Effects of early stages of prestretching on the aging kinetics in Al–Cu–Li–based alloy

A combination of high hardness values and a low energy–consumption preparation method was used to induce the precipitation of the T1 phase in the Al alloy AA2195. This combination was obtained by subjecting the alloy samples to sequential and early stages of prestretching. Hardness testing and differential scanning calorimetry were employed to explore the behaviours of the hardness and enthalpy values as functions of the prestretching level applied before aging. Results demonstrated that the optimal aging thermomechanical conditions are (1) 150 °C for 10 h after applying prestretching levels of 0%, 1%, and 2% and (2) 150 °C for 20 h after applying prestretching levels of 3% and 4%. Under these conditions, the recorded enthalpy values for the formation of the T1 phase at prestretching levels of 0%, 1%, and 2% were 2.86, 1.72, and 1.14 J g−1, respectively and those obtained at prestretching levels of 3% and 4% were 1.43 and 1.27 J g−1, respectively.


Introduction
Third-generation Al-Li alloys have several key benefits that make them candidate materials for aerospace applications [1]. These benefits stem from their lower density, excellent corrosion resistance, high fatigue performance, high strength, and toughness, in addition to considerable weight savings [2][3][4]. In this context, Al-Li alloys have been widely used in launch-vehicle liquid-propellant tanks and aerospace components [5]. A good example of a third-generation Al-Li alloy is AA2195, which is a type of heat-treatable alloy. This alloy is usually strengthened because of the presence of different precipitates in its microstructure. The microstructure of the alloy and its mechanical properties during heat treatment have been evaluated [6][7][8][9]. Recently, an improvement of the microstructure of the alloy and its mechanical properties has been realized using several engineering methods to treat the alloy's surface [10], such as prestretching before aging and the rolling control technology. Specifically, the prestretching method usually affords the promotion of the precipitation of the most hardening phase in the AA2195 alloy, that is, the T 1 (Al 2 CuLi) phase [7,8]. The quantitative relationship among the prestretching level, microstructure, and strength increment in the AA2195 alloy has been studied by Rodgers and Prangnell [11]. Moreover, the effects of the prestretching level and aging on the mechanical properties and the corrosion behaviour of the AA2195 alloy formed by extrusion after spray deposition were studied by Wang et al [12]. Generally, prestretching before the aging treatment improves the tensile strength of the alloy and inhibits the formation of grain boundary precipitates and the precipitate free zone along grain boundaries [12]. Additionally, the effects of higher prestretching levels than that used in industrial practice on the strength, microstructure, and precipitation kinetics during the artificial aging of the AA2195 alloy were investigated by Rodgers et al [11].
The scientific basis for improving the microstructure of the alloy and its mechanical properties by applying prestretching before the aging treatment can be summarized as follows. (i) An external action from applying deformation causes higher atomic movements than those caused by conventional thermal diffusion. (ii) The material is forced to undergo a state transformation similar to that that occurs at a certain effective temperature. (iii) A phase transformation in the form of formation or decomposition of a supersaturated solid solution (SSSS), dissolution of phases, disordered or ordered phases, and amorphous crystalline phases is finally achieved in the microstructure of the alloy [13]. Despite the fact that several studies have been performed on the effects of increasing the prestretching level on the precipitation strengthening in Al-Li alloys within standard limits, few works focus on applying close and sequential steps of prestretching before alloy aging. Thus, it is critical to understand the effects of a systematic increase of prestretching on the aging kinetics of Al-Li alloys.
This work aimed to systematically investigate the effect of different, sequential, and early stages of prestretching on the aging kinetics of the AA2195 alloy. To this end, hardness testing and differential scanning calorimetry (DSC) were employed to explore such systematic effects of prestretching steps on the precipitation strengthening of the studied alloy. The results were utilized to model the different hardness and enthalpy values as a function of the prestretching level applied before the aging procedure.

