Application of cast Al–Si alloys in internal combustion engine components

Engine components and requirements

The engine block and cylinder head, which are shown in Fig. 7, are the two major components of an engine; both components have historically been manufactured from cast iron owing to its inherent high-temperature strength. Nevertheless, cast iron is a dense material (∼7·5 g cm⁻³) and the engine is the single heaviest component within the powertrain group (∼14% of total vehicle mass). ⁴⁹ About 3–4% of the total mass of an average vehicle is generally assigned to the engine block. The improved specifications and legislations for fuel economy and emissions oblige car manufacturers to make a significant weight reduction in their products. It has been reported that each 100 kg in weight reduction could contribute to ∼0·5 L of petrol being saved per 100 km driven. ^49–52 As illustrated in Fig. 8, weight reduction of a vehicle by a certain amount could result in significant improvement in fuel economy. ^53,54 Social impetus, for instance the US Partnership for a New Generation of Vehicles (PNGV) programme, demands car manufacturers to produce vehicles having a fuel consumption of lower than 1 L/30 km. ⁵⁵

OPEN IN VIEWER

7.

a Cross-section of a cylinder head and b engine block (reprinted with permission from Taylor & Francis) ⁴⁰

OPEN IN VIEWER

8.

The relation between vehicle mass and fuel consumption ^53,54

Using materials with higher strength and stiffness, such as compacted graphite iron (CGI) instead of grey cast iron, contributes to increase in power and decrease in size of an engine by reducing the main bearing thickness (see Table 1). ⁵⁰ Another alternative is to replace cast iron with lightweight materials (e.g. aluminium and magnesium alloys). Owing to the considerable difference in the density between cast iron (∼7·5 g cm⁻³), aluminium (∼2·7 g cm⁻³) and magnesium (∼1·74 g cm⁻³) alloys, the substitution of cast iron by one of these alloys could make a significant weight reduction.

Magnesium alloys

Magnesium is ∼75% and ∼33% lighter than iron and aluminium, respectively. It has attracted great interest in the automotive industry. However, the specific stiffness of aluminium and iron has been reported to be slightly (∼0·69% and 3·752%, respectively) higher than that of Mg; but the specific strength of Mg is significantly greater than that of aluminium and iron (14·075% and 67·716% for aluminium and iron, respectively). ^53,56

The regular commercial cast Mg alloys (e.g. AZ91 and AM50), which are widely used in the automotive industry, suffer from poor creep resistance. ⁵⁷ The creep resistance of the magnesium alloys (e.g. AM50: Mg–5Al–0·3Mn–0·2Zn (approximate wt-%1)) has reported to be ∼15% less than that of aluminium alloys (e.g. A380: Al–8·5Si–3·5Cu–3Zn (approximate wt-%)) at 293 K, and ∼65% less at 403 K. ⁵⁸ Therefore, new Mg alloys (e.g. MRI 201, MRI 230) have been developed to improve the creep resistance and high-temperature strength. These alloys could compete with the commercial Al alloys (e.g. A380 and A319) in terms of creep resistance and high-temperature strength. ^59–61

Despite these advantages, application of magnesium alloys in the automotive industry has been very limited: the average application of Al alloys has been reported to be over 100 kg per car, while that of Mg alloys has been reported as ∼6 kg. ⁶² The higher total cost of Mg alloys is one of the major reasons for impeding their widespread application in the automotive industry. ^62–64 It is worth noting that the price of magnesium has been considerably reduced in the last few years. ⁵⁶ Lower thermal conductivity and higher thermal expansion are other disadvantages of Mg alloys compared with Al alloys. ⁵⁷

Aluminium alloys

In the late 1970s, the generation of aluminium engine blocks was introduced to be used in gasoline engines. However, because of technical requirements, application of aluminium alloys was very limited in diesel engines until the mid-1990s. Nowadays, blocks for gasoline engines are generally cast in aluminium alloys; and the use of aluminium in diesel engines is continuing to increase. Also, most cylinder heads are cast in aluminium alloys.

Substitution of cast iron by aluminium in engine blocks could result in a weight reduction of 15–35 kg. ⁶⁵ Inline cylinder blocks made in aluminium are noticeably lighter than corresponding cylinder blocks produced with CGI. For an engine weighing 35 kg in CGI, the weight of the inline cylinder block should be 28 kg using an aluminium alloy. ⁵⁰ However, if the design of the engine is adapted to CGI (V-8 instead of inline), a marginal weight saving can be made with CGI.

A comparison of some important properties of Al alloys, Mg alloy, grey cast iron (GJL-250) and CGI (CGV-400) is shown in Fig. 9. As shown in this figure, another advantage of aluminium alloys compared to cast iron is their excellent thermal conductivity, which accelerates cooling of engine. In spite of all these advantages, softening of the commercial foundry aluminium alloys at service temperature restricts their application in engine components. For instance, as shown in Fig. 10, some studies from AVL reported that the application of aluminium engine blocks must be restricted for those passenger car engines with 150 bar peak firing pressure. ^66,67

OPEN IN VIEWER

9.

Material properties of compacted graphite iron (CGI-400), grey cast iron (GJL-250), Al-A390, AlSi9Cu and Mg-MRI 230D. ^50,59,66

OPEN IN VIEWER

10.

Peak firing pressure limits for various materials in diesel engine cylinder block ^43,66

Table 2 presents the chemical compositions of the most common aluminium alloys used in engine applications. Alloys 356+Cu and 319 have been extensively studied for use in engine components, in particular in cylinder heads. For instance, they were studied by BMW, ^13,68 VAW Aluminium AG, ⁶⁹ Ford Motor Company ⁷⁰ and General Motors. ^71,72 Considering their importance, special emphasis will therefore be given to the 356- and 319-type alloys in the following sections. Hypereutectic Al–Si alloys could be another alternative for cast iron in production of engine blocks. Jorstad, ⁷³ who is often credited as the pioneer of 390 hypereutectic Al–Si alloys, has thoroughly reviewed the application of these alloys in the manufacture of engine block from inception until now. Mercedes, BMW, Porsche, Audi and Volkswagen are some of the companies which have used hypereutectic Al–Si alloys in the production of engine blocks.

Table 3 presents some major mechanical and physical properties of three Al–Si (319-, 356- and 390-type) alloys. The symbols F (as cast, without heat treatment), T₄ (quenched and naturally aged), T₅ (artificially aged after casting), T₆ (quenched and artificially aged for maximal strength) and T₇ (quenched and overaged), which represent the most common heat treatment condition of Al–Si alloys, have been designated by the Aluminium Association of the USA. ⁷⁶

The 356-type aluminium alloys present good combinations of strength and ductility, but their strength reduces rapidly above 473 K (200°C). The 319-type aluminium alloys present relatively higher yield and creep strength at elevated temperatures (∼523 K), although prolonged exposure at such temperatures could result in softening. Therefore, to achieve the increasingly exacting requirements of engine components (higher pressure and temperature) without new material inventions, the existing capabilities of Al–Si hypoeutectic alloys have to be improved by optimisation of either production process (e.g. casting and heat treatment) or chemical composition.

The rest of this review is dedicated to the characteristics of Al–Si based alloys and how their composition, their processing and their final microstructure can improve their performance in engine applications. Characteristics of these alloys are first reviewed essentially when used in cylinder heads (Sections ‘Description of Al–Si based alloys’, ‘Solidification sequence in 356 and 319 Al alloys’, ‘Effect of microstructural features on TMF strength’, ‘Strengthening of cast aluminium alloys’, and ‘Recent developments in Al–Si alloys and applications in engine components’). The topics that are more specifically addressed are solidification microstructures, TMF, strengthening mechanisms, dispersed phases, heat treating and recent developments of Al–Si alloys. The major characteristics required for cylinder head materials are TMF strength, low density and high thermal conductivity. For cylinder blocks, however, besides these characteristics, appropriate friction coefficient and wear resistance are the other two major required characteristics. Hypereutectic Al–Si alloys (e.g. A390) could be an interesting alternative for cast iron engine blocks, but they suffer from cost issues and troubles in the production process (e.g. casting and honing processes). Hypoeutectic Al–Si alloy engine blocks are subject to wear; therefore, the cylinder bore of hypoeutectic Al–Si alloy cylinder blocks is required to be protected by suitable materials. Thermal spray coating, electroplating and reinforcement by suitable particles/fibres are the main solutions to fortify the cylinder bore. The requirements and potential alternative materials in manufacturing cylinder blocks are reviewed in the ‘Characteristics of the engine block’ section.

