Print version ISSN 1516-1439
Mat. Res. vol.13 no.3 São Carlos July/Sept. 2010
Eugenio José Zoqui*; Luis Vanderlei Torres
Department of Manufacturing Engineering, Faculty of Mechanical Engineering, University of Campinas - UNICAMP, R. Mendelev, s/n, Cidade Universitária "Zeferino Vaz" CP 6122, CEP 13083-860, Campinas, SP, Brazil
This study involved a complete evaluation of the thixoformability of AA7004 and AA7075 alloys, from their microstructural characterization to their viscous behavior. The alloys were subjected to globularization heat treatments for 0, 30, 90 and 210 seconds in two conditions of solid fractions, 45 and 60%, and to viscosity assays under the same conditions. Heat treatments promote the globularization of primary phase particles; hence, the best viscosity results were achieved for alloys with low solid fractions heat-treated for 210 seconds. Alloys AA7004 and AA7075 showed an apparent viscosity of 104 to 105 (Pa.s). The behavior of materials in this range is similar to that of molten glass and they show high formability. However, the AA7075 alloy showed a better performance than the AA7004 due to the smaller size of its primary particles and original grains, their lower growth during reheating, and depending on the condition, their viscosity of 104 Pa.s, which is extremely low for thixoforming standards.
Keywords: semi-solid, thixoforming, AA7004 and AA7075 alloys
Studies for the production of near-net-shape parts (components close to the final shape) have progressed hand-in-hand with studies aimed at reducing the weight of parts on assembly lines, especially in the automotive industry. In this context, thixoforming offers an alternative to conventional processes such as casting and pressure casting. Some of the advantages of thixoforming are its lower consumption of energy, fewer production phases, and higher productivity and quality, since the final product has a high structural homogeneity, good mechanical properties, and low porosity and segregation1-4.
Aluminum alloys are the raw materials most commonly used in processes involving semi-solid materials. However, hypoeutectic aluminum-silicon alloys stand out from other alloys because they are used in about 95% of all applications5. Therefore, the aim of this study is to contribute to the overall development of the field of solid-state conformation, or thixoforming, based on an in-depth analysis of AA7004 and AA7075 aluminum alloys. This analysis involves microstructural characterizations of these alloys and their rheological behavior in the semi-solid state under different conditions of heat treatment time and temperature. The purpose is to determine whether this technology can be applied to conventionally produced alloys. To this end, their morphological evolution is mapped during reheating to the temperature corresponding to the semi-solid range and the resulting viscous behavior is evaluated. Treatment times of 0, 30, 90 and 210 seconds are used because in current industrial practice, reheating of semi-solid materials for thixoforming is done based on time-controlled rather than temperature-controlled heat treatment. This allows the raw material to reach the desired temperature within a window of time, which often means that the material is held at this temperature for longer than necessary for partial liquefaction. The AA7XXX series was chosen because of the outstanding data in the literature on the evaluation of thixoformability of these alloys, especially for the AA7075. The AA7004 alloy although presents a lack of data5-9.
2. Experimental Procedure
The aluminum alloys chosen for this study were AA7004 and AA7075 commercial alloys produced by ALCOA Alumínio S.A. and manufactured by conventional continuous casting. Their chemical compositions are listed in Table 1. The flowchart in Figure 1 illustrates the steps involved in the evaluation of the thixoformability of the two alloys. The first step consisted of a solubilization heat treatment in a muffle furnace. When the furnace reached the required temperature, the alloys were placed inside it, heated to their treatment temperatures (465 °C for AA7004 and 480 °C for AA7075), held at those temperatures for 2 hours, and then cooled in water.
After the solubilization heat treatment, the working temperatures were characterized using differential scanning calorimetry (DSC) and Thermo-Calc® software to determine the temperature interval between solidus and liquidus, and the region of equilibrium between the solid and liquid phases. For the DCS analysis, 100mg samples were placed in a NETZSCH STA 409 °C simultaneous thermal analyzer, which was heated to 700 °C at a heating rate of 5°C/min, after which they were cooled to room temperature at the same rate. Based on the DSC data, the curve of the liquid fraction vs. temperature was then constructed for each alloy, indicating the temperatures corresponding to the solid fractions of 45 and 60%, as well as the liquidus temperature of the alloys used in the globularization heat treatments and viscosity tests.
The simulations with the Thermo-Calc® software were made using the ideal compositions of each of the alloys under study, thus excluding any inclusions or residual elements, resulting in the temperature vs. liquid fraction curves of the alloys. In these simulations, the software employed a calculation routine to evaluate the solidification conditions in and out of equilibrium.
