Factors affecting thermal contraction behavior of an AA7050 alloy

Factors affecting thermal contraction behavior of an AA7050 alloy

Materials Science and Engineering A 527 (2010) 3264–3270 Contents lists available at ScienceDirect Materials Science and Engineering A journal homep...

2MB Sizes 0 Downloads 8 Views

Materials Science and Engineering A 527 (2010) 3264–3270

Contents lists available at ScienceDirect

Materials Science and Engineering A journal homepage: www.elsevier.com/locate/msea

Factors affecting thermal contraction behavior of an AA7050 alloy L. Zhang a , D.G. Eskin a,b,∗ , M. Lalpoor b , L. Katgerman a a b

Delft University of Technology, Department of Materials Science and Engineering, Mekelweg 2, 2628CD Delft, The Netherlands Materials innovation institute, Mekelweg 2, 2628CD Delft, The Netherlands

a r t i c l e

i n f o

Article history: Received 19 November 2009 Received in revised form 1 February 2010 Accepted 2 February 2010

Keywords: Aluminum alloys Casting Grain refinement Thermal contraction Hydrogen AA7050 alloy

a b s t r a c t The understanding of the contraction behavior in high-strength wrought 7XXX-series alloys is very important for the analysis of stress–strain development and modeling of hot tearing and cold cracking in direct-chill casting. The linear thermal contraction during and after solidification of an AA7050 alloy has been studied experimentally under different conditions. The experimental parameters included the sample height (metal level in the mold), grain refining, and gas content. The results showed that evolution of gas during solidification and pressure drop across the mushy zone are the main reasons for the preshrinkage expansion. The measured coefficient of linear thermal contraction decreases with the decreasing grain size at all temperature ranges. The role of gas precipitation and grain refining is discussed.

1. Introduction Direct-chill (DC) casting is the major technology for producing aluminum ingots and billets for further deformation processing. The products have been widely applied in the automotive, aerospace and construction industries. However, casting defects arising during and after solidification of metallic alloys constraint their applications. Among the most common problems that occur during casting are hot tears and cold cracks. The main cause of these defects is that stresses and strains built up during and after solidification are too high compared to the actual strength and ductility of the semisolid and as-cast material [1,2]. Cracking is particularly typical of DC casting of high-strength aluminum alloys, such as an AA7050 alloy. Generally, an AA7050 alloy appears to be extremely brittle in the as-cast condition at temperatures below 300 ◦ C [3]. Large solidification range and low thermal conductivity values [4] make the DC casting billets from this type of alloys vulnerable to both hot and cold cracking due to imposed large thermal gradients. Nonequilibrium solidification conditions, on the other hand, result in micro-segregation and formation of non-equilibrium phases (intermetallics) mainly on the grain boundaries and inter-dendritic space. Such brittle phases potentially provide favorable locations

∗ Corresponding author at: Materials innovation institute, Mekelweg 2, 2628CD Delft, The Netherlands. Fax: +31 15 2786730. E-mail address: [email protected] (D.G. Eskin). 0921-5093/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.msea.2010.02.005

© 2010 Elsevier B.V. All rights reserved.

for crack initiation and propagation. Despite considerable efforts that have been made to avoid the hot and cold cracking, this kind of defect is still a considerable problem in casting of high-strength aluminum alloys. In the manufacturing practice, grain refining, filtration and degassing are the common ways to alleviate the hot and cold cracking [1]. During casting, grain refiner is added and degassing is applied to achieve fine equiaxed grains and reduce the porosity, which can decrease the sensitivity to cracking and improve the mechanical properties of the as-cast metal [2,5,6]. But the specific effect of these methods on reducing the driving forces for cracking is still unclear and few reviews are available on this subject. As for the driving force, it is generally accepted that the inadequate feeding compensation of the shrinkage and solid contraction in the presence of thermal stresses caused by temperature gradients in the solidifying casting is the major reason for the occurrence of hot cracking [1,7]. As for the cold cracks, the accumulation and concentration of residual stresses in the casting during cooling after the end of solidification under conditions of high temperature gradients and low ductility is usually given as a main reason for the brittle failure [8]. Thermal contraction above and below the solidus is, therefore, the constitutive property that governs the formation of stresses and eventually the cracking of the casting. The understanding of the contraction behavior is very important for the analysis of stress–strain development and modeling of hot tearing and cold cracking. It was noticed that the apparent thermal contraction of the billet shell surrounding the liquid and semi-liquid sump during DC casting was larger than it could be estimated based on the coefficient of