The binary Al–Si system

The phase diagram of the Al–Si system is illustrated in Fig. 11. There is an eutectic reaction at 850·75 K and 12·6 wt-% silicon, where the liquid phase is in equilibrium with the α-Al solid solution phase and nearly pure Si (L→α-Al+Si). ^78,79 The maximum solubility of silicon in aluminium is ∼1·5 at-% at the eutectic temperature and decreases down to ∼0·05 at-% at 573 K. Generally, the morphology of the eutectic microconstituent tends to be fibrous if the volume fraction of the minor phase is less than 25%. However, in Al–Si binary alloys, the typical Al–Si eutectic morphology is usually lamellar. This could be ascribed to the low interfacial energy between Al and Si and the strong growth anisotropy of silicon. ⁷⁹

OPEN IN VIEWER

11.

The equilibrium phases diagram of the Al–Si alloy system ^78,79

The morphology of the eutectic silicon particles (i.e. particle size and shape) can appreciably affect the mechanical properties of Al–Si alloys. The coarse lamellar silicon particles, which appear under normal solidification conditions, may act as stress concentration sites and crack propagation paths. ^72,80,81 This negative effect can be alleviated by imposing higher solidification rates, ^82,83 carrying out solution heat treatment ^26,84 or by alloying with certain elements (e.g. Sr, Na, etc.), which can change the morphology of Si particles from plate-like form to fine fibrous form. ^85,86 During the solution heat treatment, the unmodified Si particles undergo: (a) necking at several places along the length of the Si particles resulting in their fragmentation, (b) gradual spheroidisation of the fragments and (c) coarsening by the Ostwald ripening process. ^26,84

There are numerous elements which can modify the Al–Si eutectic microstructure, such as Sr, ^85,87 Na, ⁸⁸ Ca, ^89,90 Sb, ⁹¹ Sc ^92,93 and several rare earth metals. ⁹⁴ It was proposed that the modifier agent is adsorbed at the silicon/liquid interface and results in the growth of twins and branching of silicon particles. ^86,95,96 Such modifications could reduce the solution treatment time and improve the overall mechanical properties. ^37,97 Nevertheless, some studies ^98,99 have shown that the addition of the modifier elements is often associated with increased porosity. Gruzleski and Closset ¹⁰⁰ and Lados et al. ¹⁰¹ stated that chemical modification by Sb and Sr did not have a considerable impact on fatigue lifetime of AlSiMg alloys; meanwhile Gundlach et al. ²⁶ reported the beneficial effect of eutectic Si modification on thermal fatigue resistance. Therefore, an optimum content of the modifier agent is required to yield an acceptable level of modification without affecting the porosity level. The optimum content can be varied depending on the constituents of each alloy. For instance, the modifying effect of Sr can be somewhat nullified by the presence of other elements, namely P, Bi, Sb ¹⁰² and Mg. ^72,102 For more details on the modification of Al–Si casting alloys refer to various publications. ^83,102,103

Silicon significantly improves castability (fluidity, metal-feeding) ^104,105 and wear resistance ¹⁰⁶ and contributes to reduce the density and the coefficient of thermal expansion of aluminium alloys. ¹⁰⁴ In addition, dissolution of Si in α-Al matrix (e.g. ∼0·7 wt-% at 773 K) can significantly improve the age hardenability of AlSiCuMg alloy by combining with Mg. ¹⁰⁷

Influence of iron as impurity

Al–Si binary alloys, even prepared from pure materials (∼99·99%), can contain more than 50 ppm of iron. The presence of iron can considerably affect the solidification process of Al–Si alloys. ⁷⁸ Iron, as the most common impurity in Al–Si alloys, strongly reduces the fluidity and the overall mechanical properties through the formation of brittle intermetallic phases. Primary Al–Si alloys typically contain between 0·05 and 0·3 wt-% Fe; but, in secondary Al–Si alloys, it can reach up to 1 wt-%. Economically, there is no known way to further reduce Fe from primary Al–Si alloys. Owing to a relatively high solubility of Fe in liquid Al, it can readily enter into the melt from unprotected steel tools, furnace equipment and addition of low-purity alloying materials. ¹⁰⁸ The amount of Fe exceeding the solid solubility limit appears in the form of iron-bearing intermetallic phases such as β-AlSiFe, α-AlSiFe and π-Al₈FeMg₃Si₆. The α-AlSiFe phase, which appears in the form of Chinese script particles, has the composition of Al₈Fe₂Si (∼31·6% Fe, ∼7·8% Si). The stoichiometry of the β-Al₅FeSi phase is Al₅FeSi (∼25·6% Fe, ∼12·8% Si), with a probable range of 25–30% Fe and 12–15% Si. The β-Al₅FeSi phase has a platelet morphology (in three dimensions), which appears as a needle in micrographs. ^109,110

Many studies ^108,111,112 found that as Fe levels increase, the ductility and tensile strength of Al–Si alloys strongly decrease; however, the yield strength remains in general almost unaffected by iron. The iron-bearing compounds are much more easily fractured under tensile load compared to the Al matrix or the modified silicon particles. Their detrimental effect is directly proportional to the morphology, size and volume fraction. The platelet morphology of β-phase explains why it is the most deleterious intermetallic phase in cast Al–Si alloys. ^113,114

The size and density of iron-bearing compounds (particularly β-phase) increase with iron content. Moreover, intermetallic phases that can form prior to (or with) the solidification of the aluminium dendrite network (pre-dendritic particles) are much larger than those that form during or after the period of Al–Si eutectic solidification. ¹¹⁵ More available time for growth at a slower solidification rate also leads to enlarged intermetallic particles. ¹⁰⁸ Furthermore, it has been reported that the amount and size of porosity in the microstructure are strongly enhanced by increasing Fe content. This behaviour is mainly related to the increased amount of β-phase, since it promotes shrinkage porosity during solidification by physically blocking the metal feeding, as shown in Fig. 12.

OPEN IN VIEWER

12.

The role of Al₅FeSi in the formation of shrinkage porosity (reprinted with permission from Taylor and France) ¹²⁴

The β-platelets are much more susceptible to crack linkage and fracture than the α-iron Chinese script particles, so the formation of the α-iron phase instead of the β-phase can be less detrimental to mechanical properties owing to its compact morphology. According to Mondolfo, ¹⁰⁹ low Mn and Cr concentration and a low cooling rate (∼0·8 K s⁻¹) are the main factors that favour the crystallisation of β-phase. Hence, chemical modification (by Mn, Cr and Ni addition), high solidification rate ^{114,116–118} and superheating of the melt ¹¹⁹ contribute to the formation of the α-iron phase. The amount of Mn needed to convert all of the β phase is not yet well known. Several researchers ^120,121 reported that an Mn/Fe ratio of 0·5 seems to be sufficient for complete substitution of Al₅FeSi by α-Al₁₅(Fe,Mn)₃Si₂ phase. However, other researchers ^108,122 stated that even at these levels of Mn addition some β-phase could still form. It should be noted that an undesired amount of Mn in AlSiCu/Mg alloy could lead to the precipitation of Al–Cu–Mn particles (T-Al₂₀Cu₂Mn₃ phase ¹²³ ) during solution treatment, which in turn decreases the Cu content in α-Al matrix. ¹⁰⁷ Kim et al. ^116,118 reported that the combined addition of Mn and Cr to modify β-phase could be more effective which considerably improved tensile properties (ultimate tensile strength (UTS) and elongation); the improved mechanical properties were attributed to the precipitation of α-Al(Mn,Cr,Fe)Si nanoparticles in the microstructure of A356 Al alloy.

356-type Al alloys

Backerud et al. ⁷⁴ studied the solidification sequence in various Al alloys using a thermal analysis technique, followed by a subsequent metallographic examination of specimens. Their results on solidification of A356·2 alloy with a cooling rate of 0·7 K s⁻¹ are summarized in Table 4. Reactions (2b) and (3b) were not observed by Arnberg et al. ¹²⁵ and Mackay et al. ¹²⁶ in their investigation of almost the same chemical composition. They stated that no pre-eutectic (Al₅FeSi) phase could be crystallized with such low Fe contents, although their specimens contained 0·08% Fe as did those of Backerud et al. ⁷⁴ Backerud et al. ⁷⁴ stated that the Fe is strongly partitioned in the liquid phase which results in precipitation of the pre- or co-eutectic Al₅FeSi phase. Subsequently, the Al₅FeSi phase is partly transformed into the Al₈FeMg₃Si₆ phase through a quasi-peritectic reaction (3b). Wang et al. ¹²⁷ confirmed the Backerud et al. ⁷⁴ results on solidification sequence by scanning electron microscopy (SEM) analysis. As illustrated in Fig. 13, the π-Al₈FeMg₃Si₆ phase was directly grown from the Al₅FeSi phase, which could imply the occurrence of reactions (3a) and (3b).