Experimental
A typical third-generation Al-Li alloy, that is AA2195 alloy, was used as the primary material. The alloy was supplied as a 20-mm thick plate. The composition range of AA2195 is provided in table 1 [14]. Alloy samples were cut and machined into the shape of tensile samples (figure 1). A 20 mm-thick alloy AA2195 (figure 1(a)) was undergone to the cold rolling by using a rolling machine, in which the texture of the rolled sheet material could be obtained ( figure 1(b)). The final obtained rolling sample has a thickness of 10 mm as shown in figure 1(c).The texture is usually represented by {hkl} [uvw], which means that the {hkl} planes of specific grains lie parallel to the sheet plane while their [uvw] direction point is parallel to the rolling direction [15]. The sheet plane is usually perpendicular to the normal direction [ND], whereas its direction is parallel to the rolling direction [RD]. The major components of the deformation texture in the Al alloys are the {112}〈111〉, it can be expected that alloy AA2195 possess a strong rolling texture in which the sheet is normal to the {110} planes and the RD is the 〈112〉. Thus, it is possible to prepare samples with a defined orientation. Tensile samples were cut and machined in accordance with a cutting machine. Figure 1(d) shows that the tensile sample has two sections: (a) a gage section and (b) a grip section and the obtained dimensions. Each test sample from the gage section underwent a heat aging treatment. Figure 2 shows the schematic for the entire thermomechanical aging process. The applied thermomechanical aging started with solution heat treatment of the specimen in a molten salt bath (75 wt% KNO 3 , 20 wt% NaNO 3 , 5 wt% NaNO 2 ) at an elevated temperature T SHT of 520°C within a singlephase region for a sufficient time (1 h) to dissolve the alloying elements, followed by rapid cooling or quenching with ice water to obtain an SSSS of solute atoms. At this stage, different prestretching levels with sequential steps of 0%, 1%, 2%, 3%, and 4% using tensile stretching were applied. The stretching was performed at a cross-head displacement of 1 × 10 −4 m min −1 . The real-time elongation of each specimen was observed through a digital display via a computer interface with an extensometer across an initial gage. Finally, the decomposition of the SSSS was controlled to form a finely dispersed precipitate by aging for convenient times. All the stretched  samples were placed on quartz ampules, which were sealed under vacuum and subsequently subjected to artificial aging with an initial heating ramp of 20°C h −1 followed by an isothermal hold at T A = 150°C for a range of times up to 20 h. The effect of prestretching on the aging kinetics was first determined by the Vickers microhardness test after applying the thermomechanical aging procedure shown in figure 2. The preparation of the specimens for the hardness test was accomplished by the standard metallurgy method, in which the alloy was mechanically ground with distilled water and SiC paper of 600, 1200, and 2400 grit size, followed by final polishing to eliminate scratches using colloidal silica on cloth. The hardness tests were performed using a 0.2 kg load applied to the sample for 10 s. The hardness value was obtained by averaging 10 measurements.
DSC was performed using a DSC-TA Discovery 250_SW instrument on samples subjected to varying degrees of prestretching and at different stages through their artificial aging treatment. This enabled the effect of prestretching on the enthalpies of the most hardened phase in the AA2195 alloy, that is, the T 1 (Al 2 CuLi) phase, to be determined by calculating the integral of the T 1 precipitation exothermic peak in the DSC curves. The DSC samples were prepared by cutting slices from the gage length of stretched tensile specimens using a diamond cutting wheel. The DSC scans were measured at a heating rate of 10°C min −1 from 25°C to 400°C. Enthalpy values for the formation of the T 1 phase were obtained by the calculation of the integral of the T 1 precipitation exothermic peak in the DSC curves using Trios v5. 1. 1.46572 software program.
X-ray diffraction (XRD) was used to determine the effect of prestretching on the precipitation strengthening of the respective alloy. XRD patterns of the thin plates of the samples were obtained using a Bruker D2 PHASER XRD, equipped with a Mythen 1 K silicon detector and a Cu anode producing mainly CuKα 1 radiation. Powder diffraction data were collected over a range of 10°-90°to identify the effect of prestretching on the presence of the T 1 (Al 2 CuLi) phase in the microstructure of the respective alloy.
Microstructural characterization was performed using transmission electron microscopy (TEM -FEI titan CT) operated at 300 kV. Electron microscope analyses and structural observations were performed on electrontransparent foils from alloy AA2195 at different prestretching levels of 0%-4%.. These foils were grinded to an approximate thickness of 150 μm, mechanically punched, and electro-polished by standard twin jet polishing using a solution of 20 vol% of nitric acid in methanol at −20°C.

