TherModynaMIc analySIS of 6xxx SerIeS al alloyS: phaSe fracTIon dIagraMS

Microstructural evolution of 6xxx Al alloys during various metallurgical processes was analyzed using accurate thermodynamic database. Phase fractions of all the possible precipitate phases which can form in the as-cast and equilibrium states of the Al-Mg-Si-Cu-Fe-Mn-Cr alloys were calculated over the technically useful composition range. The influence of minor elements such as Cu, Fe, Mn, and Cr on the amount of each type of precipitate in the as-cast and equilibrium conditions were analyzed. Phase fraction diagrams at 500 °C were mapped in the composition range of 0-1.1 wt.% Mg and 0-0.7 wt.% Si to investigate the as-homogenized microstructure. In addition, phase fraction diagram of Mg2Si at 177 °C was mapped to understand the microstructure after final annealing of 6xxx Al alloy. Based on the calculated diagrams, the design strategy of 6xxx Al alloy to produce highest strength due to Mg2Si is discussed.


Introduction
Al alloys are widely used as the structural part in transportation sector due to their attractive properties, such as high strength, good corrosion resistance, formability, weldability, and low cost.Because of the continuous demands for weight reduction and improvement of fuel efficiency of automobiles, high strength Al extruded profiles of complex shape are needed in the transportation industry.6xxx Al alloys, particularly 6063, are widely used for extrusion components in the automotive industry.
In 6xxx Al alloys, Mg and Si are the major alloying elements which trigger the formation of Mg 2 Si for low temperature age-hardening [1].In general, Si can increase the fluidity in the molten state and further the castability, improve the abrasion resistance (hard Si particles), reduce the thermal expansion coefficient, but decrease the machinability [1,2].Mg can improve the corrosion resistance and weldability [2].Cu has appreciable solubility and strengthening effect [1][2][3].The addition of Cu in 6xxx Al alloys can produce substantial solution and precipitation strengthening effects by forming Al 2 Cu, Al 2 CuMg, and Q (Al 3 Cu 2 Mg 9 Si 7 ), improve the machinability, but decrease the corrosion resistance.Cu may also have effect on the fluidity [4].
The as-cast billet of 6xxx series Al alloys cannot always satisfy the subsequent extrusion process due to the harmful precipitates formed during casting.For instance, the size and distribution of Mg 2 Si critically determine the extrusion speed, and the grain boundary particles determine the surface finish [5].Fe impurity forms Fe-bearing intermetallic phases, such as Al 13 Fe 4 , α_AlFeSi, β_AlFeSi, Al 8 FeMg 3 Si 6 , etc., since the solubility of Fe in Al solid solution is very low.Al 13 Fe 4 forms only in Al alloys with low Si content [6].β_AlFeSi is highly faceted and poorly bonded to the matrix.This phase has brittleness and abrasive nature, and always forms at grain boundaries and interdendritic regions during casting.β_AlFeSi is claimed to cause poor ductility and severely reduces the hot workability of Al alloys during extrusion process.The α_AlFeSi phase with compact Chinese script morphology and nonfaceted interface with Al matrix can improve the ductility and extrudability of Al alloys.The α_AlFeSi phase is also known to have no pickup defect during extrusion and can improve the properties and surface finish of the final products [7].The unfavourable β_AlFeSi phase in the commercial 6063 alloys can transform into the less detrimental α_AlFeSi by prolonged heat treatment process [8].However, the production cost will be increased.
Suitable amounts of Cr and Mn have been used to modify the microstructure.Dispersoid phases like Al 7 (Cr, Mn), Al 4 (Cr, Mn), Al 18 Mg 3 Cr 2 , and Al 28 Cu 4 Mn 7 [9] can form during high temperature annealing (400-J.Min.Metall.Sect.B-Metall.54 (1) B (2018) 119 -131 550 °C), and have strong influence on the recovery, recrystallization, and growth and result in a fine grain size.However, the introduction of Mn and Cr in Al alloys containing Si can dramatically increase the amount of coarse α_Al(Fe, Mn, Cr)Si phase which induces a rapid decrease of the alloy properties [10].
The purpose of the present study is the application of the CALPHAD-type [11][12][13][14][15] thermodynamic database to the design of 6xxx Al alloy.So far, no accurate thermodynamic analysis has been conducted for 6xxx Al alloys.The only relevant study for the application of the CALPHAD-type of database was carried out by Sarafoglou and Haidemenopoulos [16].They mapped the phase fractions of Mg 2 Si and β_AlFeSi for the solidification microstructure.In the present study, systematic thermodynamic analysis for the effects of common alloying elements such as Cu, Fe, Mn, and Cr, on both the solidification and equilibrium precipitates in Al 6xxx alloys was carried out using the CALPHAD-type database, and optimum compositions of high strength Al 6xxx alloys due to Mg 2 Si precipitation hardening were discussed considering the casting, homogenization, extrusion, and final annealing processes.This study is part of a large Integrated Computational Materials Engineering (ICME) project to develop a new high strength Al front rail for automotive applications.

