, particles wrapped in aluminium foil were introduced in the melt along with Keryolit (Na 3 AlF 6 ), followed by a stirring period of 14 min at 420 rpm at melt temperatures (800, 850 and 950 • C) and pouring. In the powder metallurgy method, MgO p and Al powders were mixed in ethanol for 30 min and dried, followed by uniaxial pressing and sintering at different temperature (575, 600 and 625 • C). Density measurements reported an agreement to calculated density, while powder metallurgy induced cracks and micropores were observed for higher MgO content. The authors suggest an optimal processing temperature of 625 • C and 850 • C for powder metallurgy and stir casting, respectively, for mechanical property enhancement. Stir casting relative shows some benefits in compressive strength and hardness due to lower porosity. Mazahery and Ostadshabani have fabricated A356/Al 2 O 3np composites with different reinforcement particle content by stir casting at 800 • C and 900 • C [119]. Al 2 O 3 and aluminium particles were ball milled, passed through a mesh screen, cold-pressed (200 MPa) and passed again through a mesh screen. The particle Al powder mixture was combined with 1 wt.% Mg p , and packed in the aluminium foil, being added to the melt once the vortex was formed. Stir casting procedure is characterised by the initial heating of the melt at the desired temperature under protective Argon gas, followed by a stirring period of 15 min at 600 rpm and finishing by pouring into preheating steel die. Samples microstructure displays uniform distribution of particles with residual agglomeration. Grain refinement is evident in all samples, particularly for 1.5 and 2.5 vol.%, in which the highest values of tensile and compressive strength were obtained. Although Al 2 O 3np improved ductility at a melt temperature of 800 • C, the same was verified for 900 • C. The authors suggest that the minimal increase in porosity for greater particle content can be attributed to the reinforcement higher surface area. Shabani et al. have researched the fracture mechanics of 1.8 wt.% A356/Al 2 O 3np composites produced by gravity sand casting, squeeze casting and compo casting [120]. Al 2 O 3 particles were milled with Mg and Al to promote wettability (1:1:2 Al 2 O 3 :Mg:Al) and pressed as tablets. Sand casting and squeeze casting were stirred at 720 • C melt temperature for 7 min before pouring and pressing with a 50 kg weight. For compo-casting, materials were initially melted at 720 • C and cooled down to 612 • C and magnetically stirred for 7 min before being poured into a steel mould and pressed with a 50 kg weight. Compared to sand casting, dimples observed in squeeze casting fractography were smaller and more homogenous. It is reported severe agglomeration in the sand casting samples, whereas the pressurised methods are shown to have broken dendrites and dispersed more homogeneously. The authors conclude that smaller and uniformly distributed silicon mitigate the stress concentration, thus enhancing the ultimate strength.

Wetting Agent
A wetting agent is considered an additive that can be applied to the reinforcement or melt to enhance the chemical bonding between the two phases and overcome the capillary forces that lead to non-wetting scenarios [50,84,121].
Singh et al. have studied the effects of a Mg wetting agent on the mechanical and thermal properties of an Al/SiC p composite [84]. Results report small changes in the melting point and heat fusion resulting from the introduction of ceramic particles. There is an improvement in tensile strength and hardness at cost of elongation as the weight fraction of the reinforcing phase is increased. The authors state that Mg acts as surfactant by consuming oxygen from SiO 2 and forming MgO and MgAl 2 O 3 . Despite the strengthening effect, the decrease in melt fluidity by the introduction of Mg element promotes clustering in contents higher than 15 wt.%. Without Mg, suboptimal tensile strength is reported past 25 wt.% in which the differences were pointed to the brittleness of the microstructure phase. Tahamtan et al. have studied the effects of introducing ceramic particles in composite particles for 5 vol.% of Al 2 O 3np and Al 2 O 3mp [65]. Reinforcement particles were processed through ball milling the Al 2 O 3 with Al and Mg powders, after which, some were coldpressed into tablets at 1 MPa. Stir casting was carried out at 400 and 1200 rpm for 15-30 min at 745 • C for the liquid method and 640 • C for the semi-solid method. As seen from the tensile properties, either form of the composite particles was proved as an effective measure to improve wettability as well as achieve a uniform distribution. The decrease in particle size and stirring temperature is reported as a great improvement as there is less interfacial reaction and the dendrites in the semi-solid state are fractured by the vortex. Su et al. has manufactured an 2024 alloy reinforced with 0.6 wt.% Al 2 O 3 along with Al powder as a carrier agent [122]. With an average diameter of 13 nm and 80-100 µm for Al 2 O 3p and Al p at a 9:1 wt.% ratio, respectively, particles were introduced into the melt and stirred for 10 min before pouring. The authors state that the carrier agent establishes a pre-dispersed state which later enhances the wettability between the metal and nanoparticles. Compared with the traditional stir casting, the employed method enhanced UTS, YS and hardness by 59%, 58% and 16%, respectively. Mousavian et al. has manufactured composite SiC-reinforced aluminium MMC milled with Cr, Cu and Ti by stir casting followed by hot extrusion [123]. In this work, mixed particles were encapsulated in aluminium foil packets before their introduction in the melt. Additionally, the authors refer additives as carrier agents given their functionalities in wetting prevent particle agglomeration. Even so, segregation and agglomeration could not be avoided. Cr and Cu reported as unsuitable for mechanical property enhancement and it was reported that (SiC/Ti) p was successful in reinforcing the matrix with good distribution. The authors show that particle size, process temperature and stirring time limit the manufacturing efficacy. Samiee et al. has manufactured an (Al/Al 2 O 3 ) p composite from a pure aluminium alloy by stir-casting. In their work, the authors introduced the reinforcement particle through an argon-based injection along with 3 wt.% Mg during the stirring stage. It was reported that for extended periods (10 h) of milling, processed (Al 2 O 3 /Al) p would increase in size and volume, due to collision events and their individual mechanical properties [91]. Akbari et al. have studied the wetting effects of either Al or Cu powders on the stir casting of A356/Al 2 O 3np [93]. In this work, composite particles were balled milled for different periods (1-24 h), wrapped in aluminium foils and introduced in the vortex. Stirring was performed for 12 min at 450 rpm at 850 • C. After pouring and solidification, the samples were subjected to a T6 heat treatment. The authors report a uniform distribution nanoparticle and reduction in grain size favoured by the ball milling process and mixed powders. With the increase in ball milling time, porosity increased slightly and a general decrease in mechanical properties was observed. The authors suggest that gradual oxidation of the metallic powders may have had a negative impact on the results. The superior mechanical properties reported by the Al 2 O 3 /Cu sample were attributed to the strengthening effect of copper in the alloy. Yuan et al. has used squeeze casting to reinforce an A356 alloy with (SiC/Al) p [124]. While performing a stirring period of 10 min at 180 rpm, the authors state that when the size of the composite powder is too small, the stirring stage will not disperse the particle as efficiently as expected, which will then result in agglomeration. Amirkhanlou and Niroumand have studied the effect of aluminium powder with reinforcement particles by the stir casting and compo-casting to obtain A356/SiC p composites [125]. (SiC/Al) p was processed through a 52 h low energy ball milling and later injected into the melt by means of Argon gas as a carrier agent. After a melt homogenization period of 2 min at 700 • C, particles were injected along with a stirring stage at 500 rpm. The stirring period lasted until the desired temperature of 607 • C (compo-casting) or 650 • C (stir casting) was reached. The authors report that the injection of a composite powder decreases the overall porosity, promotes homogeneous distribution, and enhances wettability. Compo-casting was shown to further reduce porosity and enhance hardness relatively to the other approaches.

