1Struers Inc., 2887 N. Southern Hills Drive, Wadsworth, IL 60083-9293, USA

2Materials Science and Metallurgical Engineering Department, The University of Oviedo, Gijon Polytechnic School of Engineering, Viesques University, Gijon, Asturias 33203, Spain

3Materials Science and Metallurgical Engineering Department, The University of Oviedo, The School of Mines, Oviedo, Asturias 33004, Spain

Abstract: The present study investigates the effect of the solidification strategy for AA 6063 alloy on the surface appearance of anodized extrusions. The microstructure of the samples was analyzed using both light optical microscopy and scanning electron microscopy. Results show that if heavy segregation occurs from rapid solidification, coarse Mg2Si particles form, thus reducing the potential for precipitation strengthening by the finer b-Mg2Si developed in the solid state. Differentially-strained regions formed during hot extrusion induce differences in particle size for magnesium silicide (Mg2Si) precipitates. Anodizing generates surface roughness due to Mg2Si particle dissolution and AlFeSi decohesion, which is related to both particle size and deformation. During anodizing, an oxide layer forms on the surface of the extruded products, which can lead to streak formation, usually a subject of rejection due to unacceptable heterogeneous reflectivity.

Key words: light optical microscopy, scanning electron microscopy, quantitative metallographic characterization, 6063 aluminum alloys, precipitation, streak formation

INTRODUCTION

Al-Mg-Si alloys are mainly used in the manufacture of extruded components. This group of aluminum alloys has an attractive combination of properties. They present a very good combination of medium strength, good formability, high corrosion resistance, good weldability, and good machinability. They are widely used in various industry sectors such as construction, transport, marine, heating, etc. (Gupta et al., 2001; Esmaeili et al., 2003; Chakrabarti & Laughlin, 2004; Tsao et al., 2006; Cai & Cheng, 2007; Meyveci et al., 2010) . Extrusion formability and properties depend on their chemical composition and degree of homogeneity (Jackson & Sheppard, 1997; Gavgali & Aksakal, 1998). However, they are not suitable for forming by extrusion, rolling, etc. in the as-cast condition and must be homogenized due to the excessively low ductility they exhibit, mainly attributable to the presence of b-AlFeSi intermetallic particles formed at grain boundaries and interdendritic regions during solidification. They usually have an acicular morphology, sometimes referred to as needle-like. Homogenization heat treatment promotes the transformation of these needles into smaller, rounded a-AlFeSi particles, thus improving the material's ductility. Homogenization will also cause a reduction in microsegregation of solute atoms and partial solution of Mg and Si from preexisting Mg2Si particles. Fine magnesium silicide precipitates (b-Mg2Si) are responsible for the potential strength of the alloy (Couto et al., 2005; Al-Marahleh, 2006; Gaber et al., 2007). The homogenization of 6063 alloys consists in an annealing at temperatures in the range of 520-600°C for a few hours. The billets are then cooled to room temperature with forced air. The homogenized bars are subsequently transported to the extrusion plant, where the bar is cut into slugs of known weights. Each slug is then preheated prior to deformation, and once a prefixed temperature and time have been reached, it is forced into the extrusion press. The extrusions may be optionally hyperquenched and aged, depending on the extrusion thickness. Finally, they are subjected to the anodizing process (Lassance et al., 2007). During the whole manufacturing process, Mg2Si particles and AlFeSi intermetallics will change shape, size, and distribution. In the present study, extrusion tests were conducted to observe the effect of Mg2Si and AlFeSi particles on the surface characteristics and appearance of the anodized extruded surfaces.

MATERIALS AND METHODS

Samples were studied in the as-heat-treated condition after extrusion and anodizing. The alloys correspond to the 6063 series, according to the standard provided by the American Aluminum Association. Their chemical compositions are given in Table 1.

