Warm Aging of Pre-aged AA6013 Sheet and Its Relevance to Room Temperature and Warm Forming Applications—Experimental and Modeling Analyses

ASTM-E8 sub-sized tensile coupons (Ref 42) were machined from the AA6013-PA sheet. The gauge length and width of the sub-sized specimens were 25 and 6 mm, respectively. The AA6013-PA samples were subjected to a secondary aging process (without deformation–deformation effects are addressed in the following section) in a fluidized sand furnace (FB-08C Fluidized Bath Furnace, Techne®) and quenched in room temperature water. Isothermal secondary aging treatments were conducted on the tensile coupons at the following temperatures: 220, 227, 235, 243, 250, 265, and 280 °C. The exposure times for these isothermal heat treatment times varied from 30 to 3000 s in duration. A subset of these tensile coupons was further treated with an additional laboratory aging process of heating at 177 °C for 30 min to simulate a paint bake cycle (PBC).

Following heat treatments, samples were tensile tested at room temperature with a 100 kN load frame (MTS Criterion™: Model 45). A 25 mm virtual extensometer was used to report strain values for stress–strain curves, acquired using digital image correlation (DIC)-based strain measurement. The cross-head speed for all tests was 0.0508 mm/s (corresponding to a nominal strain rate of 0.002 s⁻¹) and each test condition was repeated a minimum of three times. For each condition, the average yield strength and ultimate tensile strength were reported.

Stereoscopic digital image correlation was utilized for the acquisition of all reported strain measurements within this manuscript. With DIC, a temperature-stable random black on white speckle pattern was produced on the specimens using VHT Flame Proof© elevated temperature paint and a modified spray nozzle. Deformation of speckled samples was tracked by two 4 megapixel, 140 frames per second, FLIR Gazelle Camera Link cameras. The pixel density, in pixels/mm, varied as a function of the test setup, but was typically greater than 15 pixels/mm. Deformation and strain fields were analyzed using the Correlated Solutions Vic 3D© software package using a subset of 25-35 pixels, a step size of 2-4 pixels, and a Gaussian strain filter of 5 pixels. These analysis parameters influenced the minimum strain resolution and controlled the range over which strain averaging could be prominent (Ref 43-45). Sometimes referred to as the virtual strain gage length (VSGL) (Ref 46), the approximate minimum and typical regions of strain averaging in this work were 0.4 and 1.3 mm, respectively.

AA6013-PA specimens (40 mm by 40 mm by 2 mm) were also subjected to bending tests following the VDA 238-100 standard (Ref 47). Bendability, defined in ref. (Ref 48), is an important indicator of fracture resistance and is often used to assess the ability of a sheet metal to undergo hemming operations. Bendability can also be used to gauge the ability of a material to deform through progressive folding under crash loading (Ref 49). For side impact structural beams, Kurz et al. (Ref 49) have found that the VDA bend angle correlates well with the amount of beam deflection (during 3-point bending) that can be realized prior to crack initiation. Additionally, they have reported that no cracking was observed for material conditions displaying VDA bend angles above 80°-100° for side impact beams subject to 150 mm of deflection (again, in 3-point bending). As per the VDA standard, a knife edge radius of 0.4 mm was used along with a punch speed of 20 mm/min. Teflon aerosol lubricant was used to lubricate the rollers prior to each test. Further details of the apparatus can be found in the publication by Cheong et al. (Ref 48). A threshold load drop of 30 N was used as the criterion for failure during the bending test, and the associated bending angle at failure was determined using DIC data. The failure threshold of 30 N corresponds to the standard for sheet metal less than or equal to 2.0 mm in thickness (Ref 47).

Warm tensile experiments were performed on the pre-aged AA6013 samples in order to characterize the extent of aging taking place during warm forming operations. In addition, warm forming experiments were performed to assess the effect of aging on formability, as well as to provide samples for calorimetry measurements examining the impact of warm deformation on the subsequent aging kinetics.