Additives
Alternatively, authors have explored the innate wetting effects of direct melt additives to develop sustainable solutions. Early publications report on the wetting effects of Mg, Cu and Si alloying elements, by that order, through the contact angle between Al and Be, B 4 C, and graphite [126].
Zheng et al. have studied the effect of stirring time in the microstructure and dispersion of B 4 C-reinforced aluminium with the use of a Ti additive [94]. The authors report that introduction of the Ti in melt promotes the growth of Al 3 Ti films in the surface of the particle, thus, limiting the erosion reactions within a 5 min stirring time. Furthermore, it is stated that its particle incorporation threshold (16 vol.% average) is achieved for stirring periods greater than 20 min. However, for longer holding and stirring time, the Al 3 Ti phase will decompose into Al 3 BC, AlB 2 and TiB 2 , with this last one growing courser and detaching from the B 4 C p . Shahriyari et al. have fabricated Al-based MMC by incorporating (Al 2 O 3np /Cu mp ) p and studied the effect of a Sr as an additive in its microstructure and mechanical properties [127]. The authors report that at optimum 0.5 wt.% strontium, a refining effect can be observed as well as the morphological spheroidization of the Si phase. Higher content will lead to the formation of Al 2 Si 2 Sr, which, along with a brittle copper phase, will have a negative impact on its mechanical properties. It is suggested that the increase in wettability is promoted by a transition layer or a protective oxide layer in the particle/matrix interface. Mohammadpour et al. have studied the effect of introducing 1 wt.% Ca, Mg, Si, Ti, Zn and Zr alloying elements in the incorporation of 3 wt.% SiC mp in pure aluminium melt [87]. Particles were wrapped in aluminium foil and added to the centre of the vortex. Stir casting was carried out at 500 rpm, at a melt temperature of 680 • C for a period of 6 min, after which the melt was poured into a steel mould preheated at 450 • C. For an introduction of approximately 2.53 vol.%, Mg (2.46 vol.%) and Ca (1.92 vol.%) stand as the most effective element on particle incorporation followed by the less efficient Si, Zn and Zr, in that order. Among the alloying elements, samples contacting Ti and Zr substantially increase in microhardeness as a result of the formation of secondary phases (Al 3 Ti and Al 3 Zr). Panthglin et al. have studied the effect of Zr additives on an A356-SiC composite without resorting to Mg wetting agents [88]. The addition of 0.2-0.9 wt.% refined the microstructure and refine the grain by 61-63% when compared to the 54% of the Al-5Ti-B master alloy. The introduction of this additive further improved particle incorporation as well as promoted a uniform distribution of the SiC p . Additionally, Zr enhanced the mechanical properties, such as creep resistance, hardness and microhardness.