OPEN IN VIEWER

13.

SEM micrograph of A356 as-cast Al alloy showing the close association between Al₅FeSi and π-Al₈FeMg₃Si₆ phase (reprinted with permission from Springer) ¹²⁷

319-type Al alloys

The solidification sequences of two 319-type aluminium alloys with chemical compositions of (Al–5·7Si–3·4Cu–0·62Fe–0·36Mn–0·10Mg (wt-%)) ⁷⁴ and (Al–6·23Si–3·8Cu–0·46Fe–0·14Mn–0·06Mg (wt-%)) ¹²¹ are listed in Table 5. The precipitation of Al₁₅Mn₃Si₂ (possibly together with Al₅FeSi) which was observed by Backerud et al. ⁷⁴ was not detected by Samuel et al. ¹²¹ This is presumably because of the smaller Mn content of the alloy studied by the latter authors. The presence of Mg (even in the small amount of ∼0·06 wt-%) leads to the transformation of the Al₅FeSi phase to π-Al₈FeMg₃Si₆ phase as well as precipitation of Mg₂Si phase during solidification, attributed to reaction (C) in Table 5. ^121,128 Furthermore, precipitation of Q-Al₅Cu₂Mg₈Si₆ phase, corresponding to reaction (E), is caused by the addition of Mg. ^121,129 The Q-Al₅Cu₂Mg₈Si₆ phase grows out of θ-Al₂Cu particles during the complex eutectic reaction in the final stages of solidification. ^72,130 The morphology of the θ-Al₂Cu phase, which can be of blocky or eutectic form, strongly depends on solidification rate and Sr modification. It has been reported that high solidification rate leads to fine eutectic Al–Al₂Cu phases, ^121,131 while Sr modification increases the proportion of blocky Al₂Cu phase. ^132–134

Porosity

The combined effect of volumetric shrinkage and dissolved gas leads to the formation of porosity. ^111,141 In alloys with low fluidity, the shrinkage of the melt between dendrites cannot be fully filled by the liquid phase remaining, which leads to porosity being spread out along these dendrites. The only gas which is sufficiently soluble in aluminium alloys and leading to porosity is hydrogen. ^142,143 The solubility of hydrogen decreases with decreasing temperature and hydrogen atoms precipitate and form molecular hydrogen during solidification.

Porosity formation in Al–Si hypoeutectic alloys can be affected by alloying elements via a few mechanisms. Addition of Cu to Al–Si alloys assists porosity formation by increasing both the solidification range and the solidification shrinkage. ^144–146 The overall solidification shrinkage in Al–Cu binary alloys is ∼8·4% while it is ∼4·5% for Al–7% Si. ^145–147 Moreover, increasing the copper content enhances the activity coefficient of hydrogen which, in turn, decreases the solubility of hydrogen. Therefore, the alloys containing copper can be more prone to form porosity during solidification. ¹⁴⁸ Caceres et al. ^144,146 stated that ‘the addition of only 1% Cu causes the development of a significant level of porosity in comparison with the Cu-free A356·2 alloy, while increasing the levels of Cu beyond 1% and up to about 4% results in a relatively small increase in porosity level’. The iron-bearing platelets (e.g. β-AlSiFe phase) reduce permeability and restrict the flow of liquid metal at the latter stage of the solidification process, ¹⁴⁹ which was elaborated in the ‘Influence of iron as impurity’ section. Grain refinement obtained by alloying elements such as titanium and boron reduces the volume fraction and size of porosity. ^149,150 It is worth pointing out that Mg ^146,149 and Si ^145,146 can have a positive impact in reducing both pore size and density.

Tensile and fatigue properties are made significantly poorer by increasing porosity. ^36,42,151 Surappa et al. ¹⁵² found that the decrease in the elongation to fracture could be correlated to the pores on the fracture surface. Ma ¹⁵³ showed that increasing metal soundness, in terms of porosity, resulted in a higher elongation to fracture in alloys A319 and A356. The effect of porosity on fatigue strength is strongly dependent on a number of factors, such as morphology, size and position of the pores within the cast part. Skallerud et al. ¹⁵⁴ reported that a shrinkage pore could be more deleterious than a gas pore. Fatigue cracks are generally initiated from shrinkage pores at or near the free surface of a specimen. The effect of large pores far away from the free surface of specimens on the fatigue lifetime can be very small, while even a small pore (or inclusion) located near the free surface can be very deleterious to fatigue lifetime. ^36,155,156

Secondary dendrite arm spacing

In an alloy microstructure, the SDAS generally characterises the solidification rate. Increasing the solidification rate substantially improves the fatigue and tensile properties (except modulus). ³⁴ This improvement is generally attributed to the influence of solidification rate on the number density and size of porosity, and to the refinement of grains and secondary phase microconstituents. ^23,157 Several authors ^12,23,135 reported that reducing SDAS strongly decreased both the number density and the average pore size in Al–Si alloy castings. Chen et al. stated that ‘in A356·2 Al alloy as the SDAS increases from 15 μm to 50 μm, the fatigue lifetime decreases about three times under LCF and over six times under HCF’, since for the alloy with SDAS greater than ∼30 μm, pores act as fatigue crack initiation sites. ^158,159 Furthermore, the content of β-Al₅FeSi as the least desirable secondary phase was significantly reduced by the increasing cooling rate (see Fig. 14). ²³ So, one can say that the influence of SDAS cannot be separated from the influence that solidification rate has on the size and distribution of all microconstituents.

OPEN IN VIEWER

14.

Effect of cooling rate on the formation of β-Al₅FeSi brittle phase ²³

Segregation

Segregation is another important phenomenon which can considerably affect fatigue lifetime. In casting, heat is transferred through the mould walls and this causes higher volume fraction of the α-Al phase to be located in the outer surface of the casting and a larger volume fraction of eutectic phases and shrinkage porosity to be located in the centre. Consequently, local fatigue resistance could vary with the location within a casting. Seniw et al. ^155,160 reported interesting results on the effect of segregation of Si on fatigue properties of A356 cast alloy. They revealed that specimens taken from the outer surface of a cast bar, which was the first zone to be solidified, survived 10⁶ cycles without failure, while specimens taken from the part to solidify last failed after only 150 000 cycles. This illustrates how fatigue lifetime can be reduced down the solidification path.

Cracking/debonding of Si particles

Crack propagation in Al–Si based alloys depends on the size, orientation and local distribution of the Si particles. ^101,161,162 In modified Al–Si alloys (fine Si particles, size ∼1·5–2·5 μm), fatigue cracking progresses by decohesion of the Si particles from the Al matrix. But, with increasing Si particle size, the tendency to particle cracking increases, such that in unmodified alloys (coarse Si particles, ∼3–9 μm) particle cleavage is the dominant feature. ^101,163,164 Figure 15 illustrates the debonding of a Si particle from the Al matrix and a fractured Si particle caused during a fatigue test. Plastic deformation in TMF loading can cause debonding of Si particles. ^68,81,165 This is a result of significant thermal/mechanical misfit between the brittle Si particles and the surrounding ductile matrix, which leads to separation during thermal/mechanical loading. ^68,137

OPEN IN VIEWER

15.

SEM images of: a debonded (reprinted with permission from Elsevier) ¹³⁷ and b fractured Si particle (reprinted with permission from Springer) ⁸⁰

Slip bands

Several researchers ^166–169 reported that in the absence of casting defects (e.g. porosity) or in castings with small SDAS, ¹⁷⁰ cracks initiated from persistent slip bands on the surface. Nyahumwa et al. ¹⁶⁸ observed a faceted transgranular appearance on the fatigue fracture surface of hot isostatic pressed A356·2 aluminium alloy. They reported that the faceted transgranular fracture mode of some specimens was by the slip mechanism. Jiang and co-workers ¹⁷¹ observed slip band cracking only in naturally aged and underaged samples. Zhu et al. ¹⁶⁷ also reported that twin boundary initiated failures in 319 Al alloy could occur only at elevated temperature. Jang et al. ¹⁶⁶ hypothesised that, with increasing temperature, the critical effective stress for fatigue crack initiation (CES_FCI) value at slip band would be comparable to the CES_FCI value at porosity, while it is lower at porosity than at slip band at room temperature. Moreover, Jang et al. ¹⁶⁶ reported interesting results on TMF crack initiation in cast 319-T7 aluminium alloy. The crack initiation of 11 specimens (out of 29 specimens) occurred at near surface porosity, but, for those specimens with relatively small porosity near the surface, coarse transgranular facets were observed at the crack initiation site. They proposed that the slip band mechanism was responsible for crack initiation. Owing to the presence of oxide films in these transgranular facet areas, the authors ^166,172 concluded that these oxide films were formed as a result of fretting damage under fatigue cyclic loading, rather than pre-existing oxide films. However, Campbell ^173,174 criticised their idea and proposed that the oxide film on the fatigue initiation site was created as an inclusion during the solidification, and was a prerequisite for slip band crack initiation.