Warm pre-straining experiments were performed using an annular Marciniak die-set (Fig. 3) in order to examine the effect of warm deformation on the subsequent aging response (during PBC, for example). The Marciniak punch promotes a region of uniform strain at the center of the sample that was suitable for subsequent harvesting of calorimetry samples. These pre-straining experiments were performed at room temperature (RT), 170 and 230 °C and utilized heated tooling with an inner diameter of 106 mm and die entry radius of 12 mm. The Marciniak samples were rectangular of width 50.8 mm and length 203.2 mm. The Marciniak punch had a 32 mm recessed hole (Fig. 3) machined in the center to promote uniform stretching over the punch. The rectangular samples were clamped between the dies using a force of 330 kN to prevent material draw-in.

The punch speed was 1.0 mm/s and motion was interrupted at punch depth of 24-26 mm which corresponded to a major true strain of approximately 0.1 ± 0.01 in the uniform deformation region for all test conditions, with a minor true strain of approximately -0.04, as determined using the DIC system. A typical strain contour from one of the pre-strained specimens is seen in Fig. 3. Specimens were heated following the same procedure as in the Nakazima warm forming operation with a 3 min heating time with asymptotic temperature rise toward the target temperature. Following pre-straining, samples from the uniform strain regions were sectioned for calorimetry experiments to evaluate the effect of deformation on the subsequent aging kinetics of the PA temper.

To characterize the effect of deformation cycle on the precipitation process, Differential Scanning Calorimetry (DSC) experiments were conducted using a Netzsch DSC 404C apparatus and a constant heating rate of 10 °C/min within the range of 25-580 °C. The DSC tests were performed on the samples extracted from the center of the Marciniak specimens deformed at RT, 170 °C and 230 °C and the undeformed flange regions of the RT specimens with an average sample mass of 60 mg. Two repeated DSC tests were performed for each sample condition. For each testing condition, an alloy sample was placed in the test vessel and a reference pure aluminum sample of identical mass was placed in the reference vessel. The heat released or absorbed was recorded as a function of temperature corresponding to precipitation or dissolution reactions, respectively. To obtain the baseline, a second test was run with pure aluminum samples placed in both test and reference vessels. The net DSC trace was obtained by subtracting the data obtained from the reference material, i.e., pure aluminum, from the data recorded for the alloy sample.

To analyze the precipitation kinetics, Isothermal Calorimetry (IC) tests were conducted using a SETARAM C80 calorimeter, based on the methodology introduced by Esmaeili et al. (Ref 17, 24). Multiple pieces of the deformed and undeformed samples with a total mass of 930-970 mg were extracted from the center and flange regions, respectively, of the RT and warm deformed Marciniak samples. The calorimeter was first stabilized at the test temperature of 177 °C, then the samples were placed into the test vessel 20 s after data acquisition was started. Heat flow was recorded as a function of aging time at the constant temperature of 177 °C. The same IC test was completed on a pure Al sample with a total mass of 935 mg. The result of the pure aluminum run was then subtracted from the alloy sample run trace. The repeatability of the IC test results was checked by running at least two tests on each studied condition.