Carrier Agent
Additionally, the method on how the reinforcement particles are introduced is also considered an important variable that impacts the chemical bonding with the matrix. Given the popularity of the stir-casting process, most authors usually introduce the preheated particles wrapped in aluminium foil [93,[128][129][130][131][132] in the centre of an already formed vortex [68,75,79,82,90].
Ghandvar et al. have studied the impact of different steps in particle pre-processing on the reinforcement of 20 wt.% SiC p in an A356 alloy [105]. Samples include untreated particles, particles cleaned in acetone and ultrasonic bath followed by a 900 • C (2 h) heat treatment and a separate set of clean and heat-treated particles that was mixed with 1 wt.% Mg and packed together in aluminium foil. The authors report that each pre-processing stage produced a significant increase in the composite mechanical properties. The superior wettability between ceramic reinforcements and matrix is justified by the addition of Mg in aluminium foil. Samal et al. proposed a hollow spindle-stirring mechanism as an alternative for the introduction of particles in the vortex [133]. It is explained that by using a plunger rod with a mild-steel capsule at one end, particles can be directly introduced to the inner regions of the melt. For purposes of experimental validation, the authors reinforced an Al-Mg alloy with 10 wt.% SiC. Particles were wrapped in aluminium foil and plunged to 800 • C melt and followed by a 10 min stirring period at 500 rpm. From the resulting microstructure and hardness at different positions, the authors concluded that particles incorporated settled mostly in the middle and bottom section of the cast. Hanizam et al. have studied the effects of mechanical stirring and heat treatment on thixoformed A356 alloy reinforced with 0.5 wt.% multiwalled carbon nanotubes (MWCNT) [128]. Nanotubes were mixed with 0.5 wt.% Mg and wrapped in aluminium foil. After melting at 700 • C, the temperature was lowered to 650 • C for particle introduction. Similar to the previous authors, the wrapped MWCNTs were placed in the melt through a plunge. With a distinct three-bladed impeller, the melt was mixed at 500 rpm for 10 min and then poured to a preheated mould (150 • C). Afterwards, the sample was thixoformed at 580 • C, followed by a T6 heat treatment. Despite the stir-casting sample reported a 14% volume porosity, a gain in ultimate tensile strength, elongation and hardness is also verified with no brittle behaviour. Amirkhanlou and Niroumand have produced A356/SiC p through stir-casting and compo-casting [82]. MMC processing includes the injection of either SiC p , (SiC p /Al) p or (SiC p /Al/Mg) p assisted by pressurised argon at 700 • C, followed by a cooling and stirring period up until the melt temperature of 650 • C or 607 • C, after which the cast was poured into the mould. The authors state that the injection of the reinforcement particles has enhanced the particle distribution and helped minimise the cluster agglomeration. With each reinforcement particle employed, a reduction of the SiC p average diameter, secondary dendrite arm spacing and grain size through image analysis. Compo-casting was reported to display enhanced mechanical properties, despite both values had a similar impact on the SiC particle size. Prakash et al. have studied the effects of particle feeding techniques on the microstructure and mechanical properties of a 2.5 wt.% 7075/Al 2 O 3np by bottom tapping stir casting [134]. In this work feed techniques include (i) regular feeding, (ii) two-step feeding, (iii) pre-melt feeding, (iv) capsulate feeding and (v) double-layer feeding, which are then followed by a stirring period and pouring. Regular feeding refers to the introduction of the particles in a static melt. Two-step feeding consists of the regular method of half the melt charge and particle content followed by the charging of the remaining aluminium charge and reinforcement content. The pre-melt method is described as the charging of the two materials in the crucible to melt together. Capsulate feeding consists of the packaging of the reinforcement particles in aluminium foil to be later pre-heated at 300 • C and added to the melt. Similar to the pre-melt method, particles are loaded into the furnace between two layers of aluminium charges. The three first techniques employed (i-iii) report high levels of porosity, an agglomeration tendency and failure to incorporate the matrix as a considerable percentage of the particles are found at the metal surface. Optimal mechanical properties and particles distribution are achieved by double-layer and capsulate feeding, in that order, given that the last reports some agglomeration and higher porosity.

Melt Additives
The combination of multi-parameter processes comprising the introduction of an additive can frequently overlap the effects promoted by reinforcement particle pre-processing and have a significant impact on the final MMC mechanical properties. Considering that the incorporation of particles is widely recognised for refining effect, the use of grain refiners along with ceramic reinforcements is not a common procedure considering that it may be redundant.
Suresh et al. have studied the effect of a commercial Al-Ti-B master alloy and flux modifier in the reinforcement of 1-15 wt.% fly ash in eutectic Al-12Si alloy [135]. The authors report an increase in hardness, ultimate tensile strength and elongation in additions up to 10 wt.%. For higher mass fractions, both tensile strength and hardness tend to decrease, while elongation only decreases past 15 wt.%. Optimum results are achieved by samples with grain refiner, that may hinder increases of 48%, 58% and 5%, respectively, in hardness, tensile strength and elongation. The use of the flux modifier, either isolated or complemented with the grain refiner, is reported to enhance the same properties, however, still lower than the grain refiner approach. Toptan et al. have fabricated A1070/B 4 C p and A6063/B 4 C p through traditional casting and T6 heat treatment. To overcome particlematrix wettability problems, the author added K 2 TiF 6 flux [136]. Results show an improved particle wettability as the flux induces the formation of TiC and TiB 2 layer in the B 4 C interface. The formation of this reaction layer denotes no significant changes for either alloy, as-cast or T6 heat treated sample. Kalaiselvan et al. have reinforced A6063 with 4 to 12 vol.% of B 4 C by stir casting with the use of K 2 TiF 6 flux as a wetting agent [137]. Processing procedure consists of an initial melt of the alloy at 920 • C under an inert Argon atmosphere, followed by a stirring period of 5 min at 300 rpm. Particles preheated at 400 • C and K 2 TiF 6 flux were introduced into the vortex at a 0.8-1.2 g/s feed rate. The authors report an improvement in B 4 C wettability as the flux reacts with the melt inducing the formation of Ti compounds at the particle's surface. Results display a linear improvement of the ultimate tensile strength, microhardness and macrohardness for every particle amount. This is in agreement with other published work [94,129,138,139], as authors use either K 2 TiF 6 flux or Al-Ti-B master alloys to promote the interfacial Ti compounds in B 4 C particles surface. Other fluxes such as Cryolite (Na 3 AlF 6 ) [78,140] and others are employed to improve casting quality and efficiency rather than particle wettability or distribution. The same can be extended to other casting specific additives, as the hexachloroethane degassing tablets [81,[141][142][143] since no conclusion is derived concerning particle-related issues.