The purpose of globularization heat treatments is to evaluate the structural evolution of the remaining solid phase, allowing these features to be correlated with the materials' mechanical and rheological properties. The globularization heat treatments were performed in a resistance furnace, using 15 mm high by 20 mm diameter samples with a small 1.6 mm hole at mid-height into which was inserted a chromel-alumel thermocouple to monitor and record the temperature during the test. The samples were placed in the furnace after it reached the required temperature and were reheated to the temperatures of solid fraction of 45 and 60%, where they were held for 0, 30, 90 and 210 seconds and then cooled in water. After the heat treatments, the samples were sectioned longitudinally and prepared metallographically for the microstructural characterization.
For the microstructural characterization, the samples were embedded in Bakelite, sanded with 220, 320, 400, 600, 1200 grit sandpaper, and polished with 6μm diamond paste, and lastly finished with 1 μm diamond paste. After polishing, the samples were etched using Keller's reagent (2.5 mL HNO3, 1.5 mL HCl, 1 mL HF and 95 mL H2O), submerging the samples in the reagent for 10 seconds. After etching, the samples were rinsed under running water for about 30 seconds and dried.
The metallographic analysis was carried out using a Leica DM IL optical microscope. The primary particle size was measured using the Heyn intercept method, following the ASTM E112 (1996) standard. The primary globule size was counted in five different fields on each micrograph, and five images were recorded from different sections of each sample, thus making a total of 25 counts of primary particles in each sample. The value of the shape factor (roundness) of each alloy was estimated using the ImageJ 1.40 g software.
For the microstructural characterization by color metallography, the same samples used for the previous characterization were etched again, this time using 6.0% HBF4 electrolytic solution, applying a voltage of 20 V for about 90 seconds under moderate and constant agitation. The samples were then rinsed under running water and blow-dried. They were then analyzed under the same microscope, using polarizing filters to obtain the color images of their grains, so that grains with the same crystal orientation presented similar coloring, thus facilitating their identification and characterization. Heyn's intercept method was also used to determine the grain size, with 25 grain counts made for each tested condition.
The rheological behavior was evaluated based on hot compression tests, as described by Laxmanan and Flemings10 and by earlier works11-13. This characterization method was chosen because it is extremely simple and yields excellent comparative results. The tests were performed on an MTS 810 universal testing machine (load capacity of 100 kN) with a resistance furnace attached to the machine shaft (maximum working temperature of 1200 °C), using Inconel 718 parallel plates whose area was larger than the maximum area occupied by the deformed samples, thus keeping the deformed volume constant. The specimens, whose dimensions were the same as those used in the globularization heat treatment, i.e., samples are 15 mm in height (H0) and 20 mm in diameter. A constant 10 mm/s compression rate (δH/δXt) was used. Results were plotted as strain (F) vs. deformation (H) and transformed to apparent viscosity (µ) vs. shear rate γ using the Equations 1 and 2[11-13]:
where µ is the apparent viscosity (Pa.s), F is the load (N), in a time t (s), V is the sample volume (m3), considered constant, H is the instantaneous height (m) and Ho is the initial height (m).
The hot compression tests were performed under the same conditions of temperature and time as those applied in the characterization of the morphological evolution. Compressive force vs. deformation was converted to apparent viscosity vs. shear rate using the same equations and procedures as those employed in previous works11-13.
3. Results and Discussion
3.1. Characterization of working temperature
The first step was to determine the working temperatures to be employed in this study. To this end, the two methods most commonly reported in the literature were chosen to determine the solid-liquid phase transformation: differential scanning calorimetry (DSC) and simulation using the Thermo-Calc® software package. Figure 2 illustrates the solid-liquid phase transformation based on the evolution of the liquid fraction as a function of temperature.
The thermogram of the AA7004 alloy (not shown here) indicated only one endothermic peak, corresponding to melting of the alloy between 603 °C (solidus temperature and beginning of the endothermic peak) and 662 °C (liquidus temperature and end of the endothermic peak). In this case, due to the lack of sufficient copper or any other element to form a significant quantity of intermetallic compounds, only the melting phenomenon was visible. The solid fraction was determined by partially integrating the area under the endothermic peak as a function of temperature. To this end, we ignored the dissolution or conversion of all the expected phases or precipitates that did not appear clearly in thermogram, namely MgZn2 (the hexagonal Laves structure (C14) partly responsible for hardening of the AlFCC matrix) and AlMgZn5.