Thermodynamic database
There are several commercial CALPHAD-type thermodynamic databases available; FactSage-FTlite (www.factsage.com),ThermoCalc-TCAL4 (www.thermocalc.com),and Pandat-PanAluminium (www.computherm.com).Although all these databases are generally applicable for Al alloy calculations, the accuracy of these databases for specific alloy design varies because of the history of each database.
In the present study, all the CALPHAD-type thermodynamic calculations were carried out using the FactSage software [17].The thermodynamic descriptions of all the solution and stoichiometric phases related to the Al-Mg-Si-Cu-Fe-Mn-Cr system developed by the present authors were used for the thermodynamic calculations.In this database, 21 binary systems, 18 ternary systems, and one quaternary system were optimized.A detailed list of the systems optimized in the database is shown in Table 1.Other ternary and quaternary systems were treated as ideal systems (No interaction parameters for solution phases were added).
Unfortunately, there are no comprehensive experimental data for the equilibrium phase fractions or thermodynamic properties of the solid and liquid phases in multicomponent Al 6xxx alloys.However,

results and discussions
Wrought heat treatable 6xxx Al alloys for extrusion applications are produced by a sequence of metallurgical processes.Cast billet ingots are normally homogenized at 450-550 o C depending on alloy composition.Extrusion is carried out at 400-500 o C followed by air quenching.Age-hardening is carried out at 150-200 o C. In the present work, solidification calculations were carried out under the Scheil-Gulliver assumption: the diffusivities of alloying elements in the solid phases were assumed to be zero and in the liquid phase were assumed to be infinite.This assumption has certain rationality since the diffusivities of alloying elements in the liquid phase are several orders larger than those in the solid state.Equilibrium phase fractions of precipitates were calculated at 500 °C and 177 °C to represent the homogenization stage and final age-hardening stage, respectively.

Effect of alloying elements on the solidification microstructure
To investigate the effects of major alloying elements on the as-cast microstructure, the Scheil calculation was performed to compute the amount of each type of precipitate.In the calculation, the composition of each element was changed from the baseline alloy composition of Al 6063.
All the computed results are shown in Fig. 1.It should be noted that open symbols refer to the left vertical axis in logarithm scale, and filled symbols to the right vertical axis in linear scale.It can be seen from Fig. 1(a) that Cr can effectively induce the formation of Al 7 Cr when its content is above 0.12 wt.%.Although there is a slight increase in the weight fractions of α_AlFeSi, β_AlFeSi, Mg 2 Si, Si, and Al 8 FeMg 3 Si 6 , and a slight decrease in the fractions of Al 14 Fe 3 , Al 2 Cu, Al 4 (Cr, Mn), Al 28 Cu 4 Mn 7 , and Q, they are less significant.Fig. 1(b) shows the effect of Cu.The weight fractions of Cu bearing phases such as Al 2 Cu, Al 28 Cu 4 Mn 7 , and Q increase with the increasing Cu content.Since the formation of Q substantially consumes Si, the amounts of Mg 2 Si and Si decrease with Cu content.Other phases like α_AlFeSi, β_AlFeSi, Al 4 (Cr, Mn), Al 8 FeMg 3 Si 6 , and Al 13 Fe 4 show no significant difference until Cu content reaches 0.19 wt.%.As can be seen in Fig. 1(c), β_AlFeSi is forming as soon as the Fe addition to the alloy system.When the amount of β_AlFeSi reaches 0.15 wt.% (at 0.05 wt.% Fe), α_AlFeSi begins to form.It is at 0.283 wt.% α_AlFeSi (at 0.15 wt.% Fe) that the Al 13 Fe 4 phase starts since Si is not available anymore .The amounts of other phases show little difference with Fe content.The change in Mg content varies the precipitates in a more complex manner.According to the calculation in Fig. 1(d), the amount of α_AlFeSi first increases and then decreases till 0.9 wt.% of Mg.The amount of Al 14 Fe 3 increases till Mg content reaches 0.9 wt.% and then it starts to decrease.The amounts of β_AlFeSi, Al 4 (Mn, Cr), and Si decrease as Mg content increases.On the other hand, the amounts of Al 2 Cu, Mg 2 Si, Al 28 Cu 4 Mn 7 , and Q increase.The amount of Al 8 FeMg 3 Si 6 firstly decreases and then increases with increasing Mg content.According to the current calculation in Fig. 1(e), Mn has only a slight influence on the amounts of different precipitates.With the increasing Si content, the amount of Al 14 Fe 3 dramatically decreases due to the increased formations of α_AlFeSi, β_AlFeSi, Q, and Al 8 FeMg 3 Si 6 .However, when the Si content reaches 0.45 wt.%, the α_AlFeSi amount starts to decrease because the composition of the alloy shifts to the β_AlFeSi phase region.Other phases show little changes with Si content.
The calculated results are summarized in Table 2.In Table 2, increase of phase amount is indicated by "↑" and decrease is indicated by "↓".As discussed above, the change of phase amount is not always straightforward due to the interplay of the alloying elements for the formation of different precipitates.With the summary in Table 2, one can select the suitable alloy composition which can give the desired as-cast microstructure.The results of the ascast microstructure for three 6xxx Al cast billets (with composition listed in Table A1 in Appendix) are summarized in Table A2 in Appendix.The overall amounts of precipitates in as-cast microstructures are well calculated using the present database.2018