Two ingots of 178 mm in diameter, one of each alloy, were treated in a gas-fired homogenizing furnace. The homogenization temperature was set at 520°C for 5 h (Tokit et al., 2004; Al-Marahleh, 2006). Ingots were cooled, after homogenizing, first in still air and then under accelerated conditions in a forced air convection cooling chamber. Homogenized samples were cut from heat-treated ingots to conduct metallographic characterization. Samples were extracted from the periphery, the intermediate zone, and the center of such ingots. Observations were made under light optical microscopy (LOM) and scanning electron microscopy (SEM) techniques. To facilitate the identification of Mg2Si particles and AlFeSi intermetallics, polished surfaces were etched in an aqueous solution of 0.5% HF (Mulazimoglu et al., 1996; Vander Voort, 2000, 2004).

Measurements of Mg2Si and AlFeSi particles were obtained following quantitative metallographic techniques (Vander Voort, 1984; Muirhead et al., 2000; Higginson & Sellars, 2003). Determinations of the volume fraction and of the number of particles per unit area were carried out by the manual point counting method (ASTM E562-11, 2011). At least 25 micrographs were analyzed in each of the three examined zones in order to obtain a reasonable accuracy (Vander Voort, 1994). Observations were performed with a Nikon Epiphot metallographic bench connected to a Kappa ImageBase image analyzer.

Suitable amounts of the commercially produced billets, cut into slugs, were retained for subsequent extrusion. The slugs were preheated to around 490°C, prior to deformation, whereas the extrusion dies were heated to 450°C. The extrusion speed was set at 18 m/min so as to achieve a controlled exit temperature of the extruded product in the range of 520-550°C. Experimentally extruded sections had thin walls (<3 mm) and were cooled in still air. Samples were cut from the extrusions, prepared by mechanical processes, and etched in an aqueous solution of 0.5% HF for LOM observation. The extrusions were then process anodized under identical conditions: an alkaline etching stage conducted at 70°C in a NaOH solution for 10 min was applied prior to the anodizing stage. A desmutting stage right after alkaline etching was applied to remove insoluble residues from etching and deposited on the extrusion surfaces was conducted. The anodizing stage follows and was carried out in sulfuric acid at room temperature over a period of 40 min. Again, samples were cut from both at the web zone and the contiguous zone. These samples were first mechanically polished and then HF-etched for LOM observation. Additionally, samples in the anodized state were observed using a JEOL-5600 electron microscope (JEOL Ltd., Tokyo, Japan) without any mechanical or chemical preparation.

RESULTS

Surface Appearance of the Aluminum Extrusions

The extrusions from the aluminum alloys under study present different surface appearances. Figure 1 shows the surface of the anodized samples. Sample K exhibits bright streaks alternating with dull bands. These streak defects are only visible on the extruded surface above the web zone and reflect light making these zones to appear brighter than the contiguous dull ones. Sample B has a uniform brightness appearance resembling that of the streak zones in sample K. Figure 1 depicts the zones and codes that will be further studied. Four different zones were analyzed. The surfaces exhibiting bright streak defects above webs are denoted as KW , which stands for the web zone in sample K. The equivalents zones in sample B are referred to as B W . Regions adjacent to the web zone are denoted as K N and BN , where the subscript N stands for normal region.

Microstructure of Homogenized Billets

Figure 2 shows the microstructures of alloys K and B in the homogenized condition observed under LOM after etching the mechanically prepared surfaces in a reagent consisting of 0.5% HF aqueous solution to reveal the nature of second phase precipitates. At least three types of precipitates were found. Primary precipitates mainly consist of high melting temperature elements such as Fe, Mn, and Si (gray phase). These are homogeneously distributed along grain boundaries and triple points. Coarse Mg2Si precipitates (black phase) were detected in both samples. The micrographs show the presence of fine b-Mg2Si precipitates, preferentially present inside the grains. It can be seen that alloy B (Fig. 2b) presents a higher percentage of this type of precipitate. Furthermore, coarse Mg2Si precipitates can be observed in both samples, preferentially in the grain boundaries.