The evolution of the room temperature yield stress (YS), ultimate tensile strength (UTS), and toughness (integral of the tensile stress–strain response) of AA6013-PA after secondary aging at temperatures ranging from 220 to 280 °C is shown in Fig. 4. For material aged at 220 °C for longer durations of 1000-4000 s, the YS is slightly lower than the measured as-received (T6) YS (solid black symbol in figures), while the UTS is slightly higher. For short exposure times, the YS of the material does not decrease when aged at 250 °C and below, compared to the PA starting condition (open black symbol in figures). This behavior indicates that the 100 °C, 4 h PA treatment sufficiently stabilizes precipitates such that dissolution does not occur upon higher temperature exposure. For AA6013-PA exposed to secondary aging at 265 and 280 °C, there is a small decrease in the measured YS at the shortest exposure time of 30 s. Given the scatter in the data at 280 °C and 30 s, it is hard to conclude whether dissolution of precipitates occurs at this temperature. Regardless, due to the rapid aging kinetics at these temperatures, any dissolution which occurs upon exposure is quickly recovered by formation of new precipitates, as evidenced by the recovery of the YS past the original PA yield stress within 40 s. The UTS follows a similar trend as the YS at these short exposure times; however, the time required to recover the UTS is longer. While the scatter in the tensile toughness data is larger than the YS and UTS data, the trends are clear. For aging at 265 and 280 °C, the tensile toughness drops from the PA value to a level much closer to the T6 value for less than 100 s of exposure. At longer exposure times, the tensile toughness falls below that of the AA6013-T6 condition. For exposure at 250 °C, the tensile toughness after 100-400 s exposure remains higher than the T6 condition, although it does appear that at long exposure times the final toughness is equivalent to that of the T6 condition. For lower temperature exposures, i.e., 235 and 220 °C, the final tensile toughness remains greater than the AA6013-T6 condition up to the time required to reach peak strength at each temperature.

The effect of the PBC following secondary aging on the flow behavior of AA6013-PA is examined in Fig. 5. Here, a secondary aging temperature of 235 °C is considered, with secondary aging durations of 90-410 s (Fig. 5a, b, c, and d, respectively). This secondary aging temperature was selected since it offered a high level of toughness (Fig. 4c) and a strong difference in yield strength between short (90 s) and long (410 s) secondary aging durations. For reference, the stress–strain response immediately following PA, as well as that for AA6013 in the T6 peak aged condition are also plotted. The UTS following the PBC and short secondary aging times is very close to the UTS for AA6013 in the T6 condition, as illustrated in Fig. 5(a). Interestingly, the corresponding YS is relatively low compared to the T6 condition, resulting in a stronger work hardening response and higher elongation. As the secondary aging duration increases, the YS increases as a function of exposure time while the UTS stays relatively constant, thereby lessening the observed differences in the pre and post paint-baked engineering stress strain curves. Additionally, the change in strength due to the PBC alone decreases as a function of secondary aging time at 235 °C. For an initial exposure of 410 s, there is only a minor observable difference between the stress–strain responses of the material prior to and after the paint bake treatment (Fig. 5d). For all secondary aging times considered, from 90 to 410 s, the strain hardening rates and elongations of the pre- and post-paint baked specimens remain higher than the strain hardening rate and elongation in the T6 condition. This difference persists up to the peak strength condition shown in Fig. 5(d) (410 s exposure at 235 °C, plus PBC treatment). The lower yield strength levels (relative to the T6 condition) are likely a consequence of the difference in aging temperatures between the as-received T6 material, which would have been aged at temperatures lower than 180 °C, and the aging temperature used in this series of tests, i.e., 235 °C. From Fig. 5, the aging temperature strongly influences the peak YS that can be obtained, which is a direct manifestation of the effect of temperature on the precipitate population (i.e., dependence on the metastable phase diagram (Ref 50)), as well as solute diffusivity.

The effect of aging duration and paint bake cycle on subsequent sheet bendability (fracture resistance) was also examined. These samples also considered an aging temperature of 235 °C and durations of 90-410 s. The bendability results for AA6013-PA aged at 235 °C are shown in Fig. 6a. The bendability is quite high in the PA condition and then decreases rapidly during the first 200 s of exposure at 235 °C, then more gradually between 200 and 400 s, approaching the T6 value for longer exposure durations. Thus, to achieve a bend angle greater than the AA6013-T6 condition, the secondary aging (warm forming) exposure time should be limited to less than 300 s. For exposure times less than or equal to 300 s there is a clear distinction between the pre and post PBC AA6013-PA conditions. Although the post PBC AA6013 still has a final bend angle greater than the AA6013-T6, in order to obtain a final bend angle greater than 80-100°, the lower range shown by Kurz et al. (Ref 49) to suppress fracture during automotive side impact, the secondary aging (warm forming) time must be kept less than 90 s. In Fig. 6b, the bend angle has been plotted against the strain hardening potential, defined here as the difference in UTS and YS. The resulting linear trend shows a strong correlation, indicating that fracture resistance during crash can be improved by maintaining a strong work hardening potential.