On other hand, the DSC thermogram of the AA7075 alloy showed an initial peak occurring between 470 and 500 °C, which is characteristic of the dissolution of zinc compounds, notably MgZn2. Moreover, in this case, with an Mg/Zn ratio of 1:3, there is formation of Al32(Mg, Zn)49 and, because the copper content exceeds 1 wt. (%), also CuMgAl26,7. The second peak corresponds to the endothermic melting peak that occurs between 548 and 636 °C (between 530 and 642 °C; however, the object of this study is the solidification process). The temperature of 548 °C corresponds to the solidus temperature (beginning of the endothermic peak), while the temperature of 636 °C (end of the endothermic peak) corresponds to the liquidus temperature. Solid fractions are characterized in this temperature range by integrating the area under the peak, as indicated in Figure 2.
In the Thermo-Calc® simulations, the data generated were the temperature vs. liquid fraction curves for each alloy under study, starting from each alloy's original composition. Figure 2, which summarizes the results obtained from the thermograms (DSC) and simulations, indicates the temperatures corresponding to the solid fractions of 45 and 60%. Note that the simulated melt onset temperature is considerably lower than that obtained by DSC, i.e., 475 °C for the AA7004 alloy and about 430 °C for the AA7075. This difference is due to the fact that Thermo-Calc® does not actually have a routine that considers melting of the material, only thermodynamic calculations that simulate solidification. This problem is solved by converting the solid fraction vs. temperature data to liquid fraction vs. temperature, although the melting kinetics and thermodynamics may differ considerably from those of solidification. In practice, there are numerous studies and points of view about the solidification of materials, and companies and universities have entire groups dedicated to these studies. However, little information is available about inverse phenomenon, i.e., melting and partial melting.
Table 2 summarizes these data. To facilitate the analysis, each temperature is accompanied by an indication of the material's sensitivity to changes in temperature, i.e., δfs / δT, which represents the derivative at the point corresponding to this temperature. The lower this sensitivity the more suited the material is for thixoforming processes, since it could undergo increases or decreases in processing temperature without undergoing major changes of the liquid fraction, which would significantly alter its viscous characteristics. The literature reports similar values, albeit in terms of "mole fraction/°C"5. In the present case, sensitivity is measured in percentage of mass fraction of existing liquid.
In general, each 0.01%fs / °C represents a 1% change in the relative volume of the solid and liquid phases. For example, for the AA7075 alloy at the temperature of 614 °C, each degree Celsius above or below this temperature represents 2.3% more or less in the liquid fraction. Hence, if the working temperature oscillated by ± 2 °C, a solid fraction of 40.4 to 49.6% (DSC) or 41.6 to 48.4% (Thermo-Calc®) is expected. This indicates a very wide variation in the thixoforming process. The AA7004 alloy displayed an even more unfavorable behavior: to obtain 45% of solid fraction at a temperature of 647 °C and the same oscillation of ± 2 °C, one would have a solid fraction of 37.4 to 52.6% (DSC) or 38.6 to 51.4% (Thermo-Calc®), i.e., an extreme variation that would greatly compromise control of the process.
The confluence of data obtained by DSC and Thermo-Calc®, which yielded equivalent data for the AA7075 alloy, was not observed for the AA7004 alloy. For this alloy, the differences in temperature expected for the solid fractions of 45 and 60% reached 21 °C. Therefore, temperatures of 614 and 607 °C, respectively, were adopted for the AA7075 alloy. In contrast, all the temperatures found for the AA7004 alloy were tested in the globularization heat treatment. However, at the temperatures obtained by Thermo-Calc® simulations, the samples remained completely solid during the tests, for which reason they were excluded. At the temperatures determined from the DSC tests (647 and 642 °C), the samples appeared to be consistent with visible solid and liquid phases, so these temperatures were adopted as the working temperatures. It was therefore concluded that for these alloys of the AA7XXX series, DSC characterization appears to be more accurate. Overall, the two working temperatures are similar to those reported in the literature5-9.
3.2. Characterization of morphological evolution
Having determined the temperatures corresponding to solid fractions of 45 and 60%, the globularization heat treatments were carried out for 0, 30, 90 and 210 seconds to evaluate the morphological evolution of the primary phase. The structural characterizations were performed by conventional metallography, and by color metallography via polarized image, and consisted of the following procedures. Characterization of the structures of the as-cast alloys solubilized; characterization of the as-cast alloys melted and heat-treated for 0, 30, 90 and 210 seconds at the temperatures required to reach 45 and 60% of solid fraction; determination of the size of the primary phase or globule (GLS); determination of the shape factor (SF), roundness in the conventional metallography; and characterization of the grain size (GS) in the polarized color metallography.