Equilibrium phase fraction map at 500 o C
The variation of equilibrium phase distribution of Al 6063 alloy with temperature is calculated in Fig. 2. Various precipitate phases are presented.At high temperature, Al 13 Fe 4 , α_AlFeSi, and β_AlFeSi are formed first.Then, Al 13 Fe 4 and α_AlFeSi are dissolved into Al as temperature decreases, and Al 8 FeMg 3 Si 6 and Mg 2 Si form instead.At low temperature, Al 12 Mn, α_AlMnSi, and Al 2 Cu are also formed.This equilibrium phase distribution calculation can give an idea for the homogenization temperature.Because the purpose of homogenization is to dissolve the precipitates and uniform the matrix composition, the material approaches the equilibrium condition.Unfortunately, there is no temperature zone to completely dissolve all the precipitates in this case.Theoretically, to get rid of the harmful β_AlFeSi phase and minimize the precipitates, homogenization can be carried out at 590-610 °C.But this is too close to the melting temperature of Al 6063.Therefore, 460 to 500 °C can be taken as the homogenization temperature to reduce the amounts of precipitates in the as-cast microstructure.The amount of total precipitates of Al 6063 alloy (with a nominal composition listed in appendix Table A1) in the as cast microstructure is calculated to be about 0.86 wt.% as listed in appendix (see Table A2).This can be further reduced in the homogenization treatment.
The suitable extrusion temperature should be lower than the homogenization temperature.The required condition of Al alloy for extrusion includes relatively low precipitate amounts (particularly, β_AlFeSi) and high temperature for a fast diffusion of the alloying elements.In this case, it can be between 450 and 500 °C.
In the current work, equilibrium phase fractions of precipitate phases at 500 °C (the possible homogenization or extrusion temperature) were mapped in the composition range of 0-0.7 wt.% Si and 0-1.1 wt.% Mg.The effects of Cu, Fe, Mn, and Cr on the precipitate amounts were also investigated.Fig. 3(a) presents the calculated phase diagram of the Al-Mg-Si alloy with 0.05 wt.% Cu, 0.15 wt.% Fe, and 0.03 wt.% Mn at 500 °C.The equilibrium phases at a given composition can be directly read from the diagram.The stable phase region of each equilibrium phase can be defined by its zero phase fraction lines.Based on this concept, the phase diagram can be divided into 5 parts, as shown in Fig. 3(b).The red line to the up-right corner of the diagram, named region 1, is the Mg 2 Si formation region.The green line to the top of the diagram, named region 2, is where Al 8 FeMg 3 Si 6 forms.Above the blue line is the β_AlFeSi phase formation region, named region 3.The α_AlFeSi phase is only stable between the two orange lines (region 4), and the Al 14 Fe 3 phase forms below the black line (named region 5).In the present study, the phase fraction of each precipitate was calculated based on the five phase areas.
Four series of phase fraction maps were calculated at 500 °C to show the effects of Cr, Cu, Fe, and Mn on the phase fraction of each precipitate phase.Fig. 4 shows the effect of Cr on the different precipitate phases.The Cr concentrations of 0.02 wt.% and 0.20 wt.% were selected in the calculations, and the concentrations of Cu, Fe, and Mn were fixed at 0.05, 0.