The current results have potential practical implications related to the warm forming of 6000-series aluminum alloys using a pre-aged (-PA) starting temper. First, depending on the temperature, time, and strain imposed during warm forming and subsequent paint curing, UTS levels and, to a lesser extent, YS levels very close to T6 values can be readily achieved from the initial PA starting condition, followed by secondary aging (warm forming) and PBC. In a manufacturing setting, the time at temperature prior to forming, deformation during forming, and subsequent thermal treatments (such as the PBC in automotive parts) can be utilized to warm form AA6013-PA sheet and obtain parts with a high strength and parts with a high bendability (bend angle greater than 110° as in Fig. 6).

In order to examine the effect of temperature on aging at elevated temperatures, this section focuses on application of precipitation kinetics modeling to the current experiments in order to predict the increase in yield strength during pre-heating and to estimate the in situ increase in strength over the duration of the tensile tests. Here, the interaction between dislocations and rate of precipitation is neglected, essentially assuming that the precipitation process occurs under an undeformed state. The effect of warm deformation on the precipitation process is examined in Sect. 3.5.

To assess the precision of the fit kinetic parameters, as well as extend the range of applicability to include yield strength modeling of multi-step aging processes, additional heat treatments and analyses were performed.

Equation 2, 3, 4, and 5 and the kinetic parameters obtained in Sect. 3.2.2 (Table 2) are used to model the YS evolution of AA6013-PA during warm aging followed by paint bake cycling (PBC) processes. The warm aging temperatures are 227, 235, and 243 °C and the PBC has been simulated as a subsequent aging step at 177 °C for 30 min. For a multi-step aging treatment, fr is modeled using the following relationship (Ref 23):where k_i is a temperature dependent \(\left( {k_{i} = \left[ {\frac{{f\left( {T_{i} } \right)}}{n}} \right]^{1/n} } \right)\) kinetic parameter, and t_i, T_i, are the values of time and temperature at discretized points i along the non-isothermal temperature history. The modeling results, along with the measured values of yield strength are shown in Fig. 10(a), (b), and (c). For the aging temperatures of 227 and 235 °C, the yield strength predictions before and after PBC match the experimental data with very good accuracy. At the higher aging temperature of 243 °C, the model predictions prior to PBC are also of good accuracy; however, the predicted yield strength after PBC is slightly underestimated. In general, the multi-step process of warm aging plus paint baking is captured to within 10% of the experimentally measured quantities.

The elevated temperature (230 °C) engineering stress–strain response of AA6013 in the pre-aged condition is presented in Fig. 11(a) for a range of heating profiles and strain rates, while Fig. 11(b) shows the corresponding true stress versus effective plastic strain curves for each condition. These experiments were performed primarily to understand the effect of aging during pre-heating and warm forming on the mechanical response of the alloy during warm forming. Stress–strain curves for hold times of 5, 70, and 180 s were obtained for nominal strain rates of 0.01 and 0.1 s⁻¹. The different hold times prior to tensile testing serve to characterize the effect of secondary aging during heating on the “at-temperature” tensile response of the PA temper. These times reflect the range of “time at-temperature” in industrial applications in warm forming, spanning longer duration conventional furnace heating versus rapid contact or IR heating, as well as accounting for transfer times between the furnace and forming tools.

The yield strength at 230 °C is summarized in Table 3 and plotted in Fig. 12 for both strain rates as a function of hold time, from which it can be seen that the yield strength increased significantly with the hold time. In order of ascending hold time for the lower strain rate, yield strengths of approximately 130, 179, and 236 MPa were observed, with relatively low scatter as shown in Table 3. This increase in yield strength with increased hold time can be attributed to secondary aging during the heating and hold period. For comparison purposes, the predicted increase in room temperature yield strength, obtained using the kinetics model developed in Sect. 3.2.2, is also plotted in Fig. 12. While these predictions are for room temperature, it can be seen that the aging predictions scale well with the measured increase in elevated temperature yield strength as the hold time increases.