Other Methods
Recently, more authors have extended their research to other particles or alternative reinforcement methodologies. Frequently these include multi-sized reinforcement [144][145][146], in which two different magnitude-sized particles of the same compound are incorporated, or hybrid reinforcement [146][147][148][149][150][151], in which two chemically distinct particles are introduced.
Adebisi et al. studied the effect of different sized particles and multi-sized particles on the mechanical and tribological properties of 6061/SiC mp [144]. The methodology includes charging and complete melting of the Al alloy as well as the addition of 1 wt.% Mg. Melt temperature is lowered to a semi-solid state, in which pre-oxidised particles (1200 • C for 2 h) are incorporated. Afterwards, the melt is heated to 800 • C and stirred at 500 rpm for 3 min before pouring into the mould. The authors report an increase in hardness for smaller sized particles, however, best results are presented by the multi-sized particle sample. Superior impact strength is also reported by the multi-sized particle sample followed single-sized particle samples from larger to smaller. The authors attribute multi-sized particle tribological performance to its superior load-bearing capacity. Lakshmikanthan et al. have studied the effect of the incorporation of dual sized SiC particles in the microstructure, mechanical and wear properties of an A357 composite [152]. Particles of 140 ± 10 µm and 30 ± 5 µm size were configured in three different sets at a total 6 wt.% SiC. Melt was initially stirred at 550 rpm and degassed with Hexachloroethane tablets. After lowering stirring speed to 300 rpm, preheating particles (1100 • C for 2 h) were incorporated in the melt and stirred for 15 min until pouring into a steel mould. Composite favouring large-sized particles (4 wt.% 140 µm + 2 wt.% 30 µm SiC) obtained the highest hardness values, while superior tensile properties were reported for the finer-sized particles (2 wt.% 140 µm + 4 wt.% 30 µm SiC). The authors infer that the smaller particles were responsible for grain refinement and dislocation strengthening, while larger sized particles induced a greater load-bearing effect. Wear resistance was found to be correlated best with loadbearing capacity, thus encouraging larger sized particles. Kumar et al. have produced an A356 alloy reinforced with Al 2 O 3 /SiC/Gr through squeeze casting and studied its effect on mechanical properties [153]. In this work, 1 to 5 wt.% of each of particles (preheated at 450 • C) along with 1 wt.% Mg were introduced in the melt at 750 • C, which was then stirred at 250 rpm for 10 min. Afterwards, molten aluminium was poured into the hydraulic plunger with an applied pressure of 100 MPa for a period of 30 s and later the cast samples were subjected to a T6 heat treatment. Results show an increase in hardness and tensile strength up to 3 wt.% individual particle content, after which these properties tend to decrease. The authors report a reduction in elongation with higher reinforcement volume fraction. Ghanaraja [154]. Along with a 5 wt.% Mg wetting agent, particles were introduced and stirred with a coated steel two-stage impeller. The generation of nanosized Al 2 O 3 is suggested to have been generated during the milling stage, while larger particles are originated during melt processing. It is reported that the higher content of smaller particles tends to have higher ductility when compared with larger particles at lower content. Although higher reinforcement content leads to higher ductility, yield and ultimate tensile strength, these gains cease after 3 wt.% given the tendency to form clusters. Gayathri and Elansezhian studied the effect of different CuO np content along with spent Al 2 O 3 catalyst in the microstructure and mechanical properties of a hybrid LM25 alloy-based composite [146]. A 10 vol.% of 20 µm Al 2 O 3 was preheated at 500 • C for 3 h, while 0.5 to 1 vol.% 40 nm CuO were preheated at 750 • C for 6 h. After initial melting at 750 • C, the melt was cooled to 580 • C for the addition of spent Al 2 O 3 and stirred at 350 rpm for 10 min. Afterwards, the mixture was heated to 650 • C and stirred at 500 rpm for 10 min. Results report no agglomeration among the tested samples. It is reported a linear gain in hardness and tensile properties along with the increase in CuO content. The authors propose the Cu phase precipitates in the inter-dendrite region and, thus, enhance the mechanical properties. Elshalakany et al. have studied the effect of wt.% content of multiwalled carbon nanotubes (MWCNT) on an A356 alloy-based composite produced through hybrid rheocasting and squeeze casting [143]. In this work, MWCNT is mixed with Al p to form a composite billet through a low temperature compression for the introduction. A356 alloy was melted at 660 • C, degassed with hexachloroethane and protected with Argon gas. After cooling down to 601 • C, 0.75 wt.% Mg and reinforcements were introduced in the semisolid state and stirred for 1 min at 750 rpm until pouring and squeezing. The authors state that all samples achieve a homogenised distribution without the precipitation of new phases. A 50% and 60% improvement on UTS and YS for optimum content of 1.5 wt.%, while for compressive strength this was verified at 1 wt.% A356/MWCNT, which was related to lack of wettability and agglomeration. In contrast, hardness was improved up to 88 HRB to the maximum content tested (2.5 wt.%).

Stir Casting
With the incorporation of heterogeneous particles in the melt, the fluid's flow and heat transfer characteristics are altered which then influence the distribution and agglomeration of the particles. Karantzalis et al. have reported that even if the reinforcement particles are well distributed and wetted in the melt, the difference in thermal conductivity between the reinforcement and the matrix move the particle's position during solidification [155]. Mousavian et al. have verified in their work that even though a post-cast extrusion has modified the microstructure of the samples through deformation, the distribution of the nano or microparticles did not change [123]. In squeeze casting and similar processes, parameters (squeeze pressure and holding time) hold a greater impact on the materials properties rather than melt or mould temperature. Even so, the authors have stated that melt stirring overlaps its own function as it delays the particle settling before solidification [156]. In order to improve MMC casting typical problems, the authors have procured several methods to successfully disperse and/or de-agglomerate particles in the melt. Among them, the stir casting and ultrasonic melt treatment present themselves as common manufacturing options in the literature [128,142,[157][158][159][160][161][162][163]. However, it is yet a processing path difficult to be summarised by a generalised guideline. Although it is stated that mechanical stirring enhances the particle's wettability, there is also a benefit by decreasing the stirring speed in the introduction as it avoids the projection of low-density particles during high shearing flows [145]. As verified by several authors, the average dispersion of reinforcement particles tends to be more effective and efficient for smaller size particles, which also tends to produce superior mechanical properties [77,92]. Figure 4 summarises the impact of differing stirring speeds and periods of distinct sized particles on <5 vol.% composite relying only on stir casting as reported by several authors. Even though the two bar plots suggest an optimal value for each particle type and size, it should be noted that mechanical properties are still subjected to other treatments whose influence is evidenced by the standard deviation between parameter values. Therefore, appropriate selection of stirring parameters is imperative, and yet subjective, as the different particles size and density as well as pre-heating and melt temperature can interfere in its successful dispersion, unavoidably, their material properties.
squeezing. The authors state that all samples achieve a homogenised distribution without the precipitation of new phases. A 50% and 60% improvement on UTS and YS for optimum content of 1.5 wt.%, while for compressive strength this was verified at 1 wt.% A356/MWCNT, which was related to lack of wettability and agglomeration. In contrast, hardness was improved up to 88 HRB to the maximum content tested (2.5 wt.%).