Results and discussion
The effects of different, sequential, and early stages of prestretching on the aging kinetics of the AA2195 alloy can reveal the precipitation process to some extent. Figure 3 shows the hardness curves at different prestretching levels of 0%-4%. Figure 3(a) indicates that at a prestretching level of 0%, that is, the non-stretched stage, the hardness values were somewhat high at (114-149 HV) at the beginning of aging time (up to 1 h), followed by an increase to a relatively stable hardness value of 203.1 HV at an aging time of 10 h. Finally, the hardness value decreased gradually to 168.4 HV when the aging time was increased to 20 h. Similar behaviours of hardness response can be seen in figures 3(b) and (c) at deformation levels of 1% and 2%. At these deformation levels, the hardness values at the beginning of aging times (0 and 1 h) were relatively small and fell in the range of 108-143 HV, indicating that the hardness of the nonaged sample and samples aged for a short time (1 h) remained relatively stable and reached small values. However, increasing the aging time to 10 h yielded a much higher hardness value of 209.3 HV for the specimens at a 1% prestretching level and a relatively high hardness value of 177.3 HV for the specimens at a 2% prestretching level. Similar to the specimen at a prestretching level of 0%, increasing the aging time to 20 h afforded a decrease in the hardness value to 135.3 HV for the specimen at a 1% prestretching level and to 143.6 HV for the specimen at a 2% prestretching level. Increasing prestretching steps to 3% and 4% afforded a completely different behaviour of hardness response (figures 3(d), (e)). In this case, the hardness values were relatively low at the beginning of aging times (in the range of 135-144 HV at 0 h and 123-128 HV at 1 h). By increasing the aging time to 10 h, the hardness value showed a little increase to above 150 HV in both cases. Interestingly, the hardness values for the samples that were aged for a longer time of 20 h resulted in reaching peak hardness values of 209.3 HV and 215.8 HV at 3% and 4% prestretching levels, respectively. Thus, the hardening responses for the specimen at prestretching levels of 0%, 1%, and 2% followed the S type fitted curve. However, with the limited data numbers, it is not easy to confirm this fitting type. On the other hand, the linear behaviour can be observed easily at prestretching levels of 3% and 4% in the aging time range of up to 20 h with the hardness values that were linearly increased as the aging time.
To understand the variations of the hardness values as a function of prestretching levels and aging times, a deep insight into the microstructure of the alloy is required. For this purpose, XRD patterns were obtained for the specimens that show peak hardness values at different deformation levels. Figure 4 shows the XRD patterns of the prestretching alloy samples and the standard powder XRD data for the T Al CuLi 1 2 ( )phase as well as for the single Al phase. The XRD patterns for the prestretching samples at a 0% level aged for 10 h, 1% level aged for 10 h, 2% level aged for 10 h, 3% level aged for 20 h, and 4% level aged for 20 h together with the matching data patterns of the T Al CuLi ( )phase is considered to be responsible for the bulk precipitation strengthening. It has a hexagonal crystal structure with lattice parameters of a = b = 0.496 nm and c = 0.935 nm, with the symmetry P6/mmm. It forms as platelets on the {111} plane of the Al matrix [16]. Moreover, TEM micrographs of the specimens at different prestretching levels of 0%-4% were recorded and one of the obtained micrograph is shown in figure 5 . In figure 5(a), bright field TEM image for the specimens at a 3% prestretching level is shown and clearly illustrates the nucleation of the platelet precipitates on both the grain boundaries and the matrix. This observation indicates that the grain boundaries act as the sinks of solute and vacancies that cause the precipitate to concentrate on its surroundings. A magnified view is shown in figure 5(b). Precipitates were characterized using an exact [110] zone. According to obtained selected area diffraction (SAED) pattern in the [110] Al in figure 5(c) , the observed precipitates can be identified as T Al CuLi , 1 2 ( ) in the agremeet with the XRD results in figure 4.
The presence of the T Al CuLi 1 2 ( )phase as the dominant phase in the microstructure of all prestretching samples can be explained in terms of the presence of a strong rolling texture, which is possessed by the microstructure of the respective samples. As shown in figure 1, applying the rolling procedure on the alloy sample caused the presence of a sheet normal to the {110} planes and the RD in the 〈112〉. Thus, it is possible to prepare a sample with a microstructure at a defined orientation, that is, a sample with an axis parallel to the 〈111〉 direction in our case. This was accomplished by cutting the sheet sample after it has been machined into the tensile shape using a wire-sewing machine and boron-carbohydrates as grinding particles in a direction parallel to the [112]. By rotating the sample by an angle of 90°, it will be aligned with the [112] direction and have a sample axis parallel to the [111] direction. Experimental evidence of such obtained microstructure can be seen in the study done by Khushaim [14]. In this study, the results confirm the potential role of the development of texture on the strengthening mechanism. This role can be clearly seen in the interesting microstructure possessed by the rolled specimen with the uniform and homogenous distribution of different platelet precipitates. Having this final product, that is, a sample with a well-defined orientation, leads to the presence of different sets of T 1 platelets in the microstructure of the alloy. This fact has been confirmed by the obtained XRD data (figure 4), which clearly indicates the presence of crystalline peaks that correspond to the T 1 phase in all prestretching samples. Additionally, in figure 4, there is an improvement in the crystallization of the T 1 phase upon increasing the prestretching level. This is simply because the XRD measurements reveal that the peak intensity and sharpness improved by increasing the prestretching level, where an improvement of the crystallinity of the alloy microstructure by the increment in the prestretching level would be induced. Moreover, it has been noted that the intensity of XRD peaks at 38°and 45°which are corresponding to the crystallization of  the T 1 phase has been changed upon the variation of the prestretching level. Once again, this behaviour can be explained in the light of the potential role of the development of texture on the strengthening mechanism. This potential role of the texture development is to simulate the nucleation of the most hardening phase in the AA2195 alloy, that is, the T 1 (Al 2 CuLi) phase by reducing the interfacial energy between precipitate and matrix. Thus, the number of nucleation sites for the T 1 phase would be increased in which the crystallization of this phase is expected to be developed upon increasing the prestretching level.