Figure 3 shows the as-cast solubilized conditions, i.e., the starting structures for the experiments. With regard to the size of the primary phase, the AA7004 alloy was double the size of the AA7075 alloy. The microstructures of the two alloys generally displayed typically rosette-shaped grains, although the microstructure of the AA7004 alloy was coarser. The as-cast solubilized structure of these alloys therefore differs from that of cold-formed structures reported in the literature6-9.
The main effect of the solubilization heat treatment was to reduce residual stresses in the samples supplied by ALCOA, allowing for machining of the test specimens. However, it should be noted that the two materials showed well defined dendritic boundaries. In addition, the AA7004 alloy showed the presence of particles that were not completely dissolved in the treatment. This type of alloy could also be stress relieved at lower temperatures (from 130 to 150 °C).
Figure 3 also shows a micrographic view of the structures under polarized light, which is a more efficient way to determine grain size. According to metallurgical concepts, grains differ from each other due to different crystallographic orientations, so different grains display different colors under polarized light. This phenomenon facilitates the differentiation of neighboring grains, which may be mistaken for separate entities in conventional micrography. Again, what is visible initially in this figure is the fine rosette structure of the AA7075 alloy and the coarser rosette structure of the AA7004 alloy.
Figures 4 to 7 show the microstructures of the alloys treated for 0, 30, 90 and 210 seconds, respectively, for 45 and 60% of solid fraction. Figures 4 and 5 present the usual micrographs reported in the literature, while the color micrographs in Figures 6 and 7 offer a better view of the samples' grain structure.
Figure 4, which depicts the microstructures of the AA7004 alloy, reveals a relatively globular shaped primary phase, starting from the treatment for 0 second, which increases in size and sphericalness as the treatment time increases, as expected. However, this morphology is very coarse, showing much larger globules than those of the AA7075 alloy (Figure 5). Both alloys also showed innumerable small precipitates finely distributed in the aluminum matrix, which were not clearly visible by color metallography. However, there was little entrapped liquid, which is usually present in structures deformed prior to globularization heat treatments8.
A qualitative analysis of the AA7075 microstructures in Figure 5 indicates good evolution towards a globular morphology, even at 0 second of heat treatment. As expected, the size of these primary globules increased slightly when the material was held in the semi-solid range for 210 seconds. However, the primary phase showed a smaller amount of precipitates.
Figure 6 illustrates the evolution of the AA7004 alloy microstructure, showing globular grains with much larger sizes than those of the AA7075 alloy, starting from the treatment time of 0 second. The size and sphericalness of these grains also increased as the material's morphology evolved up to 210 seconds of processing time. The grain were homogeneous, and the results revealed no major differences in the alloys treated at 45 and 60% of solid fraction for 0, 30, 90, and 210 seconds. The minor difference of 5 °C in the temperature of the two tests may explain this slight variation in grain size.
Figure 7 shows the morphological evolution of the AA7075 alloy during reheating to the semi-solid state. The micrographs reveal globular primary phase grains starting from 0 second of heat treatment time, i.e., only due to the heating, with a slight increase in grain size during the material's morphological evolution up to 210 seconds of processing.
The microstructures showed highly homogeneous grains compared to the as-cast structure, with no major differences in the results obtained for the alloys treated at 45 and 60% of solid fraction for 0, 30, 90, and 210 seconds. The slight microstructural change between the as-cast and treated materials was attributed to the Ostwald ripening mechanism.
Table 3 lists the average size of the primary phase or globule size (GLS), the shape factor (SF) values obtained by conventional metallography (Figures 4 and 5), and the grain size (GS) obtained by color metallography (Figures 6 and 7). It should be kept in mind that the shape factor (SF) used in this study was the roundness parameter, i.e., the closer to "1" the more globular the structure. Moreover, the high standard deviations found in this study are completely acceptable for these casting alloys, which have highly heterogeneous particles.
The higher solid fraction, 60%, yielded larger primary phase sizes than the solid fraction of 45% due to coalescence, which is strongly favored in higher solid fractions due to the greater contact between solid particles. This phenomenon was observed in both alloys. With regard to the shape factor, both alloys also showed increased roundness, although this effect was more pronounced in the AA7075 alloy.
As has been shown previously11-13, the smaller the initial grain size of the structure to be heat-treated the stronger the effect of globularization. This fact is confirmed by comparing the microstructures of the as-cast and heat-treated alloys, especially in the case of AA7075.