Stir Casting
With the incorporation of heterogeneous particles in the melt, the fluid's flow and heat transfer characteristics are altered which then influence the distribution and agglomeration of the particles. Karantzalis et al. have reported that even if the reinforcement particles are well distributed and wetted in the melt, the difference in thermal conductivity between the reinforcement and the matrix move the particle's position during solidification [155]. Mousavian et al. have verified in their work that even though a post-cast extrusion has modified the microstructure of the samples through deformation, the distribution of the nano or microparticles did not change [123]. In squeeze casting and similar processes, parameters (squeeze pressure and holding time) hold a greater impact on the materials properties rather than melt or mould temperature. Even so, the authors have stated that melt stirring overlaps its own function as it delays the particle settling before solidification [156]. In order to improve MMC casting typical problems, the authors have procured several methods to successfully disperse and/or de-agglomerate particles in the melt. Among them, the stir casting and ultrasonic melt treatment present themselves as common manufacturing options in the literature [128,142,[157][158][159][160][161][162][163]. However, it is yet a processing path difficult to be summarised by a generalised guideline. Although it is stated that mechanical stirring enhances the particle's wettability, there is also a benefit by decreasing the stirring speed in the introduction as it avoids the projection of lowdensity particles during high shearing flows [145]. As verified by several authors, the average dispersion of reinforcement particles tends to be more effective and efficient for smaller size particles, which also tends to produce superior mechanical properties [77,92]. Figure 4 summarises the impact of differing stirring speeds and periods of distinct sized particles on <5 vol.% composite relying only on stir casting as reported by several authors. Even though the two bar plots suggest an optimal value for each particle type and size, it should be noted that mechanical properties are still subjected to other treatments whose influence is evidenced by the standard deviation between parameter values. Therefore, appropriate selection of stirring parameters is imperative, and yet subjective, as the different particles size and density as well as pre-heating and melt temperature can interfere in its successful dispersion, unavoidably, their material properties.  [10,75,76,78,79,82,95,119,123,127,164,165].  [166]. Nanoparticles are ball milled with Al or Cu powders in a 1:1 mass ratio and later wrapped in aluminium foil. Melt is heated to 850 • C and stirred using a graphite system at 450 rpm for 4, 8, 12, and 16 min. After pouring into a steel mould, samples were subjected a T6 heat treatment. Micrographs depict some cases of agglomeration, particularly in the lower section of the mould. The authors report that porosity levels increased with the volume fraction of reinforcement particles and for longer periods of stirring on all samples. Even though, optimal hardness and compressive strength, as well as minimum wear rate, were achieved for the (Al 2 O 3np /Cu) p sample after stirring of 4 min. Despite the effect of porosity, the authors attribute the relative mechanical performance to the strengthening factor and distribution of the Cu powders. Shirvanimoghaddam et al. have studied the mechanical and physical properties of an A356/B 4 C composites fabricated by stir casting with different volume fractions and melt temperatures. Further, the authors use their results to train neural networks for two different models to predict hardness and tensile strength [78]. The manufacturing procedure consisted in an initial melting of A356 ingot and introduction of 10 g Cryolite (Na 3 AlF 6 ), followed by a stirring with a four-bladed 45 • angled graphite impeller at 350-400 rpm at a melt temperature of 800 • C or 1000 • C. Particles were packed in aluminium foil and introduced into the melt in the stirring stage for 4 to 12 min with a 5 to 15 vol.% B 4 C content. The authors report an increase in wettability and mechanical properties for higher melt temperatures. Additionally, it is stated that a shorter stirring time will cause a non-uniform distribution, while for longer periods, particles tend to be projected and agglomerate in the sides of the crucible. Comparison between the experimental data and the data from the neural network revealed accurate and low errors. Lakshmikanthan et al. have studied the effects of artificial ageing temperature and particle size ratio on a double particle size SiC mp reinforced A357 composite [145]. This procedure consisted in a two-step stirring: (i) first in the raw melt was stirred at 550-600 rpm; and (ii) and the second consisted in the addition of particles at stirring speed of 300 rpm. Cast samples were finally subjected to a T6 heat treatment. Although most samples report a good distribution with a small cases of nanoparticle clustering, best distribution was achieved with 2 wt.% coarse and 4 wt.% fine particles. The authors advise that a careful selection of stirring speed needs to be taken with caution for smaller particles as it transports in relative higher momentum in the direction of the flow. It is further shown that in high shear melt flows, smaller particle velocity is restricted by larger particles, eventually settling in their vicinity. Chen et al. have studied the particle trajectory of SiC particles in the A356 melt during casting through numerical analysis and experimental data, in which filling was observed by X-ray radiography and a highspeed camera [142]. During pouring, it is reported that the particle fractions decrease for longer fluid trajectories. A relative lower particle volume fraction was reported at the surface of the cast component, to which authors have hypothesised this to be promoted by the primary α-Al pushing particles to its inner region as a result of its solidification mechanism. In a laminar flow regime, particles tend move to near-wall regions, while for turbulent flows particle trajectory is mostly dictated by eddy flow and wall collision. Sahu and Sahu have studied the flow behaviour of the melt stirring B 4 C and Fly-ash particles in an Al 7075 melt through numerical simulation and optimised its parameters for particle distribution and minimise stagnant/dead zones [138]. The authors suggested a blade angle of 30 • , impeller with a diameter that corresponds to the crucible radius and stirring speed 550 rpm for optimal stir casting under the studied conditions. Through Grey relational grade, it is reported that the mitigation of dead zones is influenced by impeller size, blade angle and stirring speed, in this order. Su et al. have studied the effect of melt flow behaviour on the distribution of the particles in the stir casting process through numeric simulation [167]. Resorting to the finite element method, the authors studied the impact of stirring-related parameters such as the blade angle, rotating speed, the diameter of the impeller, and the stirrer geometry on the effective flow parameter. The authors suggested using a multistage stirrer and high stirring speeds to obtain optimal processing conditions, showing that stagnant regions and dead zones in the crucible could be minimised. Considering that a high angle blades induce shearing flow, lower angles are preferable for particle incorporation. The author state that an increase of impeller diameter/crucible diameter ratio from 0.25 to 0.55 has a significant impact on particle distribution. However, for higher ratio, axial flows increase which then leads to excessive turbulence. Numerical results were later validated empirically for stirring speed, being concluded that the best distribution was achieved for 1000 rpm promoting particle cluster deagglomeration. Li et al. has studied the effect of the dynamically liquid-stirring periods on the resulting microstructures, and particles dispersion [94]. The authors show that the shearing and turbulent melt flow is capable of deagglomeration and dispersing B 4 C particles to sufficient state for 20 min, but a fully randomised dispersion was observed at 55 min. It is stated that for B 4 C, sedimentation and buoyancy effects can be ignored given the close density between the particle and the aluminium melt.
Although the method is not common, some authors have adopted a dynamic melt temperature cycle in their processing, often labelled as two-stage stir casting. The methodology can be described as an initial melting, followed by a simultaneous cooling and stirring stage at a lower temperature (liquidus [77,124,160] or semi-solid [129,142]). Afterwards, the melt was superheated to the designated pouring temperature (above liquidus) and held for a specific amount of time before pouring. As the particles immersion time and melt temperature changes between publications, it can be difficult to quantify the efficiency of the method relative to others even though its efficacy is reported and verified. Wu et al. processed A356/SiC np by stir casting and ultrasonic treatment (UST) [160]. SiC particles were oxidised at 850 • C for 2 h and mixed with aluminium powders through high energy ball milling. After the alloy melts, it was cooled down to 680 • C, at which the preheated particles (250 • C) were introduced into the vortex. Stirring proceeded at 250 rpm for 10 min, followed by a melt superheating to 720 • C and holding for 30 min. Subsequently, UST (2.8 kW at 20 kHz) was employed for a duration of 3 min. Composite melt was poured into a preheated steel mould (250 • C) and squeeze cast at 60 MPa. It is reported that no significant agglomeration was observed, which indicates the benefits of the UST. The authors show that ball milling SiC with the carrier agent allowed a good particle-matrix wettability. Optimal mechanical properties were with 2 wt.% SiC np and no drawback tendency reported. Hu et al. studied the effects of Ti containing powders and fluxes as well as stirring stage temperature in the fabrication of 10 wt.% A356/B 4 C [129]. Liquid processing path consisted of an initial stirring at 300 rpm at melt temperature 780 • C along with the introduction of 1 wt.% Ti powders wrapped in Al foil. Afterwards, B 4 C particles are introduced into the vortex and stirred at 450 rpm for 15 min before the cast samples are quenched in cold water. The semi-solid stirring path consisted of an initial melt of the alloys (A356 and Al-15Ti) at 700 • C, followed by cooling to a semi-solid state (590-610 • C). Subsequently, B 4 C were introduced into the vortex stirred at 350 rpm. Melt was then heated to 670 • C and stirred at 450 rpm for 150 before quenching in cold water. It is reported that the introduction of Ti-containing compounds benefit the distribution as it promotes wetting between the melt and the B 4 C, and it later nucleates a layer of TiB 2 crystals. The authors justify that the semi-solid stirring enhanced the distribution through the imposed shearing on agglomerated particles, thus inducing deagglomeration and homogenous distribution.