Gundlach et al. ²⁶ investigated TMF of 319 and 356 Al alloys and reported the occurrence of stress relaxation in 356 Al alloy on heating above ∼505 K. Takahashi et al. ²⁹ stated that stress relaxation started at ∼493 K in the TMF process of Al–6Si–2·5Cu–0·3Mg (wt-%) alloy. At this temperature, which is ∼0·56T _m,2 diffusion creep and dislocation creep can occur; ¹⁶⁹ therefore, they concluded that these creep micro-mechanisms could be responsible for softening of the alloys. ²⁹ Angeloni ¹⁷⁵ also reported that the aforementioned creep micro-mechanisms could be responsible for plastic deformation of Al–9Si–3Cu–0·3Mg (wt-%) alloy in elevated temperature fatigue tests (∼553 K).

Heat treatment of AlSiCuMg alloys

The common thermal treatments, which are generally applied for AlSiCuMg cast alloys, involve either age hardening of the as cast alloy (T₅ type) or solution treatment followed by age hardening (T₆, T₇ type). ^104,176 If peak mechanical properties are not required, castings with sufficiently high cooling rates and artificially aged (T₅ type) may meet the intended strength requirements. This allows a reduction of production costs since the solution heat treatment is not made. However, T₆ (‘peak-aged’) and T₇ (‘overaged’) are the most common heat treatments made on AlSiCuMg alloys. The T₆ heat treatment is generally used for room temperature applications, ^177,178 while for high temperature applications, and especially in the case of 319-type Al alloys, the T₇ treatment is recommended. ^22,28,179 These heat treatment processes, which involve the following three consecutive stages, have to be optimised: (1) solution treatment, (2) quenching and (3) ageing. ^180–182

Solution treatment

The solution heat treatment is achieved by heating the alloy at a temperature range between the solvus and the solidus line (see Fig. 16). The soaking period must be long enough to cause one or more constituents to enter into solid solution. Homogenisation of the alloying elements and spheroidisation of the eutectic Si particles are the other purposes of the solution treatment. ^183,184

OPEN IN VIEWER

16.

Temperature ranges for heat treatment and relevant solvus line for binary aluminium alloys ¹⁸⁴

The dissolution rate of intermetallic compounds is strongly dependent on the solutionising temperature. Samuel ¹⁸⁵ has reported that increasing the solutionising temperature from 753 to 773 K in Al–6·17Si–3·65Cu–0·45Mg (wt-%) alloy improved the yield strength from 330 to 410 MPa and the UTS from 340 to 420 MPa. On the other hand, the maximum applicable solution treatment temperature is limited by incipient melting of the last solidified phases. ^185–187 Incipient melting deteriorates the mechanical properties as a result of void formation. ^131,180 According to Samuel, ¹⁸⁵ the solutionising temperature of a cast Al–6Si–3Cu (wt-%) alloy containing 0·04% Mg can be ∼792 K, but increasing the Mg content to 0·5% restricts the solution treatment temperature to ∼778 K to avoid incipient melting. ¹⁸⁵ It has been reported that even a small amount of Mg (0·1 wt-%) can reduce the solidus temperature of a 319·0-type Al alloy down to 780 K under non-equilibrium solidification conditions. ^74,121 Moreover, Fuoco et al. ¹⁸⁸ pointed out that the solutionising temperature for AlSiCuMg alloys must not exceed 773 K to avoid incipient melting. Therefore, the melting point of the last solidified phase must be known accurately to optimise the solutionising temperature. This can be achieved by using a microsegregation model or by conducting a thermal analysis.

Sokolowski et al. ^189,190 reported that single-step solution treatment of Al–7Si–3·7Cu–0·23Mg (wt-%) alloy, which must be at less than 768 K, is neither able to maximise the dissolution of Cu rich phases nor able to homogenise the microstructure and modify the Si particles. As a result, they proposed a two-step solution treatment (i.e. 8 h at 768 K+2 h at 793 K). By doing so, the Cu-containing phase (Q-Al₅Cu₂Mg₈Si₆) with the lowest melting point (∼780 K) ^185,191 would be dissolved at the first step. The higher solutionising temperature of the second step could dissolve the remaining Cu-bearing phase and further homogenise the microstructure. ^190,191 Nevertheless, some authors reported the stability or very slow dissolution rate of Q-Al₅Cu₂Mg₈Si₆ phase at ∼773 K (500°C) ^192,193 when the magnesium content is sufficiently high. The holding time period of the first step and the solution temperature of the second step are very critical parameters to avoid incipient melting. ^180,194,195 Therefore, to achieve an effective dissolution while avoiding coarsening of the constituents, the solutionising parameters (namely time and temperature) have to be optimised. ^196,197 In this regard, differential scanning calorimetry (DSC) and electron probe microanalysis are powerful tools, which are discussed in more details below.

Wang et al. ¹⁹¹ used DSC analysis to optimise the solutionising treatment of Al–11Si–4Cu–0·3Mg (wt-%) alloy. Figure 17 displays the DSC curves of the alloy for different solution times at 773 K. Peaks (1), (2) and (3) correspond to the following reactions:

OPEN IN VIEWER

17.

DSC curves of AlSiCuMg alloy solution treated at 773 K for different times (reprinted with permission from Elsevier) ¹⁹¹

Reaction of peak (1): α (Al)+Al₂Cu+Si+Al₅Cu₂Mg₈Si₆→Liquid

Reaction of peak (2): α (Al)+Al₂Cu+Si→Liquid

Reaction of peak (3): α (Al)+Si (+Al₅FeSi+…)→Liquid

As illustrated in this figure, with increasing solution time, the height of peaks (1) and (2) gradually decreased. After 10 hours of solution treatment, peak (1) completely disappeared, which indicates the complete dissolution of eutectic phases (α-Al+Al₂Cu+Si+Al₅Cu₂Mg₈Si₆). Therefore, the temperature at the second step of solution treatment could be increased up to the onset temperature of peak (2) (∼793 K) to quickly dissolve the remaining Cu-rich intermetallics. The temperature of the second solution treatment step should be lower than 793 K in order to avoid incipient melting of (α-Al+Al₂Cu+Si) eutectic phase. ¹⁹¹ It is noteworthy that increasing the Mg content enhances the stability of Q-Al₅Cu₂Mg₈Si₆ phase in AlSiCuMg alloy system. ^193,198 For instance, Lasa et al. ¹⁹³ reported that the Q-Al₅Cu₂Mg₈Si₆ phase in Al–12·5Si–4·5Cu–1·3Mg (wt-%) alloy was almost completely undissolved after 24 hours of solution treatment at 773 K.

Dissolution of Cu phases (e.g. Al₂Cu/Al₅Cu₂Mg₈Si₆) which increases the Cu content in the α-Al matrix is one of the major purposes of the solution treatment. In order to determine the efficiency of a specific solution treatment, Sjolander et al. ¹⁸³ and Han et al. ^192,199 proposed to measure the Cu distribution in the α-Al matrix by means of line scans in electron probe microanalysis. For instance, the solutionising time of Al–8Si–3Cu (wt-%) alloy with different SDAS (10, 25, 50 μm) was studied by Sjolander et al. ¹⁸³ Figure 18 illustrates the concentration of Cu in the dendrite arms for various specimens with different solution time (0, 10, 60, 180, 360, 600 min). Homogenisation in the dendrite arms occurred very fast (within 10 and 60 min), but the concentration of Cu was strongly dependent on the microstructure and solutionising time. For the finest microstructure (SDAS of 10 μm), 10 min of solutionising time seemed to be enough; but for the very coarse microstructure (SDAS of 50 μm), even 10 h of solutionising time at 768 K was not sufficient. ¹⁸³

OPEN IN VIEWER

18.

a SDAS 50 μm, b SDAS 25 μm and c SDAS 10 μm (reprinted with permission from Elsevier) ¹⁸³