L. Zhang et al. / Materials Science and Engineering A 527 (2010) 3264–3270

Fig. 1. Schematic view of the experimental setup. Dimensions are given in millimeters.

thermal expansion [9]. This happens because the internal volume of the billet, being semi-liquid and soft, offers little resistance to the thermal contraction of the external shell and the rate of contraction becomes higher as compared to the completely solid billet. The “additional” thermal contraction increases when the billet sump becomes deeper, i.e. with the increased casting speed and billet diameter [9]. These early observations drew attention to the relationship between the inhomogeneity of solidification development and the thermo-mechanical behavior of the real casting. In the past decade, an experimental technique for measuring the various contraction parameters has been developed and employed [10]. It was shown that the geometry of the sample and structure of the alloy affect the contraction behavior in the solidification range. It was also shown that the same technique can be successfully used for estimating the linear thermal contraction coefficient at high temperatures. This paper presents the results on the effect of such process parameters as grain refining and gas saturation on the contraction behavior of an AA7050 alloy at temperatures above and below the solidus. 2. Experimental procedure The experimental setup for measuring the linear contraction at temperatures above and below the solidus is described in detail elsewhere [10]. It is based on the idea suggested by Novikov [6] and comprises the following parts: a T-shape graphite mold with one moving wall, a water-cooled bronze base which provides high cooling rate that is similar to the conditions in DC casting, a linear displacement sensor (linear variable differential transformer (LVDT)), a K-thermocouple, and a computer-based data acquisition system. Fig. 1 shows the schematic view of the experimental setup. The cross-section of the main cavity is 25 mm × 25 mm with a gage length of 100 mm. The cross-section of the T-shape cavity is thinner than that of the main cavity, which allows the melt to solidify faster. So the solidifying sample can be fixed on this side. On the opposite side, a metallic rod is fixed in the moving wall to attach the solidifying metal, so the position of the moving wall can be detected by LVDT, which is accurate to 6 ␮m or 0.006% of the gage length. The temperature is monitored with a 0.1-mm thin, opentip K-thermocouple standing vertically in the center of the mold at about 1.5 mm above the mold bottom to avoid the problem of filling the gap between the thermocouple tip and the mold bottom. During the experiments, the temperature and displacement are recorded simultaneously by the data acquisition system. Based on the previous experience [10], a refractory paint (bone ash) was applied onto the internal walls in the central part of the

3265

Fig. 2. Examples of data obtained from experiments. lexp is the preshrinkage s−neq onset is the temperature of the contraction onset; lcontr is the amount expansion; Tcontr of contraction in the non-equilibrium solidification (NES) range, and TCC is range where the linear thermal contraction coefficient (TCC) is calculated.

mold, as shown in Fig. 1, to equalize cooling rates, reduce temperature gradient, and assure that the first contact between opposite coherent parts of the sample occurs in the area where the temperature is measured. The melt temperature was 720 ◦ C in all studied cases. After having acquired the primary data, the temperature was plotted versus time in order to derive the cooling curves from which the cooling rate and characteristic temperatures in the solidification range were obtained. After this, the data were reconstructed to find the direct dependence of displacement on temperature. Fig. 2 demonstrates an example of the contraction curve for an AA7050 alloy. From this curve, the temperature of the contraction onset, the preshrinkage expansion, the amount of contraction in the solidification range and the thermal contraction coefficient (TCC) at subsolidus temperature can be extracted directly. The preshrinkage expansion is calculated in the temperature range between the temperature of expansion onset and the temperature of contraction onset. The linear solidification contraction (the amount of contraction in the solidification range) is determined as follows [6,10]: εs =

ls + lexp − lf × 100%, ls

(1)

where ls is the initial length of the cavity (100 mm); lf is the length of the sample at the temperature of non-equilibrium (NEQ) solidus, and lexp is the preshrinkage expansion. It is worthy to note that at the used cooling rates (5–10 K/s), the solidification range was considered as NEQ, hence the solidification ended at the eutectic temperature. The temperature of non-equilibrium solidus (NES) for the AA7050 alloy is about 465 ◦ C, which is observed from the cooling curves and agrees well with literature data and Thermocalc calculations [11]. The average TCC is calculated as follows: TCC =