Figure 3 presents the quantitative determinations of the volume fraction and number of particles per unit area of coarse Mg2Si precipitates and AlFeSi intermetallics. Measurements were carried out in the three selected zones of the cross section of the homogenized ingots.

Microstructure of the Extruded Products

Figure 4 shows the cross-sectional microstructures of the aluminum extrusions both at the web zones and at the normal zones too, after etching in 0.5% HF aqueous solution. Fe-rich precipitates are observed in light gray, whereas coarse Mg2Si precipitates appear in black.

Microstructure of Anodized Extrusions

Figure 5 shows the LOM microstructures corresponding to the surface of the anodized extrusions after removal of most of the anodized layer by mechanical polishing, followed by light aqueous HF etching. It can be seen that precipitation in alloy B is similar and homogenously distributed in all zones (Figs. 5c, 5d). In alloy K, however, precipitation in the web zones (Fig. 5a) is more abundant than in the other zones, where it barely occurs (Fig. 5b).

Figure 6 shows the surface irregularities of samples in the anodized condition without further mechanical preparation or chemical etching, as observed directly in the scanning electron microscope. Etch pits and grain boundary grooves are evident in both alloys. The morphology, size, and depth of etch pits describe the surface topography of the extruded surface and relate to the occurrence of Mg2Si and AlFeSi particles observed in Figures 4 and 5.

DISCUSSION

In Al-Mg-Si alloys, Mg and Si in solution combine in the presence of aluminum to form Mg-Si particles, which are the primary hardening phase of 6xxx alloys when precipitation occurs in the solid state. However, Mg2Si starts precipitating in the solid + liquid stage during solidification. This precipitation sequence has been reported as (Edwards et al., 1998)

L r a-Al + b-AlFeSi + Si (587°C/578°C) (1)

L + a-AlFeSi r a-Al + b-AlFeSi + Mg2Si (576°C) (2)

Supersaturated solid solution

Clusters of Si atoms and clusters of Mg atoms r

GP zones

Intermediate precipitate b''

Intermediate precipitate b '

Equilibrium phase b-Mg2Si. (3)

Mg2Si is thus a ternary peritectic reaction product according to reaction (2). But, when represented on a pseudoeutectic binary diagram, Al-Mg2Si is usually referred to as a eutectic product (Fig. 7). Such eutectic corresponds to 13.9 w t% of Mg2Si.

The aluminum alloys investigated in this study contain 0. 44 wt% Mg and 0.45 wt% Si for alloy K and 0.51 wt% Mg and 0.45 wt% Si for alloy B (Table 1). Assuming the stoichiometric ratio, Re, of Mg2Si precipitates to be

Re(Mg2Si) = 2 X Aw(Mg)/Aw(Si) = 1.73, (4)

Where Aw(Mg) and Aw(Si) are the atomic weights of Mg and Si, respectively. Given the weight percents of Si and Mg in alloys K and B, the above ratio can be determined to a first approximation by

Rk (Mg2Si) = 0.99

Rb (Mg2Si) = 1.13, (5)

which denotes that both compositions are substoichiometric with regard to the Re(Mg2Si) ratio above defined, i.e., all Mg combines to give Mg2Si.

Si is present in the as-cast condition as Mg2Si particles and AlFeSi intermetallics. Considering that all the Mg is present as Mg2Si precipitates, we can thus calculate the total weight percentage of silicon as Mg2Si precipitates and the total amount of Mg2Si expressed in weight percent (Mrowka- Nowotnik, 2010) as

[Si(Mg2Si)] = 0.58 X [Mg] inwt% (6)

[ Mg2Si] = [Mg] + [Si(Mg2Si)] in wt%. (7)

The maximum weight percentage of Mg2Si precipitates in K and B alloys is 0.69 and 0.80, respectively. Furthermore, the volume fractions of the potential b-Mg2Si hardening phase calculated from the weight percentage of such precipitates (assuming the specific weight of Mg2Si to be 1.99 g/cm3 and that of the 6063 alloy ~ 2.69 g/cm3 ) (ASM Specialty Handbook, 2002) are 0.93 vol% for alloy K and 1.08 vol% for alloy B.