For all hold times, the higher strain rate led to an increase in YS relative to the lower strain rate that evolved from 2 to 14 MPa for the shortest to the longest hold time. A positive difference in YS for the imposed increase in strain rate indicates a positive strain-rate sensitivity at yielding at 230 °C that was most pronounced for the longest hold time.

The hardening behavior after yielding for the longest hold time of 180 s also exhibits positive strain rate sensitivity, as seen in Fig. 11b, which is consistent with other stable (solution strengthened or peak aged) aluminum alloys at elevated temperature (Ref 4, 5, 7). Of particular interest in the current results, the strain hardening response at strains beyond yielding for the shorter 5 and 70 s hold time experiments exhibits a seemingly negative strain rate sensitivity. To elaborate, in Fig. 11b, the lower rate curves at 5 and 70 s exhibit a higher degree of hardening, overtaking the strength of the higher strain rate curves after approximately 3-4% tensile strain. To further examine the effect of strain rate on hardening rate, the hardening exponent (n-value) was calculated for all test conditions for effective plastic strains in the range of 0.04-0.06 and is plotted in Fig. 12. For a hold time of 5 s, the average n-values for the 0.01/s and 0.1/s rates were 0.27 and 0.20, respectively, which again reflects the negative rate sensitivity of the hardening response (lower hardening at higher strain rate) observed in Fig. 11. For the 70 s hold time, the n-values were somewhat lower and still exhibited a mild negative rate sensitivity, with n-values for the lower and higher strain rates being 0.16 and 0.14, respectively. For the longest hold time of 180 s, the calculated n-values were both effectively 0.08 for the lower and higher strain rates; thus, the positive strain rate sensitivity observed at yielding (Fig. 11) is preserved up to the UTS for both rates for this condition.

The decrease in the degree of hardening observed at higher rates of strain for shorter hold times is somewhat surprising since aluminum alloys typically exhibit an increase in work hardening with increases in strain rate at elevated temperatures. The overall hardening response captured by the calculated n-values in Fig. 12 has several components, the first being conventional work hardening due to plastic strain accumulation. There will also be an age hardening contribution during the tensile test that will be larger at lower strain rates due to the associated increase in the test duration (Table 3) leading to longer times at elevated temperatures. This effect is most pronounced for the PA temper and shorter hold times for which the extent of secondary aging prior to the start of the tensile test is lower, as reflected by the YS data in Fig. 12. The effect of in situ aging during the elevated temperature tensile tests can be estimated using the kinetics model (Eq 2, 3, 4, and 5) and kinetic parameters obtained in Sect. 3.2.2. For instance, for the tensile tests using a 5 s hold time, the aging contribution to the room temperature yield strength due to in situ aging during testing at 0.01 s⁻¹ would be 9-10 MPa. This increase in stress at least partially accounts for the greater degree of hardening observed in the lower strain rate data in Fig. 11. In contrast, the in situ aging over the short duration of the higher rate test (0.1 s⁻¹) would be negligible.

The formability of AA6013 was characterized in the PA and T6 tempers under isothermal conditions at room temperature, 150, and 230 °C, at a stroke rate of 1.0 mm/s, using dog-bone samples meant to approximate a plane strain loading condition (which is typically the lowest point on a forming limit curve). Representative strain paths taken at room temperature for each temper are plotted in Fig. 13(a). The strain paths were acquired from the necking region on each specimen over an approximate area of 1 mm² using DIC techniques and are plotted up to the image before fracture. Both strain paths demonstrate typical Nakazima characteristics, starting with biaxial bending as the blank wraps around the hemispherical punch geometry, followed by near-plane strain stretching controlled by the blank width and lubrication condition.