The morphology showed a visible evolution in response to the treatment in the semi-solid range. The SF of the AA7004 alloy with 60% of solid fraction varied from 0.08 to 0.25, with the greatest variation occurring in response to 210 seconds of heat treatment. It can therefore be concluded that globularization occurs naturally in response to longer heat treatments, mainly due to Ostwald ripening and coalescence. However, these values were very low. The SF of the AA7075 alloy also showed little variation, i.e., 0.17 to 0.28, due to the same mechanisms, although this material showed extensive interlinking. The material appeared globular under low magnification, but high magnification revealed its interlinked structure (Figure 8).
The shape factor (SF) value indicates the structure's globularity. In general, the structure of the AA7075 alloy was almost globular, with an SF value of about 0.20. The condition that yielded the highest globularity was the solid fraction of 60%, which showed values of 0.20 to 0.30, especially in heat treatments of 210s (SF = 0.28). The micrographs indicate that the AA7004 alloy had a rosette to globular-shaped microstructure characterized by low shape factor values, especially in the solid fraction of 45%.
Table 3 also presents the average grain sizes (GS) of the alloys in each tested condition. Note that the high standard deviations found here are completely acceptable for these casting alloys, whose particles are expected to be highly heterogeneous. However, the evolution of grain growth at high temperatures was very moderate, since the AA7004 alloy showed a grain size of 280 to 307 μm, i.e., a growth of only 3.4%, while the AA7075 alloy showed grain sizes of 124 to 137 μm, indicating a grain growth of about 10%. This growth was lower than expected when compared with previous works11-13.
In previous work12,13, a comparison of the RQI values and apparent viscosity of the structures led to the conclusion that, at RQI values exceeding 0.35, the morphology can be expected to show good formability in the semi-solid state. At an RQI of less than 0.30, the structure showed an irregular globular morphology, which was attributed to the greater interaction among primary globules and their neighbors. The Rheocast Quality Index (RQI) indicates how similar the grain size and globule size are; the more similar they are the less complex and more globular the structure is. The AA7075 alloy with 60% of solid fraction showed the closest approximation to this criterion, with values close to 0.30, especially in the test condition of 210 seconds, which yielded an RQI of 0.28. However, the AA7004 alloy showed the lowest RQI values, particularly in the condition of 45% solid fraction, and the treatment time of 0 second yielded an RQI of 0.14, which is expected to show poor viscosity.
In Figure 9, the RQI values of the AA7004 and AA7075 alloys showed an upward trend starting from the treatment time of 0 second to that of 210 seconds, indicating the effects of coalescence and Ostwald ripening. After coalescing, these particles develop a globular morphology, causing the RQI to increase. The alloys with a solid fraction of 60% showed a better performance in terms of the RQI, since higher solid fractions facilitate coalescence. However, although greater globularization is considered a necessary condition, it does not suffice for thixoforming. In other words, in addition to the globularity of the primary particle, thixoforming requires a structure with the smallest possible grain size. Therefore, the structure that would probably show the best performance in terms of viscous behavior is the AA7075 alloy, since it presented the lowest grain/globule size ratio. It is important to notice that most of the industrial semi-solid processing are time-dependent rather than temperature-dependent, i.e., the lower variation of the RQI at the semi-solid range, showing a stable structure, favors the industrial application of both alloys.
Lastly, it was found that the two alloys showed low mean RQI values and that their morphology is not very favorable for thixoforming. However, this finding does not, per se, mean that these structures cannot be used as semi-solid raw materials. To confirm that these alloys can actually be used as semi-solid pastes, they were subjected to viscosity tests, whose results are presented and discussed below.
3.3. Characterization of the rheological behavior
The characterization of the viscosity of materials in the semi-solid state is essential to understand their rheological behavior when they are subjected to compressive stresses, which is the basis of all thixoforming processes. The characterization of this behavior, performed here via hot compression tests, aimed to approximate solutions in terms of stress vs. strain and apparent viscosity vs. shear rate for thixoforming, especially for thixoforging operations. Both these data can be used in simulation software to evaluate the filling of dies and the forces required for this purpose. Note that the compression speed used here was 10 mm/s, aiming to simulate a behavior as similar as possible to practical thixoforging conditions.
The stress-strain curves for each case are presented first, followed by the apparent viscosity vs. shear rate curves. Figures 10 and 11 represent the stress-strain curves, as well as the viscosity vs. shear rate behavior of the AA7004 and AA7075 alloys reheated to solid fractions of 45% and 60% for 0, 30, 90 and 210 seconds, respectively.