Figure 7 shows the equilibrium pseudobinary Al-Mg2Si phase diagram (Zhang et al., 2001), illustrating the possibility of formation of coarse eutectic Mg2Si under nonequilibrium conditions due to accelerated cooling typical of 6063 semicontinuously cast billets. The degree of displacement of the solidus line (Callister, 2000) determines the length of the AO segment, and thus of the corresponding weight fraction, WW, of nonequilibrium eutectic phase formed [given approximately by WW (Eutectic) = AO/EO]. The eutectic is composed of a-Al and Mg2Si, and their respective weight fractions depend on the total amount of eutectic formed. Hence, the bigger the fraction of the coarse intergranular eutectic Mg2Si, the lower is the amount of b-Mg2Si available for precipitation strengthening of the a-Al matrix. Also, depending on the cooling rate in the continuous casting mold, microsegregation in a-Al grains could increase, enabling the formation of a cored structure in a-Al dendrites. In view of the above, two fractions of Mg2Si could be expected at room temperature:

[Mg2Si-total] = [Mg2Si-coarse] + [b-Mg2Si] in wt%. (8)

Coarse Mg2Si precipitates form under nonequilibrium solidification conditions due to rapid cooling rates. According to the aforementioned calculations, equations (6) and (7), an average value of 0.7 wt% of Mg2Si precipitates can be assumed for both alloys. Homogenization takes place at a temperature of 520°C above the equilibrium precipitation temperature, Tp (Fig. 7), sufficient to dissolve the finer fraction of the b-Mg2Si precipitates. However, it is insufficient to dissolve the coarse fraction of these precipitates.

An evidence of the process of nonequilibrium solidification is the macrosegregation phenomena occurring in alloys K and B, which can be appreciated in the results of the particle volume fraction presented in Figure 3. This constitutes inverse radial-type segregation. Despite the fact that the Mg- and Si-enriched liquid able to start Mg2Si precipitates is formed at a certain distance from the ingot wall, it undergoes exudation of such enriched liquid through the weak a-Al grain boundaries toward the periphery regions, due to high metallostatic pressure typical of vertical castings. This allows the opening of certain grain boundaries so that this liquid can migrate to the subcutaneous grains located at the ingot skin, forming at these sites the highest volume fractions of coarse Mg2Si precipitates (Gruzlesky, 2000) .

Determinations carried out using quantitative metallography techniques with LOM micrographs only allow us to assess coarse Mg2Si precipitates. The results obtained for both samples (Fig. 3) were, on average, 0.92 vol% for alloy K and 0.56 vol% in alloy B after heat treatment. These results indicate a greater presence of the ternary peritectic Mg2Si precipitates in alloy K in the homogenized state. According to the principles of nonequilibrium solidification (Gruzlesky, 2000) , this is indicative of a more rapid rate of solidification of alloy K. The total theoretical fraction of Mg2Si precipitates is similar in both materials. Moreover, as the fraction of coarse Mg2Si precipitates in the homogenized material K is almost double that in material B, a percentage of fine b-Mg2Si precipitates can be expected, given by


[b-Mg2Si]K = 0.93 - 0.92 ~ 0.01 vol%

[b-Mg2Si]B = 1.08 - 0.56 ~ 0.52 vol%. (9)