A further clarification of the effect of applying prestretching sequentially on the aging kinetics of the AA2195 alloy can be obtained by performing DSC for all specimens subjected to varying degrees of prestretching and through their artificial aging treatment. The results are shown in figure 6. The presence of exothermic peaks can be seen in the temperature range of 220°C-280°C for all specimens at different prestretching levels and aging treatment. Comparison with the ternary Al-Li-Cu phase diagram [17] and with the observations of previous work [18] suggests that the main observed exothermic peak in the temperature range of 220°C-280°C corresponds to the reprecipitation of the T 1 phase under dynamic heating conditions in DSC. To attribute an observed peak of the DSC curve at each applied prestretching to the precipitation sequences, a series of DSC signals for the specimens at prestretching levels of 0%, 1%, 2%, 3%, and 4% were obtained and are shown in figures 6(a)-(e), respectively. It is clearly seen in figure 6(a) that for the specimen at a prestretching level of 0% and aged for 1 h, the peak area for an exothermic reaction peak, which corresponds to the crystallization of the T 1 phase, is wider than those corresponding to the T 1 crystallization exothermic reaction peaks at aging times of 0, 10, and 20 h. Moreover, identical peak areas can be observed at aging times of 0 and 20 h while at an aging time of 10 h, a relatively small peak area was recorded. These variations in the peak areas of the crystallization exothermic reaction peaks for T 1 at different aging times might be attributed to the absence of the prestretching procedure for the specimen at a 0% deformation level. Thus, there is no external force that would be responsible for phase transformation, and the presence of the T 1 phase in the microstructure of the alloy is usually caused by the presence of the Al Zr 3 b ¢ ( )phase, which acts in most cases as a nucleation site for the T 1 phase [19]. Notably, after the solution aging treatment, most of the Zr is found within small spherical Al 3 Zr dispersoids that are approximately 20 nm in diameter [20]. However, the aging kinetics of the precipitation of the T 1 phase at the heterogeneous nucleation sites of the β′ particles vary with aging time. This can be concluded from the variation of the peak areas by changing the aging times ( figure 6(a)). Approximate behaviours of DSC signals can be observed for the specimens at prestretching levels of 1% and 2% (figures 6(b) and (c)). Careful inspection of the T 1 crystallization exothermic reaction peaks reveal identical peak areas for the specimens aged for 0 and 1 h. Moreover, these areas are relatively high compared to those for the specimens aged for 10 and 20 h, in which the peak areas were observed to be small at an aging time of 20 h and diminished at an aging time of 10 h. Once again, applying prestretching before the artificial aging usually induces phase transformation, in which a small shift of atoms would be sufficient rather than a long mass transfer. Moreover, increasing the aging times usually afforded increasing diffusivity, in which the completion to reach maximum precipitation rates would be faster. Thus, the small peak areas of the T 1 crystallization exothermic reaction peaks are expected at an aging time of 20 h. However, increasing the prestretching levels to 3% and 4% afforded different behaviours for the DSC signals (figures 6(d) and (e)). In this case, the peak areas of the T 1 crystallization exothermic reaction peaks are huge and almost identical for the specimens aged for 0 and 1 h, while they almost diminished for the specimens aged for 10 and 20 h. Increasing the prestretching levels to 3% and 4% leads to increasing dislocation density, and hence, the overall number of nucleation sites. Thus, the saturated matrix will require additional heat to nucleate or incorporate through a dislocation-particle interaction [14]. However, increasing the aging times to 10 and 20 h causes annealing out of the quenched-in excess vacancies formed during the original aging, and hence, slowing the aging kinetics for the T 1 phase at those times.