The measured plane strain limit strains, as a function of forming temperature and initial temper, are plotted in Fig. 13(b). Also plotted are the dome heights at onset of fracture for each case. At room temperature, the major true limit strain of the PA temper was 0.28 ± 0.01, which is, on average, a 55% improvement over the RT limit strain for the T6 temper. This significant formability improvement associated with the PA temper is attributed to the much stronger work hardening (n-value) relative to the T6 temper, as is evident in Fig. 5.

The T6 temper shows very little difference in formability with temperature for the range of test temperatures, as is the change in dome height. Note that the scatter in the test results (n_min = 3) was low for all conditions and testing of the T6 baseline temper was generally repeatable.

In general, the elevated temperature formability of the PA temper was found to decrease relative to RT while the repeatability also decreased, as evidenced by the increase in the scatter of the measurements. The final dome heights prior to fracture further confirm this trend, with a decrease from 30 mm to approximately 23 mm from RT to 230 °C. In Sect. 3.3, it was observed that the hardening rate of the PA temper at 230 °C varies significantly as a function of time at temperature and that a short heating cycle might be leveraged to increase formability; however, in the warm formability testing conducted in the present study, heating rates were limited to the contact heating rates of the tooling. In total, 3 min was required to heat the specimens to the target temperatures, with 60-90 s of asymptotic heating. Thus, the current 230 °C warm formability experiments on the PA temper reflect a condition that likely falls somewhere between the 70 and 180 s hold times in the warm tensile tests for which n-values of approximately 0.16 and 0.08 were measured, respectively, as opposed to the 5 s hold time for which an effective n-value of approximately 0.27 was measured. The higher n-value for short heating times suggests that improvements in warm formability may be possible using duration heating; however, fast heating for the formability experiments was not available in the current work.

Warm forming was not found to be beneficial for the PA temper using the current specimen heating approach (duration); however, the room-temperature formability of the PA temper was excellent and showed a significant improvement over the peak aged formability. From Fig. 5 it can be seen that the PA temper, after exposure to a thermal cycle of 410 s (6.8 min) at 235 °C, will have a stress–strain response in-line with the baseline T6 temper (independent of a subsequent paint-bake.) Thus, the PA starting temper, when combined with a relatively short secondary aging cycle after forming, offers a T6 final strength level with a 50% improvement in formability. This cycle could be further optimized using the calibrated kinetics model in Sect. 3.2 based on desired final strength and ductility characteristics or for different heating cycles based on furnace capabilities.

The effect of deformation on the aging kinetics occurring during warm forming was examined by relative comparison of DSC traces of pre-strained AA6013-PA samples, deformed at RT, 170 and 230 °C, against a DSC trace of AA6013-PA without pre-strain or temperature exposure. Deformed samples were extracted from the center of the pre-strained Marciniak samples, while the undeformed sample was extracted from the flange of an RT Marciniak specimen. The DSC traces are presented in Fig. 14 for which the exothermic events have positive (upward facing) peaks and are associated with the formation of new precipitates, while endothermic events have negative (downward facing) peaks and indicate dissolution of existing precipitates. The exothermic peak between 200 and 300 °C (peak A) represents the precipitation event associated with the β″ phase which is the primary hardening phase in artificially aged AA6xxx alloys (23, 50). Three key observations can be made from examination of the DSC curves in Fig. 14. First, peak A is not present in the DSC trace of the sample deformed at 230 °C. This absence is attributed to the fact that aging which occurred during sample heating and throughout the warm forming process at 230 °C may have brought the sample into a condition close to the peak aged (or T6) temper. As such, the DSC thermogram for the sample warm deformed at 230 °C only displays a broad endothermic trace and peak at ~ 500 °C, corresponding to dissolution of the equilibrium β (Mg₂Si) phase. As a result, for the specimen deformed at 230 °C, any potential effects due to deformation cannot be readily discerned from the DSC trace alone. Second, for the samples deformed at room temperature and 170 °C (dashed gray and red lines in Fig. 14), the temperature associated with the maximum heat flow of peak A (and hence maximum rate of reaction) was shifted to a lower temperature relative to the DSC trace for the undeformed sample (black solid line). This temperature shift indicates that the precipitation process in the deformed samples was accelerated relative to the undeformed baseline. Finally, the absence of a secondary exothermic peak (peak B) in the deformed samples relative to the undeformed sample (secondary peak at temperature of ~ 280 °C) indicates that in addition to rate effects, deformation (at room temperature and warm temperatures) may alter the ratio of precipitate types which form during the aging process (β″ and Q′). A similar observation has been made for copper-free 6xxx alloys pre-deformed at room temperature, see Yassar et al. (Ref 36, 37).