In conclusion, this fraction may be considered negligible for material K, as it could be seen in Figure 2a, where no traces of fine precipitates can be observed within a-Al grains. The finer fraction of b-Mg2Si precipitates of alloy B is one order of magnitude higher than that of alloy K as calculated in equations (9). These b-Mg2Si precipitates occur both by homogeneous precipitation in the bulk of a-Al grains and heterogeneously nucleated at the preexisting coarse Mg2Si particles. The later are favorable sites for nucleation because both nucleant and nuclei possess identical crystal lattice (Gruzlesky & Closset, 1999). Furthermore, it should be noted that the homogeneously nucleated fraction is the only one with the possibility of redissolving during the short time of dwell at the heating temperature prior to extrusion and constitute the precipitation fraction capable of providing the most relevant age hardening in the extruded product. However, higher temperatures are needed to dissolve the b-Mg2Si fraction heterogeneously nucleated at preexisting coarse Mg2Si due to its lower surface-to-volume ratio. Accordingly, the precipitates originating from material K can be expected to offer negligible precipitation hardening.

The extrusion process commences with the heating of the slug, the aim being to dissolve the Mg and Si originating from the fine b-Mg2Si precipitates. The finer the precipitates, the greater their surface/volume ratios and the faster they are brought into solution (Vermolen et al., 1998). The Mg/Si ratio influences solution and precipitation temperatures of b-Mg2Si. Increasing percentages of Mg in solid solution shift the precipitation temperature toward higher values (Rivas et al., 1999). Diffusivity is favored by high temperatures. According to this reasoning, the higher percentage of Mg in solid solution in alloy B, compared to that of alloy K, will facilitate the precipitation kinetics of b-Mg2Si in the extruded product, which is beneficial for the extruded mechanical strength. During aluminum extrusion, the dynamic softening mechanisms are only of the dynamic recovery type, as corresponds to high stacking fault energy alloys. The presence of coarse precipitates, however, restrains dislocation mobility, and this is shown by an increase in the stress level of the corresponding hot deformation s - e (stress-strain) curve (Fig. 8). As a consequence, alloy K with a higher volume fraction and particle count of insoluble Mg2Si particles and AlFeSi intermetallics (Fig. 3) will be harder and more difficult to extrude and point to the need to reduce the extrusion speed significantly to avoid unsuitable surface quality (Zhu et al., 2011). Consequently, the volumetric energy, U, given by the area under the corresponding stress-strain curve, is greater for alloy K than it is for alloy B:

U(K) > U(B). (10)

Assuming the true deformation e at the web to be bigger than that at its contiguous zone (normal zone) (Zhu et al., 2011) , we can write

«W > «N. (11)

Then from Figure 8,

Area(OA'K'C') > Area(OA'KC) r U(KW) > U(KN). (12)

In summary, the volumetric energy in alloy K at the web zone (KW ) will be the highest.

For a frictionless deformation process in the absence of heat transfer (quasi-adiabatic process), the maximum increase in temperature is given by (Dieter, 1988)

Hence, a greater increase in the quasi-adiabatic temperature, Atq_a , can be expected in sample K. This temperature rise, Atq_a , is added to the preheating temperature, Tpreheat,, thus resulting in a higher capacity for solution of b-Mg2Si precipitates.

The higher strain at the web zones, compared to normal zones, causes thermal gradients to be developed (Tabrizian et al., 2010). Accordingly, within the extruded bars, Atq_a is highest in the intensely deformed zones, i.e., the KW regions. We can state that

Moreover, this temperature, T (KW ), may be sufficient to ensure partial dissolution of the Mg2Si heterogeneously nucleated at coarse Mg2Si in these zones. Cooling rates higher than 60°C/min (Ali et al., 2011) allow for late in situ precipitation of the prior fraction of b-Mg2Si. In agreement with the former, the microstructure presented in Figure 4a corresponding to the web zone in sample K shows more abundant b-Mg2Si than in the normal zone (Fig. 4b) The precipitation in sample B (Figs. 4c, 4d) is similar in all zones. This indicates that Atq_a is insufficient to achieve solution of the b-Mg2Si precipitates heterogeneously nucleated and thus the absence of further reprecipitation in the B W zone.