Evaluation of stresses during the conformation process is crucial for the design of tooling and of presses or injection machines to be used. It also allows for numerical simulations of the filling of these dies. Note that extremely low stresses were observed in all the tested conditions. However, the heat treatment time appears to be a determining factor in evaluating the compression behavior of these alloys. The values of stress declined as the morphological evolution progressed to a higher degree of globularization, decreasing gradually from the holding time of 0 second to the time of 210 seconds. This fact was due to the longer heat treatment times, which gave rise to more globular structures as well as a better distribution of liquid phase in the material, improving its flow.
At 60% of solid fraction, the two alloys showed higher stress than the same alloys at 45% of solid fraction, which is explained by the larger amount of solid phase in the structure, and hence, greater flow resistance, since deformation of the solid phase requires greater effort. The final peaks at end of deformation, which are visible in all the curves, are due to end of flow, i.e., the expulsion of the liquid of the material under deformation. However, these peaks were more pronounced in the curves of the 60% solid fraction, since the larger amount of solid was naturally more resistant to deformation during the hot compression test.
Another aspect worth pointing out are the initial deformation peaks, especially of the 60% solid fraction, which all occurred before the strain of 0.1, i.e., at the beginning of compression. These peaks indicate the material's thixotropic behavior, showing their initial resistance due to a three-dimensional network of globular solid particles scattered throughout the material, whereby the semi-solid phase supports its own weight and can be handled as a solid1-3,5. Because of the higher concentration of solid phase at 60% fs, and therefore, a more complex skeleton of solid globular particles permeating their structure, the two alloys present much higher peaks, i.e., a greater inertia or flow resistance at the beginning of compression.
Some of the strain curves in the range of 0.15 to 0.5% reveal the presence of a threshold where the material shows continuous deformation with little or no variation in stress as a function of continued deformation. At this stage the liquid acts as a lubricant, facilitating the movement of primary phase globules and enabling deformation without increasing the applied stress. The final peaks in all the curves, which begin after a strain of 0.5%, are explained by the model proposed by Kang16. According to this model, at a certain degree of compression after the liquid phase is forced out to the periphery, the solid particles begin to touch each other and become deformed, thereby strongly increasing the alloys' yield stress. Table 4 summarizes the main mechanical properties obtained from the hot compression tests, namely mean stress, maximum stress and average apparent viscosity. The values considered average were obtained from the stress-strain curves at the point of deformation equal to 0.3.
As can be seen in Table 4, the yield stress tended to decline with increasing heat treatment time. This decline is mainly due to the fact that longer treatment times lead to greater globularization, and may also lead to reduced interaction among particles. This finding suggests that the adoption of longer treatment times will result in lower stresses for the onset of deformation of semi-solid material. Table 4 also indicates that, regardless of test conditions, the two alloys showed very similar stress values, demonstrating that all the tested conditions can be applied in the thixoforming process.
With regard to the rheological behavior, the AA7004 and AA7075 alloys showed very similar characteristics, as illustrated in Figures 11 and 12. These figures clearly show a gradual decrease in apparent viscosity at longer heat treatment times, as well as a decline in the decrease of solid fraction from 60 to 45%. The initial peaks of the curves correspond to the material's initial resistance due to the previously described three-dimensional skeleton of solid particles. The data showed here are consistent with the results obtained by Kim et al.17,18: for the AA7075 the author reaches Maximum Stress of 3.5 MPa at 600 °C and Viscosity ranging from 104 up to 107 Pa.s at Shear rate ranging from 0.03 up to 2 s-1 in a close die compression test.
Note, also, that the apparent viscosity decreases markedly in response to longer heat treatment times. This behavior is very similar to that described for the mean and maximum stress and reported in previous work12-14. The higher solid fraction resulted in higher apparent viscosity, due to factors similar to those discussed earlier for the mean and maximum stresses of these alloys.
All the apparent viscosity vs. shear rate curves showed an almost constant threshold at which the viscosity remained unchanged while the shear rate increased, particularly between the points corresponding to a shear of 1.0 to 3.0 s-1. At this threshold, the material behaves almost like a Newtonian fluid, i.e., its viscosity remains constant despite variations in the shear rate. This phenomenon has been reported previously12,14, when it was also observed that the greater the globularity of the semi-solid material the lower the decrease in apparent viscosity with increasing shear rates.