Other Forms of Stirring
Although stir casting is commonly performed through impellers as it is the most accessible option, other authors have employed alternative methods of melt agitation [64].
Kaisarlis et al. have produced coal fly ash composites from an A380 alloy by gravity casting [168]. The authors have developed an apparatus based on a high-energy ring-mill pulverizer to produce ultra-fine particle sizes (<1 µm). Additionally, an oscillating microgrid melt mixer implemented in a piston-rod assembly is used to impose shearing forces onto the melt with de-agglomeration purposes. With the increase in coal fly ash content up to 10 wt.%, it is reported an enhancement in tensile strength, hardness and wear resistance along with a decrease in ductility and toughness. Lal et al. have characterised the microstructure, thermal and mechanical properties of a 7075/(Al 2 O 3 /SiC) mp composite [169]. Alloys were charged into the crucible and melted at 715 • C, being degassed for 10 min with argon. Different particles compositions (5-20 wt.%) were mixed in a 1:1 ratio and preheated at 900 • C. Afterwards, particles were introduced in the melt through a formed vortex and continuously electromagnetically stirred for 6 min. Cast samples were to cool in the air. The authors report uniform and homogeneous distribution in all samples with no formation of Al 4 C 3 brittle phase. Differential thermal analysis shows a negligible material loss in the composite. Optimised tensile properties were reported for 15 wt.%, while superior hardness was achieved at 20 wt.%. Elongation and impact energy tends to decrease as the reinforcement content increases. Golak and Dyzia developed a form of centrifugal casting with an interior channel of a concrete mould subjected to electromagnetic stirring [170]. Induction coils are positioned inside the mould (7 mm from the wall). An AlSi 12 CuMg is incorporated with SiC mp through stir casting at 720 • C and poured to a 220 • C preheated mould. Once the mould is filled, the inductor is activated for 7 s, after which the casting is allowed to cool. From image analysis, the authors report a constant volumetric mean and standard deviation of reinforcement particles up to 4 mm from the inner wall. Yang et al. have used intensive melt shearing to study the dispersive effect on different sized particles on A356/SiC p composite and fluidity tests were performed to attest liquid flow and solidification behaviour [79]. MMC processing consisted of initial heating to 800 • C to melt the alloy, after which the melt temperature was lowered to 600 • C in which the particles were introduced. The melt was then stirred by an impeller at 650 rpm for 5 min followed by a second stirring stage using two distinct approaches: (i) impeller stirring at 650 rpm for 5 min at 730 • C; or (ii) intensive metal shearing at 5000 rpm for 5 min at 730 • C. The authors report that the melt shearing processing improves both mechanical and tribological properties when compared to the second step impeller stirring. These gains seem more noticeable for particles size of 4.5 µm and 9.5 µm rather than 12 µm. It is hypothesised that these volume fractions promote a reduction in melt viscosity.