During anodizing, aluminum is converted via an electrochemical reaction into a porous oxide film of Al2O3 . This transparent layer is several hundred times thicker than the natural oxide. Light is scattered, refracted, and reflected differently from this film depending on the compositional variations, incorporated particles, and microscopic roughness of the metallic substrate surface. However, during the alkaline etching process, prior to anodizing, further surface defects are also generated due to uneven chemical attack on the microstructure. Etching pits are one of the most common surface defects created during this etching process. They are created on the extrusion surfaces due to different reaction rates between the inclusions or intermetallic particles, and the aluminum matrix (Zhu et al., 2010). The Mg2Si precipitates have a lower electrochemical potential than the aluminum matrix. During etching, these precipitates act as anodes. Accordingly, these particles dissolve with preference with respect to the aluminum matrix. This leads to the formation of deep etch pits with sizes that vary depending on the original size of the Mg-Si particles (Figs. 5, 6). On the other side, AlFeSi intermetallics acts as cathodes with respect to the a-Al matrix in which they are embedded (Zhu et al., 2009). Figures 3c and 3d show that both volume fraction and number of particles per unit area of these intermetallics are higher in sample K than in sample B. Due to the difference in the electrochemistry, the a-Al matrix surrounding these particles dissolves away during etching, allowing decohesion of AlFeSi. Therefore, coarser pits are developed. Both deep and coarse etch pits contribute to the differences in the surface roughness. If the surface roughness is greater than 0.2 microns and concentrates in certain bands of the surface in the extruded sections, the difference in intensity and diffuse reflectance of light between these bands and the surrounding material may be large enough to be detected by the naked eye, and hence streak defects may be differentiated (Zhu et al., 2010). In agreement with the results of the quantitative metallography (Fig. 3), the observations in Figures 5 and 6 reveal that the differences in surface roughness between W-zones and N-zones are more pronounced in sample K than in sample B. The explanation for these observations is to be found in the density and size of the etch pits. The former results in the streaks being observable only in material K. Streaks in Figure 1 appear brighter in K W than in KN . The grains of the matrix in K W have flat zones free of etch pits that explain the higher reflectivity of these bands.

CONCLUSIONS

Too high cooling rates during solidification of 6063 alloys induce the formation of a greater volume fraction of ternary peritectic Mg2Si particles as assessed by quantitative metallographic techniques on LOM micrographs. This volume fraction increases with the rate of solidification in the casting process. The formation of these particles diminishes the possibility of structural hardening by b-Mg2Si precipitation in the solid state from a saturated a-Al matrix. On the other hand, a high presence of coarse Mg2Si and AlFeSi particles impedes dislocation mobility and dynamic recovery processes during hot extrusion. This entails an increase in the resistance to deformation by extrusion of the billet and the resulting increase in temperature of the billet during the forming process.

In a material in which virtually all the Mg2Si is present as coarse precipitates at the heating stage prior to extrusion, it is only possible to form b-Mg2Si strengthening precipitates by high temperature solutionizing apt for decomposing the heterogeneously nucleated b-Mg2Si. Temperatures conducive to this mechanism are achieved in the zones exposed to severe strain during the extrusion process. Subsequent reprecipitation takes place in the form of the b-Mg2Si compound homogeneously nucleated, as observed in anodized samples.

The formation of streak defects of different brightness in anodized extrusions is evident as a result of the different surface roughness between highly deformed zones and adjacent zones with lesser deformation. Dull areas were observed to correspond to coarse etch pits networks that overlap each other.

REFERENCES

ALI, A.H.A., HASSAN, J.M.G., MARTIN, G. & GHOSH, K. (2011). Development of alba high speed alloy. Light Metals 2011: Proceedings of the TMS 2011 Annual Meeting & Exhibition, S a n Diego, CA, Feburary 27-March 3, 2011, Lindsay, S.J. (Ed.), pp. 803-807. Warrendale, PA: Minerals, Metals, and Materials Society.