Discussion
From the above Scheil and equilibrium calculations, the effect of major alloying elements on the as-cast microstructure and equilibrium microstructure at homogenization and in extrusion process can be understood.This information can be very helpful in new alloy design.
In the extrusion process, for example, the amount of β_AlFeSi is considered as a major constraint due to its brittle feature.To minimize the harmful effect of β_AlFeSi, we can choose suitable alloy composition to control the as-cast and equilibrium amount of β_AlFeSi, and employ the homogenization temperature to let the β_AlFeSi transform to α_AlFeSi.According to Table 2, increasing Mg content and decreasing the contents of Si and Fe can effectively decrease the amount of β_AlFeSi in ascast microstructure.On the other hand, the equilibrium amount of β_AlFeSi at 500 °C can be reduced by increasing Cu and decreasing Cr, Fe, Mn, and Si.An alternative way is to choose suitable homogenization procedure.As discussed before, the β_AlFeSi phase can completely transform to α_AlFeSi, if the alloy is homogenized at 590 °C for a certain time as can be seen in Fig. 2. But this may be less applicable in the real practice.

S. Cui et al. / JMM 54 (1) B (2018) 119 -131
Another important issue considered in high strength 6xxx series Al alloy design is the formation of maximum amounts of metastable β' and β" phases at the age-hardening stage.However, the Gibbs energies of various metastable phases in the 6xxx series Al alloys are currently not implemented.Instead of calculating the amounts of metastable β' and β" phases directly, the present work studies the possible formation of Mg 2 Si.The equilibrium amount of Mg 2 Si can be an indicator of the ability of precipitation hardening by the metastable β' and β" phases which can be transformed to Mg 2 Si by prolonged heat treatment.The Mg 2 Si phase already formed before final heat treatment cannot be effective for precipitation hardening.Therefore, the difference of Mg 2 Si amount between the heat treatment temperature and extrusion/homogenization temperature can be a useful indicator for the precipitation hardening ability.In the present study, 177 °C, which is the typical ageing temperature of Al 6063 alloys, was used for calculation.The equilibrium amount of Mg 2 Si in Al-Mg-Si alloys with 0.02 wt.% Cr, 0.05 wt.% Cu, 0.15 wt.% Fe, and 0.03 wt.% Mn at 177 °C is calculated in Fig. 8.According to the calculated results, the Mg 2 Si phase has its maximum amount along the Mg/Si weight ratio of 1.731:1.Addition of Mg cannot increase the amount of Mg 2 Si, and the addition of Si can decrease the amount of Mg 2 Si formation.The red line indicates the limit composition for the Mg 2 Si formation at 500 °C, and the green line indicates the zero phase fraction line of Al 8 FeMg 3 Si 6 at 500 °C where the β_AlFeSi phase fraction begins to decrease rapidly.The green colored rectangule indicates the alloy composition range of commercial 6063 alloy, which is close to the Mg/Si ratio of 1.731:1.Even though there will be a certain amount of Mg 2 Si (<0.47 wt.%) formation at the homogenization temperature in the composition range, the ability of precipitation hardening is still not significantly decreased.Whent he Mg/Si ratio changes to 1:1, which contains excess amount of Si, the ability of Mg 2 Si formation shows a certain decrease compared to Mg/Si ratio of 1.731:1 at 177 °C.However, as there is no Mg 2 Si formation along the 1:1 ratio line at 500 °C, Al alloy with Mg/Si ratio of 1:1 can still produce a large amount of β' and β" precipitation hardening phases as indicated by Gupta et al. [42].
Considering the minimization of the β_AlFeSi phase, reasonably low amount of the Al 8 FeMg 3 Si 6 phase at extrusion/homogenization temperature (500 °C), and maximum amount of precipitation hardening ability at 177 °C, the most optimum alloy composition range can be suggested as shown in the hatched area in Fig. 8.Of course, the annealing time for the maximum precipitation hardening should be optimized by experiment.In addition, the effect of solutes, Mg, Si, Cu, and Mn, on the extrusion flow behavior and solid solution hardening should be further considered to narrow down the alloy composition for new high strength Al 6xxx alloys.The rectangular region (composition range of commercial 6063 alloy) in Fig. 8 is very wide, and can produce quite different amount of Mg 2 Si, which can significantly change the mechanical properties of the finally products.It should be noted that the design strategy can be altered if the thermodynamic descriptions of the metastable phases are available in the database.The amount of each type of metastable phase could be directly mapped.In that case, the prediction could be more applicable.