To quantify the extent of kinetic enhancement as a result of warm deformation, isothermal calorimetry was conducted on the same series of samples deformed with the Marciniak punch at room temperature, 170 and 230 °C and compared with corresponding undeformed samples using the protocol defined in (Ref 17). All IC tests were conducted at a test temperature of 177 °C, to reflect typical paint curing temperatures used in automotive BIW applications. Representative IC traces, given as normalized heat flow (W/g) as a function of test time, are shown in Fig. 15. For a precipitation reaction, the released heat gives rise to an exothermic heat event during the IC run and is measured during the IC test. As the reaction proceeds to completion, the rate of the heat release approaches zero. In the current work, reactions were recorded as complete at time t_f when the measured rate of heat flow dropped below 10^–5 W/g. For precipitation hardenable Al alloys, it has been reported that the values of t_f obtained from IC tests are consistent with the times required to reach the peak-aged condition (Ref 17). Additionally, the area under each IC trace shows the amount of heat released due to the precipitation process and can be used to compare the extent of precipitation for specimens of similar alloy chemistry.

The average reaction completion times, t_f, for the room temperature undeformed and deformed samples were 7 and 6 h, respectively; highlighting that the RT pre-straining influenced the subsequent aging kinetics. This result is consistent with the DSC-based observations that the presence of dislocations in the deformed material enhanced the precipitation process by shifting the exothermic peak to lower temperatures. While the difference in t_f associated with a change in kinetics was approximately 14%, the total heat released differed by less than 6% for both cases. This small difference in total heat released may again be associated with the ratio of precipitate types that form during the aging process (β″ versus Q′). For samples exposed to the 170 °C warm thermal cycle, undeformed and deformed completion times decreased to 5.8 and 5 h and further decreased to 3 and 2.6 h, respectively, for the 230 °C pre-aging condition. The significantly shorter t_f values for both the undeformed and deformed the 230 °C conditions relative to RT and 170 °C suggests that considerable aging occurred during thermal cycling of the 230 °C conditions. Furthermore, the relatively minor difference in t_f between the undeformed and deformed states shows that temperature exposure was the primary driver for the reduction in time to the peak-aged condition for the 230 °C thermally cycled samples during IC analysis. The IC and DSC results indicate that more precipitates have been formed during the warm forming cycle at 230 °C, which leaves fewer solute atoms available for precipitation during the IC/DSC runs compared to the other conditions.

Although the multi-step aging process consisting of (i) pre-aging at 100 °C, (ii) a short temperature exposure between 220 and 280 °C (i.e., the warm forming cycle), and (iii) a PBC, results in slightly superior toughness and bendability properties than the AA6013-T6 condition, it should be noted that the sensitivity to the actual thermal history during the elevated temperature cycle will need to be addressed by automotive parts manufacturers aiming to produce large numbers of components with tight quality control tolerances. The aging model developed as part of this study can be a useful design tool, as shown by its predictive capability in Fig. 10. Specifically, it can be used to predict the evolution of the volume fraction of precipitates and yield strength at warm forming temperatures, which will be useful when determining heating and paint curing cycles, to a first order of accuracy. As shown through differential scanning and isothermal calorimetry experiments on deformed AA6013-PA material, deformation enhances subsequent precipitation aging kinetics. Consequently, future work should consider deformation effects on aging response in order to develop a more precise prediction of the post-warm formed alloy properties. An example of such approaches accounting for aging in deformed samples is described by Poole et al. (Ref 52) and would be a logical extension of the model developed here.