AL-MARAHLEH, G. (2006). Effect of heat treatment on distribution parameters and volume fraction of Mg2Si in structural Al alloy 6063. Am J Appl Sci 3, 1819-1823.

ASM SPECIALTY HANDBOOK (2002). Properties of wrought aluminum alloys. In ASM Aluminum and Aluminum Alloys, Davis, J. R., Davis & Associates (Eds.), pp. 686-687. Materials Park, OH: ASM International.

ASTM STANDARD E562-11 (2011). "Standard test method for determining volume fraction by systematic manual point count." West Conshohocken, PA: ASTM International. Doi:10.1520/ EO562-11.

CAI, M. & CHENG, G.J. (2007). Microstructure-properties relationship in Al-Mg-Si alloys subject to a combined process of extrusion and ageing. JOM 59, 58-61.

CALLISTER, W.D. (2000). Development of microstructurenonequilibrium cooling. In Materials Science and Engineering: An Introduction, Anderson, W. (Ed.), pp. 253-255. New York: John Wiley and Sons Inc. CHAKRABARTI, D.J. & LAUGHLIN, S.D. (2004). Phase relations and precipitation in Al-Mg-Si alloys with Cu additions. Prog Mater Sci 49, 389-410.

COUTO, K.S., CLAVES, S.R., VAN GEERTRUYDEN, W . H . , MISIOLEK, W. Z. & GONCALVES, M. (2005). Treatment effects of homogenization on microstructure and hot ductility of aluminum alloy 6063. Mater Sci Tech 21, 263-268.

DIETER, G.E. (1988). Fundamentals of metalworking. In Mechanical Metallurgy, pp. 524-526. London: McGraw Hill Book Co.

EDWARDS, G.A., STILLOER, K., DUNLOP, G.L. & COUPER, M.J. (1998). The precipitation sequence in Al-Mg-Si alloys. Acta Mater 46, 3893-3904.

ESMAEILI, S., WANG, X., LLOYD, D.J. & POOLE, W.J. (2003). On the precipitation-hardening behavior of the Al-Mg-Si-Cu alloy AA6111. Metall Mater Trans A 34, 751-763.

GABER, A., MOSSAD, A., MATSUDA, A . K . , KAWABATA, T., YAMAZAKI, T. & IKENO, S. (2007). Study of the developed precipitates in Al-0.63Mg-0.37Si-0.5Cu (wt%) alloy by using DSC and TEM techniques. J Alloy Compd 432, 149-155.

GAVGALI, M. & AKSAKAL, B. (1998). Effects of various homogenization treatments on the hot workability of ingot aluminum alloy AA2014. Mater Sci EngA 254, 189-199.

GRUZLESKY, J.E. (2000). Segregation phenomena. In Microstructure Development during Metalcasting, AFS (Ed.), pp. 117-131. Des Plaines, IL: American Foundrymen's Society Inc. GRUZLESKY, J.E. & CLOSSET, B.M. (1999). Grain refinement. In The Treatment of Liquid Aluminum-Silicon Alloys,AFS (Ed.),pp. 127- 142. Des Plaines, IL: American Foundrymen's Society Inc. GUPTA, A.K., LLOYD, D.J. & COURT, S.A. (2001). Precipitation hardening in an Al processes 0.4% Mg-1.3% Si-0.25% Fe aluminum alloy. Mater Sci EngA 301, 140-146.

HIGGINSON, R.L. & SELLARS, C.M. (2003). Worked Examples in Quantitative Metallography. London: Institute of Metals, Maney Publishing.

JACKSON, A. & SHEPPARD, T. (1997). Extrusion limit diagrams: Effect of homogenizing conditions and extension to productivity analysis. Mater Sci Tech-Lond 13, 61-68.

LASSANCE, D., FABREGUE, D., DELANNAY, F. & PARDOEN, T. (2007).