LT2 − LT1 Lgage (T2 − T1 )

,

(2)

where T2 and T1 are the temperatures (T2 > T1 ) below the solidus; LT2 and LT1 are the positions of the displacement sensor at T2 and T1 , respectively; and Lgage is the gage length of the sample. Experiments were performed with an AA7050 alloy, which composition is given in Table 1. The melt was prepared in the electrical resistance furnace in graphite crucibles. When studying the effect of grain refinement different amounts of Ti (from 0.01 to 0.2 wt%) in a form of an Al–8% Ti master alloy were added to the melt at 740 ◦ C.

3266

L. Zhang et al. / Materials Science and Engineering A 527 (2010) 3264–3270

Table 1 Chemical composition of experimental alloy (in wt%). Alloy

Zn

Cu

Mg

Si

Fe

Zr

Ti

7050

5.66

2.25

1.84

0.04

0.15

0.11

0; 0.01; 0.02; 0.06; 0.14; 0.2

When studying the effect of gas saturation, three charges were prepared in the following manner. (1) A steel container with water was placed in the furnace along with the crucible with the melt. Water evaporated and created humid atmosphere in the closed furnace. Before contraction experiments, the melt was exposed for 30 min to the humid air in the furnace. (2) The same amount of melt was prepared under dry (normal atmospheric) conditions. (3) The melt prepared under dry conditions was further degassed by ultrasonic treatment at a temperature of 740 ◦ C. An ultrasonic transducer with an input power of 4 kW and a preheated niobium sonotrode were used. The alloys in this series of experiments were grain refined by Ti. The saturation of the metal with gas was indirectly assessed by measuring the density of the solid samples by Archimedes’ method. Although we have not measured the gas concentration directly, we can suggest (based on the literature [8,12]) that the hydrogen concentration was about 0.8–1.0 cm3 /100 g in the case of humid air; 0.3–0.4 cm3 /100 g in the case of dry air; and 0.2–0.25 cm3 /100 g in the case of ultrasonic degassing. Also the effect of the sample height (the metal level in the mold cavity) was studied. In this case the cavity of the mold was either fully filled (melt level 25 mm) or halffilled (melt level 12 mm). At least three samples were prepared for each condition to make the results more statistically valid. Selected samples were cut in the middle close to the position of the thermocouple and their structure was examined in an optical microscope after being ground, polished and electro-oxidized. Grain size was measured using the linear intercept method and the statistical analysis of the results was performed.

Fig. 3. Effect of sample density on the preshrinkage expansion in a grain-refined AA7050 alloy. Sample height 25 mm.

3. Results and discussion 3.1. The effect of gas saturation and grain refinement on expansion and contraction behavior of an AA7050 alloy during solidification Different degree of gas saturation was obtained by exposing the melt to water vapor environment (0.8–1.0 cm3 /100 g), dry-air environment (0.3–0.4 cm3 /100 g) and ultrasonic treatment (0.2–0.25 cm3 /100 g). The resultant solid materials exhibited different densities ranging from 2.76 to 2.83 g/cm3 as shown in Fig. 3. Gas saturation and different density are also obvious from the amount of porosity in the samples as demonstrated in Fig. 4. The results show that the temperature of expansion onset increases dramatically with increasing gas saturation, from 606 ◦ C (in the case of dry air) to 674 ◦ C (in the case of water vapor environment). The preshrinkage expansion also increases with gas saturation (decreasing density) as shown in Fig. 3 for the 25-mm high samples. The temperature of the contraction onset, however, is not affected by the gas saturation and remains stable at about 538 ◦ C. The increasing temperature of expansion onset and preshrinkage expansion can be explained by the evolution of the gas (mainly hydrogen) at the onset of solidification and upon further solidification until the rigid solid skeleton is formed in the sample and the rigidity temperature is reached, which is in good agreement with literature [6]. Another important factor that can influence the preshrinkage expansion is the pressure drop across the transition, two-phase zone [13]. The solidifying sample is not homogeneous as the solidification develops from the edges and the bottom of the mold with the semi-liquid core [10]. This inhomogeneity results in