Ageing

During ageing of Al alloys, solid solution strengthening gradually disappears and the coherent structure of Guinier–Preston (GP) zones leads to an intense strain field in the surrounding area. ^204,205 The mechanisms contributing to increase the yield strength by the motion of dislocations through precipitates may include chemical, stacking fault, modulus, coherency and order strengthening. ^206,207 These mechanisms were thoroughly reviewed by Ardell. ²⁰⁸ Ageing is performed by holding the supersaturated solid solution at temperatures below the solvus line to form a fine distribution of precipitates from a supersaturated solid solution (see Fig. 16). The thermodynamically unstable supersaturated solid solution will reach equilibrium conditions by ageing at room temperature (natural ageing) or with a precipitation heat treatment (artificial ageing). Time and temperature are the two main parameters of ageing which affect the strengthening mechanisms. Higher ageing temperature accelerates the ageing process by increasing nucleation and growth rates. ^180,207,209

Several investigations have been carried out to understand the effect of underaging, peak ageing and overaging on: hardness, ^210–212 tensile strength, ^180,201,211 crack propagation behaviour, ^171,213 TMF behaviour ²¹⁴ and cyclic stress–strain response of AlSi(Cu,Mg) alloys. ²¹⁵ The sequence of precipitation of θ-Al₂Cu begins by GP zones, which are thermodynamically the least stable but kinetically the most favoured phases: α-Al→GP zones (plate-like)→θ″ (plate-like)→θ′ (plate-like)→θ (Al₂Cu). GP zones and θ″ are fully coherent with the α-Al matrix, θ′ particles can be either coherent or semi-coherent, while θ particles are incoherent. ^216–218 GP zones with 3–5 nm diameters consisting of localised concentrations of Cu atoms have been observed in specimens aged at 373 K for 2·5 h. ²¹⁹ The required ageing time at 373 K was reported to be at least 1000 h to obtain a microstructure where plate-shaped Cu-rich particles (GP zones) predominate. ^179,219 However, some authors have stated that GP zones undergo dissolution at temperatures higher than 373 K; ^77,187 the presence of GP zones after ageing at 403 K for 16 h ²²⁰ and the coexistence of ‘GP zones and θ″’ after ageing at 423 K for 3·5 h ²¹⁹ have also been reported. The peak strength is influenced by the amount, size and site density of θ″ and θ′ phases. ²⁰⁵ According to reports, ^179,219 the reason for softening with overaging in 319-type Al alloys can be attributed to the coarsening of the θ′ phase. The transformation of θ′ to θ occurs only when ageing at 523 K (or higher) and for time periods greater than 1000 h. ^179,219

Two different combinations of precipitates have been observed in peak-aged condition of the AlSiCuMg alloy system: (1) precipitation of β″ (based on Mg₂Si) and/or θ′ and (2) precipitation of Q″ and/or θ′, where the θ′ phase only appears for a high concentration of Cu (≧1 wt-%). ^187,221,222 In several studies, ^221,222 no θ-Al₂Cu phase has been reported during artificial ageing of AlSiCuMg alloys when the Cu content was less than 1·0 wt-%. Figure 19 illustrates the DSC curves of as quenched and aged Al–7Si–3Cu–0·4Mg (wt-%) alloy with a 10 K min⁻¹ heating rate. Formation and dissolution temperature of GP zones, Q′ phase, θ′ phase and θ phase were found to be at about 303–493, 493–543, 543–633 and 633–733 K. ¹⁰⁷ It is worth mentioning that the temperature at which a given peak occurs increases with increasing scan rate. ^107,223 An exothermic peak corresponding to GP zone formation was only detected for the as quenched specimen. In the alloy with some impurities (e.g. 0·6 wt-% Fe and 0·5 wt-% Mn), GP zones could not be detected at all; instead, the precipitation of θ′ phase appeared at earlier stages. ¹⁰⁷

OPEN IN VIEWER

19.

DSC curve of Al7Si3Cu0·4Mg alloy, solution-treated 10 h@773 K, water-quenched and aged for different times at 443 K (reprinted with permission from American Foundry Society) ¹⁰⁷

The ageing time to reach peak strength is longer for AlSiCu alloy than for AlSiCu(Mg) alloy. ¹⁰⁷ The required time period to obtain peak strength in AlSiCu(Mg) alloy varies from 30 h up to 120 h and even longer at a lower temperature (433 K). ^107,224,225 The addition of Mg accelerates and intensifies the precipitation-hardening process of AlSiCu alloys. ^107,209,224 Kang and co-workers ¹⁰⁷ reported that not only was the peak hardness obtained for AlSiCu alloy lower than that obtained for AlSiCuMg alloy, but also the ageing time required to reach peak hardness for the former was ten times longer than for the latter. On the other hand, Wang et al. ²²¹ stated that Cu addition to AlSiMg alloy not only increases the age hardenability, but also extends the time (from about 700 to 3000 min) required to reach the peak hardness.

The large discrepancy between the thermal expansion coefficients of α-Al matrix (23·5×10⁻⁶ K⁻¹) and Si particles (9·6×10⁻⁶ K⁻¹) generates a lot of dislocations during quenching around the Si particles and makes these locations become a preferential site of nucleation for the θ′ phase. ^107,182,217 On the contrary, the Q′ phase can nucleate at locations of lower surface energy since this phase is assumed to have a better coherency (semi-coherency) with α-Al. Therefore, Q′ can precipitate on a dislocation located anywhere in the matrix giving a more homogeneous distribution of these precipitates. The lengths of diffusion are reduced and then less time is required to reach peak hardness. ^107,226 This could explain the higher age hardening rate of AlSiCuMg alloy relative to AlSiCu alloy. Nevertheless, it has been reported that at elevated temperature the Cu-containing θ″–θ′ phase can be much more stable than the Mg-containing λ′–λ (Al₅Cu₂Mg₈Si₆) and β″–β′ (Mg₂Si) phases. ^107,227 ; also, S′–S (Al₂CuMg) phase has been reported to be more stable than β″–β′ (Mg₂Si) phase. ²²⁸

In addition to the presence of β-Mg₂Si and Q-Al₅Cu₂Mg₈Si₆ phases in an aluminium 319 alloy, S-(Al₂CuMg) phase was also identified by Medrano et al. ¹⁸² The authors stated that β-Mg₂Si and S-(Al₂CuMg) phases probably precipitated during solidification, and still remained undissolved after solution treatment. According to Mondolfo, ¹⁰⁹ in AlCuMg alloy with the ratio of Cu to Mg between 4:1 and 8:1, the ageing agent would be both the Al₂Cu phase and Al₂CuMg phase. In the case of AlSiCuMg alloy with high silicon content, the S-(Al₂CuMg) phase is not usually found, but it can be seen in small amounts owing to compositional heterogeneities. ¹⁹³ Nevertheless, the presence of S-(Al₂CuMg) phase in AlSiCuMg alloys has been observed by some authors. ^182,209,219 Ma et al. ¹⁷⁸ pointed out the presence of Al₂Cu and Al₂CuMg phase in Al–11Si–2·5Cu–0·4Mg (wt-%) alloy. Reif et al. ^225,229 likewise reported the presence of S′-(Al₂CuMg) phase with increasing Mg addition to AlSiCu alloy.

It is worth noting that increasing the Mg level beyond 0·3 wt-% in 319-type Al alloys does not significantly change the alloy strength, ^230,231 but it can considerably reduce the ductility of the alloys. In 356-type Al alloys, increasing the Mg content up to 0·5 wt-% enhances the strength, while further increasing Mg content can have a negative effect on the strength of the alloys. ¹⁸⁷ Wang et al. ¹³⁶ reported that the fatigue lifetime of A357 alloy (with 0·7 wt-% Mg) was lower than that of A356 (with 0·4 wt-% Mg). In alloys with high Mg content, a large fraction of the π-Al₈FeMg₃Si₆ phase could be formed which is stable during the solution treatment. ^187,231

The small precipitates/zones which are cut by the dislocations in motion lead to a maximum yield stress once the dislocations pass through them. This causes the local work hardening to be small and the plastic deformation to be restricted on a few active slip planes, which would probably be very deleterious to fatigue lifetime. ²²⁰ On the other hand, for large particle size/interspacing, bypassing particles by dislocations results in rapid work hardening and the plastic strains are distributed throughout the specimen. However, because of weak strengthening of these precipitates, the yield stress is not high enough. Fine ²²⁰ stated that ‘the interesting possibility is to have a dispersion of two kinds of second phase particles, small closely spaced particles to give high yield stress plus large particles to distribute the plastic deformation throughout the material’.

Therefore regarding the operating condition, the aluminium alloys might be used after peak strengthening with metastable microstructure (T₆) or after overaging with equilibrium microstructure (T₇). For engine components which are exposed to TMF, T₇ condition seems to be more appropriate than T₆, since:

a.