Table 4 also presents the average viscosity values measured at 2.0 s-1 of shear rate, most of which fall within a range of 103 to 105 (Pa.s). According to Fleming10, materials in this range of viscosity behave similarly to molten glass and show high formability. The samples heat-treated for 0s showed a higher average viscosity than those treated for longer times, especially samples with 60% of solid fraction, due to the direct effect of the lower degree of globularization and lower dispersion of liquid in the material.
A comparison of the two alloys indicates that the AA7075 alloy performed better in terms of apparent viscosity, achieving 1.0*104 Pa.s in the heat treatment time of 210 seconds in 45% of solid fraction. However, this value is very similar to that of the AA7004 alloy (1.1*104Pa.s), despite the considerable difference in grain and/or primary particle size. Nevertheless, it should be emphasized that both the working temperature evaluation methods used here, DSC and Thermo-Calc® software, can lead to misinterpretation, i.e., the real solid fraction at the temperature under study may be higher or lower than the predetermined 45% and 60%. Minor temperature variations in these materials lead to major variations in solid fraction, which implies that the viscosity determined in this particular case may be higher or lower. In this case, the temperature in the compression test varied by +/- 2 °C, a temperature range that is difficult to maintain during 210 seconds of heat treatment.
Figure 12a illustrates the relationship between the apparent viscosity observed and treatment time with a given fraction solid, while Figure 12b shows the morphology expected from this treatment (in terms of the RQI). Note that increasing the treatment time leads to a marked decrease in viscosity. However, despite this decrease in viscosity with increasing RQI, the variation of this factor is too small to cause such an abrupt decrease. In other words, the apparent viscosity is more affected by temperature (which defines the solid fraction) and by time (which defines the distribution of liquid phase in the material) than by the morphology. In fact, the morphological evolution with RQI varying from 0.15 to 0.30 is too small to explain this decrease, especially when one considers that the size of the still solid particle remaining in the AA7004 alloy is about twice the size expected for AA7075. Apparently, the AA7004 alloy contains a larger amount of liquid than the 55% expected at 646 °C, i.e., the two alloys displayed a similar behavior but appear to contain different amounts of solid fraction. It should be kept in mind that this material is highly sensitive to temperature variations, as indicated in Table 2.
Lastly, it was found that the two alloys showed a rheological behavior similar to that of alloys analyzed previously11-13. However, the AA7075 alloy showed a better performance than the AA7004, since its original globules and grains were smaller and showed a lower growth rate during reheating and, depending on the condition, its viscosity was almost 1.0*104 Pa.s, which is extremely low for thixoforming standards.
It can be stated that in general, and particularly in the case of the AA7004 and AA7075 alloys, the working temperatures obtained by DSC tests were more suitable for testing the globularization heat treatment of this alloy series. The primary solid particles of these alloys underwent growth and globularization during their morphological evolution. The effects of coalescence and Ostwald ripening were found to depend on the holding time of the alloys at high temperatures.
Structure with larger original grain size, the AA7004 alloy, showed a higher growth than the AA7075 due to coalescence among primary particles. The solid fraction and treatment times were found to exert a significant influence on the morphology of primary globules/grains. The longer the dwell time of the alloy and the higher its solid fraction, the better the results in terms of primary globule size and shape, shown by the Rheocast Quality Index achieved. But the RQI varies only from 0.14 up to 0.25 for the AA7004 and from 0.18 up to 0.28 for the AA7075 alloy showing stability of those structures. It is important to notice that most of the industrial semi-solid processing are time-dependent rather than temperature-dependent, i.e., the lower variation of the RQI at the semi-solid range, showing a stable structure, favors the industrial application of both alloys.
Despite the solid fraction of 60% presented the better Rheocast Quality Index (RQI), (because the larger solid fraction favored contact among particles and a higher degree of globularization of the grains than the solid fraction of 45% due to coalescence), the compression tests indicated that both alloys are suitable raw materials for thixoforming due to their low stress and low viscosity presented, especially in the condition of 45% of solid fraction. The AA7075 alloy with 45% of solid fraction showed lower stress and viscosity than the AA7004 alloy in the same condition. The values of mean stress and average apparent viscosity generally decreased as the morphology evolved due to longer heat treatment times, leading to rounder structures, which translates into better flow behavior in hot compression tests.
As expected, tests with 60% of solid fraction presented higher stress values than tests with a solid fraction of 45%. This is explained by the greater amount of solid phase in the structure of these alloys, leading to a higher flow resistance since deformation of the solid phase requires more energy than deformation of the liquid phase.