Micromechanics temperature fracture in 6xxx Al alloys. Prog Mater Sci 52, 62-129.

MEYVECI, A., KARACAN, I., CALIGULU, U. & DURMUS, H. (2010). Pin-on-disc characterization of 2xxx and 6xxx aluminum alloys aged by precipitation age and hardening. J Alloy Compd 491, 278-283.

MROWKA-NOWOTNIK, G. (2010). Influence of chemical composition variation and heat properties of 6xxx alloys. Arch Mater Sci Eng 46, 98-107.

MUIRHEAD, J., CAWLEY, J. & STRANG, A. (2000). Quantitative aspects of grain size measurement. Mater Sci Tech-Lond 16, 1160-1166.

MULAZIMOGLU, M . H . , ZALUSKA, A., GURZLESKI, J.E. & PARAY, F. (1996). Electron Al-Fe-Si intermetallics in 6201 aluminum alloy. Metall Mater Trans A 27, 929-936.

RIVAS, A . L . , MUNOZ, P., CAMERO, S. & QUINTERO-SAYAGO, O. (1999). Effect of the microstructure on the mechanical properties and surface finish of an extruded AA-6063 aluminum alloy. Adv Mat Sci Tech 2, 15-23.

TABRIZIAN, N . , HANSEN, H . N . , HANSEN, P.E., AMBAT, R. & MOLLER, P. (2010). Influence of annealing and deformation on optical properties of ultra precision diamond turned and anodized 6060 aluminum alloy. Surf Coat Technol 204, 2632-2638.

TOKIT, Y., GAVGALI, M., SALENDER, R. & KAYMAZ, I. (2004). The effect of the microstructural difference between surface and center of the workability of AA6063 homogenized ingot. J Adv Mater 36, 53-59.

TSAO, C.S., C H E N , C.Y., JENG, U.S. & KUO, T.Y. (2006). Precipitation kinetics and transformation of metaestable phases in Al-Mg-Si alloys. Acta Mater 54, 4621-4631.

VANDER VOORT, G.F. (1984). Metallography: Principles and Practice. New York: McGraw-Hill.

VANDER VOORT, G.F. (1994). Precision and reproducibility of quantitative measurements. In Quantitative Microscopy and Image Analysis, Diaz, D.J. (Ed.), pp. 21-34. Materials Park, OH: ASM International.

VANDER VOORT, G.F. (2000). Specimen preparation for image analysis. In Practical Guide to Image Analysis, ASM (Ed.), pp. 35-74. Materials Park, OH: ASM International.

VANDER VOORT, G.F. (2004). Metallographic techniques for aluminum and its alloys. In Metallography and Microstructures, vol. 9, Vander Voort, G.F. (Ed.), pp. 711-751. Materials Park, OH: ASM International.

VERMOLEN, F., VUIK, K. & VAN D E R ZWAAG, S. (1998). A mathematical model for the dissolution kinetics of Mg2Si phases in Al-Mg-Si alloys during homogenization under industrial conditions. Mater Sci EngA 254, 13-32.

ZHANG, J., FAN, Z., WANG, Y.Q. & Z H O U , B.L. (2001). Equilibrium pseudobinary Al-Mg2Si phase diagram. Mater Sci Tech 17, 494-496.

ZHU, H., COUPER, M.J. & DAHLE, A.K. (2011). Effect of process variables on Mg-Si particles and extrudability of 6xxx series aluminum extrusions. JOM 63, 66-71.

ZHU, H., ZHANG, X., COUPER, M.J. & DAHLE, A.K. (2009). Effect of initial microstructure of anodized aluminum extrusions. Metall Mater Trans A 40, 3264-3275.

ZHU, H., ZHANG, X., COUPER, M.J. & DAHLE, A.K. (2010). The formation of streak defects on anodized aluminum extrusions. JOM 62, 46-51.