conclusion
In the current work, a newly developed thermodynamic database for Al alloys has been used for the comprehensive thermodynamic analysis of Al 6xxx alloy development.Influence of various alloying elements such as Fe, Mn, Cu, and Cr on secondary phase formation in the Al-Mg-Si alloy was studied using Scheil cooling calculation for as-cast microstructure and equilibrium calculation for The calculated phase fractions in the as-cast microstructure of all four alloys are listed in Table A2.It is obvious that the precipitates cannot form if S. Cui    there is no segregation in the liquid phase during solidification.The calculated liquid composition as a function of solidification is shown in Fig. A1(c).The compositions of Mg, Si, and Fe show strong segregation in the liquid phase during the solidification process.It can be seen from the diagram that all the other precipitates except for the fcc_A1 phase only start to form when solidification reaches almost 90 percent where the liquid phase is enriched with Mg, Si, and Fe.Similar to the 6063 alloy, the simulated solidification paths of the three as-cast alloys are presented in Fig. A2.The precipitation of these three alloys are quite similar.The Cr containing alloys (Alloys 1 and 3) show Al 7 Cr formation before the fcc_A1 phase.As shown in Table A2, the amounts of Q, Al 2 Cu, and Mg 2 Si in the three alloys are calculated apparently higher than those of 6063 alloy.
Fig. A3 shows the microstructure of the three cast alloys.Fig. A3(a) is the back scattered electron (BSE) micrograph of alloy 1.The Chinese script shaped α_AlFeSi, very narrow needle like β_AlFeSi, and spherical shaped Q phase were detected.Similar microstructure can be seen in Fig. A3(b) and Fig. A3(c) for alloys 2 and 3, respectively.It worth noting that the Al 8 FeMg 3 Si 6 particle with the similar shape as the Q phase was also observed in alloy 3. Other phases were not marked since the particles are very small.The primary Al 7 Cr phase was not clearly observed in the micrograph shown in Fig. A3.However, a lot of fine particles rich in Cr were detected in the Al matrix in the enlarged image of alloy 1 (Fig. A3(d)).The precipitates excerpt α_AlFeSi, β_AlFeSi, and Q were not detected under SEM in the present study.Most probably, the size of the precipitates is too fine to be detected.The α_AlFeSi phase has a composition of about Al-(9.5-14.7)wt.% Fe-(5.2-10)wt.%Si, and β_AlFeSi is Al-(21.9-23.9)wt.% Fe-(7.1-9.8)wt.% Si.ImagJ software was utilized to analyze the area fraction of the precipitates in Figs.A3(a) to (c).A large area is considered during the image analysis to give a reasonable representation of the microstructure.Since it was difficult to separate the bright phases α_AlFeSi, β_AlFeSi, and Q, a total area of precipitates was measured.The volume fraction was calculated from the area fraction assuming particles are spherical.The measured total volume fractions of Q, Al 8 FeMg 3 Si 6 , α_AlFeSi, and β_AlFeSi are listed in Table A3 compared with the Scheil simulation results.In the Scheil simulation results, only Q, Al 8 FeMg 3 Si 6 , α_AlFeSi, and β_AlFeSi were computed for comparison.Molar volumes of all the phases were assumed to be the same.It can be seen from Table A3 that the experimental values are quite close to the simulated results.It should be noted that the Scheil calculation always overestimates the amounts of precipitates than reality due to its extreme assumptions.And the measured total volume of precipitates as listed in Table A3 also have considerable experimental errors.This is because the number of sections measured is insufficient and the measurement method itself is not advance enough.But, it can be concluded that the present simulation results are reasonable.

Figure 2 . 4 ,Figure 4 .
Figure 2. Calculated equilibrium phase distribution of 6063 alloy as a function of temperature.

Fig. 6 (
Fig.6(c), the amount of Mg 2 Si can be increased in the Si rich region and decreased in the Mg rich region.Fig.7compares the calculated results of 0.03 wt.% Mn with 0.1 wt.% Mn.The amounts of Al 13 Fe 4 , Mg 2 Si, Al 8 FeMg 3 Si 6 , α_AlFeSi, and β_AlFeSi change slightly but negligible with the variation of Mn.At 0.1 wt.% Mn, the Al 12 Mn phase becomes stable in the entire composition range investigated.

Figure A1 .
Figure A1.(a)Calculated solidification path of 6063 alloy, (b) calculated phase fraction, and (c) calculated micro-segregation of alloying elements in the liquid phase.

Table 2 .
) 119 -131 Effect of alloying elements addition on the secondary phases formation during solidification.