Ultrasonic Melt Treatment
Ultrasonic melt treatment is a current popular option for its ability to promote grain refinement, equiaxed grains as well as melt degassing [57,171]. The acoustic pressure waves have been reported to be able of fragmenting dendrites and induce a vortex-free flow. The use of the ultrasonic treatment in manufacturing has been reported to transfer its enhanced tensile strength and elongation onto metal matrix composites along with improving the particle's wettability [160,172,173].
Jia et al. have experimented and modelled the effect of ultrasonic cavitation in the production of 6061 MMCs reinforced with 1 wt.% Al 2 O 3 or SiC nanoparticles. Modelling analysis includes a fluid flow and particle distribution during particle injection and the ultrasonic probe is positioned at the bottom or the top of the crucible [174]. Results report a residual change in UTS but significant increase in elongation, which was justified by observation of agglomeration and microporosity. Furthermore, it is suggested that the presence of nanoparticles induce cross slip in non-basal slip planes that may have increased the sample's ductility. The CFD modelling results show no significant improvement in either setup, to which should remain true as long as the flow remains strong enough to disperse the nanoparticles. Jia et al. have experimented and modelled the effect in ultrasonic cavitation in the dispersion of Al 2 O 3 or SiC particles in an A356 melt [83]. The authors report that both MMC samples promote grain refinement and dendrite fragmentation under ultrasonic treatment. Some minor agglomeration is reported in the numeric and experimental results which could be originated during the ultrasonic processing or solidification. Vorozhtsov et al. have studied the effect of ScF 3np on the microstructure and mechanical properties of A356 alloy-based MMC [175]. The introduced reinforcement consists of 20 wt.% ScF 3np mixed with Al p and later wrapped in aluminium foil. Melt procedure can be summarised with UST degassing for 1 min at 760 • C, followed by particle introduction and UST for 2 min at 720-740 • C, after which, the melt was poured at 710 • C into a metallic mould. Compared to the as-cast A356 alloy, results report an increase in hardness, UTS, YS and elongation for 0.2 wt.% but slightly less for 1 wt.% ScF 3np . The authors state that the clusters formed along the grain boundaries may have affected the properties of the MMC. Idrisi and Mourad have compared the effects of conventional stir casting and the ultrasonic-assisted stir casting techniques on AA5083/SiC MMC [173]. Re-sults show that deployment of the ultrasonic probe enabled the enhancement of composite tensile and compressive strength, hardness as well as the reduction in porosity, particularly for higher weight percentage. Although some clusters were detected, the authors explain that, through the acoustic transient cavitation of microbubbles, clusters are shattered and their gas layer removed, thus, improving the particle/matrix wettability.

Hybrid Processing Stir
Authors have stated that UST coupled with mechanical stirring improves the wettability and distribution of the particles reinforcements [68].
Aybarc et al. have studied the effects of a hybrid ultrasonic-assisted stir-casting of Al 2 O 3np and Al 2 O 3mp in an A356 melt [68]. Although reinforcement addition shows an improvement in mechanical properties, the use of lower content of nanoparticles reports the optimal enhancement with minimum porosity percentage. Even so, the authors state that particle size and its higher thermal conductivity play a greater role in grain refinement and porosity mitigation. Yuan et al. have studied the effect of different ultrasonic treatment periods in A356/SiC np MMCs processed through squeeze casting [77]. After melting at 720 • C stirring was performed and particles were introduced in the vortex. The melt was cooled to 680 • C and continued to be stirred at 180 rpm for 15 min. Afterwards, the melt was superheated to 720 • C, held for 30 min and degassed for 10 min with argon gas. Ultrasonic melt treatment (2.8 kW at 20 kHz) was employed for different time periods (0.5-5 min) before being poured into a preheated steel die at 250 • C and squeeze cast. It was reported that its maximum refining effect is achieved at 2-3 min, after which the gains fade for longer treatment time. Although by itself ultrasonic treatment does not significantly reduce secondary dendritic arm spacing, the authors state that the grain refinement effect of the nanoparticles is only noticeable when accompanied by ultrasonic-assisted process. Despite the results, the effects of the UST combined with SiC np on the microstructure is not clear as particle distribution are difficult to observe, even though agglomeration is reported. Yuan et al. have produced (SiC/Al) p reinforced on A356 alloy by squeeze casting [124]. After a melting of the alloy at 720 • C, the melt was cooled to 680 • C and particles were introduced, followed by a stirring period of 10 min at 180 rpm. The melt temperature was superheated to 720 • C for a holding stage of 30 min, after which the ultrasonic treatment was applied for 3 min and then poured into the mould. It is reported that at 0.5 wt.% the reinforcement effect was negligible given the insufficient content for a uniform distribution. The authors state that for 1 and 2 wt.% residual agglomeration was observed. Li et al. have studied the effect of the period of ultrasonic melt treatment in a SiC p in Al-5Cu composite [176]. Composite granules with a size of 1-2 mm were produced by milling 40 nm SiC p (6 wt.%) and 30 µm Al p . Materials were melted at 750 • C, stirred for 10 min at 120 rpm under a protective atmosphere and later degassed by Argon for 10 min. Afterwards, the melt was poured to a metal cup in a holding furnace for UST at different time periods, after which, it was poured into the mould and squeezed with 50 MPa pressure. The authors have verified that a 1 min treatment was sufficient for the de-agglomeration into smaller aggregates, however, particles are uniformly dispersed for a 5 min treatment period. The UST refinement effect on the α-Al and Al 2 Cu showed no adverse effect for the tested time periods. Furthermore, it is stated that the UST degassing effect enabled the decrease in porosity from 4% (as-cast MMC) to 0.53% (UST 5 min). Results report a 7.6%, 6.8% and 29% increase in UTS, YS and elongation for 5 min treatment compared to the as-cast MMC.

Recommendations
Considering the integrated analysis of this review and the results from these publications that are detailed in the discussion section, the recommended process parameters for stir casting or UST casting for the manufacturing of Al-Si based MMCs are provided in Table 3.