Fig. 4. Microstructures of as-cast AA7050 alloys with different amount of gas: (a) melted in the vapor-enriched atmosphere, corresponding to 2.76 g/cm3 ; (b) degassed by ultrasonic treatment, corresponding 2.83 g/cm3 .

L. Zhang et al. / Materials Science and Engineering A 527 (2010) 3264–3270

3267

Fig. 5. Effect of grain refining on the preshrinkage expansion and the contraction onset temperature of an AA7050 alloy. Sample height 12 mm.

non-uniform volume shrinkage and causes the pressure drop in the two-phase zone in the longitudinal direction of the mold. As a result, there is a pressure-induced flow directed from the center toward the edges of the sample. This flow occurring above the rigidity point creates the momentum that contributes to the preshrinkage expansion. The contribution of this factor should be related to the kinetics of the solidification development, and especially to the development of rigidity, when the influence of the shrinkage flow on the expansion will become negligible. Fig. 5 demonstrates for the 12-mm high samples that the preshrinkage expansion becomes more pronounced with grain refining and, at the same time, the thermal contraction starts at a lower temperature. In the case of grain-refined alloys, the structure consists of small, equiaxed grains that form rigid skeleton at a relatively low temperature (contraction onset shown in Fig. 5). As a result there are more paths and time for the pressure-induced flow. In addition, finer grains provide more interfaces that facilitate gas precipitation from the melt during solidification and, therefore further increase preshrinkage expansion. 3.2. The effect of sample height, grain refining and gas saturation on the contraction behavior of an AA7050 alloy below the solidus In our earlier work, the effect of sample height on the TCC was studied [10]. The results showed that the TCC decreases with the decreasing height of the sample, which was in line with the argumentation that was given in Introduction and was related to the inhomogeneity of the solidifying sample [9]. In our recent experi-

Fig. 6. Effect of sample height on the linear thermal contraction coefficient in an as-cast AA7050 alloy with (0.05% Ti) and without grain refiner (a) sample height 12 mm and (b) sample height 25 mm.

ments, another interesting result was observed, which is shown in Fig. 6. In the case of a 12-mm high sample (Fig. 6(a)), the TCC increases with the increasing temperature in both the grain-refined alloy and the non-grain-refined alloy, which is in good agreement with previously reported date [4,6,14]. But in the case of a 25-mm high sample (Fig. 6(b)) and when the grain refiner (GR) was used, the TCC

Fig. 7. Schematic view of solidification progress in the mold at a certain time: (a) sample height 12 mm and (b) sample height 25 mm. A—liquid zone; B—transition zone; C—solid zone.

3268

L. Zhang et al. / Materials Science and Engineering A 527 (2010) 3264–3270

decreases with the increasing temperature. While in the not grainrefined alloy (NGR) the TCC still increases. This unusual change of TCC probably reflects the inhomogeneity of the solidifying sample and is related to the interaction between the structure and gas precipitation as will be discussed below. Fig. 7 gives a schematic view of solidification progress at certain times after pouring. It is important to note again that the thermocouple is standing vertically in the center of the mold at about 1.5 mm above the mold bottom, which means that we measure the temperature reflecting the onset of bridging between two opposite walls (in the longitudinal direction) of the mold. After the melt is poured into the mold, it begins to solidify from two edges (where the sample is fixed) toward the center of the mold. Then at a certain time (as shown in Fig. 7), the alloy in the point where we measure temperature (reference temperature) must reach the solid fraction at rigidity (rigidity temperature [10]), but above this point the rest of the sample is still in the semi-liquid or liquid state. There are two factors acting here. On the one hand, because of inhomogeneity, the whole sample needs some time to reach the rigidity point and start to contract. Usually, this time is very short and can be neglected, such as in the case showed in Fig. 6(a) and (b) without the grain refiner. With grain refining, however, the fraction solid at which the coherency point is reached increases significantly from 0.25 to 0.52 [15]. This difference is translated to the delayed rigidity [10,16], which means that the volume of the sample, which does not undergo contraction at relatively high reference temperatures, increases with increasing the degree of grain refining. But the