T₆ condition can cause localised deformation; ²²⁰

b.

prolonged exposure at service temperature leads to higher thermal growth3 in T₆ condition. The thermal growth of W319 Al alloy was 0·045% and 0·006%, respectively, in T₆ and T₇ conditions; ¹⁷⁹ and

c.

T₇ shows more stable microstructure and higher TMF lifetime than T₆. ²⁸

Dispersion hardening

Trying to improve the elevated temperature strengths of aluminium alloys has involved continuing efforts for more than three decades. ^28,220 Before going further, it could be worthwhile to consider the reason for successfully engineered Ni-based superalloys being mechanically stable at high temperatures (exceeding 0·75T _m). ²³² The interesting mechanical properties of Ni-based superalloys at elevated temperatures can be mainly related to the presence of very large volume fractions of fine γ′-Ni₃(Al,Ti) precipitate with L1₂ structure, which is coherent–coplanar and moderately ductile. ^65,220,233 The term ‘coherent–coplanar’ means the precipitate/matrix interfacial energy is very low and the tendency for coarsening/coalescence of the precipitate is very small. To develop an effective high-strength high-temperature Al alloy, it can be useful to remember the characteristics of this precipitate in Ni superalloys. ²³⁴

Softening of the precipitation hardened Al alloys (e.g. AlSiCu) is the major problem at elevated temperatures because of the dissolution/coarsening of the metastable precipitates. A high-strength high-temperature Al based alloy must have a distribution of fine precipitates/dispersed phases, which must be thermodynamically stable, coherent–coplanar and ductile. ^22,105,220 A low solid solubility as well as limited diffusivity of the solutes in α-Al at the intended service temperature, which is essential to retard volume diffusion, controls the rate of dissolution and coarsening of the precipitated phases. ^220,234,235 Moreover, the larger the interfacial energy, the higher the driving force for coarsening/coalescence of the precipitates (Ostwald ripening). Therefore, the required driving force for coarsening can be very small in the coherent–coplanar precipitates. ²¹⁷ Zedalis ²³⁶ stated that the coarsening rate of the tetragonal Al₃Zr dispersed phase (D0₂₃ with semi-coherent interface) is 16 times higher than that of the cubic modified one (L1₂ with coherent interface). Furthermore, the coherency of the precipitate/matrix interface magnifies the strengthening efficiency of the dispersed phase. Accordingly, precipitated phases with a similar crystal structure and a low lattice parameter mismatch with the α-Al solid solution are preferred. ^220,234,236

Among the transition elements, only the first element of the third group (i.e. Sc) exhibits a high symmetry L1₂ trialuminide (Al₃Sc) structure which is an ordered fcc lattice of the Cu₃Au type of structure. ^234,237 Group 4 (Ti, Zr, Hf) and group 5 (V, Nb, Ta) elements crystallise with the body-centred tetragonal D0₂₂ and D0₂₃ (Al₃M) structures, as shown graphically in Fig. 20. The brittle low symmetry tetragonal structure (of Al₃M; M = Ti, Zr, Hf, V, Nb) can be transformed to the cubic structure (L1₂) by alloying. ²³⁴ Furthermore, it has been stated ^238–241 that the precipitation sequence in the ageing treatment of supersaturated Al–Ti, Al–Zr and Al–Hf solid solutions occurs initially by the formation of a metastable cubic L1₂ (Al₃M) phase. The overall sequence of precipitation in Al–Zr and Al–V systems has been reported ²⁴² to be: (supersaturated solid solution)→(cubic spheres and rod L1₂)→(tetragonal plates D0₂₃/D0₂₂). Long term exposure (hundreds of hours) at high enough temperatures (>450°C) is required to transform these metastable phases to the equilibrium tetragonal (Al₃M) structure. In other words, these phases are thermodynamically metastable (Gibbs free energies of formation of the tetragonal (D0₂₃) and cubic phase (L1₂) of Al₃Zr are −40·75 and −38·35 kJ mol⁻¹, respectively ²⁴³ ), but kinetically stable at elevated temperatures even close to 673 K, because of the extremely slow diffusion rate of these transition elements in α-Al. Moreover, some alloying elements can reduce much more the rate of this transformation. Zedalis ²³⁶ stated that ‘addition of V to Al–Zr alloy led to a reduction of the precipitate-matrix mismatch for both phases, and also retarded both coarsening as well as the cubic to tetragonal transformation’. Litynska ²⁴⁴ wrote that the addition of 0·2% Zr to Al–1Mg–0·6Si–1Cu–0·4Sc (wt-%) retarded the coarsening of Al₃Sc phase and restricted the size of Al₃(Sc,Zr) precipitates to about 20–40 nm, which were fully coherent with the matrix. The retardation of coarsening of Al₃V phase by Zr addition was also confirmed by Fine et al. ²⁴²

OPEN IN VIEWER

20.

a L1₂, b D0₂₂, and c D0₂₃ crystal structures (reprinted with permission from Elsevier) ²⁴⁶

To have a coherent/coplanar interface, dispersoid phases with small lattice parameter mismatch are preferred. For the transition elements Hf, Zr, Sc, Nb, Ti, V and Ta the lattice parameter mismatches between the precipitate (pure binary Al₃M (L1₂) trialuminides) and α-Al matrix at room temperature are 0·04%, 0·75%, 1·32%, 1·49%, 2·04%, 4·44% and 5·26%, respectively. ^234,245

Because of the low volume fraction of the dispersoid phases in Al alloys, the precipitates should be very small and resistant to coarsening. Therefore, for an alloy subjected to prolonged exposure at elevated temperatures, slow diffusion kinetics is required to maintain strength. Figure 21 compares the calculated diffusivities of different solute elements in α-Al at three different temperatures (i.e. 573, 673 and 933 K). It has been reported that elements belonging to the same group might be assumed to show similar diffusion kinetics in α-Al (e.g. ). ²³⁴

OPEN IN VIEWER

21.

Calculated diffusivities for different solute elements at 573, 673, and 933 K (T _m of Al) (reprinted with permission from Carl Hanser Verlag) ²³⁴

Of the transition elements, Zr seems to be one of the most promising for the design of lightweight high-strength high-temperature Al alloys. ^227,234 Al₃Zr phase not only impedes the dislocation motion but also refines the microstructure of Al alloys. With the addition of 0·15 wt-% Zr to Al–2% Cu alloy, the columnar grain structure changed to equiaxed structure. ²⁴⁷ Fasoyinu et al. ²⁴⁸ studied the effect of Zr, Sc and a combination of both on grain refinement of 356 alloy; effective concentration ranges of Zr and Sc of 0·37–0·69 and 0·39–0·75 (wt-%), respectively, are required to achieve a considerable grain refinement. Nevertheless, the phase and microstructure evolution of different Al based alloys (binary AlZr or multicomponent AlSiCuMgZr alloys) in the presence of this element has been keenly disputed.

Mahmudi and co-workers ^249,250 investigated the effects of 0·15 wt-% Zr addition on the mechanical properties of A319 Al alloy. The hardness and wear resistance of the A319+Zr alloy were improved by 10% and 60%, respectively, compared to the A319 alloy, which were ascribed to the presence of the Al₃Zr phase. Garat et al. ²² observed the presence of fine, semi-coherent ternary (Al–Zr–Si) dispersoids in the α-Al dendrites of (A356+Zr) alloy, which were formed during solution treatment above 773 K. They observed no binary Al₃Zr phase in the microstructure. Ozbakir ²⁵¹ also reported that with 0·15 wt-% Zr addition to A356 alloys, the eutectic ternary ϵ-(Al–Si–Zr) phase was formed instead of the peritectic binary Al₃Zr phase. Prasad ²⁵² observed the presence of both Al–Zr–Si and Al₃Zr phases. Iveland ²⁵³ reported the presence of rod-shaped AlSiZr and AlSiZrTi precipitates in the heat treated microstructure of A356 alloy containing Zr and Ti. Recently, the presence of relatively coarse Al₃Zr particles (diameters ∼600 nm) in (A356+Zr) as cast alloy was reported by Baradarani et al. ²⁵⁴ After solution treatment, very fine Al₃Zr particles were observed in the microstructure, which led to the conclusion that either the Al₃Zr particles were not completely dissolved during solution treatment or the particles re-precipitated after dissolution. Baradarani et al. ²⁵⁴ and Srinivasan et al. ²⁴³ stated that the dissolution–precipitation mechanism was promoted by the motion of grain boundaries, which activates dissolution ahead of the advancing boundary and precipitation behind.