The Rheocast Quality Index (RQI) reflects the flow behavior better than the shape factor (SF), since it takes into account the structure's grain and globule sizes. The smaller the grain size the higher the shape factor, and the higher the RQI the lower the viscosity. However, the effect of the solid fraction is greater than that of the morphology.
The authors gratefully acknowledge ALCOA Alumínio S.A. for its cooperation, and FAPESP (Fundação de Amparo a Pesquisa do Estado de São Paulo) (Project 2008/03946-4), CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico) (Project 470081/2006-6) and CAPES (Coordenação de Aperfeiçoamento de Pessoal do Ensino Superior) for their financial support of this work.
1. Garat M, Maenner L and Sztur CH. State of the Art of Thixocasting. In: Proceedings of the 6th International Conference on the Semi-solid Processing of Alloys and Composites. 2000; Turin, Italy. p.187-194. [ Links ]
2. Figueiredo A, Apelian D, Jorstad J, DasGupta R, Flemings MC, Alexandrou NA et al. Science and Thecnology of Semi-Solid Metal Processing. Rosemond, USA: Ed. NADCA - North American Die Casting Association; 2001. [ Links ]
3. Hirt G and Kopp R. Thixoforming - Semi-Solid Metal Processing. Weinhein, Germany: Wiley-Vch Verlag GmbH & Co.; 2009. [ Links ]
4. Chiarmeta G and Zanardi L. Production of Structural Components by Thixoforming Aluminum Alloys. In: Proceedings of the 3th International Conference on the Semi-solid Processing of Alloys and Composites. 1994; Japan. p. 235-244. [ Links ]
5. Camacho AM, Atkinson HV, Kapranos P and Argent BB. Thermodinamic predictions of Wrought Alloy Compositions Ameable to Semi-Solid Processing. Acta Materialia. 2003; 51: 2319-2330. [ Links ]
6. Liu D, Atkinson HV, Kapranos P, Jirattiticharoean W and Jones H. Microstructural evolution and tensile mechanical properties of thixoformed high performance aluminum alloys. Materials Science and Engineering A. 2003; 361:213-224. [ Links ]
7. Chayong S, Atkinson HV and Kapranos P. Multistep induction heating regimes for thixoforming 7075 aluminum alloy. Materials Science and Technology. 2004; 20(4):490-496. [ Links ]
8. Chayong S, Atkinson HV and Kapranos P. Thixoforming 7075 Aluminum Alloys. Materials Science Engineering A. 2005; 390:3-12. [ Links ]
9. Atkinson HV, Burke K and Vaneetvelt G. Recrystallisation in the semi-solid state in 7075 aluminum alloy. Materials Science Engineering A490. 2008; 266-276. [ Links ]
10. Laxmanan V and Flemings MC. Deformation of Semi-solid Sn-15%Pb Alloy. Metallurgical Transactions A. 1980; 11:1927-1937. [ Links ]
11. Zoqui EJ, Shehata MT, Paes M, Kao V and Es-Sadiqi E. Morphological evolution of SSM A356 during partial remelting. Materials Science and Engineering A. 2002; 325(1-2):38-53. [ Links ]
12. Zoqui EJ, Paes M and Robert MH. Effect of macrostructure and microstructure on the viscosity of the A356 alloy in the semi-solid state. Journal of Materials Processing Technology. 2004; 153-154(1-3):300-306. [ Links ]
13. Paes M and Zoqui EJ. Semi-solid behavior of new Al-Si-Mg alloys for thixoforming. Materials Science and Engineering A. 2005; 406:63-73. [ Links ]
14. Zoqui EJ, Paes M and Robert M. Rheological behavior and microstructural evolution of semi-solid A356 alloy produced by different routes. Diffusion and Defect Data Pt.B: Solid State Phenomena. 2006; 116-117:565-568. [ Links ]
16. Kang CG, Choi JS and Kim KH. The Effect of Strain Rate on Macroscopic Behavior in the Compression Forming of Semi-solid Aluminum Alloy. Journal of Materials Processing Technology. 1999; 88:159-168. [ Links ]
17. Kim WY, Kang CG and Kim BM. The effect of the solid fraction on rheological behavior of wrought aluminum alloys in incremental compression experiments with a closed die. Materials Science and Engineering A. 2007; 447(1-2);1-10. [ Links ]
18. Kim WY, Kang CG and Lee SM. Effect of viscosity on microstructure characteristic in rheological behaviour of wrought aluminium alloys by compression and stirring process. Materials Science and Technology. 2010; 26(1):20-30. [ Links ]
Received: December 2, 2009
Revised: April 9, 2010