Fig. 8. The effect of grain size on linear thermal contraction coefficient (TCC) in an as-cast AA7050 alloy.

time, when the point at which we measure temperature reaches the rigidity point, does not depend on the sample height. So the higher the sample is, the more time it needs to activate the whole sample’s contraction, and the less the calculated TCC at high temperature is.

Fig. 9. Microstructures of as-cast AA7050 alloys with different amounts of Ti: (a) no grain refiner; (b) 0.02%Ti; (c) 0.06%Ti; and (d) 0.2%Ti.

L. Zhang et al. / Materials Science and Engineering A 527 (2010) 3264–3270

This trend becomes more intense when a larger amount of grain refiner is employed. On the other hand, when the grain refiner is employed, the semiliquid part of the solidification (transition) range will enlarge as compared to the case of non-grain-refined material, especially in the case of 25-mm high sample. This provides more time for the gas evolution and expansion, which offers the resistance to contraction and should be more pronounced in the grain-refined material as there is more interfaces available for gas precipitation (see Fig. 5). In this case, the measured thermal contraction coefficient is apparent and reflects the inhomogeneity of the solidifying sample. This trend can be confirmed by the amount of contraction in the solidification. In the case of a 25-mm high sample, the average εs decreases from 0.31% without grain refiner to 0.18% with the grain refiner, while the average εs with a 12-mm high sample remains almost the same in both cases, being about 0.32%. It is noteworthy that the situation we discussed above occurs only at high reference temperatures, before the whole sample reaches the rigidity point and begins to contract. At lower reference temperatures, the TCC will increase with the height (compare Fig. 6(a) and (b)), which is in a good agreement with previous data [10]. In this case, the higher the sample is, the lower the calculated TCC is at high temperature, and then the greater the calculated TCC is at low temperature, as shown in Fig. 6(a) and (b) with grain refining. This unusual change is evidently due to the inhomogeneity of solidification and can be avoided by decreasing the height of the sample, as illustrated in Figs. 6(a) and 7(a). The results obtained for the 12-mm high sample are closer to the real material properties. Yet another interesting effect of grain refining can be noticed in Fig. 6(a): the TCC at any temperature range is lower for the grainrefined alloys as compared to the non-grain-refined alloy. This effect is especially important because grain refining is known to decrease the hot and cold cracking susceptibility of the alloys. These effects are usually explained from the improved feeding of solidification shrinkage (lower coherency and rigidity temperatures) and enhanced mechanical properties [1,2]. But the specific effect of grain size with respect to the thermal contraction coefficient is virtually unknown. It is worth to note that there is no direct comparability between the TCC we measured in this work and the reference data on the linear thermal expansion coefficient (LTEC), because of essentially different experimental conditions [10]. We suggest that the TCC we measured during cooling after solidification of a 12-mm high sample is closer to the actual as-cast material data. Fig. 8 demonstrates the TCC for AA7050 alloys with different amounts of grain refiner in dependence on the grain size in the case of the 12-mm high sample. The coefficients were calculated in four temperature ranges (as can be seen from the legend in Fig. 8). The TCC increases with increasing the temperature range, which is a good agreement with literature [1,4]. And TCC decreases with decreasing grain size in all temperature ranges which further confirms the observation in Fig. 6(a). Fig. 9 illustrates how the grain size changes with different amounts of grain refiner. The lower TCC of grain-refined alloys can be explained by the precipitation of gas in the solid state. Fig. 10 demonstrates the TCC for AA7050 alloys cast under different conditions. The TCC in the case of dry-air atmosphere is higher than that in the case of vaporenriched atmosphere in all temperature ranges, which means that the precipitation of the gas after solidification still offers a noticeable contraction resistance. The grain-refined structure provides more interfaces for the gas precipitation and also makes it more uniform, which increases the contraction resistance. Therefore a lower thermal contraction coefficient observed during cooling of the as-cast grain-refined alloy which results from gas