Hypereutectic Al–Si alloy cylinder block

Use of monolithic hypereutectic Al–Si alloy cylinder blocks was the earliest alternative approach to eliminate cast iron liners. Porsche, as a pioneer, used a hypereutectic Al–Si alloy cylinder block in the 1960s. The first application of a hypereutectic Al–Si alloy in a commercial engine was in 1971, when Reynolds A390 aluminium alloy was used by General Motors in the liner-less Vega block. ^73,272 Hypereutectic Al–Si engine blocks have significantly evolved since the Vega. Some car manufacturers which are using Al–Si hypereutectic alloys are Mercedes, Audi, BMW, Volvo and Honda. ^73,270 Many different hypereutectic Al–Si alloys for cylinder blocks are produced in the world market with specific trade names (e.g. Alusil®, Lokasil®, Silitec®, DiASil, Mercosil, ALBOND®). For instance, Mercosil, as a copper-free hypereutectic Al–Si alloy, is utilised in Mercury Marine’s cylinder blocks. More uniform distribution of primary silicon particles in Mercosil gives better tribological characteristics compared to 390 Al–Si alloy. ²⁷⁰

Chemical composition and mechanical properties of 390·0 hypereutectic Al–Si alloy are presented in Tables 2 and 3, respectively. It is well known that by adding Si to Al alloys, beyond the eutectic composition (at ∼12·6 wt-% Si; see Fig. 11), primary silicon particles crystallise out in the microstructure of the Al–Si hypereutectic alloy which consequently enhances the hardness of the alloy and provides the required tribological surface characteristics. In the absence of silicon particles, premature wearing of Al cylinder bore surfaces by the piston rings could rapidly occur; consequently, the compression of the engine will be lost which results in engine failure. Therefore, a uniform distribution of primary silicon particles in the alloy is the key criterion for wear resistance of cylinder bores. ²⁶⁸

The piston ring slides over a layer of both the primary silicon particles and lubricant without touching the Al matrix in a hypereutectic Al–Si alloy cylinder bore. To make sure that the piston ring rides over the primary silicon particles and not on the Al matrix, a chemical etching process is generally applied to the cylinder bore surface. The etching process leads to silicon particles protruding a slight distance above the aluminium cylinder surface. ^270,273 However, Dienwiebe et al. ²⁷⁴ stated that the height of both Si grains and Al matrix are at the same level after a short running time. They proposed that wear particles from various sources (e.g. silicon fragments, iron oxide and iron nitride particles from the piston ring, and carbonaceous particles) are mixed and placed in the soft Al matrix; the piston ring slides over a layer of wear particles and primary Si phase.

Some mechanical/physical properties of A390 Al alloy are compared with other competing engine materials in Fig. 9. Several benefits are associated with hypereutectic Al–Si alloy cylinder blocks compared to cast iron block/liner. These alloys provide not only the required tribological surface characteristics but also improved thermal conductivity (up to three times more) compared to cast iron. The HCF strength of Mercosil hypereutectic Al–Si alloy was ∼90 MPa, which was ∼50% higher than that of hypoeutectic Al–Si alloy. ²⁷⁰ Low coefficient of thermal expansion, high hardness and being highly recyclable are the other major characteristics of the hypereutectic Al–Si alloys. ²⁶⁸ The Cagiva group reported that the Husqvarna engine equipped with Mercosil alloy showed approximately 6% more horsepower than the honed Nikasil® electroplated bore. ²⁷⁰

In spite of all of these advantages, certain characteristics of the hypereutectic Al–Si alloys prevent their widespread application in the manufacturing of engine blocks. Their applications have been mainly limited to German engines. ⁷³ The casting process of the hypereutectic Al–Si alloys with homogeneous and uniform properties is really difficult. The tribological surface characteristics of hypereutectic Al–Si cylinder bores are strongly dependent on the even dispersion of Si particles. Additionally, manufacturing of hypereutectic Al–Si alloy cylinder blocks is an expensive process; these alloys require the addition of expensive elements (e.g. Ni) to improve the mechanical properties. They also need to be chemically etched to protrude primary silicon particles from the Al cylinder bore surface, which increases processing time and costs. ^275,276

Reinforced Al–Si alloy cylinder block/bore

Use of metal matrix composites (MMCs) is the second possible alternative for replacing the traditional aluminium cylinder block equipped with cast iron liner. MMCs can be used to produce either the entire cylinder block, or only a cylinder liner to be placed into a hypoeutectic Al–Si alloy cylinder block. ^265,277 An MMC is made up of at least two distinct phases, properly distributed to realise unique combinations of properties which are not attainable by the individual components. It is composed of a matrix and reinforcements (e.g. fibrous or particulate phase), in which the reinforcements are surrounded by the matrix. ²⁶⁵

Extrusion of an Al–Si alloy in which reinforcement ceramic particles are already dispersed, and infiltration of liquid metal through a ceramic preform by squeeze casting are the two common methods to produce MMC cylinder liners. In the squeeze casting process, pressure is applied on the liquid metal in such a way that it is forced to infiltrate porous fibre preforms.

Al–Si alloys can be reinforced by both hard and/or soft phases. The major purpose of adding hard ceramic phases is to improve material’s hardness, with the further benefit of enhancing the wear performance. SiC and Al₂O₃ are the two most common hard phases used to reinforce Al–Si alloys. As illustrated in Fig. 23, addition of 6 vol.-% SiC appreciably improves the wear resistance of Al–12Si (wt-%). ²⁷⁸ Soft secondary particles/fibres (e.g. graphite and MoS₂) can also be added into the matrix to decrease the friction, consequently resulting in a potential reduction of wear. Figure 24 illustrates the normalised wear data of Al alloy–graphite composites; the wear rate is considerably reduced with increasing graphite content up to 3 wt-%, after which it remains almost constant. ²⁷⁹ Furthermore, cylinder bore surfaces strengthened with hybrid phases of alumina and carbon can experience even better wear resistance. ^280,281 As illustrated in Fig. 25, the wear resistance of an Al–12Si (wt-%) alloy reinforced with 17 vol.-% short alumina and 7 vol.-% carbon fibres was equal to that of a cast iron cylinder bore; and was significantly higher than that of an Al–Si hypereutectic cylinder bore. ²⁸²

OPEN IN VIEWER

23.

The effects of addition of 6% SiC particles on wear loss of Al–Si alloy ²⁷⁸

OPEN IN VIEWER

24.

ageing wear rate of Al–graphite composites ²⁷⁹

OPEN IN VIEWER

25.

Wear amount of different cylinder bore material ²⁸²

Toyota, as a pioneer in the use of MMCs in cylinder blocks, successfully manufactured MMC lined cylinder bores. The preform of the MMC liner was composed of mullite particles and alumina–silica fibres. ^261,283 Honda also produced a hybrid MMC preform composed of short carbon and alumina (Al₂O₃) fibres, which was subsequently infiltrated by aluminium. ²⁷⁹ Takami et al. ²⁸³ compared the characteristics of two different types of engine block made by Toyota: a cast iron lined Al engine block (1ZZ-FE engine) and an MMC lined Al engine block (2ZZ-GE engine). The bore-to-bore distance was decreased from 8·5 mm in the 1ZZ-FE engine to the shortest possible distance of 5·5 mm in the 2ZZ-GE engine. Using the new MMC lined cylinder block (2ZZ-GE engine) allowed an enhancement of the piston bore diameter and reduction of the piston stroke, which help to improve the engine cycling capacity. Consequently, the horsepower and the torque of the MMC lined cylinder block (2ZZ-GE engine) were, respectively, 26·17% and 4·65% superior to those of the cast iron lined cylinder block version (1ZZ-FE engine). ²⁸³

MMC lined cylinder blocks provide several advantages such as being lighter, showing lower thermal deformation, being more thermally conductive and having lower bore-to-bore distance compared to cast iron liners. The weight of a cast iron liner and MMC liner was reported to be 9 and 6·5 kg per block, respectively. ⁵⁸ A Honda cylinder block produced using Al alloy fortified with MMC was ∼50% and ∼20% lighter than the cast iron engine block and the Al alloy engine block equipped with cast iron liner, respectively. ²⁷⁷ Moreover, MMC lined cylinder blocks can be more affordable and easier to process than hypereutectic Al–Si alloy cylinder blocks (e.g. Alusil®). Nevertheless, the higher production costs, uneven distribution of reinforcement, difficulty in melt infiltration, post-casting machining and recyclability problems prevent the mass production of MMC cylinder blocks. ^277,283

Electroplating Al–Si alloy cylinder block wall

The third potential alternative to provide the necessary surface requirements for cylinder bores is to deposit high hardness, wear resistant materials by electrochemical processing. The most common electroplating processes, which are either currently being used or being evaluated, are nickel ceramic composite coating (NCC) ⁴⁶ and plasma electrolytic oxidation coating (PEO). ^284,285 The PEO process is also known as micro-arc oxidation (MAO). ^286,287 The details of the NCC and PEO/MAO processes have been described by Shrestha ²⁸⁸ and by Yerokhin ²⁸⁶ and Krishtal, ²⁶⁶ respectively.