The recorded behaviours of the hardness response at different degrees of prestretching specimens are shown in figure 3; the variations in the aging kinetics for the T 1 phase in figure 6 have been influenced by the enthalpy values for the formation of the T 1 phase in all prestretching specimens. Table 2 lists the enthalpy values. It is clear from this table that the enthalpy values for the formation of the T 1 phase were small for the 0%, 1%, and 2% prestretching specimens aged at 10 h with values of 2.86, 1.72, and 1.14 J g −1 , respectively. For the 3% and 4% prestretching specimens, the smallest enthalpy values of 1.43 and 1.27 J g −1 , respectively, were recorded at aging times of 20 h. Careful inspection of figure 6 reveals the diminished crystallization exothermic reaction peaks for the T 1 phase for the 0%, 1%, and 2% prestretching specimens aged at 10 h and for the 3%, and 4% prestretching specimens aged at 20 h. Thus, the observed small enthalpy values for these peaks are expected. Moreover, the maximum hardness values for these specimens with small enthalpy values for the crystallization exothermic reaction peaks for the T 1 phase were recorded ( figure 3). The listed enthalpy values in table 2 have been plotted in figure 7, in which the relations between the ageing times and enthalpy can be observed clearly.
The hardness behaviour and enthalpy values for the 0%, 1%, and 2% prestretching specimens aged at 10 h and 3% and 4% prestretching specimens aged at 20 h are shown in figure 8. In figure 8(a), the maximum hardness values can be achieved at an aging time of 10 h for the 0%, 1%, and 2% prestretching specimens and at  an aging time of 20 h for the 3% and 4% prestretching specimens. Such optimal aging kinetics of the T 1 crystallization are expected at these aging times. At these aging times, the diffusion of solute atoms to fully fill the areas of high-energy sites, dislocations, and vacancies in the microstructure of the alloy can be achieved with a low consumption of heat flow, as can be seen from the enthalpy values at these aging times in figure 8(b). Notably, the obtained enthalpy values for the T 1 phase in table 2 and figure 6 are in a good agreement with the calculated values obtained by Dorin et al [21]. In this context, it can be also noted that applying prestretching procedure resulted in more uniform nucleation site density, and thus both compressive and tensile strains in Al matrix and the T 1 platelet precipitates are expected [22]. Hence, the variations in the enthalpy values as illustrated in table 2 would be reasonable. Hence, the optimal thermomechanical conditions for applying the sequential and early stages of prestretching are as follows. (i) The 0%, 1%, and 2% prestretching specimens should be aged at 150°C for 10 h. (ii) The 3% and 4% prestretching specimens should be aged at 150°C for 20 h. In this context, the criteria for selecting the best optimal thermomechanical conditions are low consumption energy and cost effectiveness, along with superior mechanical properties. Finally, an investigation on hardness and enthalpy values as functions of the prestretching level applied before aging revealed a truly enhanced welldeveloped microstructure of the respective Al alloys or any light alloy systems.

Conclusions
The aim of this study is to systemically investigate the different and sequential early stages of prestretching on the aging kinetics of the T 1 (Al 2 CuLi) phase. The specimens subjected to prestretching levels of 0%, 1%, 2%, 3%, and 4% were selected, which are commonly used in industrial practice. The results demonstrated that at prestretching levels of 0%, 1%, and 2%, the hardness values as a function of aging time at constant temperature followed the Gaussian distribution while at prestretching levels of 3% and 4%, the hardness values showed a linear increase with increasing aging time at constant temperature up to 20 h. For each deformation level, optimal aging conditions were selected based on the ability to produce a superior hardness response by applying low energy consumption. On this basis, at prestretching levels of 0%, 1%, and 2%, the best conditions correspond to aging the specimen at 150°C for 10 h, in which the enthalpy values for the formation of the T 1 phase were recoded to be 2.86, 1.72, and 1.14 J g −1 , respectively. However, at prestretching levels of 3% and 4%, the optimum aging conditions correspond to 150°C for 20 h with enthalpy values for the formation of the T 1 phase of 1.43 and 1.27 J g −1 , respectively. Such a systematic investigation of the effect of prestretching on the hardness response and acceleration of the aging kinetics of the precipitation of the T 1 (Al 2 CuLi) phase can be expanded to other intermetallic phases that exist in the microstructure of the alloy, and hence, an intelligent microstructural design and optimal surface treatment procedure could be achieved.