3269

Fig. 10. The effect of gas saturation on the linear thermal contraction coefficient (TCC) in a grain-refined as-cast AA7050 alloy. Sample height 25 mm.

precipitation at high temperatures in the solid state may be one of contributors to the lower cracking susceptibility. 4. Conclusions 1. The amount of gas dissolved in the melt influences the temperature of expansion onset and the preshrinkage expansion, while the pressure drop across the mushy zone is another important mechanism of preshrinkage expansion. Grain refining facilitates the preshrinkage expansion. 2. The effect of the inhomogeneity of the solidifying sample on the measured thermal contraction can be alleviated by decreasing the melt level in the experimental mold. The experimentally determined TCC obtained with the lower melt level are closer to the material properties, whereas those obtained with the high melt level reflect the situation close to DC casting practice. 3. The effect of grain refining on the TCC was studied. The results show that TCC decreases with decreasing grain size at all temperature ranges. It is suggested that the precipitation of hydrogen from the solid solution during cooling offers the resistance to the thermal contraction. This may be one of the mechanisms of the lower cracking susceptibility of grain-refined alloys. 4. There is no direct comparability between the TCC measured using the applied technique and the reference data on the LTEC. However, our technique is highly reproducible and reflects the behavior of as-cast material during cooling after the end of solidification. Acknowledgements This work was partially done within the framework of the research program of the Materials innovation institute (former Netherlands Institute for Metals Research), project 05237. Authors would like to thank MSc student I. Alastuey whose results triggered this investigation. References [1] [2] [3] [4]

J. Campbell, Castings, Butterworth-Heinemann, Oxford, UK, 1991. D.G. Eskin, Suyitno, L. Katgerman, Prog. Mater. Sci. 49 (2004) 629–711. M. Lalpoor, D.G. Eskin, L. Katgerman, Mater. Sci. Eng. A 497 (2008) 186–194. J.R. Davis (Ed.), Aluminum and Aluminum Alloys. ASM Specialty Handbook, ASM International, Materials Park, OH, 1993, pp. 696–697. [5] R. Nadella, D.G. Eskin, L. Katgerman, Metall. Mater. Trans. A 39A (2008) 450–461.

3270

L. Zhang et al. / Materials Science and Engineering A 527 (2010) 3264–3270

[6] I.I. Novikov, Hot Shortness of Nonferrous Metals and Alloys, Nauka, Moscow, 1966. [7] D.G. Eskin, L. Katgerman, Metall. Mater. Trans. A 38A (2007) 1511–1519. [8] V.A. Livanov, R.M. Gabidullin, V.S. Shipilov, Continuous Casting of Aluminum Alloys, Metallurgiya, Moscow, 1977, pp. 62–98, 110–113. [9] V.I. Dobatkin, Continuous Casting and Casting Properties of Alloys, Oborongiz, Moscow, 1948, pp. 66–71. [10] D.G. Eskin, Suyitno, J.F. Mooney, L. Katgerman, Metall. Mater. Trans. A 35A (2004) 1325–1335. [11] N.A. Belov, D.G. Eskin, A.A. Aksenov, Multicomponent Phase Diagrams: Applications for Commercial Aluminum Alloys, Elsevier, Amsterdam, 2005, pp. 199–202.

[12] G.I. Eskin, Ultrasonic Treatment of Light Alloy Melts, Gordon and Breach Science Publishers, Amsterdam, 1998, pp. 87–100. [13] G.A. Korol’kov, G.M. Kuznetsov, Sov. Castings Technol. (6) (1990) 1– 3. [14] A.E. Vol, Handbook of Binary Metallic Systems. Structure and Properties, vol. 1, Israel Program for Scientific Translations, Jerusalem, 1966. [15] L. Arnberg, L. Bäckerud, G. Chai, Dendrite Coherency, in: Solidification Characteristics of Aluminum Alloys, vol. 3, American Foundrymen’s Society, Des Plaines, IL, 1996. [16] A. Stangeland, A. Mo, D. Eskin, Metall. Mater. Trans. A 37A (2006) 2219– 2229.