In an Ni/SiC composite, Ni adheres to the cylinder bore surface to hold and support the silicon carbide particles. These reinforcement particles provide the required tribological surface characteristics and hardness for the cylinder bore. ^46,289 The size of SiC particles is 3–7 μm and corresponds to a volume fraction ranging between 8 and 17% of the coating. The hardness and thickness of the coating are usually 500–600 HV and 70–75 μm, respectively. ^48,284 In addition to SiC, BN and Si₃N₄ are other ceramic particles that can be incorporated in coatings produced using the NCC process. Nikasil®, as a trademark of Ni/SiC composite, has been used by some automakers (e.g. BMW and Jaguar). ^46,290 However, nickel based coatings are highly sensitive to the presence of sulphur in gasoline; over time, sulphur can break down the coating. Therefore, BMW and Jaguar both decided to switch from the Nikasil cylinder liners to cast iron liners. ²⁹⁰ Nevertheless, cylinder bores coated with Ni/SiC composite have been successfully employed where fuel quality is not a concern; ^284,290 for instance, they have been used in Italy (Ferrari), Germany (for sports cars at a level of 90 000 vehicles per year) as well as in France (Citroën). ²⁶⁶

In the PEO process, aluminium transforms to aluminium oxide when a high voltage (>100 V) bipolar pulsed current circulating in an alkali electrolyte produces a plasma discharge on the surface of the substrate. The coating is mainly composed of α-Al₂O₃, γ-Al₂O₃ and 3Al₂O₃⋅2SiO₂ (mullite) phases with two distinct layers. The outer layer, with a microhardness ranging between 500 and 1000 HV and having a porosity level >15 vol.-%, is generally removed by diamond honing; otherwise it could dramatically affect the friction coefficient and wear resistance. ⁴⁸ The inner layer (with a thickness of 15–30 μm) has excellent wear and corrosion resistance, a very hard surface (microhardness of 900–2000 HV) and superior adhesion property to the substrate. Previous studies illustrated that PEO coating on Al–Si alloys leads to lower friction and superior wear resistance in comparison to cast iron ²⁹¹ and Nikasil. ²⁸⁴

The main advantages of the electroplating process are: low friction coefficient, high corrosion resistance, high hardness, high wear resistance, possibility of decreasing the distances between cylinder bores and stronger bonding with the substrate material. ⁴⁶ Besides all of these advantages, a complex and costly process is the major limitation of electroplating processes. The necessity for selective area coating on the cylinder block requires more complicated equipment to carry out localised deposition or masking. Moreover, the high hardness of the applied coating requires specific machining tools for the honing process; it is worth mentioning that honing is an essential process to reach an appropriate coating thickness. ^48,260 As mentioned earlier, nickel based coatings can be susceptible to the presence of impurity elements in fuels (e.g. sulphur). Furthermore, the cost of raw materials (in particular Ni) is another main limitation for the NCC process which hinders the feasibility of mass production. ^260,290

Thermal spray coating of Al–Si cylinder block wall

Another approach to improve the wear resistance of the cylinder bore is to deposit a layer of an appropriate material onto the cylinder bore wall with a thermal spraying process. Atmospheric plasma spraying (APS), high velocity oxy-fuel spraying and electric wire arc spraying are the major thermal spraying processes able to apply a coating on cylinder bores. The APS process has been utilised in mass production to coat cylinder bores over the past 10 years. ²⁹² A description of thermal spray processes of cylinder bores and coating materials has been elaborated upon elsewhere. ^{271,292–294}

Coating of a cylinder bore wall by the thermal spray process is implemented in several steps. After casting the Al–Si cylinder block with a rough diameter, the cylinder bore is pre-machined to a certain diameter. The surface is thoroughly cleaned to remove oil and residues from the machining process. Thereafter, one of the following three processes is applied to activate the bore surface: grit blasting, high pressure water jet or mechanical activation. ²⁹² The purpose of the activation process is to break the aluminium oxides from the surface and to obtain a sufficient roughness. ²⁹⁵ Since there is no metallurgical bond between coating and substrate, this roughening can increase the bond strength. ²⁹² The required bond strength value has been reported to be more than 30 MPa. ^292,296 The advantages and disadvantages of the aforementioned activation processes have been compared elsewhere. ^267,292

The next step is to deposit the coating materials by means of one of the aforementioned thermal spray processes. Some of the most commonly used APS coatings, which are generally iron based materials, are listed in Table 7. Depending on the working condition of the engine, one of these coating materials could be an appropriate choice. Adding ceramics (e.g. Al₂O₃/ZrO₂) to the ferrous coating materials improves the scuffing and wear resistance; however, they do not affect the hardness. This class of coatings, which are MMCs, is useful when high abrasive wear resistance is required. 14Cr–2Mo steel (XPT627) as the base metal of MMC is suitable for corrosion resistant cylinder surfaces. The corrosion resistance of coating materials can be suitable where the presence of impurity elements in fuels (e.g. sulphur) is a concern. Mixing the alloy powders with carbides can considerably enhance the hardness. To produce hard and corrosion resistant coatings, carbides can be mixed with the corrosion resistant matrices. Honing with diamond ledges to reach the appropriate diameter is the final step of the thermal spray coating process. The details of the process, coating materials, microstructure and thickness of the coatings have been elaborated by Ernst et al. ²⁹²

The results of testing an MMC coating material at AVL Austria in a four cylinder diesel engine (50 kW L⁻¹, 150 bar) are summarised in Figs. 26 and 27. As illustrated in Fig. 26, the oil consumption (test duration up to 250 h) can be decreased appreciably by MMC plasma coating of the cylinder bore. In Fig. 27, the radial wear of the plasma coated cylinder bore and the first and second piston rings is compared with that of those parts made with cast iron. In the engine equipped with the plasma coated cylinder bore, the measured wear in all parts was considerably lower.

OPEN IN VIEWER

26.

Comparison of the oil consumption with standard cast iron liners and plasma coated cylinder liners ²⁹⁷

OPEN IN VIEWER

27.

Radial wear on 1st and 2nd piston ring as well as cylinder bore at the top ring position compared for cast iron and plasma coated cylinder liners ²⁹⁷

The major advantages of thermal spray coated Al–Si cylinder bores (compared with cast iron liners) can be summarised as follows: ^271,295

The contact between the Al cylinder bore surface and the coating materials is mainly by mechanical bonding. Therefore, if the thermal spray coating is not appropriately implemented, the coating will be delaminated over time and subsequently separated from the Al cylinder bore surface. ²⁹⁸ Thermal spray coatings are susceptible to the presence of impurity elements in fuels (e.g. sulphur) which introduce the possibility of corrosive attack. ²⁹⁹ Overheating of the Al cylinder bore is another concern during the high velocity oxy-fuel thermal spray process which might damage the substrate. Finally, it has to be taken into account that thermal spray coating equipment needs high initial investment. Despite all of these limitations, thermal spray coating processes are the most recent and the most promising surface modification technologies for Al–Si alloy cylinder blocks, and are potentially suitable for mass production. ^260,271 An overview of qualified engines with plasma coating is given in Table 8.

Comparative analysis of the alternative approaches for iron lined cylinder blocks

A more thorough review on alternative approaches for iron lined cylinder blocks has been done by Lenny. ²⁶⁰ In order to rank the four mentioned alternative approaches and select the best choice, he compared ten different characteristics: ‘Previous Applications, Wear Resistance, Scuffing Resistance, Thermal Conductivity, Friction between cylinder block and piston rings, Fuel Economy, Engine Emissions, Manufacturing Costs, Engine Performance, and Mass Production Feasibility’. Comparison results illustrated that a hypoeutectic Al–Si cylinder block equipped with an iron liner is still the best alternative since it is the most affordable choice. However, social impetus for fuel economy and emissions reduction necessitates application of other alternatives instead of cast iron lined Al–Si cylinder blocks. Hypoeutectic Al–Si cylinder blocks equipped with thermal spray coating proved to be the most promising alternative approach in comparison, which is potentially suitable for mass production. Electrochemical deposition coating on hypoeutectic aluminium bore surfaces, hypereutectic Al–Si alloy cylinder blocks and fibre reinforced aluminium matrix composite cylinder blocks are other promising alternatives. ²⁶⁰