Particle Pre-Processing and Introduction
As seen previously in Figure 3, a higher particle temperature can lead to an improvement in the particle wettability. However, prolonged exposure to high temperatures promotes an oxidising atmosphere and can lead to the formations of new phases. For B 4 C, these oxide films are brittle in nature and tend to reduce the mechanical properties. However, for Al 2 O 3 and SiC, these oxide films can improve their wettability. Another recommended method of improving the interface wettability consists of the coating with a differing element such as Cu and Ni, as these are reported to improve the incorporation and inhibit agglomeration. The addition of the particles to the melt is a crucial stage in the MMC processing as particles can trap oxides formed at the melt surface. To prevent these issues, the authors have introduced the particles packed in aluminium foil into the vortex or through injection assisted by an inert gas. Although both methods are effective, packing in aluminium foil can induce an early agglomeration, particularly for nanoparticles [124]. In both a melt additive and ball milling powder approaches, the use of 1-2 wt.% Mg contributes to the reduction of interfacial energy [82,95,177]. Higher Mg content (>2 wt.%) will lead to the formation and coarsening of Mg 2 Si and Mg 5 Al 8 as well as the increase in porosity [178]. Additionally, its positive effect on melt fluidity can also be considered to improve the filling capacity.

Processing Temperature
Similar to particle preheating, the melting temperature can be intimately related to the contact angle of the reinforcement. Even though higher temperatures can improve wettability, a melt temperature higher than 900 • C will result in detrimental effects as an aluminium affinity for oxygen is prone to form oxides in the casting. Considering the thermal properties of the ceramic particles, the heat loss from the particle will be significantly lower than the melt, leading to longer solidification times and relatively inferior tensile properties. Considering their individual wettability, a melt temperature of 750 • C is recommended for Al 2 O 3 and SiC, along with 850 • C for B 4 C particle. While using steel die, it is recommended to pre-heat it between 250 • C and 300 • C, while higher temperatures are not reported as being beneficial to enhance mechanical properties or wettability [179,180].

Melt Stirring
Considering a particle behaviour in the melt, longer stirring periods are often desirable as it increases the chance of a successful interfacial wetting [173]. However, as discussed in previous chapters, stir casting parameters are subject to change as its ability to dispersing particles trajectory, without projecting to the crucible walls or allowing them to settle, is found to be highly dependable on its individual particle density and size. Independently of the size, B 4 C does not require a high stirring speed as its density (2.52 g/cm 3 ) is close to that of most aluminium alloys and can endure longer immersion periods without any negative effect. Therefore, a stirring period of 10-15 min at 300-350 rpm is recommended. Although longer stirring periods would improve the contact angle between the melt and the reinforcement, publication regarding SiC composite seem to report higher tensile properties when a short timed high-speed stirring methodology is adopted. Considering the several publications, a short stirring period of 5-10 min is recommended at 650 rpm or higher. Similar to the previous, Al 2 O 3 does not require an extended stirring period (10 min) as prolonged immersion times led to an increase in gas accumulation and porosity which verified that a stirring speed of 450 rpm is optimal to achieving a homogenous distribution. Regarding the stirring apparatus characteristics, very few publications were found, usually addressing these issues empirically or through numeric simulation. Conclusions converge to the idea that a 3-4 bladed multi-stage impeller with a blade/crucible diameter of 0.5 is indicated as optimal to mitigate stagnant zones and thus avoid particles settlement. Likewise, a 30 • -45 • blade angle is recommended as it creates sufficient shearing force and upward fluid motion that ensures the suspension of the reinforcement particles.

Ultrasonic Melt Treatment
Additionally, ultrasonic melt treatment can be recommended as a complementing method for its grain refining as well as the wetting effect [173], distribution and deagglomeration impact in the MMC processing. When relying fully on UST, it is suggested to introduce particles directly under the acoustic horn, the area of effect in order to maximise its de-agglomeration effects [173]. Despite its benefits, it is advised to complement the technique with mechanical stirring as the achievable dispersion tends to be more homogenous [68]. Although UST is generally a quick procedure (<1 min), a treatment period between 2 and 3 min is recommended before the gains per treatment start decreasing [77,124].

Conclusions
A wide range of methods and techniques were presented and discussed in the context of ceramic particle wettability, distribution and de-agglomeration. As composite properties are dependent on particle type and volume fraction, current systematic focused on Al 2 O 3 , B 4 C and SiC. Likewise, parameter recommendation is formulated for these particles.
Current MMC research tends to be focused on material property enhancement. However, it is suggested that nanoparticles may have a lower reinforcement content threshold as the presence of a large number of nanoparticles tends to promote agglomeration and induce porosity. In contrast, micro-reinforcement has a higher chance of success at higher volume fractions, allowing both particle incorporation and mechanical properties [90,181]. For each particle type and size, some authors have suggested there is an optimum amount for each particle and when this threshold is overlapped the volume of porosity and agglomerated clusters cannot be ignored [55,84,182].
Among the other MMC-manufacturing techniques, stir-casting is restrained in its reinforcement volume fraction, and the course of action for both nano-and micro-particles is to enhance mechanical property/content ratio efficiency. This comprises the mitigation of porosity levels in the final MMC cast samples, and extend the optimum reinforcement to higher volume fractions. As discussed in previous chapters, the fabrication of hybrid composites and the incorporation of multiple sized particles is the most promising step. However, there is no current sufficient literature to show the influence on mechanical properties and its correlation with optimal processing parameters.
Additionally, a crucial parameter that is not sufficiently addressed in the literature is the analysis of the particle-matrix reinforcement ratio (e.g., combining wettability factor [89] and the reinforcement incorporation factor [90]), which quantifies the volume fraction of particles that effectively act as reinforcements in the final cast components. Funding: This work was supported by PTDC/EMEEME/30967/2017 and NORTE-0145-FEDER-030967, co-financed by the European Regional Development Fund (ERDF), through the Operational Programme for Competitiveness and Internationalization (COMPETE 2020), under Portugal 2020, and by the Fundação para a Ciência e a Tecnologia-FCT I.P. national funds. Also, this work was supported by Portuguese FCT, under the reference project UIDB/04436/2020, Stimulus of Scientific Employment Application CEECIND/03991/2017, research doctoral Grant 2020.08564.BD.